’9'" ’U‘V‘vv" v

N ‘ .b- L.

All l.‘\,‘..;u .

9"‘V‘ ‘\~f\ not“.
'..I* A y y¢

nuau rhly . bu.-

!\! T

r..\ J.

‘ t
«‘1 {ECE“‘ r15

0‘.

"twmw'
"'.. ‘ F.“ ‘
~~‘_ '4‘. h18 D'P!
‘ ~‘.
5

r" ‘I ‘ \
- hfi‘ \
s,“ . H h ‘
“‘ NU 5‘1 u "'GiA»
"’ \J

”-2., .‘i'. " ALI ;
‘ ““h «S Cr‘e '
‘

 

 

 

ABSTRACT

THE INCOME REDISTRIBUTIONAL EFFECTS OF
TAXES AND PUBLIC EXPENDITURES IN THAILAND:
AN INTERTEMPORAL STUDY

BY

Medhi Krongkaew

A recent rise in interest in the economics of income
distribution has prompted much research on income distribu-
tion in both developed and underdeveloped countries. This
study, which is one among many, is concerned with the sta-
tistical estimation of the income redistributional impact of
taxes and public expenditures in Thailand in, and between
1963 and 1969. Knowledge of this so-called "fiscal incidence"
is important in evaluating and recommending government poli-
cies regarding income redistributional issues in Thailand.

The study begins with the estimation of money income
distribution in 1963 and 1969 using the results of the 1962-
1963 household expenditure survey and the 1968-1969 socio-
economic survey as the main sources of data. The money in-
come concept is then adjusted for underreported income, im-
puted rents, income in kind, net corporate savings and in-
direct taxes. The new, more comprehensive income concept is
called "adjusted income," and both income concepts are used

in computing the effective rates of taxes and expenditures of

.1 rare fiscal 5';
........ms at e c;
:;;;::em which w;
resent: tax ar.

‘xzfi a?" Edi

" ifi wily

ure

((4.-

.

fiwv-‘van

.....__:-_ as ar ..
.

:.:;;‘-.‘._\ $ ‘ ' ~
"-‘vlcn 3.0, tnl b

 

 

\"\~ ‘
'-..~ r,
~ I
“in
nyet b3
L2)—
" ‘4‘
\: ‘\..'_ |
"f; t \
.‘L
85 Bra“!
' QVKG‘
"&
"°'~6d i.
lam“
”~U,.
. is
I» Ni,
V "\
.st A
var‘
‘ kl“? -
“ U‘la
Midi»
‘L n
KICFM
\ ‘ e b’r
‘ I
‘u n
h |
v Cr
4‘. ‘
‘& th‘
{‘q L‘

Medhi Krongkaew

the entire fiscal system. These money and adjusted income
distributions are called the "pre-fisc" income distribution,
the pattern which would result if the government replaced
the present tax and expenditure program with a proportional
tax and expenditure program. The tax incidence is then in—
terpreted as a reduction from, and the expenditure incidence
an addition to, this pre-fisc income.

The estimation of tax incidence by income classes in
1963 and 1969 is carried out conventionally, that is, first
certain shifting assumptions are decided upon, and, second,
the absolute tax burdens are allocated to different income
brackets according to certain allocation criteria. On the
expenditure side, the accounting approach, where total expendi—
ture benefits are assumed equal to total costs, is used to
value the benefits. Expenditure data by functional categories
disaggregated down to departmental levels help in the alloca-
tion of benefits to various income groups.

The incidence of taxation in Thailand expressed
through effective tax rates in 1963 and 1969 shows a generally
regressive pattern under a money income base, with the lowest
income bracket bearing the highest relative tax burdens and
the highest bracket the lowest relative tax burdens. Under
the adjusted income base, however, the tax incidence becomes
alumst proportional due to the fact that the income levels of
tflue lower income brackets under this income base are relatively

Inuch higher than the income levels under a money income base.

flaring the tax 5

‘ , ‘
b 1“.~.I-|O
....., L‘
.:;.': -6 9 Sins”

5:232:11; than the

- 7... ~ ‘1’
2:312:31 a‘...;CDt 1..

_::::'.::3. ‘ncier an
:.--: J. 1969' hla’fi'c

.‘A ‘l a
:-.:.-1' .333 thVSe

N
u»...
‘I‘QWQYPN ‘
witr' t:‘
" 3 ‘11:.
‘\I‘ W“. ‘
I‘Licn 1n
"“ "5:3“, ~ ‘
1 mail.
"a;

 

\..
:‘I‘ A:
' ‘1" AL‘
\u‘la‘~ .
o flees 1n
~L.
._: ‘

Pu -
f‘~. ‘
2‘: ~V~

“DAVIL
"l N.-
UquOV-
H
":5 y
‘\ D
“.9. v.
i ”Ste:
5

 

 

Medhi Krongkaew

Comparing the tax systems of 1963 and 1969, the latter ap-
pears to be slightly less regressive, thus less income de-
equalizing than the former. Expenditure incidence assumes
a pattern almost identical to tax incidence, that is, it is
generally regressive under a money income base and almost pro-
portional under an adjusted income base. The expenditure pro-
grams of 1969, however, appear to be much more regressive or
pro-poor than those of 1963, this enabling the net benefits
(expenditure benefits minus tax burdens) in 1969 to accrue to
lower-income households relatively more than the net benefits
in 1963. Therefore, while the net fiscal incidence pattern
in 1963 ranged from slightly regressive to slightly progres-
sive under money and adjusted income, respectively, the net
fiscal incidence in 1969 was regressive under both income
bases, indicating that fiscal programs in 1969 helped improve
the distribution of income of the Thai households, while fis-
cal programs in 1963 left the distribution of income in that
year more or less unchanged.

However, the degree of improvement in income equality
(or a reduction in income inequality) from 1969 to 1963 was
still very small. In terms of the Gini concentration ratio,
which is used extensively in this study to evaluate the de-
gree of changes in the patterns of pre-fisc, post-tax, post-
benefit, and post-fisc income distribution, the post-fisc in-
come distribution in 1969 improved by about 2.3 to 3.8 percent
over the post-fisc income distribution in 1963. The government

can effect this much improvement simply by imposing an extra

;.i:e::e:1t tax en t:

...' 2”er ,

.. ; u-
._...:.u utlflg tut

V... .- A a k 5". .
" " Y?,,V1'b’ hue i

no- 0".va bu
n

. -.
.'-:‘ :nA :1 r'-,-\ r
u ...¢..A.Au . - :1...‘ &l

”:3:““€flq ftat

"‘“""“UB-V .9"\-‘b ‘
u

. .4
\L‘I

F
i
u
I:
(D
*4
'1
(I)
O I?)
0
>8

. . "-~... ‘
"H 4 “A
u. .«utta tin :‘
. l
2’ _. v ‘
.n- -.~h R"- ‘ ‘
"‘ '1. ”A
"o Abd=LIAU¢M
'II- ...’
" \

9
H
. «A .
~u b a “h
VA "ublf
.:_I
°.. a“.
“a. Na ‘_ ‘
Uu‘ . Q ' P .
““- CE’l,~
‘
V
.k
I '

-.:' ‘v-_
' y L- A“

~.\... A“

. ~~ith O: ‘L‘

k
5‘1

..
"‘\~'-' A.

1- r3

 

 

Medhi Krongkaew

2.5 percent tax on the income of the highest income bracket
and distributing the tax revenue to the poorest income class.
More important, the above conclusion that the fiscal programs
in Thailand helped reduce income inequality is true only under
the assumption that a budget deficit, which was present in
both 1963 and 1969, adds directly to the income of households
without their suffering any reduction in real income through
possible price increases due to the existence of such a defi-
cit. When this assumption is changed so that the budget
deficit causes the price level to rise, which adversely af—
fects each household equally, then the fiscal programs in
both 1963 and 1969 could be shown to be income de-equalizing
instead of income equalizing.

In conclusion, depending on certain assumptions con-
cerning budget deficit, the Thai government could be shown to
have contributed very little or none at all to the reduction
in income inequality in the six years from 1963 to 1969. In
the presence of the still considerable income disparities,
policy changes are strongly advocated in the areas of basic

tax reforms and increased specific and transfer expenditures.

"2": “an RE
O

sun t‘VVow

RED FUEL]

 

 

 

 

THE INCOME REDISTRIBUTIONAL EFFECTS OF TAXES
AND PUBLIC EXPENDITURES IN THAILAND:

AN INTERTEMPORAL STUDY

BY

Medhi Krongkaew

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Economics

1975

©Copyright by
MEDHI KRONGKAEW
1975

-
0-.

To my Children, Bee and Ben,

and my Wife, Kanlaya

ii

I wish to

3’:°;..~ ;. v. k
u“ ‘_ d lvll-v
““: ‘ ‘ ‘ 3 ti.

“triifiim an': tn

u

:" u.- '
... r. ‘ + - ‘
-wir . chalca‘

..
~ "-g
'--.'~“ fiv-
-."-.u‘
' I

T - ‘ .
‘ an l“Gest

L.
AS I H
Etaa~*ed :1:

'=-.YE:S‘: I".
"3! t0 Dr .

 

- I ‘5 tall and AC
37.5: 1
u
3. ‘m 1 .
v Jr. Lily Y
54:: Mean “d
5 Se
.fffih
m-
D! £4Adm'zasé
:‘ In. a
. 4,915 ‘
3‘ fatil
Eight
'1',
”I SPECial
‘“*a:; .
Coilcan
‘4
. _, .-
.5 Drlginal -
I‘AU ‘
‘1 1’
s, ‘ .
"'~‘
9-.

ACKNOWLEDGMENTS

I wish to express my sincere gratitude to Professor
Milton C. Taylor, my thesis advisor, for his invaluable
advice, encouragement and guidance, kindness and under-
standing during the difficult time when this study was being
undertaken; and to Professors Warren J. Samuels, Byron W.
Brown, and Daniel H. Saks, members of my thesis committee,
for their technical assistance, constructive comments and
suggestions, and encouragement.

I am indebted to Thammasat University whose scholar-
ship has enabled me to pursue my study at Michigan State
University; to Dr. Puey Ungphakorn, Acharn Sangwian
Intravichai, and Acharn Montri Borisuth, who were instru-
mental in the establishment and operation of this scholar-
ship; to Dr. Lily Kosiyanon and Acharn Temchai Suwannathat,
present dean and secretary, respectively, of the Faculty of
Economics, Thammasat University, who readily made available
all kinds of facilities while I was doing my research in
Bangkok.

My special thanks go to Dr. Oey Astra Meesook, my
friend and colleague, who liberally and selflessly provided
numerous original data from her study for use in this study,
and to Dr. Ammar Siamwala who gave several helpful suggestions

and comments.

iii

Finally, I v
21:52:53, and moral

2:23:3‘ he time I

.. . ‘
:09;Q-0q“ - N a ’
uu.v.¢uuatel}’ qu-

 

Finally, I wish to thank my family for their love,
patience, and moral support both during the time I was with
them and the time I was away; To them, this work is

affectionately dedicated.

iv

LIES? TABLES . .

"u ..
‘ t

“'i .1

OI

*ISURES .

0"“ f‘ y-nm'r‘"
““P‘UDUki l.J-‘

Purpose 0
Interest
The Case

The Govern
Tax and 2..
Fiscal Inc'

Elites
Bureauc

ix;endi

O
A

99. w-

T
g

.-
~

.‘._.

. \
—

 

 

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . .

LIST OF FIGURES . . . . . . . . . . . . . . . .

Chapter

I.

II.

INTRODUCTION . . . . . . . . . . . . . .
Purpose of the Study . . . . . . . . .
Interest in Income Distribution . . . .
The Case for Economic Equality . . .
The Government as Income Redistributor
Tax and Expenditure Incidence Studies
Fiscal Incidence Study in Thailand . .

Elites . . . . . . . . . . . . . . .
Bureaucracy . . . . . . . . . . . . .
Expenditure Policies . . . . . . . .
Development Policy . . . . . . . .
Banking . . . . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . .
Procedure of the Study . . . . . . . .
General Procedure . . . . . . . . . .
Specific Procedure . . . . . . . . .
Limitations of the Study . . . . . . .
Plan of the Study . . . . . . . . . . .
THE DISTRIBUTION OF INCOME IN THAILAND:

1963 AND 1969 . . . . . . . . . . . . .

Conceptual and Statistical Problems in
Estimating the Income Distribution .
Past Studies of Income Distribution in
Thailand . . . . . . . . . . . . . .
Income Distribution in 1963 . . . . . .
Nature of Data and Method of Study
The Income Distribution Estimates . .
Income Distribution in 1969 . . . . . .
Nature of Data and Method of Study .
The Income Distribution Estimates . .
Comparison between the 1963 and 1969
Distributions of Income . . . . . . .

Page
ix

xiii

33

34

39
41
41
43
53
53
56

61

Incone Di
1963 an
Statistic
Inequal
AdjuStments
Personal
Income 1:

.2- 7\“Yv\—‘. ‘
““4 AJK, lainlgz

Theoretical
Studies
AGQI‘aral ‘

Tile kfialyS
Iniivid;
CDIPOra:
Business

SeleCt
3181::
5919::
lipart

 

I?
>1 m'EJ >1.)

'9) fl-

III.

IV.

Income Distribution by Income Classes,
1963 and 1969 . . . . . . . . .

Statistical Measures of Income

Inequality . . . . . . . . . . . .
Adjustments in the Income Concept . . . .
Personal Income . . . . . . . . . . . .
Income in Kind . . . . . . . . . .
Net Corporate Retained Earnings . .
Indirect Taxes . . . . . . .
The Adjusted Income Concept . .
Summary and Conclusion . . . .
THE INCIDENCE OF TAXES . . . . . . . . .
Theoretical Problems in Tax Incidence
Studies . . . . . . . . . . . . . . .
A General Methodology . . . . . . . .
Selection of Taxes . . . . . . . . .
Allocation of Tax Burdens . . . . . . .
The Analysis of Taxes . . . . . . . . . .
Individual Income Tax . . . . . . .
Corporate Income Tax . . . . . . .

Business Taxes and Stamp Duties;
Selective Sales Taxes; Import Duties
Business Taxes . . . . . . . . . . .
Selective Sales taxes . . . . . ._. .
Import Duties . . . . . . . . . . .

Export Taxes . . . . . . . . . . . .

Taxes on Property . . . . . . . . . . .

Royalties and Permits; Government Sales
and Services: Government Enterprises

and MonOpolies; and Other Miscellaneous

Revenues . . . . . . . . . . . . . .
Empirical Results of Tax Analysis . . .
Effective Tax Rates by Income Classes
The Effects of Tax Programs on the
Distribution of Income . . . . . .
Summary and Conclusion . . . . . . . .

THE INCIDENCE OF PUBLIC EXPENDITURES . . .

Theoretical Problems in Expenditure

Incidence Studies . .‘. . . . . . . . .

Money Flow versus Benefit Approach . .

Direct and Indirect Recipients of
Benefits . . . . . . . . . . . . .

The Measurement of Benefits . . . . .
Expenditure Beneficiaries and the
"Spendee" Concept . . . . . . . . .

vi

Page

62

66
72
73
77
79
81
82
89

99

100
103
104
109
112
112
114

117
119
119
120
120
122

123
126
129

140
145

152
153
154

155
157

161

A General
Select:
Allocat

Group I: I
Agricul
Power :.
Transp '
Other i.

Group 2:
Primar‘
Second:
Higner
Vocati

Serv:

GIOup 3;

Group 4;

 

 

 

A General Methodology . . . . . . . . . .
Selection of Public Expenditures . . .
Allocation of Public Expenditures

Group 1: Economic Services . . . . . . .
Agriculture . . . . . . . . .1. . . . .
Power and Fuels, and Industries . . . .
Transportation and Communication . . .
Other Economic Services . . . . . . .

Group 2: Educational Services . . . . .
Primary Education . . . . . . . . . . .
Secondary Education . . . . . . . . . .
Higher Education . . . . . .
Vocational and Other Educational

Services . . . . . . . . . . .

Group 3: Health and Social Welfare . . .
Group 4: General Services . . . . . . .

Defense . . . . . . . . . . . . . . . .
Law Enforcement . . . . . . . . . . .
Public Administration . . . . . .

Other Miscellaneous Expenditures
Group 5: Interest Payments on the

Public Debt . . . . . . . . .
Empirical Results of Expenditure
Analysis . . . . . . . . . . . . . .
Effective Expenditure Rates by Income
Classes . . . . . . . . . . . . . . .

The Effects of Expenditure Programs on
the Distribution of Income . . . .
Summary and Conclusion . . . . . . . .
V. THE NET FISCAL INCIDENCE . . . . . . .

Effective Fiscal Rates or Net Fiscal

Incidence by Income Classes . . . .
The Effects of Fiscal Activities on the

Distribution of Income . . . . . . .
Budget Deficit in Fiscal Incidence Studies.
Summary and Conclusion . . . . . . . . . .

VI. SUMMARY, CONCLUSIONS, AND POLICY
IMPLICATIONS . . . . . . . . . . . . . . .

Summary of Purpose and Procedures . . . . .
Conclusions . . . . . . . . . . . . . . . .
Tax Incidence . . . . . . . . . . . . . .
Expenditure Incidence . . . . . . . . . .
Net Fiscal Incidence . . . . . . . . . .
Policy Implications . . . . . . . . . . . .
On the Tax Side . . . . . . . . . . . . .
On the Expenditure Side . . . . . . . . .
Suggestions for Further Research . . .

Vii

Page

163
165
169
169
169
174
175
182
182
184
185
185

186
187
189
190
191
192
193

194
196
200

212
215

225

226

235
241
251

258

258
263
263
264
266
270
270
271
273

.u-‘qvt‘n
Staten... o I t O

‘.

..
“to“

. -
.uuuloa
-.

at:

:“N.
. . “ ’
u...) B‘DYY N“-

The Estimati
in 1963 .

Basic Data f
Income Dis

5' .
228 Estimat;

Distributiv:
Expenditu:

I‘— 2- .
me Act F15

Maren:

Ud‘ drlx‘h‘:

 

Page

APPENDICES . . . . . . . . . . . . . . . . . . . . . 277
Appendix
A. The Estimation of the Number of Households
in 1963 . . . . . . . . . . . . . . . . . . 277
B. Basic Data for the Estimation of the Money
Income Distribution in 1969 . . . . . . . . 280
C. The Estimation of Income in Kind in 1963 . . 286

D. Distributive Series for Use as Tax and
Expenditure Allocation Bases . . . . . . . 290

E. Disaggregated Tax Data . . . . . . . . . . . 299
F. Disaggregated Expenditure Data . . . . . . . 306

G. The Lorenz Functions of Various Distributions
Of Income 0 O I O O O O O O O O O O O O O O 312

H. The Net Fiscal Incidence Based on the Aaron-»
McGuire-Maital Method . . . . . . . . . . . 315

SELECTED BIBLIOGRAPHY . . . . . . . . . . . . . . . 322

viii

.
I: -
.L.C

by Inc
Househ

. .
8L1:

Distri

‘stribut

.0‘.

‘

House}

UlS‘v'

.H‘ C
r h.
t I
S
l

Table

2.1

LIST OF TABLES

Page

Percentage Distribution of Families

by Income Classes from the 1962—1963

Household Expenditure Survey . . . . . . . 45
Distribution of Money Income by Annual

Income Classes from Household

Expenditure Survey, 1962-1963 . . . . . . 46
Distribution of Total Families and Total

Family Money Income, by Income Classes,

1963 O O O O I O O O O O O O O O O O O O O 52
Distribution of Money Income of Urban

Households, by Income Classes, 1969 . . . 58
Distribution of Money Income of Rural

Households, by Income Classes, 1969 . . . 59
Distribution of Money Income of National

Households, by Income Classes, 1969 . . . 60
A Comparison between Distribution of

National Money Income in 1963 and 1969,

by Income Classes . . . . . . . . . . . . 63
Distribution of Adjusted Income, by Income

Elements, by Income Classes, 1969 . . . . 83
Distribution of Adjusted Income, by Income

Elements, by Income Classes, 1963 . . . . 84
Total Adjusted Taxes and Other Government

Revenues, by Type, 1963 and 1969 . . . . . 108
Distribution of Absolute Tax Burdens by

Income Classes, 1963 . . . . . . . . . . . 127
Distribution of Adsolute Tax Burdens by

Income Classes, 1969 . . . . . . . . . . . 128

ix

L4 Effective It
by Incone

Income, I

Effective Te

Income, t
Generalized
Percentage E

Post-Tax
Income C}

a . .
. .1.
(—1)

Pr€~Fisc and

Inequalit

Total Aij‘ds:
Thailand,

u U-GV‘A' ‘
X C..‘lture

 

CH!

‘1

‘ C
a,” M I

I,‘
~“E’C‘1'”e *—
B1,,

Elfin L

01y“ndlt

II My”

J x.

Table

3.4

3.5

3.7

3.8

3.9

3.10

Effective Tax Rates Based on Money Income,
by Income Classes, 1963 . . . . . . .

Effective Tax Rates Based on Money Income,
by Income Classes, 1969 . . .

Effective Tax Rates Based on Adjusted
Income, by Income Classes, 1963 . . .

Effective Tax Rates Based on Adjusted
Income, by Income Classes, 1969 . . .

Generalized Tax Incidence Functions

Percentage Distribution of Pre-Fisc and
Post-Tax Money and Adjusted Income by
Income Classes, 1963 and 1969 . . .

Pre-Fisc and Post-Tax Gini Concentration
Ratios and Percentage Changes in Income
Inequality, 1963 and 1969 . . . .

Total Adjusted Public Expenditures of
Thailand, by Functional Categories,
1963 and 1969 . . . . . . . . . . . .

Distribution of Highway Benefits Assigned
to Trucked Transport, by Allocation
Bases, 1963 and 1969 . . . . . . .

Expenditure Benefits, by Types of
Expenditures, by Income Classes, 1963

Expenditure Benefits, by Types of
Expenditures, by Income Classes, 1969

Effective Expenditure Rates, by Types of
Expenditures, by Income Classes, Based
on Money Income, 1963 . . . . . .

Effective Expenditure Rates, by Types of
Expenditures, by Income Classes, Based
on Money Income, 1969 ’. . . . . .

Effective Expenditure Rates, by Types of
Expenditures, by Income Classes, Based
on Adjusted Income, 1963 . . . . . . .

Page

130

131

132

133

139

141

143

168

181

198

199

201

202

203

. _
\L“

O
D

a
1...)

Effective E
Expendit
on Adjus

Generalizei

:3 Percentage

Post-Ben
by Incom

Pre-Fisc an
Ratios a
Inequali

Absolute 59

Effective F
Adjusted

Assumpti
amneralized

Percentaae
POSt'Fis
Income C
ASSUfipti

 

 

 

 

Table Page

4.8 Effective Expenditure Rates, by Types of
Expenditures, by Income Classes, Based
on Adjusted Income, 1969 . . . . . . . . . 204

4.9 Generalized Expenditure Incidence Functions . 210

4.10 Percentage Distribution of Pre-Fisc and
Post-Benefit Money and Adjusted Income,
by Income Classes, 1963 and 1969 . . . . . 213

4.11 Pre—Fisc and Post-Benefit Gini Concentration
Ratios and Percentage Changes in Income
Inequality, 1963 and 1969 . . . . . . . . 214

5.1 Absolute Net Fiscal Incidence, 1963 and
1969 . . . . . . . . . . . . . . . . . . . 227

5.2 Effective Fiscal Rates Based on Money and
Adjusted Income, 1963 and 1969:
Assumption I . . . . . . . . . . . . . . . 228

5.3 Generalized Net Fiscal Incidence Functions . 234

5.4 Percentage Distribution of Pre-Fisc and
Post-Fisc Money and Adjusted Income, by
Income Classes, 1963 and 1969:
Assumption I . . . . . . . . . . . . . . . 236

5.5 Pre-Fisc and Post-Fisc Gini Concentration
Ratios and Percentage Changes in Income
Inequality, 1963 and 1969: Assumption I . 238

5.6 Effective Fiscal Rates Based on Money and
Adjusted Income, 1963 and 1969:
Assumption II . . . . . . . . . . . . . . 246

5.7 Percentage Distribution of Pre-Fisc and
Post-Fisc Money and Adjusted Income, by
Income Classes, 1963 and 1969:
Assumption III . . . . . . . . . . . . . . 247

5.8 Pre-Fisc and Post-Fisc Gini Concentration
Ratios and Percentage Changes in Income
Inequality, 1963 and 1969: Assumption
III C C O O O O O O O O O O O O I O O O O 250
A.1 Estimated POpulation in 1963 . . . . . . . . 278

A.2 Estimated Number of Households in 1963 . . . 279

xi

I I
-
k)

{‘4 ,
tAJ

Number of H;
Areas, 15

Percentage I
Househol:

Percentage I
by IDCOI‘A

Percentage l
by Incoz.

Average Urb.
Classes,

Average Rur.
Classes,

km Average

 

 

 

 

Table Page

B.l Number of Households in Urban and Rural
Areas, 1969 . . . . . . . . . . . . . . . 281

B.2 Percentage Distribution of Urban and Rural
Households by Regions, 1969 . . . . . . . 281

3.3 Percentage Distribution of Urban Households
by Income Classes, 1969 . . . . . . . . . 282

8.4 Percentage Distribution of Rural Households
by Income Classes, 1969 . . . . . . . . . 283

B.5 Average Urban Annual Money Income, by Income
Classes, 1969 . . . . . . . . . . . . . . 284

B.6 Average Rural Annual Money Income, by Income
Classes, 1969 . . . . . . . . . . . . . . 285

C.1 The Average Income in Kind, by Locations, by
Income Classes, 1969 . . . . . . . . . . . 287

C.2 The Estimated Income in Kind Regression

Functions . . . . . . . . . . . . . . . . 288
C.3 Total Income in Kind, 1963 . . . . . . . . . 289
0.1 Distributive Series, 1963 . . . . . . . . . . 291
D.2 Distributive Series, 1969 . . . . . . . . . . 292
D.3 Total Household Expenditures, 1963 . . . . . 293
D.4 Total Household Expenditures, 1969 . . . . . 294

E.1 Disaggregated Taxes and Other Revenues,
1963 and 1969 . . . . . . . . . . . . . . 299

F.1 Disaggregated Public Expenditures, 1963
and 1969 . . . . . . . . . . . . . . . . . 306

G.l Lorenz Functions of Various Distributions
of Income, 1963 . . . . . . . . . . . . . 313

G.2 Lorenz Functions of Various Distributions
of Income, 1969 . . . . . . . . . . . . . 314

H.l Comparison of the Post-Fisc Distribution of

Income Based on Aaron-McGuire-Maital
Method and Methods Used in This Study . . 320

xii

al!
N)

Distributi;
1963 an.
EffECtiVe '

by Inco:

by Inca:
7.“: ‘.
“*~eCt‘Ve 1

Income,

Effective 1

Income,
Effecti‘re I
by ITICQf

n:-

u«:ectiY'Ve I

 

Figure

3.1

3.2

4.2

5.1

5.2

LIST OF FIGURES

Distribution of Money Income in Thailand,
1963 and 1969 . . . . .

A Hypothetical Lorenz Curve . . . . . . . . .

Distribution of Adjusted Income in Thailand,
1963 and 1969 . . . . . . . . . . . . .

Effective Tax Rates Based on Money Income,
by Income Classes, 1963 and 1969 .

Effective Tax Rates Based on Adjusted Income,
by Income Classes, 1963 and 1969 . . .

Effective Expenditure Rates Based on Money
Income, by Income Classes, 1963 and 1969

Effective Expenditure Rates Based on Adjusted
Income, by Income Classes, 1963 and 1969

Effective Fiscal Rates Based on Money Income,
by Income Classes, 1963 and 1969 .

Effective Fiscal Rates Based on Adjusted
Income, by Income Classes, 1963 and 1969 .

xiii

Page

65

69

85

136

137

207

208

231

232

3345 known as a
I::I: .A‘ ' . 1 ‘ : '
.uv:_:~n “1&1 c ‘

' z."fl~ D“
r“~.d' e.
4

'rhaSis

0- I.‘.; ‘
~., '..‘:: g0".°V:-1:‘ ‘
': l‘icg‘c “Ah/W
a-~ a.n'\";“ J
:Z::‘n ‘
'i.‘-I§b‘res' es:
t
h. '
‘v” ‘G
l a the fact
VE‘:\‘ ‘
1‘: fire Gena}
"it in.
t \.reases
:- ‘ .
‘ srCSPEr
\z.
«II ‘1 C.t .
5‘ . ‘
“ ls ¥}
:‘I‘2‘E‘H‘mA
"‘5'“ v
I 90 Jer:
:Q
.‘ ‘ .
“swig?

CHAPTER I

INTRODUCTION

Purpose of the Study

 

The purpose of this study is to estimate the income
redistributional impact of the government tax and expendi-
ture policies in Thailand in 1963 and 1969 and between these
two years. In the economics of public finance, this is basi-
cally known as a fiscal incidence study. However, the present
research will differ from the conventional incidence study by
placing emphasis more upon the interyear income redistribu-
tional effects than upon the effects of government fiscal ac-
tivities within a single year.

There is a need to understand more about the extent
to which governments have helped equalize the distribution
of income among individuals or households through taxes and
expenditures, especially in underdeveloped countries. In
view of the fact that income distribution obviously has not
become more equal in many developing countries, despite ap-
parent increases in the national income and other indices of
economic prosperity, certain questions have been asked: To
what extent is the public sector responsible? Is there a
deliberate government policy not to improve the distribution

of income? Or is the government merely disinterested in or

"=Cti'Je concerni

5:22": will attexp

1:21 governmen

Before pr:
grassnt research,
Lrationale of ‘
tires: in income <
3:11 fecus on the
;::::e equality oz

3 3115 issue,

In recent

 

 

inactive concerning the income distribution issues? This
study will attempt to answer these questions in the context
cfi'Thai government policy in and between the years 1963 and
1969.

Before proceeding to the scope and procedure of the
present research, this chapter will discuss the background
and rationale of the study. Following an account of the in-
terest in income distribution in recent years, discussion
will focus on the case for action being taken in regard to
income equality or redistribution and the role of government

in this issue.

Interest in Income Distribution

In recent years there has been a tremendous increase
in interest in the economics of income distribution. This
Shift from a preoccupation with resource allocation and
9T0Wth is quite timely because the problem of persistent and
W'OISening inequality in the distribution income throughout
the'world has begun to cry for attention. An understanding
ofthe nature of income distribution, the causes of maldis-
tribution, and possible solutions is sorely needed.

Unlike the study of functional income distribution
orthe distribution of factor shares, the systematic study
C” Personal distribution of income by size has a relatively
recent history. Work began in the 19405 with a few studies
in England and the United States. Soon after, an interest

inunderdeveloped economies began to emerge. The year 1954

gritzei one of the
::::e distributi
£52331 Address t
51:: Kuznets exp
i;s::i':ution in d
3:3;i0neerin-g st
‘; Liferdevelopcd

‘u.

-.-r. that in deve

N. .....

..;...:s in income
etczczic developrfi

Interest L
13535, but re-emez
153‘ Era, the Worlr

x; .n the ecm

""33351‘] altere

7.11.5: interest

 

)1 1972
The st:

 

marked one of the earliest attempts by an economist to study
imxmm distribution in underdeveloped nations. In his Presi-
dentfifl.Address to the American Economic Association in 1954,
Shmanuznets expressed his concern about the state of income
cfistribution in developed as well as underdeveloped countries.
Ifis pioneering study, which concluded that income distribution
blunderdeveloped nations was less equal (or more unequal)
than that in developed countries, paved the way for further
research.by Irving Kravis in 1960 and Harry Oshima in 1962.2
B11963 Kuznets published another study dealing with the
Changes in income distribution through historical phases of
economic development, which is now a classic in the field.3
Interest in income distribution lagged in the late

19608. but re-emerged in the early 19705. As if heralding a
new era, the World Bank, which previously had been concerned
onlY‘mith the economic growth aspect of poor nations, con-
SEficuously altered its attitude and policy and manifested
greater interest in more equitable income distribution among
'thOSe<mountries. This change was reflected in various speech-
es by RObert McNamara, President of the World Bank. In the
Years 1971, 1972 and 1973 McNamara made the following remarks:

The state of development throughout most

of the developing world is unacceptable...

because hundreds of millions of people are

living at levels of deprivation that simply

cannot be reconciled with any rational defi-

nition of human decency....Current develop—

ment programs are seriously inadequate be-

cause they are not significantly reducing

the poverty which shapes and limits these

lives. And though the matter is complex,
basically we know why...the developing

count
to IE
ineql

Polic
eCCHC
trieS
perCE
tion
has t
set t

When

despa
9a? t
imprc
befor
twee:
polit

These 8
._.eztly by a 9‘:
.r. :r.:e:e distr
:ne2een carri‘
.3' :h‘a .

.. “ternatlona

D
;:er:ati

onal Le

 

countries are not moving decisively enough

to reduce the severe social and economic

inequalities among their own people.4

Policies aimed primarily at accelerating

economic growth, in most developing coun-

tries, have benefited mainly the upper 40

percent of the population and the alloca-

tion of public services and investment funds

has tended to strengthen rather than to off-

set this trend.5

When the highly privileged are few and the

desparately poor are many...and when the

gap between them is worsening rather than

improving...it is only a question of time

before a decisive choice must be made be-

tween the political costs of reform and the

political risks of rebellion.6

These sentiments have been echoed more and more fre-

<Illently’byagrowing number of economists,7 and more studies
(nxinccnne distribution, especially in less developed countries,
lmvekmeen carried out.8 Apart from the World Bank, other ma-
jorzhrternational or national organizations, among them the
Intemuational Labor Organization (ILO), the U. S. Agency for
Infiernational Development (AID), the Brookings Institution,
and the Economic Commission for Asia and the Far East (ECAFE),
either have their own research programs on income distribution,
Support such studies, or are actively involved in actual in-
mem distributional policy programs. In a word, what began
as a tli‘ickle in the 19503 and 19608 has becomea flood in

the 19705.

The Case for Economic Equality
To some, it may appear that any effort to make the
dlStribution of income and wealth more equal is intuitively

£3.900d thing, but this is by no means a universal sentiment.

:fact, the i7“
53115280 thrOUSh"
The pri
52:1"; and David
faras such ine‘
":35 by labor-t.
at: work longer
_::::'it. Karl M.
is value was a
£15 not earn by
Lsais to inequa]

.15 em theory

‘ u
7:“!

3~Larshall

h
.
‘m.

 

 

 

In fact, the income equality versus inequality argument has
existed throughout the history of economic thought.

The pure labor theory of value, as formulated by Adam
Smith and David Ricardo, justified inequality of income inso-
far as such inequality reflects differences in efforts meas-
ured by labor-time used in production. In other words, those
who work longer and harder would receive a higher reward or
Profit. Karl Marx, however, argued that the profit or sur-
Plus value was an unearned increment which the capitalist
did not earn by his own labor. A profit economy, therefore,
leads to inequalities which are unfair by the standard of
the labor theory of value.9 The marginal utility theory of
Alfred Marshall and A. C. Pigou, which assumed the cardinal
measurement of utility, diminishing marginal utility and
interpersonal comparison of utility, justified progressive
taxation of the wealthy for redistributional purposes. The
reasoning was that an adverse effect on the wealthy would be
less than the positive benefits accruing to the poor. But
this theory suffered a setback when most of its assumptions
came under heavy attack from the new welfare economics.

The marginalist or neoclassical value theory has re-
newed the justification for economic inequality. According
to this theory, if each person's income in a competitive mar-
ket is equal to his marginal value product, then his income
pOSj-tiOn is a result of .his "productive contribution." How—
ever, Strong opposition also has been voiced to this theory

by Various groups, particularly Marxists and radical economists.

7222;: argue that .
if :hey did, pro‘
rarginal product
232 that indivi.
i;::i".-'e contribu

Purely h

:3: the major co.

:-
(J

.rgreater econ.
meson-mic grO'
gilztical struct‘
_:::::iple that a
$3 f-iatt'new Arnol:

"-
I...

513e, ineCua:

 

 

‘-

"o

..:‘y.',

“ 1‘

‘."'|"‘.
“~le d1
ue

.1
2. :‘

“i

 

They argue that competitive conditions do not exist, and even
if they did, production techniques could be varied so that the
marginal product moves to haphazardly that no claim could be

made that individual income is a reward for individual pro-

ductive contribution . 10

Purely historical or theoretical issues, however, are
not the major concern here. Suffice it to say that the case
for greater economic equality can be supported on both ethical
and economic grounds. Inequality in the economic, social, or
political structure would, to varying degrees, violate the
PrinCiple that all men are entitled to equal opportunity.

As Matthew Arnold said at the turn of the century, "on the
one side, inequality harms by pampering; on the other by vul-

garizing and depressing. A system founded on it is against

nature, and in the long run, breaks down."11 R. H. Tawney

also has explored the danger of the "religion of inequality"
associated with the distributive concepts of laissez-faire

in his thoughtful book Equality.12

There are, of course, those who believe that measures
to reduce inequality in market economies could have numerous
undeSirable effects.l3 H. G. Johnson, for example, maintains
that the ethically motivated social concern about inequality
Should properly focus on inequalities of opportunity and on
develoPing the knowledge and resources required to equalize
them°l4 Milton Friedman believes that income inequality is
inevitable due to different rewards for risk-taking and per-

Sonal Preferences.l5 John Rawls is more sympathetic to the

32216121. In h:
fairness, he a]
5.331 in their
Eds,social a
:22th in conpe
articular the
More in
2:: concerning
least two point
1515 fear that
:i:.‘. of income

”You. ‘
1

,L-tn. They a
53 ;r:gressive
:L‘thries will
333.38, Which i:

.4. ~' 17
“Is, :‘C oth.

 

problem. In his recent work on the concept of justice as
fairness, he argues that although men cannot be perfectly
equal in their income and wealth, at least everyone must have
a strictly equal assignment of basic rights and duties. To
Rawls, social and economic inequalities are just only if they
result in compensating benefits to those less fortunate, in
particular the least advantaged members of society.16

More immediately relevant to this study is the argu-
ment concerning economic growth versus economic equity. At
least two points are at issue. First, many development econo-
mists fear that a policy which fosters a more equal distribu-
tion of income will conflict with the process of economic
growth. They are concerned that such redistributive measures
as Progressive taxes on income and wealth in less developed
countries will stifle the savings propensity of high income
groups, which in turn will reduce capital formation and econo-
mic growth.17 This need not always occur, however, since sav-
ings Could come from the public sector as well as other unin-
corPOI'ated household sectors. It also can be argued, as do
W. Paul Strassmann and W. G. Tyler, among others, that under
certain conditions economic growth will depend on a more
rather than less equitable distribution of income.18 Empiri-
cally, W. R. Cline has shown that in many Latin American
countries, although an .income redistribution policy may have

depreseing effects on savings and growth of the economy, such

effECt—S are small and do not do irreparable damage to economic

-0» On thE

.._'b¢‘.

ajfhture incc
3’le: grOWth 53‘:
:th he popul

22:1:er be b

(J I

D
o
f
(D
H
t 0
1
0.

i;s:::bution Wi‘

The sec:
:2: :~ concernec
Etical’iy will t

..'.:‘e:~' Kw h
w; uanLS

:~.-::::y moves f r

.
u.‘

 

 

growth. On the contrary, the gains in terms of immediate
and future income of the majority of people more than offset
such growth sacrifice. Cline believes that the large majori-
ty of the population in Argentina, Brazil and Mexico would
definitely be better off if the strategy of redistribution
with slower growth were chosen over the option of a skewed
distribution with somewhat faster growth.19

The second point involves the argument that one need
not be concerned about economic inequality because it auto-
matically will be reduced when the economy is more developed.
Indeed, Kuznets has convincingly demonstrated that as an
economy moves from the underdeveloped to the developed stage,
inequality first worsens and then improves as the economy at-
tains a higher level of development.20 This conclusion seems
to be historically true, but it is barren of practical mean-
ing- It does not indicate how long the worsening in inequali—
ty will continue or when it will end for each particular
economy; it does not address the problem of whether the ma-
jority of the population which suffers will tolerate this
Widegjpg income gap.21 Many economists believe that income
distribution need not conflict with economic growth and

ShOllld not be given a secondary role.22

Elie Government as Income Redistributor

In a world where the distribution of income is exces-

s ively unequal, how may such inequality be reduced? The various

grants of SOClé
heron-can, o!
atong thense
rater, this is
_:::':ably will be

at. the authorit

:1 well-known th

1

.n“
‘V

”have, the rol

U \
:"‘o-h‘l f
“‘Ob‘~
.

.y (and 65
32:: fiscal tools
.1 also would ens
exalted fully 0
ill be aided thr
Blot all g

:i.‘ " ‘
-.:-.1’12‘itl'-Je fur:

. t}
I
.N ‘
“‘Nslness
. The

 

segments of society--the rich, the middle income strata, and
the poor--can, of course, voluntarily redistribute their in—
come among themselves. In the context of modern society,
however, this is rather unlikely. The redistributive task
probably will, be left with the government, which is vested
with the authority to govern and to regulate. According to
the well-known theory of public finance propounded by R. A.
Musgrave, the role of government may be divided into three
functions: resource allocation, economic stabilization and
income distribution.23 The government performing this latter
role not only will see to it that the economy's output is
equitably (and efficiently) distributed (with the help of
such fiscal tools as taxation and expenditure policies), but
it also would ensure that the least advantaged, who have not
benefited fully or adequately from the distribution process,
will be aided through income transfers.

Not all governments can be expected to emphasize this
distributive function as much as they do the allocative and
Stalbilization roles. The classical arguments in the tradi-
tion of Adam Smith and his followers emphasized the defense
of the nation, the maintenance of law and order, the regula—

t ‘ . . .
1011 of business conduct, and the provision of public utili-

t - . ' -
les - The view of government's role increased substantially
w-
1th the introduction of Keynesian economics, as it was shown
th .
at the public sector is the best agent for controlling and

ma .
liaging aggregate demand and its resulting income and em-

Pl
QX’Inent problems. Under modern capitalism, the nature and

212 sf governnzei
:azitulistic sta‘
:2: the governm

ease of consu:

24 .
the groups. .

tabssic functic
1:51; in the accu:
thir-aze" it by :

2:25 and service

‘‘‘‘‘

The L’nitec

Lida“ A n
.2..:..€d the f 1y

5
/
.,. u

_ 5
In the 16

”A.
’Q

m. is Very muCr

1
~

ECSt poor na
......;:h power is
..... a structure
MLIOHEd by, ar
{much
,.grossly

. u
p

“‘tli‘kl '

 

‘3 w”~“'1‘tries,

"‘jilkgl I
L

“ Y to

v.

 

‘33:." p
. VP-
«VCeSS
.
VDQ

10

role of government has changed considerably. In an advanced
capitalistic state such as the United States, it is often said
that the government favors the modern corporate sector at the
expense of consumers and workers, or the lower to middle in-
come groups.24 J. A. O'Connor, for example, argues that the
two basic functions of the capitalistic government are to
help in the accumulation of private capital and then to "le-
gitimize" it by incurring social expenses in the form of pro-
jects and services which are required to maintain social har-
mony. The United States welfare system exemplifies programs
designed chiefly to keep social peace among unemployed work-
ers.25 In the less developed capitalistic economy, the situa—
tion is very much the same. As T. E. Weisskopf has pointed
out, most poor nations are characterized by a class structure
in Which power is highly concentrated among a small elite;
Such a structure in turn results in a state apparatus largely
controlled by, and responsive to, the interests of that elite.26
Although grossly oversimplified, this short description is
undeniably accurate in regard to many, if not most, develop-
1119 countries, including Thailand.
When the government tends to serve the interests of
a pCDWerful minority group or class, redistribution of income
is unlikely to be a majoropolicy' of the country, if indeed
there is such a policy at all. The government in such a set—
ting usually facilitates the maximization of profits and the

9 . .
rowth process in the modern corporate and industrial sectors

a
nd hopes that the increased wealth and income will "trickle

53:- to those 18
5::etizes govern?
axe: ly to keep
Leas: docile. Th
size accumulati
:22‘52l3n, leaves
s:::c:ic deprivat
3:1: developed an
3L:.a:ion continu
2 have to be u.“
Farrah Galbraith

i:uoao"n .
‘nuw‘y Interest

.3. .27

"fit—I

 

Bf? 7‘)
eo.e y“

I‘m-Ia: at iss‘de iE

 

 

11

down" to those less fortunate in the lower income brackets.
Sometimes governments adopt a minimal transfer policy design-
ed merely to keep the underprivileged mass content or at
least docile. The rush to produce and to grow, while it helps
in the accumulation of wealth, profits, and technological in—
novation, leaves a trail of problems, among them poverty and
economic deprivation. The same situation applies equally for
both developed and underdeveloped economies. If the above
situation continues, a new policy or a new course of action
may have to be undertaken along the lines suggested by John
Kenneth Galbraith: The state may have to be freed from the
economic interest and control of the business "planning sys—
tem."27
Before proceeding, it must be re—emphasized that the
point at issue is not the government's capability to redistri-
bute income, but either its inefficiency or its unwillingness
to do so. Therefore, changing either the government's admin—
istration or its policies is one of the most important first
steps toward a more equitable distribution of income. There
are at least two schools of thought concerning how the above
should be accomplished. The radical school favors a complete
uprooting of the existing capitalistic structure and its re-
placement with a new economic organization which has greater
concern for equality of income and wealth. The liberal school,
on the other hand, believes that it is possible to change the
system or the policy in a gradual, less fundamental way. Al-

though a government may be influenced or controlled by a small,

healthy elite,
Lined that it
with both the i
iiszributive po
:1: the rajority
rugs of the in

:e liberal em:

I." Ar:

=:; Change .

' o
" ~n~’~
' ‘1‘.
A ~'

" “““ :9 59817.8

‘ 9
‘v-R D
:‘VyObiJ.

"“5 in at least

a.‘ ‘A
"heft taxes f r

a, ._
#4.}

1 Sizes in su

. .

lay,
“"ner .‘
' 5

'he taxes

.fl'fi ‘ '
5:11:5510“ .

Secor

)-
‘-"I‘:‘"‘g
~I
i“‘¥

5' Spend

1..
~ "‘ V .
dwfi'n

“mtures fa”

 

 

12

wealthy elite, this stranglehold is not permanent. It is be-
lieved that it is possible to have a democratic government
with both the intention and power to adopt deliberate re-
distributive policies. After all, this government is elected
by the majority of the people, who are themselves on the lower
rungs of the income distribution ladder, Both the radical and
the liberal envision the same goal, but differ as to the strat-
egy of change. And unless proven ineffective, the liberal way
of change seems an attractive first choice to the existing

problem.

Tax and Expenditure Incidence Studies

 

Assuming that a government is interested, willing, and
able to use its fiscal power to redistribute income, it can do
this in at least two direct ways. First, it can impose and
collect taxes from individuals and households of various in-
come sizes in such a way that the higher the income, the
higher the taxes paid. This method is known as progressive
taxation. Second, the government can, with part of its tax
revenues, spend money in such a way that the benefits of such
expenditures fall on the lower income brackets relatively more
than on the higher income groups. These are known as regres-
sive or "pro-poor" expenditures.28

In an effort to learn more about the impact of govern-
ment fiscal policies on the size distribution of income, public

finance economists have conducted what are called tax and ex-

penditure incidence studies. These empirical studies quantify

he: the burde:
marious inco:
rid nce assum}
iifficulties, m<
fies researchers
aria: tax and e:
1552: ace of e1
*‘ a fiscal
arc-re complete

.

...: " '

..:::'.nut1ve l2:
L

if: rascal syster

..E'«':::.er, Helen

 

 

 

 

13

where the burden of taxes and benefits of government spending
to various income classes fall under certain tax and benefit
incidence assumptions. Due to conceptual and statistical
difficulties, most research is done only on the tax side. A
few researchers have attempted to do both, that is, to con-
sider tax and expenditure incidence together and net out the
difference of effects between the two across income classes.
Termed a fiscal or budget incidence study, this is obviously
a more complete and satisfactory way to learn about the re-
distributive impact of government policies.

Among early researchers into tax incidence in develop-
ed fiscal systems the following are most noteworthy: Mabel
Newcomer, Helen Tarasov, R. A. Musgrave and Associates, and
R. S. Tucker on the United States tax system,29 and Tibor
Barna, Findley Weaver, and A. M. Cartter on the British tax
system.30 The general conclusions drawn from their work were
that the tax systems were generally progressive, thus redis-
tributive in the desired direction (from the rich to the poor).
During the 19605, incidence studies in the United States be-
came more widespread, and interest also shifted from tax in-
cidence alone to complete fiscal incidence. This emphasis
is evident in the work of O. H. Brownlee, W. Irwin Gillespie,
N. M. Singer and the Tax Foundation.31 Again, the net in-
cidence result was found to be generally regressive or pro-
poor.

In the area of less-developed fiscal systems, a few

isolated studies were conducted in such Latin American countries

as Guatemala,

;: was not unt

tries have bec
:51 problems.
:pcrtant and
txnand analy
iadoping cou
gerditure inci

II.

.- been mixe
Letics of each
1515: and on t
While

3,4,, '

'"fi 15 an it?
:~‘-'EII‘.TLent See.
9.",- l u
are, this is i
""‘ sUC De “ml;
33%;:
mfirs of i'

:v4._
I ‘ N
“at

critiCS_

 

h.
C‘ "In."
sJ-O‘y o
“5 lhc
:L
“In H».

14

as Guatemala, El Salvador and Venezuela in the 19505,32 but

it was not until the mid-19605 that interest became wide-
spread. By the l9705, fiscal incidence studies in poor coun-
tries have become quite popular despite formidable statisti-
cal problems. Perhaps the income distribution issue is too
important and urgent to wait for improvement in data collec-
tion and analysis. In the past ten years, no less than 20
developing countries have been the subject of tax and/or ex-
penditure incidence case studies.33 The results, however,
have been mixed and inconclusive, depending on the character—
istics of each country, on the availability and quality of
data, and on the assumptions and techniques used in the studies.
While it is generally believed that a fiscal incidence
study is an important first step in evaluating the role of the
government sector in the matter of income redistribution which
may lead to policies for a more equitable distribution of in-
come, this is by no means generally agreed upon. R. M. Bird
and Luc De Wulf, for example, while among the most prolific
compilers of incidence studies in poor countries, are also
ardent critics. Regarding tax incidence studies, Bird and
De Wulf doubt the meaningfulness of the exercise and go so
far as to say that the usual statistical calculations of tax
burdens are virtually without merit as a basis for policy
formation. On the expenditure incidence side, they question
the validity of conventional approaches in allocating benefits
to various income groups. All in all, they are skeptical

about the usefulness of budget incidence studies for policy

clementation.

Obvious
:2: he myriad
tired, but the
;:ssible applic
::Lntries Where
515 and poses

:
‘fl-Q ‘

.-...., .ne least
72.2;129 nation
stated in quant

:."'.L" a
I. ‘ ”Neil

g: thESe

5331?? device,

:I‘; .

 

 

 

15

implementation.34

Obviously, Bird and De Wulf are correct in pointing
out the myriad of conceptual and statistical problems in-
volved, but they possibly are incorrect in prejudging the
possible application of careful and conscientious studies in
countries where grossly unequal distribution of income ex-
ists and poses an economic, social, or political threat.
Among the least of the problems faced by researchers in de-
veloping nations is the overconfidence bred by having results

35 If

stated in quantitative terms, as De Wulf has charged.
anything, these results should be used only as another per-
suasive device, for many politicians require considerable
evidence on certain policies. If the results from a fiscal
incidence study could be of some use in making important de-
cisions in an economy where the redistribution of income is
an important objective, such a study should be encouraged.
Despite its many shortcomings, an incidence study is still
useful, as Musgrave has noted:

This state of affairs [a tax incidence

study] is far from perfect and subject

to much improvement. Yet no apology is

required. This kind of analysis is need-

ed for the simple reason that distribu-

tional considerations are and should be

an important factor in tax policy and that

the economist's informed guess, based on

explicit and reasoned hypothesis, is to

be preferred...to the implicit and hap- 36

hazard assumptions of the practical man.

In summary, the income distribution issue is important,

and a government can use its tax and expenditure policies to

improve a maldistribution of income. The way to estimate the

3331: of St

.221: malca

izne its par

Fi
This
.I tax burde:

i..-.o:.e clg

3.5 of this 1'

L5": conside

Effie, this 5'

EN ‘ . ‘
‘Nsd’lbutlo.
‘

:th35ek

1 1
“Gigs-

"“5 grO‘Jps
5:" l”).
'~“-:J€Ctec t
A: S
.cb3. G" ‘
‘!‘era :
‘
«-
“.:' .L‘N! ‘
c~ u‘uyot‘plesi
:1
“~01:

. l
U“U._;‘I\
t» Struct
t‘fi
"“ C‘l
ad a
n e
“573$:
u
\;.(“
“$.3 L;
d huat an”
.‘J
‘Q
‘QtEr

l6

impact of such policies is through fiscal incidence studies,
which indicate whether and to what extent the government has

done its part in redistributing income and in which direction.

Fiscal Incidence Study in Thailand

This study is expected to yield quantitative estimates
of tax burdens and expenditure benefits to Thai households,
by income classes in 1963 and 1969. Although the main empha-
sis of this investigation is empirical, it is also within the
scope of this research to hypothesize future developments.
After considering many general facts and much supporting evi-
dence, this study hypothesizes that the fiscal activities of
the Thai government in 1963 and 1969 probably had few desirable
redistributional effects upon the existing income distribution
of households. In other words, the income positions of lower
income groups relative to those of higher income groups are
not expected to improve as a result of government fiscal poli-
cies. Several tentative reasons may be advanced to support

this hypothesis.

Elites

As in many other underdeveloped countries, the socio-
economic structure of Thailand exhibits what T. E. Weiskopf
would call an elite-centered system--a small elite group in-
fluences or controls government policy. Therefore, it is un—
likely that any pro-poor redistributive policy will be allowed
to materialize or to cause real deterioration in the income

positions of the upper-income, or elite groups.

 

Thailar
are and ineffj
and its adminis
:zindirect ta):
are relatively
51:23: tax law
:::;‘.iance and
2:55. for its 1.
Tzerefore, it 53
3:33: to fall re

2: "7+

' “‘»Ua

 

l7

Bureaucracy

 

Thailand's bureaucratic system is generally regres-
sive and inefficient. This is also true of the tax system
and its administration. The tax system relies excessively
on indirect taxes as a major source of revenue because these
are relatively easy to collect. Enforcement of the existing
direct tax law is weak, giving rise to widespread non-
compliance and tax evasion. The Thai tax system is also
noted for its lack of property, estate or inheritance taxes.
Therefore, it should not be surprising if the tax burden is
shown to fall relatively more heavily on the poor than on

the rich.

Expenditure Policies

 

The expenditure policies of Thailand can be described
as either unplanned, uncoordinated or haphazard. Moreover,
expenditures aimed at improving the social welfare of the poor
comprise only a small portion of the total budget. Coupled
with inefficient administration, the benefits from these ex-
penditures will prove to be less redistributive than they

should be, and a portion is simply wasted ineffectually.

Development Policy

 

The government's development policy is decidedly pro-
growth, with little or no concern being given to equity. For
example, in the early 19603 the government launched an indus-
trial promotion campaign aimed at encouraging foreign as well

as domestic investment. The incentives were overwhelmingly

attractive:
xrtduties <
firforeign 1

Lab-c: legisla

a . l,
2: isomer ' s

“4 Product

 

18

attractive: no tax on profits for five years, reduced im-
port duties on raw materials, certain immigration privileges
for foreign personnel, and so on. Yet, during this period
labor legislation regulating minimum wages, hours of work,
other employment benefits and compensation was virtually non—
existent. While the promoted firms experienced rapid success,

the worker's income position was only slightly improved.

Banking

In areas in which the government should exert its in-
fluence, it does not do so. The banking system is an example.
While one of the most important factors in promoting agricul—
tural production in Thailand is the availability of agricul-
tural credit, this is, ironically, one of the areas most neg-
lected by the government. The Bank of Agriculture and Agri-
cultural Cooperatives, the only government-owned bank dealing
directly with agricultural credits, is much too small. Pri—
vate commercial banks are uninterested in giving agricultural
credits; in 1963 and 1969 for example, the percentage of loans
to the agricultural sector by commercial banks was only 3.9
percent and 2.8 percent of total loans, respectively.37 The
government could have imposed a rule that a certain percentage

of such loans must be channelled to the agricultural sector,

but this was never done.

Summary

Although the above brief discussion does not offer

exhaustive explanations as to why the role of the government

regarding imprc
lazd can be ex;
’5 lead credem
Etpirical resul

“serif? or ref ut

6

Zereral Procedu
The usu
Stu-dies can be
estmte of inc
:3: be obtaine
itch the tax a
Second,

3"; a 1

‘ “*lOCatinc f

Y
3%
.‘~

ses that c,
71)
1. 1r;

V.uer to a:

m .
ihlrd’
«7
an

Elba ' .
~13 51m:

 

title L
in; Cla SS E :

C, .
“Cllri
‘
‘

Es are

 

l9

regarding improvement in the distribution of income in Thai-
land can be expected to be negligible, the information tends
to lend credence to the hypothesis that has been advanced.
Empirical results presented later in this study will either

verify or refute the hypothesis.

Procedure of the Study

 

The procedure used in this study is similar to that
followed in other fiscal incidence studies in most respects.

This general procedure is outlined below.

General Procedure

 

The usual procedure followed in most fiscal incidence
studies can be viewed as a five-stage process. First, an
estimate of income distribution of individuals or households
must be obtained or developed for use as the income base upon
which the tax and expenditure incidence is computed.

Second, the tax incidence by income class is estimated
by allocating the actual tax burdens to the different income
classes that contributed such revenues to the government.

This necessarily involves certain tax shifting assumptions
in order to allocate the tax burdens.

Third, the estimation procedure for expenditure inci-
dence is similar to that for tax incidence. Actual expendi-
tures are allocated to actual or expected beneficiaries by
income classes. In many expenditure categories, the bene-
ficiaries are specifically identifiable; in others, especially

those of a public good nature, identifying the beneficiaries

computing
Less some simp
Fourth
benefits (that
23.. ‘3 .come clas
:;::t '.e incor.
if taxes and e
called effecti
Ripective 1"
Fifth,
33-1 teat is th
:5 EXPGUditur.
33.15 rate will
.5 a net gaine.

.3,“
governmen‘:

TEN-4',“ .

 

20

and computing the benefits received by them is impossible un-
less some simplifying assumptions are made.

Fourth, the absolute tax burdens and expenditure
benefits (that is, the burdens and benefits in money terms)
by income classes are then expressed as percentages of re-
spective income levels to arrive at the relative incidence
of taxes and expenditures by income classes. These are often
called effective tax rates and effective expenditure rates,
respectively.

Fifth, the net fiscal incidence or the effective fis-
cal rate is the result of subtracting the tax incidence from
the expenditure incidence of the same income class. Overall,
this rate will show whether or not a particular income group
is a net gainer (positive rate) or net loser (negative rate)
from government fiscal activities. A more egalitarian re-
distributional fiscal activity would generally result in posi-
tive net fiscal incidence in the lower-income classes and

negative net fiscal incidence in the upper-income classes.

Specific Procedure

 

The foregoing has described the general procedure to
be used in this study, which is the one used in most other
fiscal incidence studies. .However, the methodological pro-
cedure used here differs in some respects from that used in
other research, and these differences deserve special mention.

First, this study is an intertemporal, not a single-

period, incidence study. Instead of selecting one reference

-‘
.u‘
1"

a: and a

LS1
ecf

1

is

iture
x4
that pa:
absent

5 for

ch 581
:eni

r on ea.

etiod met}

iitiona

1!

La;
S
If
.

‘L: .blLv "fi

0'." A
v“ c
-A v
‘
‘53.
"A‘-
I.”
i“
.
.
caon

21

year and assessing the impact of government taxes and expendi-
tures for that year, which is the traditional way, this re-
search selects Ewg reference years, estimates the tax and ex—
penditure incidence using the same assumptions and allocation
rules for both years, and compares the fiscal impact of each
year on each corresponding income distribution. The obvious
advantage of this method over the traditional one-period
method is that, in addition to the incidence pattern pertain-
ing to the year in question, it is possible to know the change
in that pattern over time. Moreover, the adoption of this two-
period method substantially weakens one criticism against tra-

ditional studies, that is, that they assume, in effect, that

the existing income distribution will remain the same in the
absence of the government.

Second, rather than use one income concept as the
base of the tax and expenditure incidence, two income concepts
are used in this study. One is money income, which includes
only household receipts in cash form, and the other is the
adjusted income concept, which includes not only money income,
but also nonmoney income, imputed income, and all other in-
comes which can be identified and calculated.\\‘

Third, the taxation and expenditure of the central
government will be comparable in both reference years. This
will be done by making income groupings, tax categories, and

expenditure classifications the same for both years.

A ca‘
:5 the study
is stated Wit
the most part
:cnsici-er the
expenditure b
be taken as o
got'erment fi:

Theore
313.273; is still
especially thc

Eve not been

 

22

Limitations of the Study

 

A caveat must be stated regarding the interpretation
of the study's conclusions. Although the fiscal incidence
is stated with mathematical precision, and the study is, for
the most part, quantitative in nature, the reader should not
consider the results to be an exact measure of tax burdens or
expenditure benefits. At best, this incidence estimate should
be taken as only a careful approximation of the effects of
government fiscal activities on income distribution.

Theoretically, the methodology of a fiscal incidence
study is still not fully developed. Many conceptual problems,
especially those concerning expenditure incidence, either
have not been resolved or must be dealt with in a less than
satisfactory way. In addition, the availability of statisti-
cal information in Thailand is far from ideal. Many statisti—
cal series must be "created" using logical assumptions and
available data. Also, to keep the study within manageable
proportions, three major limitations have been imposed on its
scope.

First, the study takes into account only the income
distributional effects of government taxes and expenditures;
it does not consider the potential distributional effects of
such nonbudgetary policies as price controls, land reform,
minimum wage legislation, interest rate regulation, popula-
tion control measures, foreign exchange rates, and other

micro-or macroeconomic policies.

l
O

s
Th
:1?

Sec
1
icultur.
”prese.
;act of a
‘rirar

:tain f

~t of f

‘3

break down
into class

:2

.. l 2. t

5 £ a S .1 C .c .
I aux e a i .L a .l
a t n; .l r g a F. :4 Rs - Aa
. . t ._ .d P. .. .

so I t m 4mm 8 .2 E n a A; D; km
a. . nl :4 1.... x a. .l S H u y .. .3. .l.

. his 5. E, U ._. . e . E C E . .. a

5‘. n x .\.v “Us c fl - ’K ‘5 H..‘ \ ‘ ..\

23

Second, the study considers only the distributive im-
pact of fiscal activities at the national level; it does not
break down the impact into regional or provincial levels or
into classifications such as rural/urban, agricultural/non-
agricultural, and so on. Also omitted is the incidence of
certain fiscal activities of local governments and of foreign
aids and loans.

Third, as is well known, this type of incidence study
represents partial equilibrium analysis, where the fiscal im—
pact of a certain period is restricted to that period only,
discounting previous developments as well as future repercus-
sions. Furthermore, the study is only concerned with the
primary or direct impact of fiscal policies; all secondary
or spill-over effects (externalities) are ignored.

Some explanation is appropriate as to why such limita-
tions are necessary. First, the study is confined to govern-
ment taxation and spending policies in order to keep the re-
search as compact as possible. To include other nonbudgetary
policies would be desirable but it would be impossible with
constraints of time and resources. In fact, the income dis-
tributional impact of these nonbudgetary policies would rep-
resent another entire area of study.

Second, although a series of income distribution data
could be broken down into four geographical regions (North,
Central, Northeast and South) and into municipal and non-
municipal subcategories within each region, the study is con-

fined to the national level because no tax and expenditure

5m are availé
3213' the natior
32:1 expenditure
Less, of course
:: which parts
: the several
gecement fisc
:ities, and the

the servicc

0
(ll

Ln:

.5 for them.

5:.
r1
«1)

likely to t
$51: to these
354" l ‘
because t

"‘>\
".38.

:.
t
E cat

-«.._8 Of a C

In
.~‘av

‘

Ml equlm

.Q

. .
.

1.: '7

...butive C
3‘.“ c
“e 0. the 5.

RA“":
W‘s. af“ -
VHUall“

 

 

5'5 50:1
e at th:

:I“

“QJJSis
‘

h. '
I wnlc

_ \
s,“

"at: 4;,
““ IV

‘U‘ V

as A

 

t;n~
“a|'al inc .

L'nt
Whir,
eff.
of)
St;
3‘]:

K.

24

data are available which are broken down on the same basis;
only the national data are available. Therefore, these taxes
and expenditures cannot be allocated at any other level, un-
less, of course, one is willing to make risky assumptions as
to which parts of the taxes and expenditures may be attributed
to the several regions. As for the exclusion of certain local
government fiscal activities, these are small, fragmented en-

tities, and they tend to exist on a quid pro quo basis, that

 

is, the services they provide are available only to those who
pay for them. As a result, the income distributive effects
are likely to be neutral. However, central government grants-
in-aid to these local governments were considered in this
study because they were part of general government expendi-
tures.
Finally, it is recognized that the true distributive

effects of a fiscal policy can be measured only through a
general equilibrium analysis which takes into account all
distributive components and reactions. Given the present
state of the art, however, it is not conceivable that one
could actually measure all such effects. The most that can
be done at the present time is to use partial equilibrium
analysis, which considers only primary and direct impact,
other things remaining constant or neutral. De Wulf, a critic
of fiscal incidence studies, has realized this problem:38

Until general equilibrium models are built

which allow researchers to trace the full

effects of wholesale budget substitution

on income distribution, partial incidence

studies are much more reliable guides than
system incidence studies.

The r
yereference
hrshich rea
:mnin Thail
tereferehce
justzehts in
is: such an e
alseries of
ixa,althcug

:2: 1962 and

iata after 19

 

25

Plan of the Study

 

The main reason for selecting 1963 and 1969 as the
two reference years was that these were the only two years
for which reasonable data for estimating the income distribu—
tion in Thailand were available. It is possible to stretch
the reference interval farther than six years by making ad-
justments in the income data, but there is no valid reason
for such an expansion at the expense of abandoning two origi-
nal series of data. Moreover, the 1963 tax and expenditure
data, although imperfect, are decidedly better than those
for 1962 and earlier years. In addition, the details of some
data after 1969 were unavailable. Therefore, the choice of
1963 and 1969 appears to serve the purpose of this study well.

The remainder of this study is divided into five
chapters. Chapter 2 will be concerned with estimating the
income distribution of Thailand in 1963 and 1969. Different
income concepts will be explored and adjustments made so as
to have a more complete income base for the study. Chapter
3 will consider the tax incidence in Thailand in 1963 and
1969, and Chapter 4 will compute the incidence or benefits
of public expenditures for these years. Chapter 5 will con-
solidate Chapters 3 and 4, and the net results of the entire
effects of the Thai fiscal system for the two reference years
will be estimated. In this chapter, attention will be given
to such matters as deficit financing, which have been neg-

lected in many incidence studies, but which will receive

relativelY mor
52 possible to
iistribution 0
aresult of <30
521:5 subst at
has done littl
varicus househ
its upper to t

In the
sizzle study wi
53: reducing i

:e-w.‘ '
:.a.Cfl dire-c

 

 

 

 

26

relatively more attention here. From this chapter, it will
be possible to determine whether any change occurred in the
distribution of income between the two reference periods as
a result of government budgetary policies, or whether the re-
sults substantiate the hypothesis that the Thai government
has done little or nothing in redistributing income among
various households in the desired direction, that is, from
the upper to the lower income classes.

In the last chapter, the substantive content of the
whole study will be summarized, certain policy implications
for reducing income inequality will be discussed, and future

research directions will be indicated.

1See 5:

Kibdtiontn A71
ER. Heraftei
trzzation by 51

2Irving
Discritutiori o:
42 {Haverber ,
IrenitiOhal (
:xeswith Spec
xiStatistics.

. Simon
.xsdlof Kati:

-‘A I
”‘vn2r

1 " ,
-c betel:

3 H
.. n U
u h"fir‘ '
5;::E£'(“35
Jk‘fi‘kr“
shynficp q ‘
‘rin "L, 19,1]
.s‘ql‘tlon apd I
V ‘
‘5}an fl ‘
-. r r REVISE
V

 

 

:" it-

.u‘k‘7~k

"5...; vdSed T

“‘1‘“ Av. ‘
s‘V 1‘

gay _ Jfient a'.

 

 

 

FOOTNOTES

lSee Simon Kuznets, "Economic Growth and Income Dis-
tribution," American Economic Review, 45 (March, 1955), pp.
1-28. Herafter, by income distribution is meant income dis-
tribution by size unless otherwise specified.

 

2Irving B. Kravis, "International Differences in the
Distribution of Income," Review of Economics and Statistics,
42 (November, 1960), pp. 408-416; Harry T. Oshima, “The
International Comparison of Size Distribution of Family In-
comes with Special Reference to Asia," Review of Economics
and Statistics, 44 (November, 1962), pp. 439-45.

 

 

3Simon Kuznets, "Quantitative Aspects of Economic
Growth of Nations: III, Distribution of Income by Size,"
Economic Development and Cultural Change, 2 (January, 1963),
pp. 1-80.

 

4Robert McNamara, "The Environmental Dilemma" in War
on Hunger, (Washington, D. C.: U. S. Department of State,
December, 1971) as quoted in Richard Weisskoff, "Income Dis-
tribution and Economic Growth in Puerto Rico, Argentina, and
Mexico," Review of Income and Wealth, 16 (December, 1970), P.
327.

 

 

5Robert McNamara, "Address to the Board of Governors,"
IBRD/IMF Annual meeting, Nairobi, September 24, 1973, p. 11.

6Idem., "Address to the Board of Governors" IBRD/IMF
Annual meeting, Washington, D. C., September 25, 1972, p. 15.

7One of the most forceful arguments in favor of
equity-based development policy is found in Mahbub ul Huq,
"Employment and Income Distribution in the 19705: A New
Perspective," a statement prepared for the Society for Inter-
national Development 12th World Conference, Ottawa, May 16-
19, 1971.

8See, for example, Irma Adelman and Cynthia Taft
Morris, "An Anatomy of Patterns of Income Distribution in
Developing Nations," paper prepared for USAID, Washington,
D. C., February, 1971; Felix Paukert, "Income Distribution
at Different Levels of Development: A Survey of Evidence,"
International Labour Review, 108 (August-September, 1973),
pp. 97-125; Irving B. Kravis, "A World of Unequal Incomes,"
The Annals of the American Academy of Political and Social

 

27

3:1 sce3409 (E
mien for
Pr-c‘ections, E
asaagkok, l972
Some Dimensior
Eollis Chener)
Chapter 1.
9See v
("re .lsrals of

1s to retrace
gznal producti
3. real terms:

.19 rate Of pr

:1: 53' K:

"sz
9 Y equals

3’03 which it

 

 

28

Science,409 (September, 1973), pp. 61-80; U. N., Economic
Commission for Asia and the Far East, Interregional Trade
Projections, Effective Protection, and Income Distribution:
(Bangkok, 1972); Montek S. Ahluwalia,“Income Inequality:
Some Dimensions of the Problem," Redistribution with Growth,
Hollis Chenery et a1. (London: Oxford University Press, 1974),
Chapter 1. —_ __
9See Walter A. Weisskopf, "The Dialectics of Equality,"

The Annals of the American Academy of Political and Social
Science, 409 (September, 1973). P. 167.

10Edward J. Nell has shown that one way to prove this
is to retrace the usual neoclassical exposition of the mar-
ginal productivity theory. Start with the income identity
in real terms:

 

 

Y = wN + rK (l)

where Y is output; w, the wage; N, the number of workers; r,
the rate of profit; and K, the total capital stock. Divide
(1) by K:

y = wn + r (2)
where y equals Y/K and n equals N/K. Differentiate:

dy = ndw 4 wdn + dr (3)
from which it follows that

w = dy/dn (4)

or the wage will equal the marginal product of labor if and
only if -n equals dr/dw, that is, all sectors in the economy
must have the same capital to output ratios which is a most
unlikely circumstance. In general, with a given technique

or if the techniques vary, the wage may not equal the marginal
product of labor because the real wage rate and profit rate
will move haphazardly as the relative scarcity of labor to
capital varies. See Edward J. Nell, "The Fall of the House

of Efficiency," The Annals of the American Academy of Political

 

and Social Science, 409 (September, 1973). pp. 102-111. For
a larger discussion on this "reswitching" point, see George
C. Harcourt, Some Cambridge Controversies in The Theory of
Capital (Cambridge: Cambridge University Press, 1972).

11Matthew Arnold, Lecture on "Equality" in Mixed
Essa 5, ed. 1903, p. ix, quoted by R. H. Tawney, Equality
4th ed.. (London: George Allen and Unwin, 1964), p. 33.

12R. H. Tawney, supra.

 

 

 

 

l3See
tic. (Cambrid
Fulton Fried:
of Chicago Pr.

14Bar
2: Income and
Acaderv of Po

15 .
Mil
Distribution .
I.‘-.u;ust, 1953

1

i6TO
heds X and
tion. x is
:cral worth
isles, and
consumption
2113: be inher
56259 that no
520 that X ta-
393. that is |

‘AOJ‘C

mu

33: c.
“twr of:

.th

.‘v

.. how 1

 

 

 

29

13See Bertrand de Jouvenel, The Ethics of Redistribu-
tion (Cambridge: Cambridge University Press, 1951) and
Milton Friedman, Capital and Freedom (Chicago: University
of Chicago Press, 1962).

14Harry G. Johnson, "Some Microeconomic Reflections
on Income and Wealth Inequalities," The Annals of the American
Academy of Political and Social Science, 409 (September, 1973),
p. 60.

 

 

 

 

15Milton Friedman, "Choice, Chance, and the Personal
Distribution of Income," Journal of Political Economy, 61
(August, 1953), 277-90.

16To further elaborate upon Rawls' View, two sets of
goods X and y, may be conceived in a consumer's utility func-
tion. X is a set of primary social goods where liberty and
moral worth (basic rights and duties) are often cited as ex-
amples, and y is a set of ordinary commodities for physical
consumption and enjoyment. Rawls' system requires that x
must be inherently equal for everyone in the society in the
sense that none in society has more than the least endowed,
and that x takes precedence over y in a lexicographical man-
ner, that is to say, an individual with (x2 Y1) is always

I

better off than with (x1 Y2) no matter how small (x2 - x1)

 

or how large (Y2 - Y1) is.

The inequality in income and wealth or difference

in Y2 and Y1 in the above model is "just" only when it is

agreed upon or sanctioned by a democratic approval within a
hypothetical "original situation," that is, the situation
where no one has a preconceived idea of where he stands in

the society or in Bronfenbrenner's interpretation, none knows
either his relative standing in such attributes as IQ and
physical stamina or what the relative social standing of his
particular race, sex, religion or age group may be. Then,

the next requirement in Rawls' system is that the least ad-
vantaged members of society deserve to be compensated first
and foremost. See John Rawls, A Theory of Justice (Cambridge,
Harvard University Press, 1971), p. 14, 60, 302-3, passim.

See also Martin Bronfenbrenner, "Equity and Equality," The
Angals of the American_Academy_of Political and Social Science,
409 (September, 1973), pp. 9 - 23,

17One of such view is found in W. Arthur Lewis, The
Theory of Economic Growth (London: Allen and Unwin, 1955).

18See W. Paul Strassmann, "Economic Growth and Income
Distribution," Quarterly Journal of Economics, 70 (August,
1956), pp. 425- 40; William G. Tyler, "A Model of Income Dis-
tribution and Economic Development," Weltwirtschaftliches
Archiv, (1973). PP. 321-36.

 

 

 

 

 

feund in Alb.
3:22 Inequal
"arterlv JO‘
L____+_

II

ct.

ZZHO:
‘hat while t]
i‘ith equalize
:crthodox "5
stiztlate grc
tribution anc

p
Pb K 1

[ 'A'n‘ V‘wn. ,
a. «CIC‘O‘J.:EI

23Ri<

26..
3" 1110.
." H. a
.I “Q L1
'~\-. T ‘ Jtur'

nyuliczlcs ar;

A;

’ !

 

30

19William R. Cline, Potential Effects of Income Re-
distribution on Economic Growth: Latin American cases (New
York: Prager, 1972).

20Kuznet, "Quantitative...," pp. 78-80.

21An interesting analysis of this tolerance level is
found in Albert 0. Hirschman, "The Changing Tolerance for In-
come Inequality in the Course of Economic Development,"
Quarterly Journal of Economics, 87 (November, 1973), pp. 544-
66.

2However, W. R. Cline in his recent article contends
that while the orthodox fear of a decline in the savings rate
with equalization has been found empirically doubtful, the
unorthodox "structuralist" hope that redistribution would
stimulate growth has also been found dubious. See his "Dis-
tribution and Development: Survey of Literature," Journal
of Development Economics, 1 (February, 1975), pp. 359—400.

23Richard A. Musgrave, The Theory of Public Finance
(New York: McGraw Hill, 1959).

 

24See Ralph Miliband, The State in Capitalist Society
(New York: Basic Books, 1969); Paul Baran and Paul Sweezy,
Monopoly Capital (New York: Monthly Review Press, 1966);
Richard C. Edwards, Michael Reich, and Thomas E. Weisskopf,
The Capitalist System (Englewood, N. J.: Prentice-Hall,
1972); James A. O'Connor, The Fiscal Crisis of the State
(New York: St. Martin's Press, 1973); and somewhat in dif-
ferent context, J. Kenneth Galbraith, The New Industrial State,
2nd ed., rev. (Boston: Houghton-Mifflin, 1971).

25

 

 

 

O'Connor, pp. 6-7.

26Thomas E. Weisskopf "Capitalism, Underdevelopment,
and the Future of the Poor Countries" in Jagdish Bhagwati, ed.,
Economics and World Order (London: MacMillan, 1972), p. 40.

27John Kenneth Galbraith, "Power and the Useful Econo-
mist," American Economic Review, 63 (March, 1973), pp. 10-11.

28Some authors use the word "progressive expenditure"
to mean the expenditure which benefits the lower income
groups relatively more as the income level falls. Therefore,
this has the same meaning as "regressive expenditure" used in
this study. This latter choice of word is preferable because
it is consistent with the treatment on the tax side.

29Mabel Newcomer, "Estimates of the Tax Burden on
Different Income Classes": Studies in Current Tax Problems
(New York: Twentieth Century Fund, 1937), pp. 1-52; Helen

 

 

 

pp. 1'
of .Ex Z'Pa. mem
fiaticnal Tax C
Journal. 4 (5‘

OTibc
Public Finance
.. leave}
Kingdom," Rev:
1353). p.073:
here in Post
1.__________

.....

310. l

Tuesand Pub]
J...s ness ‘

eDistrit
Lsss in Fisc
Hsslnstituti
:xstr ib utior

 

 

31

Tarasov, "Who Pays the Taxes?" TNEC - Monograph No. 3, 76th
Congress 3rd Session, Washington, D. C., 1941; Helen Tarasov
"Who Does Pay the Taxes?" Social Research, Supplement IV
(1942), pp. 1-79; Richard A. Musgrave et a1., "Distribution
of Tax Payments by Income Class: A Case Study for 1948"
NationaltTax Journal, 4 (March, 1951), pp. 1-53; Rufus S.
Tucker, "Distribution of Tax Burdens in 1948," National Tax
Journal, 4 (September, 1951), pp. 269-85.

30Tibor Barna, The Redistribution of Incomes Through
Public Finance in 1937 (Oxford: The Clarendon Press, 1945);
Findley Weaver, “Taxation and Redistribution in the United
Kingdom,” Review of Economics and Statistics, 32 (August,
1950), pp. 201-13; Allan M. Cartter, The Redistribution of
Income in Postwar Britain (New Haven! Yale University Press,
1955).

 

 

 

 

 

 

 

 

 

31O. H. Brownlee, "Estimated Distribution of Minnesota

Taxes and Public Expenditure Benefits," Studies in Economics
and Business No. 21 (Minneapolis: University of Minnesota,
1960); W. Irwin Gillespie, "Effects of Public Expenditures

on the Distribution of Income," in Richard A. Musgrave,ed.
Essays in Fiscal Federalism, (Washington, D. C.: The Brook-
ings Institution, 1965), pp. 122-86; Neil M. Singer, "Income
Redistribution through the Federal Budget in 1959f unpublished
Ph.D. dissertation, Stanford University, 1965; The Tax Founda-
tion, Tax Burdens and Benefits of Government Expenditure by
Income Class, 1961 and 1965 (New York: The Tax Foundation,

_ 1967). 3 -

 

 

 

 

 

32John Adler, E. R. Schlesinger, and Ernest Olson,
Public Finance and Economic Development in Guatemala (Stanford:
Stanford University Press, 1952); Henry Wallich and John Adler
Public Finance in a Developing Country: E1 Salvador-~A Case
Study (Cambridge: Harvard UniversiEy Press, 1951); Carl S.
Shoup 33 a1., The Fiscal System of Venezuela: A Report
(Baltimore: The John Hopkins Press, 1959).

33These countries include: Argentina, Brazil, Chile
Colombia, Jamaica, Panama, Peru, Puerto Rico, Surinam, India,
Greece, Indonesia, Kenya, Lebanon, Malawi, Pakistan, Philippines,
Tanzania, Turkey, and West Malaysia. For comprehensive sur-
veys on tax, expenditure and budget incidence studies in less
developed countries see Richard M. Bird and Luc De Wulf, "Taxa—
tion and Income Distribution in Latin America: A Critical
Review of the Empirical Studies" International Monetary Fund
Staff Papers, 23 (November, 1973). Pp. 639-682; L. De Wulf,
"Fiscal Incidence Studies in Developing Countries: Survey and
Critique," IMF Staff Papers, 22 (March, 1975), pp. 61-131; see
also Jacob Meerman,fiFiscal Incidence in Empirical Studies of
Income Distribution in Poor Countries" AID Discussion Paper
No. 25, (Washington, D. C., December, 1972). '

 

 

 

 

 

 

31

if the Ta:
tion in L.
are held
1365), p.

3.
kcerber,

35

3C
:‘j‘k;

~~HblstiCE

32

34Bird and De Wulf, p. 663 and De Wulf, passim.
35De Wulf, p. 48.
36Richard A. Musgrave. "Estimating the Distribution

of the Tax Burden" Chapter 2 in Problems of Tax Administra-
tion in Latin America: Papers and Proceedings of a Confer-

ence held in Buenos Aires, Argentina, October, 1961 (Baltimore,
1965). P. 31.

37Thailand, Bank of Thailand, Monthly Bulletin,

December, 1964, and December, 1970.

38De Wulf, p. 2.

 

 

39The two sources of data are (1) Thailand, National
Statistical Office, Reports of the Household Expenditure Sur-
vey, 1962-1963, 7 vols. (Bangkok, 1968); (2) Idem., Report

of the Socio-Economic Survey, 1968-1969 (Bangkok, 1973).

CHAPTER II

THE DISTRIBUTION OF INCOME

IN THAILAND: 1963 AND 1969

The object of a fiscal incidence study is the compu-
tation of the effective rates of tax burdens and public ex-
penditure benefits. To do so, an income base is needed as
a denominator for the values of the taxburdens and expendi-
ture benefits. It is obvious that the selection of an in-
come base is very important; if an incorrect or inaccurate
base is used, the pattern of fiscal incidence could be dis-
torted. Therefore, a precondition for a successful fiscal
incidence study is the determination of a proper income base
for each income class.

But the determination of an income base by size is
not at all easy. It usually involves many conceptual dif-
ficulties, such as a definition of income, the difference in
income received at different periods, the selection of the
individual or the family as a unit of analysis, and so on.
In an underdeveloped country such as Thailand, the problem
is even more serious; in addition to the familiar conceptual
problems, the necessary statistical data often do not exist

and if they do, they often are unreliable.1

33

:ibution 5
cf the inco
:Lued as i
uquns
ltst econc:
futons, who
'1? the mar
fl) th cha
between the

l
J. 3381' WC

3. :et weal-

15'
..

fo
.-

Inco:
119's econo:
13? to these

q
.2...

_ C
is): 0‘ none

34

Conceptual and Statistical Problems
in Estimating Income Distribution

The two major conceptual problems in an income dis-

‘tribution study are the concept of income and the selection

Choosing what is to be in-

Cm'the income recipient unit.
Income can exist in

Cluded as income is not always simple.

lmany forms and may be differently defined in various settings.
.Most;economists, however, prefer the definition of Henry C.
Sinuons, who defined personal income as "the algebraic sum of
(l) the market value of rights exercised in consumption, and
(2) the change in the value of the stock of property rights
between the beginning and the end of the period in question."2
Itrl other words, it is the sum of consumption plus the change

j~rl net wealth. A similar definition was proposed by R. M.
Income is "the money value of the net accretion to

Eiéijxg:
Accord-

C>Il€e's economic power between two points in time."
J‘IISJ to these definitions, income is more than the flow con-

C:E31?t of money income; it also includes, among other things,

1*r1<}ome in kind, imputed rental income, corporate savings,

t“Tansfers and gifts, and capital gains.
In national income accounting, where terminology is

rtIQII'—‘e or less standardized, aggregate income is defined as
the payments, disbursed or accrued, to factors of production
:tf<:,1: services rendered in a given period. Personal income in-
<:=:LV\1des all payments which are actually disbursed to the fac-
.t:<:’1=s of production plus various kinds of transfers, but cor-

<:’1=ate savings are excluded because this kind of income is

 

KCERd, r
25; rsone
:stofte
geriirg u;
stnies, 1
hearaile
imestigat

H g. ‘ A '
{1.1.0.53}, J

P:

~earninc

’:
r- .
" due 'wdl

::r m
V ““6:
1
as
a).
“0351-1

13" 'r' ‘
‘3lcua1

{7‘

.2..Sed On

4.. ,

“v.1"

35

accrued, not disbursed. Disposable income is that portion

of personal income remaining when all personal taxes have

Thus, in the area of income taxation, the con-

been paid.
cept of taxable income may be very narrow or very broad de-

pending upon a nation's tax law. In income distribution

Studies, the concepts used are not uniform but depend upon
the availability of data and the aim or preference of the

investigator. Assuming that the statistical data pose no
Problem, if one wanted to measure the distribution of rewards

for employment in different occupations, then income defined
as earnings before taxes would be the most relevant concept.
If one wanted to determine the relationship between the dis-

tribution of income and earning abilities, then all types of

c3Ompensation in money and in kind would be more appropriate.

For other purposes, there are other income concepts.

As noted earlier, in addition to the problem of

choosing an appropriate income concept, there is the problem

of selecting the income recipient unit. The choice of the

individual as the unit of study is desirable if attention is

chusedon the relationship between the distribution of pro-

dllctive earnings and current output or the distribution of

income and earning abilities. If the purpose is to examine

the changes in consumption and saving as a result of certain
E Qlicies, or to measure the changes in the "welfare" of the
Q‘Ql'isumers, then the family. unit is more appropriate. How-

Q\’er the type of unit may be determined by the availability

3 data. If income data are obtained from a population census,

‘

A
far esti:

'flh‘

,
._.at 01:6

 

('2 ‘

36

then the recipient unit is likely to be the individual; of

they are obtained from household surveys, the recipient unit

is likely to be the family. It is possible, of course, to

study both units if the data permit, but the outcome of the

analysis would be different.
A point referred to above was the statistical source

The four data sources

for estimating income distribution.
(1) a population cen-

most often used by researchers are:
(3) results from sample surveys,

Sus, (2) income tax returns,
arui (4) composite sources based on national income data and

Other related information .
Let us first consider the population census. If the

Census is complete, and if it contains information on the in-

cOmes of individuals, it is probably the best data source

for computing income distribution. But in many countries,

EDEizrticularly the less-developed, population census data often

Eilree incomplete and, as a rule, almost are devoid of the need-

In such cases, the census would

ea¢1_ information on income.
C311£Ly be useful in providing information on the size of the

EJ<:irpulation or the number of households.
Income tax returns also could be used to compute the

ciii~£3tribution of income of individuals, but these alone are

jLJL"issufficient. In most countries the taxable income concept

usually narrow and is likely to omit persons in the lower-

5L453
1L”‘3l<:ome brackets. Other income elements might escape measure-

‘Eilnt through such tax loopholes as exemptions, preferential

tax tre
evasior

fore, 1

could i

may 5

.K 1
#335.)
which t

Giten t

The Pro

 

37

I

tax treatment, and capital gains, not to mention the tax
evasion which is pervasive in less-developed nations. There-
fore, the taxable income base pep s3 is inadequate.

Household sample surveys, the third source of data,
could be household income surveys or household expenditure
surveys including questions on family income. The personal
income concept used could follow that of Simons and Haig as
closely as one wished. The surveys also could be designed
to illuminate the structure of the very poor or the very
wealthy. What is most crucial, however, is the manner in
which the survey is conducted, rather than its design.

()ften the surveying agency is constrained by a low budget,
vihich results in too small a sample size, by limited man-
Power, incompetent interviewing, and a lack of cooperation
from the households. Nevertheless, the survey method is
Still the most flexible.

Finally, composite sources have no single origin.
The process of compiling such information normally would be-
Siin with finding the reference levels of national income
31>roken down into income from wages and salaries, self-
Eamployed income, transfer payments, income from prOperty, and
£30 on. Then the distribution of each of these separate cate-
Syories would be estimated using information from other
Siources. For example, the distribution of salaried income
(:ould be estimated from industrial and business censuses or
Ilabor surveys; self-employed income, in the case of most urban

Inouseholds, from tax records, or for rural households from an

agricultural
registers.
avoids the U
the higher i
satple surve
:age, howeve
incompatible
iifficult.
Prev

‘
H
“(an

‘3'“ and undl

 

 

 

 

38

agricultural census; and transfer income from social security
registers. The advantage of this type of data is that it
avoids the underreporting of most income classes, especially

the higher income brackets, which is a problem in household

sample surveys and tax return data. Its greatest disadvan-

tage, however, is that data from different sources may be

incompatible and adjusting separate income groups is thus

difficult.4

Previous income distribution studies in both devel-
<5ped and underdeveloped countries most frequently have used
613 a single data source the household income and expenditure

ssurvey. This survey alone may be used or it may be supple-

nuanted by other sources, such as tax data, to confirm the
Ebccuracy of the distribution of the top income groups. One
Ineason for the popularity of this method may be that it is
Ikelatively easier to gather information about income dis-
t:ribution through surveys than through other collection

ITlethods. Both the accuracy and reliability of the results

Could be improved through better technical and administra-

't:ive management and increased budgets. True, the problem

<2>f sampling error always is present, but its advantages are

ESufficient to offset this drawback. Besides, sampling errors

<::an be reduced through sound statistical design and execution

‘:>f this survey.

tween 1962 a:

werweflnnlv c
A....C..ud£.--J s

.,_.
.311 SLEI'JE‘.’ ‘n
‘

Other

..
"ng V
a .

“MES on till

.-
‘6’.

....s will be

..I‘ A ‘ g
...9 2:80 1 O I‘

 

 

39

In Thailand, there is no choice among alternatives:

The only data available are from household surveys. As of

1974, four national household expenditure (and income) sur-
veys had been conducted by the National Statistical Office

The first was conducted during 1958, the second be-

(NSO).
the third during 1969, and the fourth

tween 1962 and 1963,

in 1971. Only the 1962-1963 and 1969 surveys were usable
here; the 1958 survey was too small and was limited to a few

nonrandomly selected municipal areas, and the results of the

.1971 survey were not yet published.

Other researchers have made estimates of or conducted
These ef-

Studies on the distribution of income in Thailand.
and

ftirts will be reviewed briefly, and their weaknesses

time need for a new study will be discussed.

Past Studies of Income Distribution in Thailand

Until early 1973, only four scholars were seriously

it'l‘volved in the study of income distribution in Thailand and
Harry T. Oshima,

had published their findings in some form.

whose article was published in the October 1970 issue of

EEEEge Malayan Economic Review, compared the sources of income
5

zidrrequality in Thailand with those of other Asian countries.

'Z\~J~though not an exhaustive examination, this study provid-

‘Ei'ii a penetrating insight into the relationship between in-

QQtrne inequality and economic growth in Thailand.
Udom Kerdphibule of Kasetsart University was probably

jtllfile first Thai scholar to undertake an intensive study of the

distribution
tonic Connis
‘he income d
sible for th
in the 1971 1
later in Vol.
Ed the Far 1
Effective prt
K
1972.6 In D.
Ether article
:ission for ;
Distribution
Fiblished in
Thaw Ker (1

1.1g efforts,

 

40

distribution of income in Thailand. As a member of an Eco-

nomic Commission for Asia and the Far East organized to study

the income distribution of the region in 1971, he was respon-

sible for the report on Thailand. His work first appeared

in the 1971 Economic Survey of Asia and the Far East, and
later in Volume III of another Economic Commission for Asia
and the Far East publication, Interregional Trade Prg'Lection,
E_f :fective Protection, and Income Distribution, published in

19 72.6 In December of the same year, Kerdphibule wrote an-

other article using the same material from the Economic Com-
mission for Asia and the Far East entitled "Income and the
Distribution of Income of the Agricultural Sector," it was

Published in the Thai Journal of Agricultural Economics (in

Kerdphibule should be much credited for his pioneer-

Thai) .7
His

ing efforts, but his work also had many deficiencies:
a~11alysis, for the most part, was sketchy; certain methodolo-

gies were questionable; and there were a few careless mistakes.

Therefore, the conclusions reached, if not misleading, must

be interpreted with caution.
In August, 1972, William A. McCleary of Thammasat

University published his study on income distribution in Thai-

1 and in a Thammasat University Discussion Paper. His empha-

% is was not so much on income estimation and its adjustment
8 on the sources and magnitudes of income changes between
1 9 62-1963 and 1968-1969, using the existing income distribution

Q C I
ls‘l‘ies. His work was more or less an exten31on and expan31on

Pi.

As
:5 income c'
:creestima
fily, by a
:c:;letenes
iiates wil

“‘hn ‘
can. 46"
‘

an MC.

~38 Of E
As H
a: di;.rih

 

41

of Oshima's article, with an emphasis on Thailand.

Finally, in an ongoing project, Oey Astra Meesook,
also of Thammasat University, is attempting to give special
attention to the consistency in the estimation process and
to incorporate into the money income concept certain income
adjusting elements, especially income in kind. When com-
pleted, her findings will add substantially to the scant lit-
erature on income distribution in Thailand.10

As there exists no systematic and adequate account
of income distribution in Thailand, this study will attempt
to reestimate income distribution in 1963 and 1969. Hope-
fully, by avoiding previous pitfalls and by emphasizing the
completeness and consistency of the estimates, these new es-

‘tinmtes will provide much improved income bases for the sub—

sequent incidence study.

Income Distribution in 1963

Pfliture of Data and Method of Study

 

As mentioned earlier, the main source of data on in-
CCHne distribution in 1963 is the 1962-1963 Household Expendi-
t1ire Survey.11 The results were published in seven volumes,
Orua for the whole country, one for the municipal sections of
Bangkok and Thonburi, and one for each of the five regions:
thenorth, northeast, central plain, east, and south. The
survey further classified each region into towns and villages,

with the exception of Bangkok and Thonburi, where only the

23.7.5 were su:
are, first, 1
getterns of m
rifle informat:
ani services
basis for calc
Sample
airatified sar

3;}:5 or dis

.L _
50‘

(I)
U)

ECOHd St:

k

:"“hu

1
krdur’ and tr
~‘Afi.

"~29? .‘v 1
I Cl»‘l bl

32:3 Selected

”129: the }
":..~:S Was US
I120 ‘f‘n

'.
'3

 

1.:
“ but
3" L." er
I
\A‘ C
k“ ‘w Q
“HHQ +

50 E;

‘2.

 

42

towns were surveyed. The two main objectives of the survey,
were, first, to study the income, expenditure, and savings
patterns of noninstitutional households and, second, to pro-
vide information on the relative importance of various goods
and services in household spending patterns to be used as a
basis for calculating a consumer price index.

Sample households were selected by a three-stage
Stratified sampling technique. The first stage selected the
amphurs or districts from over 400 amphurs in the country,
the second stage the city blocks or villages from the chosen
amphur, and the third stage the final households from the
Chosen city blocks and villages. At each stage, sample units
Were selected systematically with probabilities proportional
to size; the household information from the 1960 population
C3ensus was used for this purpose. The total sample selected,
6 a 420 households, was divided into 2,310 households in towns
and 4,120 households in villages.

The unit of study was the economic family, defined
either as a single person who was financially independent of
other family groups or as a group of persons who lived and
ate together, pooling their income and drawing from this com-
mon fund to pay for food, housing and other living essentials.
One or more of these families living together formed a "house-
hgld." In other words, there could be more than one economic

fanlily in a household, although almost all households were of

the one-family type.

Th

CCZEI whic

‘r ‘1
FL,
H
(L)
k 1;
a)
’3
r+

53:1". sourC
mdbfoker
assist me
and the va
Th
iistributi
tion of fa
each locat
it was ass
the entire
class was <
weighing
correspond:
could be ft
in each cle
into corres
Lat change
"0115 give

T: ‘54
t. LS tot

33‘s ’
‘ 1“ thi

 

43

The income concept used was total family money in-
come, which included wages and salaries, profits from self—
employment or family enterprises, income from rent and from
such sources as pensions and annuities, interest, dividend
and brokerage fees, money received through public and private
assistance, lottery prizes and net winnings from gambling,
and the value of food received as pay.

The general methodology used in estimating income
distribution in 1963 was as follows: the frequency distribu-
tion of families as recorded in different income classes in
each location and each region was taken from the survey, and
it was assumed that these families were representative of
the entire country. After the mean or average income of each
class was determined, the distribution of income was obtained
by weighing the average income in each income class with the
corresponding household share. This sample distribution
could be further "blown up" by multiplying the average income
in each class with the number of actual families classified
into corresponding income classes. The multiplication would
not change the pattern of income distribution; rather, it
would give the level of the total income of the whole country.
To this total (money) income, various adjustments would be

made in this study to make it as complete as possible.

The Income Distribution Estimates

 

Two pertinent types of information concerning family

income distribution were gathered in the 1962-1963 survey.

First, the
was obtain
Second, th
regions, e
five incom
2.1 and 2.

At
of income c

the average

“0 Sets 0‘
l.‘
k

 

(

(

:1 .
(D

J

—

 

 

 

44

First, the frequency distribution of families in all regions
was obtained and was classified into ten income classes.
Second, the distribution of average monthly incomes for all
regions, except Bangkok and Thonburi, was classified into
five income classes. This information is presented in Tables
2.1 and 2.2.

At first glance, it would seem that the distribution
of income could be easily estimated simply by multiplying
the average monthly income in Table 2.2 with the relevant
corresponding percentage share of households in each income
bracket in Table 2.1, after reducing the income classes in
Table 2.1 from ten to five. In fact, the matter is not so
simple because of the inconsistent use of the family income
concept by the NSC. In the survey, both the annual and
monthly income questions were asked separately, resulting in
two sets of different income figures, both of which were pub-
lished without adjustment. Whereas the families were classi-
fied into annual income classes as in Table 2.1, the average
monthly income for each of these classes was not obtained by
converting the average annual income into the average monthly
income; the latter figure was the one obtained directly from
the questionnaires. In other words, the NSC used the annual
income data to group the distribution of families, and then
used the monthly income data from the survey responses to
compute the average monthly income for each of these family

classes as the average monthly income figure. As a result,

Enigma UMBDHUCnxwrfl
"£97993: MCTAINcmH 3:; SE; Trummnéu QZOCCH \nnu TAJANECL k0 :CfiUZQwMUQHQ .n.Kt...LCUUH._:|IN .N U~.~£~.

45

 

.3 d zooms £99.55 éooofix 323

 

 

 

 

 

 

 

 

 

 

 

 

 

“mooflumofl ago mfiufimea 328:8 mfi mo uuommm define HBUflpfim HEQsz 6539? "8.38
no mo ~.o do To o4 m4. o.~m fimm m.mm o.ooH 590m
m.o m.o m.o m.o w.o o.m m.m m.h~ o.mm v.5m o.ooa Hmuucmo
do do No I To m4 o.» TR mom oém o.ooH pmmm
do .I ~.o do do m.o dm odd mow o.mo o.ooH 582
nu I do do do mo m4 mo «.2 mos o.ooH ummmfiuoz
do do do do m.o «4 dm moa oém Tom o.ooH "moofldt,
EH mA 54 dm o.m To dam oém mos no o.ooa £50m
v.H m.o ¢.H N.m m.m m.v H.va N.Hv >.ma m.HH o.ooa Hmuucmo
o.o ~.H o.~ m.m o.m m.o m.oa o.>m o.oH ~.mH o.ooH ammo
dH o.o o4 o.o o4 m.m dos o.R odd «.2 o.ooH fiuoz
o4 To o4 o.~ Em do 5.3 mdm Ema ooa o.ood $8532
«4 mo EH Tm 5m o.m 5.3 mom 5: TS o.ooH "938.
mgr. ~.~ o.m oé m.m mos TS Tam o.mH oé o.ooa fiéncofi.
uxflmfim
o.o mo o.o To o4 Tm Em 93 mom doo o.ooH 283.9395
gone... 83% mod: moo.mm moo.o~ moo.mm mooga mamas oom.m ooo.m coo
ooo.oo nooo.ov uooo.om uooo.om uoooJN uooo.oa nooo.~H uooo.o nooo.m H85 Ham
Hanna

exam: mommmanomsoocH

 

 

 

 

39389. mooaamofl mfi 58m mommmd QBBH 3 mmfigm mo Sflfiflhfio mmflfiafluudm 3nd.

mm>nsm onsuflucmmxm

Table 2.2--[

Income Clas
(Baht)

 

 

 

46

Table 2.2--Distribution of Monthly Income by Annual Income
Classes from Household Expenditure Survey,
1962-1963 (Amounts in Baht)

 

 

Income Classes North Bangkok-
(Baht) East North East Center South Thornburi

 

 

 

 

 

Urban Areas

 

Under 3,000 357 273 305 368 466
} 428

3,000 -. 5,999 491 428 585 490 653
6,000 - 11,999 794 802 792 769 766 756

12,000 - 17,999 1,341 1,339 1,359 1,375 1,222
} 2,470

18,000 and over 2,935 3,211 2,413 3,000 2,882
Average 1,089 834 1,055 1,090 1,206 1,519

 

 

 

 

 

 

Rural Areas

 

Under 3,000 151 252 318 311 .334 --
3,000 - 5,999 377 437 469 494 487 --
6,000 - 11,999 704 639 850 916 858 --

12,000 - 17,999 1,358 1,344 1,236 1,698 1,474 --
18,000 and over 1,653 2,294 1,794 2,261 2,093 --

Average 249 378 612 718 619 --

 

 

 

 

 

 

 

Source: Adapted from Thailand, National Statistical Office,
Reports of the Household Expenditure Survey,
1962-1963, 7 volumes.

 

3051; month]
the two 10"
'jl'ZPed" ti
which n0 96
whereas the
iirectly fr
As a result
:cnthly inc
limits, or
classes whe

:1
3418 .

 

47

most monthly average family incomes, particularly those in
the two lower brackets, when converted into an annual figure,
"jumped" the class limits into odd arrays of figures from
which no general conclusion could be drawn. Furthermore,
whereas the average monthly incomes of families were available
directly from the survey, the average annual incomes were not.
As a result, one is faced with the dilemma of accepting the
monthly income figures, some of which would jump the class
limits, or accepting the distribution of families by income
classes where the annual income class means were not avail-
able.

Finally, although the survey obtained information on
income in kind (or nonmoney income), the information was in-
completely presented. Only the average monthly income in
kind for each region as a whole was shown, not the average
monthly income in kind broken down by income classes. This
average nonmoney income figure thus proved of very little use
for this study except perhaps as a check against a similar
figure derived from other sources.

In view of these deficiencies, a new approach for
dealing with the existing data was necessary. After careful
consideration, this study decided not to use the mean monthly
income published in the survey reports (Table 2.2), but to ac-
cept the distribution of families by annual income class
(Table 2.1). Having so decided, the next task was to find
ways of obtaining average annual incomes that were consistent

12

with the class limits. The estimating technique chosen is

iescribed l
1. For all
to: classes

33.1131 incc

2. As the
regions, t)
lose its me

hculd be re

ail the rec

3' F0r the

’s

Lzahit 18 ' 00

is

(.1)

EC to fin.
v.
“‘9 method 1

V-
a.
u‘st

ributior

 

48

described below:
1. For all classes in Table 2.1 except the top and the bot-
tom classes, the class midpoints were used as the average

annual income.

2. As the midpoints would be the same in each class for all
regions, the analysis of income distribution by region would
lose its meaning. In this case, only the national distribution

would be relevant.13

Still the town-village or urban-rural data
presentation would be retained because in most cases it was
possible to have national figures only through summation of

all the regional or locational estimates.

3. For the two open-ended classes, that is, the lowest in-
come class (under Baht 3,000) and the highest income class
(Baht 18,000 and over), three alternative methods could be
used to find their midpoints, namely, the arbitrary method,
the method based on information from other reliable income
distribution data, and the method based on some mathematical
functions, particularly the Pareto distribution function.
These alternatives are discussed below:

(a) The Arbitrary Method. In using this method, a re-
searcher simply arbitrarily assigns the midpoint values to
the top and the bottom open-ended brackets. Usually no rea-
sons are given for why these values are selected; they nor-
mally come from the subjective or intuitive judgment of the

individual researcher.

(b)
informatiu
compute t1
year in g
tribution
that the 1
brackets :?

1111969 ar

C )
r—o.

the bot
next highe
the corres
average in

tides that

Vould be u:

 

49

(b) The Related Distribution Method. This method uses
information from the income distribution in other years to
compute the mean incomes to the open-ended brackets of the
year in question. In the case of Thailand, the income dis-
tribution data for 1969 could be adapted for 1963 assuming
that the pattern of average income in the two open-ended
brackets in relation to their adjacent brackets is the same
in 1969 and 1963. The proportion of the 1969 average income
of the bottom (top) bracket, which is known, to that of the
next higher (next lower) bracket could be used to compute
the corresponding mean incomes for 1963. For example, if the
average income of the top income bracket in 1969 was four
times that of its next lower bracket, then this proportion
would be used to multiply the known midpoint of the second
top bracket in 1963 to obtain the estimated midpoint of the
top bracket for that year.14 This is certainly an improve-
ment over the first method, because it is not unreasonable to
assume that the average income of one bracket relative to that
of the next bracket would be the same from year to year. Or,
if there should be any changes, it is unlikely that they would
be so great as to render this assumption totally unrealistic.

(c) The Mathematical Function Method. This method is
usually based on the Pareto function, which stipulates that
the percentage of peOple with income higher than any given in-
come level will fall at a constant rate as the income level

15

increases. In mathematical form, the function is given as:

where N i
Y
higher the

'53: persons

which rep:

cone recei

 

 

50

_ -b
Ny — Ay , y 1 YO > 0, and b > 1,

where Ny is the number of persons or families with income
higher than the level y; Yo is the minimum income received
by persons or families; A is the location or monetary unit
parameter; and b is the Pareto coefficient or the parameter
which represents the percentage change in the number of in-
come receivers to the percentage change in the income level.
If personal income is believed to be distributed
according to this Pareto function, various information on in-
come distribution could be derived from manipulating the func-
tion. If the 1963 mean income of the bottom bracket were to
be estimated using the Pareto function, the procedure would

require, first, the estimation of the minimum income, Y and,

or
second, use of the midpoint between this income and the floor
of the second bottom bracket as the mean income for the bot-
tom bracket. For the mean income of the top bracket, the
function would be used to estimate the total income attribut-
able to this bracket. By dividing this total income by the
total number of persons or families in the bracket, the mean
income of the bracket would result.16
Although this method of estimating the mean incomes
of the open-ended brackets is elegant and impartial, it has
several drawbacks. First, that the Pareto coefficient is al-

17 Instead, the co—

ways a constant, was empirically refuted.
efficient was often found to vary across an income range.

Therefore, to use a constant relationship between income

1
.; the

1
though

i'come c

“9 QiStri
the 10m
Igtal

51

recipients and the level of income to find the mean incomes
of the open-ended brackets is quite unwarranted. Second, al-
though the Pareto function fits the upper tail of the actual
income distribution quite well, it does so poorly for the
lower tail. As a result, the minimum income derived from the
Pareto function would almost always overestimate the true dis-
tribution. Third, the Pareto function is also known to have
overestimated the mean income of the tOp bracket.18

Given these reservations concerning the Pareto method,
and since the arbitrary method was inappropriate because the
1963 data should be on a comparable basis with the 1969 data
(for the purpose of this study at least), the related dis-
tribution method was chosen for use here. This meant that
the distribution patterns of the 1969 data was employed to
find the mean incomes for the two open-ended classes. When
all the midpoints of the 1963 income classes were multiplied
by the number of families of the respective income classes
(see Appendix A), the results obtained were the total family
money income classified by size of income. This money income
is similar in concept to the personal income used in the
national income accounts.

From Table 2.3 it can be seen that the inequality of
the distribution of income is quite striking. The families
in the lowest income bracket constituting almost half of the
total families receive less than 13 percent of the total

family money income, whereas the families in the top income

Table 2. 3-'

Income Cla
(Baht)

 

Eider 3,0(
3,000 —

6,000 - 1
1L000 - 1
18.000 - 2

'003 and
‘111 Clds
\
Wur\C9: CC
Be

a

 

52

Table 2.3--Distribution of Total Families and Total Family
Money Income, by Income Classes, 1963

 

 

 

 

Mean Total Families Total Income
Income Classes Income

‘Baht’ ”11322:?“ (Thousand> viii; 14913123..) iii;
Under 3,000 1,800a 2,489.0 48.1C 4,480.1 13.0
3,000 - 5,999 4,500 1,074.0 20.8 4,832.9 14.0
6,000 - 11,999 9,000 1,010.9 19.5 9,098.4 26.4
12,000 — 17,999 15,000 300.8 5.8 4,511.7 13.1
18,000 - 23,999 21,000 126.0 2.4 2,646.0 7.7
24,000 - 29,999 27,000 50.6 1.0 1,366.3 4.0
30,000 - 35,999 33,000 37.5 0.7 1,238.6 3.6
36,000 - 47,999 42,000 33.6 0.6 1,412.1 4.1
48,000 - 59,999 54,000 18.3 0.4 987.1 2.9
60,000 and over 117,500a 33.4 0.6 3,918.0 11.4
All Classes 6,666b 5,174.1 100.0 34,491.2 100.00

 

 

 

 

 

 

Source: Computed from Thailand, National Statistical Office,
Reports of the Household Expenditure Survey, 1962-

1963, (Bangkok, 1966), 7 volumes.

 

aFor the calculation of midpoints of the two Open-
ended brackets, see text.

bThis figure is the average annual family income
for the whole country.

cSlightly different from the original percentage of
48.4 due to rounding errors in computation.

bracket c1
all famil.
income. 1'

discussed

fiature o f
R

1968-1969.‘
was Sdbstar
a; ‘h

~‘- uOUSEhQ]
Survey

.I hon.

.1; 1n Prev

 

53

bracket comprising less than two-thirds of one percent of
all families receive almost 12 percent of the total family
income. A more exact measure of income inequality will be

discussed later in this chapter.

Income Distribution in 1969

 

Nature of Data and Method of Study

 

Similar to the situation in 1963, the main source
of data for estimating income distribution in 1969 also came
from a household sample survey, the socio-economic survey of

1968-1969.19

The objective of this and the 1962-1963 survey
was substantially the same, that is, to study the composition
of households and their spending patterns. The 1968-1969
survey, however, emphasized an important aspect which was lack-
ing in previous surveys: explicitly to obtain information

on household income, sources of income, and distribution of

20 As a result,

income both in municipal areas and villages.
the 1968-1969 survey provides much more complete, consistent,
and reliable information on the distribution of income in
Thailand than any other source. Although there are many de-
fects in the data, the quality is much improved over the
1962-1963 material.

The survey procedure was detailed in a single volume,
Report of the Socio—Economic Survey, 1968/69, published by
the National Statistical Office in 1974. Sample households

were again selected by a three-stage stratified sampling

technique.
1962-1963.
Thorburi w
:hangwads

schregions
villages t;
The third 1
3! sample 1
from each c_
its to £01.11
1'5“) Units
January an ‘
zen-ts in th
£530 hQ‘dSe

.
Ҥ
Vs

s
‘1

(C
u srom 6,0

‘ -
“170%-

sauents)

The

,..3 who 1 i“

'v

 

Nir' alti
ritmy

5‘ . Stl
a she 196
”scenes the

‘\

sg 1.
“Q ‘569

54

technique, which was somewhat different from that used in
1962-1963. In the first stage, in addition to Bangkok and
Thonburi which were selected as self-representative, two
changwads or provinces were selected from each of the nine
subregions of the country. Sample city blocks and sample
villages then were selected at random from each changwad.

The third stage classified households in each sample block

or sample village into different occupational groups, and
from each group at least two households were selected at ran-
dom to form the final sample households. Interviews with the
1,500 units selected were carried out in three rounds between
January and September 1969, with partial household replace-
ments in the second and third rounds. In all, approximately
4,500 households were canvassed in each round. At the end

of the survey period, some 13,500 questionnaires were obtain-
ed from 6,000 sample households (the original number plus re-
placements).

The unit of study in this survey was the household,
defined as a person or a group of related or unrelated per-
sons who live, eat, and share other living essentials to-
gether, although some persons may keep their finances sepa-
rately. Strictly speaking, this is not the same unit study
in the 1962-1963 survey, which used the economic family. But
because the majority of households in Thailand are one-family
households. the difference is unimportant; the household in

the 1969 survey can be considered identical to the economic

that used
overtime 1

snares of

received 1

attention
1'1 the ea r

55

family of the 1962-1963 survey.21

The 1968-1969 survey's income concept is the same as
that used in earlier projects. It includes wages, salaries,
overtime payments, bonuses, net profits from self—employment,
shares of profit and interest, pensions and annuities, money
received for rents, and income from other sources.

One other difference between the two surveys deserves
attention here: the composition of urban and rural areas.

In the earlier survey, both municipalities and sanitary dis-
tricts were both classified as urban areas (or towns), where-
as the villages outside municipal or sanitary districts were
regarded as rural areas. In the 1968-1969 survey, only the
municipalities were designated as urban areas, and both sani-
tary districts and villages were combined as rural areas.
The main reason for this reclassification, according to an
NSO official, was that the sanitary districts in 1969 exhibit-
ed characteristics more closely resembling villages than mu—
nicipalities, hence the change. This could also mean that
the municipalities had become much more "urbanized" in the
intervening six years, much more so than the sanitary districts.
The discrepancy, however, should not have any effect on the
analysis in this study, especially since the main emphasis is
at the national, not the locational or regional level.

The general methodology for estimating the national
distribution of income in 1969 was similar to that used in
1963. The income distribution pattern determined by the sur-

Vey was considered representative of all households in the

gantry: 5
Cf househc
lers are e
;resentati
survey rep
used, and

given. MO.
ended inco:
are also a:

;olate sucl

fit.
NOE

3lever A
urban and r

 

 

56

country, and the figure was "blown up" by the distribution

of households by corresponding income classes.22 Fewer prob-
lems are encountered in the 1969 than in the 1963 data. The
presentation of household distribution of income in the 1969
survey report is consistent: the annual income classes are
used, and the average annual income for each income class is
given. Moreover, the actual average incomes for both open-
ended income classes--the top and the bottom income brackets--
are also available, thus eliminating risky attempts to inter-

polate such mean incomes.

The Income Distribution Estimates

 

Most information needed for estimating income distri-
bution appears in the ten published tables in the 1968-1969
survey report. The data are presented in a rather odd way,
however, Although the income distribution is given for both
urban and rural areas in all of the four geographical regions
(the previously separated center and east are combined into
one region), the national income distribution is omitted.
Although one could combine the total families and incomes of
urban and rural areas to arrive at this figure, this process
is not easy because the income class intervals for urban and
rural households are different. .Within each urban area, the
income level is divided into 12 brackets, ranging from under
Baht 3,000 per year to over Baht 60,000 per year. Within

each rural area, the income level begins with the same lower

income br
incore br

ed in the

should be
TL-
Avg

mayas' rut

Seated in

3111 that i;

“V93 a:

e an:

 

 

57

bracket, but then progresses by smaller intervals until the
top, open-ended eleventh bracket of Baht 33,000 and over is
reached. To obtain a national income distribution, the urban
income brackets must be compressed to coincide with the rural
income brackets. As a result, only seven brackets are obtain-
ed in the national income distribution.{ There is no system-
atic way short of retabulation of the entire survey data to
expand the income brackets of the rural households to fit the
wider urban income brackets. This is a serious fault which
should be corrected in the next survey.

The distribution of households and income for urban
areas, rural areas, and the entire country for 1969 are pre-
sented in Tables 2.4, 2.5, and 2.6 respectively.

Before discussing these tables, it should be pointed
out that in estimating these income distribution series, the
average annual income by income classes used is not that pub-
lished in the NSC report. Rather, the figures were supplied
by Dr. Oey Meesook of Thammasat University. (See Appendix
B for details.) Sensing that there might be some inconsist-
encies in the final tabulation of results, Dr. Meesook ob-
tained the original survey tapes and ran them through the
computer to check the reliability of the official results.
She found that, indeed, some adjustments were made in the
final tabulations, especially in the two open-ended brackets.23

The average income of the bottom bracket in the N80 series

was generally larger than Meesook's recomputed figure, and

H

Income
(85

 

Under 3
3,000
6,000
9,000-
12,000 .
15,000 .
18.000 -
24,000 -
30.000 -
36.000 -
49.000 -
60.000 a

All C1

 

58

Table 2.4--Distribution of Money Income of Urban Households,
by Income Classes, 1969

 

 

 

Average
Income Classes Annual Total Total
(Baht) Income Families Income
(Baht) (Thousand) (Baht Million)
Under 3,000 1,731 6.9 12.0
3,000 - 5,999 4,421 29.6 131.1
6,000 - 8,999 7,373 55.9 412.3
9,000 - 11,999 10,281 84.0 863.9
12,000 - 14,999 13,171 85.9 1,131.3
15,000 - 17,999 16,167 74.7 1,207.2
18,000 - 23,999 20,232 139.6 2,824.0
24,000 - 29,999 26,062 83.8 2,184.2
30,000 - 35,999 32,233 51.1 1,647.4
36,000 - 47,999 40,575 54.9 2,225.5
48,000 - 59,999 52,752 34.0 1,791.6
60,000 and over 116,742 61.7 7,201.5
All Classes 28,385a 762.1 21,632.0

 

 

 

 

Sources: Computed from retabulated results of the 1968-1969
Socio-Economic survey supplied by Dr. Oey Astra

Meesook of Thammasat University, and from Thailand,
National Statistical Office, Report of the Popula-

tion and Housing Censds, 1970, (Bangkok , 1974).

 

 

aThis figure is the average annual money income for
the whole urban households.

Table 2

Income
(Ba

 

1‘
U3 Jrh

\es:

 

59

Table 2.5-—Distribution of Money Income of Rural Households,
by Income Classes, 1969

 

 

 

Average
Income Classes Annual Total Total
(Baht) Income Families Income
(Baht) (Thousand) (Baht Million)
Under 3,000 3 1,701 1,416.6 2,409.1
3,000 - 4,499 3,629 797.7 2,894.9
4,500 - 5,999 5,128 579.0 2,969.4
6,000 - 7,499 6,591 505.8 3,334.0
7,500 - 8,999 8,141 384.9 3,133.4
9,000 - 10,499 9,624 325.7 3,134.4
10,500 - 11,999 11,130 177.0 1,969.7
12,000 - 14,999 13,113 312.8 4,101.8
15,000 - 17,999 16,009 197.9 3,168.7
18,000 - 32,999 23,002 334.7 7,699.4
33,000 and over 70,255 114.2 8,205.1
All Classes 8,359a 5,146.4 43,020.0

 

 

 

 

Sources: Computed from retabulated results of the 1968-1969
Socio-Economic survey supplied by Dr. Oey Astra

Meesook of Thammasat University, and from Thailand,
National Statistical Office, Report of the Egpula-

tion and Housing Census, 1970, (Bangkok, 1974).

 

 

aThis figure is the average annual money income for
the whole rural households.

Table 2.

(1

Income C

a"
o
C)
O
l

 

 

60

Table 2.6--Distribution of Money Income of National House-
holds, by Income Classes, 1969

 

 

 

 

Average Total Families Total Income
Income Classes Annual
(3..., £332? (8......) 2:3; M8123... 2:3;
Under 3,000 1,701 1,423.5 24.1 2,421.0 3.7
3,000 - 5,999 4,263 1,406.4 23.8 5,995.4 9.3
6,000 - 8,999 7,268 946.6 16.0 6,879.8 10.6
9,000 - 11,999 10,172 586.7 9.9 5,968.0 9.2
12,000 - 14,999 13,125 398.7 6.8 5,233.1 8.1
15,000 - 17,999 16,052 272.6 4.6 4,375.9 6.8
18,000 and over 38,650 874.0 14.8 33,778.7 52.3
All Classes 10,942a 5,908.5 100.0 64,652.0 100.0

 

 

 

 

 

 

Sources: Computed from retabulated results of the 1968-1969
Socio-Economic survey supplied by Dr. Oey Astra _
Meesook of Thammasat University, and from Thailand,
National Statistical Office, Report of the POpula-
tion and Housing Census, 1970, (Bangkok, 1974).

 

 

aThis figure is the average annual money income for
the whole national households.

that in
other di
were min

come dis

tile meflia:

 

 

61

that in the top bracket smaller. Apart from this, there were
other differences in the average income levels, but these
were minor and would not affect the general pattern of in-
come distribution.

The distributions of income in both urban and rural
areas, as shown in Tables 2.4 and 2.5, are quite skewed.
The mean income is Baht 28,385 for urban households and Baht
8,359 for rural households, but the median income is esti-
mated at only Baht 19,890 and Baht 5,696 for these house-
holds, respectively. This skewness also is reflected in the
national figure, where the mean income is Baht 10,942 and
the median income is estimated at Baht 6,395. In the national
distribution of income series shown in Table 2.6, over one-
half of the total families, bunching into the first two lower
brackets, received only less than 4 percent of the total
money income, whereas less than 15 percent of the total fami-
lies comprising the top income bracket receive over one-half
of the total money income. Again, a more concise measure of
this apparent income inequality will be dealt with shortly.

Comparison between the 1963 and
1969 Distributions of Income

 

In comparing the distribution of income between the
two years, one must always keep in mind the definitional and
procedural differences between the two estimates pointed out
earlier, particularly the difference in the classification

of urban and rural households. At the national level, however,

there ar
In the f<
rributior
cal measr

Income Di
19363 and

2.3 and 2

lies had a

 

 

 

62

there are few difficulties in comparing the two estimates.
In the following section, the two estimates of income dis-
tribution by income class will be compared, and the statisti-

cal measurement of income inequality will be discussed.

Income Distribution by Income Classes
1963 and 1969

 

 

Table 2.7, which compares the estimates of income
distribution in 1963 and 1969, is taken partially from Tables
2.3 and 2.6. In 1963, an estimated 48.1 percent of all fami-
lies had an annual money income of less than Baht 3,000 per
family. This figure dropped to 24.1 percent in 1969. The
fact that many families moved up the income scale by 1969
also is reflected in the percentage of families in the top
income bracket: the figure increased from 5.6 percent in
1963 to 14.8 percent in 1969. However, one must be aware
that this was in part attributable to about 15 percent in-
crease in the price level from 1963 to 1969. In both years,
families in the top bracket, as a group, received the largest
share of total money income compared to any group below it,
regardless of size.

~ To facilitate comparison of two or more distribution
patterns, researchers often resort to the graphical technique
of the Lorenz curve, introduced by M. O. Lorenz in 1905.24
In this case, this method requires that the cumulative dis-
tribution of families be plotted along the bottom horizon-
tal axis of a unit square as well as the cumulative percent-

age distribution of income along the vertical axis on the

left. A curve connecting all the points which correspond to

 

.4J1C kt‘ LEC..~_ :2 C.— CECCE~ J».

IFMZCSr Cw mlflzgfii trainegb 1.; CF:
.moma ECG mwma a: mECUCH \fucg 130.4

xlnlflllrllll
#62 NO CCfiuDQflhumwQ SUDEQQ Cmvmflgnunv Quinn .N QNQQE

 

63

 

 

 

 

 

 

 

 

 

 

.0.N can. m.~ $.38. 88w Eng 8058
0.00H 0.00..” 0.00H 0.03 0.30.00 m.000.m «400.3. H65...“ mommmHU H2
m.~m 0.3 m.mm 0.m 5.2.5.3 0.3.0 H.009: «dam 5.6 98 000.0."
0.0 0.0 060m; 0&5. 000.5 I 000.3

imam 0.m A “.43."; 0.00m A
.70 0.0 H.mm~.m 5.0% 000.3” I 000.~.n
«.0 0.0 0.000;.“ Womm 000?: I 000.0
0.0NA mdafi 0.000.0A 0.08::
0.0a 0.3 0.03.0 0.03 000.0 I 000.0
m6 0.m~ 0.3 0.0m 0.30:.“ 0.0004 0.~mm.v 0.3.04 000:.“ I 000.m
5m .73 0.mH adv 0.HN¢.~ m.m~¢.a H.000; 0.mmv.m 000.0 “8.95
8.85 mmwagm 9505 mowadfim 9.85 83gb 9805 «mags—mm
35m:
003 M03 003 $3 mommmau NEH
8.0350me 9505

 

 

 

 

3an «o 20332 5 859a 628st fl 83255 @838 mg .3
:32 can 83 5 QaQH $82 Habflmz mo Sflaflfima 89.9mm somflmefio «lbw flame

the coorc
the Lore:
tion of a
both 1963

the two I.

the line .
:30th ale;
05 familie
cllrve that
PC‘Uld {near
the dista:
Clcser to
31119.): Lore

pagein th.

¥Q$v v
“ween the

 

64

the coordinates of these two cumulative frequencies is called
the Lorenz curve. This curve represents the income distribu-

tion of a given group of families.25

Using this technique,
both 1963 and 1969 income distribution series are shown by
the two Lorenz curves in Figure 2.1.

In Figure 2.1, the straight diagonal line is called
the line of perfect equality, or egalitarian line, since any
point along this line would correspond to the same percentage
of families having the same percentage of income. A Lorenz
curve that traced a distribution pattern away from this line
would mean that the distribution is more unequal the greater
the distance. With two or more Lorenz curves, the one lying
closer to the line of perfect equality and not crossing any
other Lorenz curves would depict a more equal distributional
pattern than the curve lying farther away. The similarity
between the Lorenz curves of 1963 and 1969 in Figure 2.1 is
quite extraordinary; they are virtually identical. Perhaps
the 1969 income distribution could be seen to be very slightly
more equal since its Lorenz curve lies closer to the egali-
tarian line than the 1963 Lorenz curve around the fourth quin-
tile of the household distribution. Visually, these two
curves still lie quite far away from the line of perfect equal-
ity, indicating that the money income distributions in both
years are still quite unequal. A more precise measure of in-

come inequality is considered in the next section.

God

00

’

00
.‘OUCH .HO in I>dOCH3§U

ON

.lIb

Finite 2

65

 

 

 

 

 

 

o
3
01963 . .5627
8I
0
3
0°
C‘O
H
~6’
R
O
>
.4
753.
5
0
0L
N
O 20 “O 60 80 100

Cumulative % of Families

Figure 2.1--Diatribution of Honey Incone in Thailand: 1963 and 1969

Statisti
_____,__

to the e
measure

Furtherrn

91‘ :
"'ereJ-Ore

P
L

 

66

Statistical Measures of Income Inequality

 

While the Lorenz curve is a useful device for describ-
ing the pattern of income distribution and depicting the ex-
tent of its inequality according to the proximity of the curve
to the egalitarian line, it is not a convenient practical
measure of income inequality because of its graphical nature.
Furthermore, in comparing two Lorenz curves, if one curve is
not completely inside the other curve, but crosses it at some
points, visual examination alone is inadequate or even impos-
sible in determining which distribution is the more equal.
Therefore, some measures or indices of income inequality are
necessary. I

There are more than a dozen income inequality measures
from which to choose. Among these are the Pareto coefficient,
the Gini concentration ratio, the coefficient of variation,
the standard deviation of the logarithms of income, the rela—
tive mean difference, the Elteto-Frigyes index, Atkinson's
measure, Oshima's index of decile inequality, Theil's index,
and so on.26 One of the most widely used measures, however,
is the Gini concentration ratio or Gini coefficient, first
developed by C. Gini in 1905.27 This ratio is superior to
many other measures, but not necessarily the best, for as D. G.
Champernowne has argued, there is no single "best" index or
coefficient for every aspect of income inequality.28 The
Gini concentration ratio has been widely adopted because it
is easy to estimate from existing income distribution data,

and the meaning of the coefficient itself is easy to

=

 

67

understand. Moreover, it is directly associated with the con-
cept of the Lorenz curve, which is widely used, making it an
even more attractive measure.

Using the Lorenz diagram as a reference, the Gini con-
centration ratio measures the proportion of the area bounded
by the egalitarian line and the Lorenz curve to the total area
under the egalitarian line. The ratio ranges in value from
zero (perfect equality) to one (perfect inequality). In the
former case, the Lorenz curve will coincide with the egali-
tarian line so that there is no bounded area. In the latter
case, the Lorenz curve will trace the lower horizontal axis
and the right vertical axis of the unit square, making the
bounded area and the total area identical, so the ratio is
one. Numerically, the Gini concentration ratio can be calcu-

lated from the following formula:
n
_ _. l .. -
G “ l 2 [i “1 fi-l) (Vi-1) + “£1 fi-l) (Yi yi-l)] '

where G is the Gini concentration ratio; fi and y1 are the ith
observations of the cumulative fraction of the recipients or
families and the cumulative fraction of income, respectively;
and i = l,2,.......,n, where the nth observation is the final
point in the Lorenz curve. This formula seems a cumbersome
expression of a very simple idea, but the computation is
straightforward and simple, although tedious.

However, the above formula shows only an approximation
of the area between the Lorenz curve and the egalitarian line,

which tends to understate the actual inequality; when a straight

line is
curve.

the SHIT.
{1.1) 01
the Lore
the Gini
approxirr.

5:34 the

integrati
area unde
:HRCtion

L#336112 Cu

 

 

68

line is drawn between two observation points on the Lorenz
curve, this line will lie above the actual Lorenz curve. So,
the sum of all trapezoidal areas from point (0,0) to point
(1.1) of the Lorenz box will be greater than the integral of
the Lorenz curve between the same interval. Subsequently,
the Gini ratio of concentration derived from this trapezoidal
approximation (TA) method will be smaller than that derived
from the integration method.

If possible, the TA method should be replaced by the
integration method, but to be able to integrate the entire
area under the Lorenz curve, one must know the mathematical
function of that curve. There are several ways in which the
Lorenz curve function can be generated, but the method adopted
in this study is the one devised by N. C. Kakwani and N.
Podder of the World Bank.29

According to Kakwani and Podder, if F(x) is the pro-
portion of units (families) that receive income up to x, and
F1 (x) is the proportion of total income received by the same
units, the Lorenz curve is then the graphical representation

of the relationship between F(x) and F (x). This curve is

1
shown again in Figure 2.2.

Supposing that P is any point on the curve with co-
ordinates (F,Fl), the line from this point perpendicular to
the egalitarian linelcalled n, would have the length
l/JE (F-Fl), and the segment of the egalitarian line from
the origin to n, called fl, would have the length l/fE (F+Fl).

The equation of the Lorenz curve in terms of fl and n could

now be written as:

69

 

 

P(F,F1)

 

 

Figure 2.2-4 Hypothetical Loren: Curve

where

the c1

where

tion,

’D

n
U)

‘ {~-

tight:

Estita

.3.
Q:+QY~H'
"'c

'5‘

CM
‘u

MCi) f.

70

n = f(fl). (l)
whereilvaries from zero to J3. Specifically,the equation of

the curve proposed by Kakwani and Podder is of the form

n = afia (J3 — fl§, a >o, a > o, and B > o, (2)
where a, a, and B are the parameters to be estimated.
From grouped observations of a given income distribu-

tion, it is possible to obtain r and yt as the consistent

t
estimators of fit and nt, reSpectively, where
= P + q
rt t t
{E
and
= P ‘ q
Yt t t,

«[3

where pt and qt are the observed cumulative proportions of

income receiving units and the observed cumulative proportion
of the total income of those units, respectively. The para-
meters of the chosen Lorenz curve function (2) can then be
estimated from the following log-transformed linear regres-
sion equation:

log yt = a + a log r + 8 log ( J3 - rt) + wlt’ (3)

t

where a = log a, and wlt = random disturbance.

When the parameters are known and the Lorenz curve
determined, the Gini concentration ratio can be estimated

from the following integral:
.5
G = 2 (I) f(‘fl) d1},

which for the Lorenz curve of Equation (2) is

9\

 

C!»

A

.t t
A V

b
‘\

ta

A x.
A:

a! ‘

phd

M.

71

f2-
G=2éafia(\/_2--fl53 an

2a< (3)1” +8 B(l+0L, 1+8),

where B(l+a, 1+8) is the Beta function, whose value could be
obtained from the Beta function table or could be recomputed.

Of course, the parameter fl is to be replaced in actual
computation by p. As expected, the Gini concentration ratio
obtained from the Kakwani-Podder (KP) method shows a greater
degree of income inequality than the estimate from the TA
method.

Some may question the advantage of using the compli-
cated KP method over the simple TA method in computing the
Gini concentration ratios, especially when the income distri-
bution observations are as few as five to seven, as in the
case of Thailand in 1963 and 1969. It might appear that the
new estimate is not worth the effort. Since the estimated
Lorenz curves for both years have an extremely high goodness
of fit and very low standard errors despite few observa-
tions, and since the estimation procedure is greatly facili-
tated by high-speed computers, there is good reason to se-
lect the KP method in computing the Gini concentration ratios.
What is more important is that the KP method makes the com-
parison of the Gini concentration ratios between 1963 and
1969 much more compatible and more reliable because the prob-
lem of having to approximate and then compare the ratios from
two different observations with different frequency distribu-

tions of households in the same income class is eliminated.30

that p
somewh.
from l<
in Fig
centrat
1963 an
these 17:

0f inco;

 

 

72

At the end of the previous section it was mentioned
that perhaps the 1969 money income distribution might be
somewhat more equal than the same distribution in 1963 simply
from looking at the free-hand drawing of the Lorenz curves
in Figure 2.1. This belief is confirmed when the Gini con-
centration ratios have been estimated at .5627 and .5550 for
1963 and 1969, respectively. Also, the Gini coefficients of
these magnitudes are regarded as depicting a very high degree

of income inequality in the country.31

Adjustments in the Income Concept

 

An important question arises: How complete and re-
liable is the estimate of money income distribution presented
above? An honest answer would be that it is not as complete
and reliable as one would wish. Certainly, it could not be
compared to the well-known income concept used by Joseph
Pechman and Benjamin Okner in their tax incidence study in
the United States,32 and there is room for improvement.

It is appropriate that the income concept of Pechman
and Okner should serve as a model which this study strives
to emulate. After careful investigation, it was found that
various adjustments could be made to the Thai money income
estimate to arrive at the equivalent of "adjusted family in-
come" which Pechman and Okner used as the final income base
in their tax incidence estimation. In this study, a similar
income concept will be referred to simply as "adjusted in-

come" in contrast to the previously obtained "money income."

This adju
(2) nonmo
tained ea
this inco
nition of
statistic

jUStment

I
timates f
Character
naticmal
tia'flates f
to be tak
in this s
PErSOnal
income es

S6v966.5

73

This adjusted income will be composed of (1) personal income,
(2) nonmoney income or income in kind, (3) net corporate re-
tained earnings, and (4) indirect taxes.33 When completed,
this income concept will be as close to the Simons-Haig defi-
nition of income as one possibly could be under the present
statistical conditions. In the following sections, the ad-

justment procedures will be discussed in some detail.

Personal Income

 

It was pointed out earlier that the money income es-
timates from the 1962-1963 and 1968-1969 surveys had the
characteristics of personal income as defined in terms of
national income accounting. But if the personal income es-
timates from the national income accounts of Thailand were
to be taken as reliable references, the money income estimate
in this study would, to a great extent, understate the true
personal income of households. In 1963 and 1969, personal
income estimates from the national income accounts were Baht
56,966.S million and Baht 102,615.0 million, respectively.34
The money income in the same period was estimated in this
study at Baht 34,491.2 million and Baht 64,652.0 million
leaving the underreported personal income of Baht 22,475.3
million in 1963 and Baht 37,963.0 million in 1969. This is
quite understandable and in fact to be expected because in-
come information obtained from surveys is often incomplete
due to the forgetfulness or the attempt to cover up actual

income on the part of. the individuals being interviewed. The

next imp<

ed income

this incc

90 percer

group wou
group . 31

taing , th

30719 POsi1

 

74

next important question is how to deal with this underreport-
ed income.

First, it was arbitrarily decided that 10 percent of
this income be allocated to the top income bracket alone and
90 percent to all income brackets according to the pattern of
the distribution of money income in each year. One might ask
whether the extra 10 percent allotted to the highest income
group would over—estimate the actual income position of that
group. Most likely the answer would be negative. If any-
thing, this figure is even insufficient to cover the real in-
come position of the top income group since individuals in
this group are likely to receive their income from various
sources, some of which might not be accounted for at all.

At any rate, this percentage was chosen because it would give
a reasonable bias toward a more logical distribution of in-
come. The distribution of this underreported personal income
in absolute terms is presented in column 2 of Tables 2.8 and
2.9.

The already determined money income plus underreported
income should be equal to the personal income estimate of the
national income account. However, further adjustments are
needed owing to the apparent underestimation of the imputed
rents of the owner-occupied homeowners given in the official
statistics.

In the national income accounts,both actual and im-
puted rents of households were presented under the industrial

category of "Ownership of Dwelling" in the gross national

product ta
hm census
census, re}
each month
rent class
mnunted tc
Baht 175 a
merage rem
rural avera
holders are
reported as
houses in t]
d’w‘Ellings a]
Mar-0cm“
Gillivalent i
We of 170118
tainly Can 13

with

f
.01‘ Urban ho

 

75

product table. However, the recent 1970 population and hous-
ing census casts serious doubt on these estimates.35 In that
census, renters were asked about the actual rents they paid
each month, and the frequency distribution of the renters by
rent class was recorded. The average rent paid by a renter
amounted to about Baht 263 a month for an urban dwelling, and

36 While the urban

Baht 175 a month for a rural dwelling.
average rent was credible and indeed quite acceptable, the
rural average rent was not; it was too high. Most rural house-
holders are farmers who own their own homes. Those who were
reported as rural renters in the census are likely to rent
houses in the sanitary districts or village centers, where
dwellings are more similar to urban rented houses than rural
owner-occupied homes. If one assumes that the actual rent is
equivalent in value to the imputed rent regarding the same
type of house, then the urban rent figure from the census cer-
tainly can be used, but not the rural rent figure.

With the above assumption, the total imputed rents
for urban households would be Baht 1,470.2 million per year,

compared with the official figure of only Baht 934.9 million.

Total rents, actual and imputed, for urban households alone

 

would be Baht 2,405.1 million, or 93 percent of the total

value added for ownership of dwellings for the whole country
given by the national accounts in 1969. In short, one is faced
with a choice of accepting the estimates of the national ac-
counts, which is likely to understate the actual rental income,

or accepting the rent figure computed from average urban rent

from the 1970
obviously is
In so
be one-fourt‘r
per month or
occupied rura
Puted rents
3,876.9 mill
the figure f
S’lbtracting
the amount ,
The
tents to Va
5110Cation

dCCoIding 1

76

from the 1970 population and housing census. The latter choice
obviously is preferred.

In so doing, the average rural rent is then assumed to
be one-fourth the amount of the urban rent, or about Baht 65
per month or Baht 730 per year. With the total owner-
occupied rural houses numbering 4,970,356, the estimated im-
puted rents for rural households for 1969 would amount to Baht
3,876.9 million. Together with the total urban imputed rents,
the figure for the whole country becomes Baht 5,347.1 million.
Subtracting the official total rent of Baht 2,570.0 million,
the amount of understated rent would be Baht 2,769.1 million.

The next question is how to allocate these imputed
rents to various income brackets. There appears no better
allocation rule than to assume that such rents are distributed
according to the distribution of income patterns in urban and
rural areas, the reason being that the income positions of the
owners, to a great extent, determines the quality (and value)
of the houses they own.

Lack of data in 1963 forced an assumption that the
rent pattern was the same in 1963 as in 1969, but the level
of average rents was deflated 15 percent, to about Baht 224
per month for an urban house and Baht 55 per month for a rural
house. The proportions of rented to owned houses in both
urban and rural areas also was assumed to be the same, that
is, about 67 percent of urban families owned their own homes,
while about 97 percent of rural families were owners. The

number of owner-occupied houses in 1963 was estimated at

346,945 3‘“
tively' M
rents! the
for urban Z
holdS- Th
2,700.4 mi

same way a

is Shotw'n i}

The
cannot be 0'
the majoritj
part of the
produced for
care in kind

the lower in

some in kind

 

77

546,945 and 4,036,170 in the urban and rural areas, respec-
tively. Multiplying these figures by the average imputed
rents, the total imputed rents become Baht 1,824.6 million

for urban households and Baht 2,663.9 million for rural house-
holds. The underreported rent, which was estimated at Baht
2,700.4 million, was then distributed to households in the
same way as was done in 1969. The result of the distribution

is shown in column 3 in Tables 2.8 and 2.9.

Income in Kind

 

The importance of income in kind in an income base
cannot be overemphasized. In underdeveloped countries, where
the majority of households are in the farming sector, a major
part of their income is often in the form of food grown or
produced for their own consumption. By including this "in-
come in kind" in the income concept, the income positions of
the lower income brackets are considerably improved. But in-
come in kind is not necessarily restricted to the lower in-
come groups; the higher groups could also receive income in
kind in the form of fringe benefits, living amenities, stock
options, real estate appreciation, and so forth.

There is no published information available at present
on income in kind distributed by income classes. Fortunately,
the 1968-1969 socio-economic survey did include a question
about the goods and services that were produced at home or were
received free. Although the responses to this question were
not reported in the published statistics, they could be ob-

tained from the original data tapes. Again, Dr. Oey Astra

Meesook wa
tion from
tinated va
ceived fre
C. Total
her of hou
for 1969 i:
The
Ageneral I

total" incc

Onthe 1969

tEChni

911e,

 

78

Meesook was instrumental in extracting this important informa-
tion from the raw data. The average income in kind or the es-
timated value of goods and services produced at home or re-
ceived free by families by income classes is shown in Appendix
C. Total income in kind is estimated by multiplying the num-
ber of households by the average income in kind. The result
for 1969 is shown in column 4 of Table 2.8.

These 1969 data can be applied to the 1963 situation.
A general methodology requires that (1) some relationship be-
tween money income and income in kind in 1969 be established,
and (2), by assuming that the same money income/income in kind
relationship also existed in 1963, some adjusting factors
which reflect this relationship be applied to the money income
in 1963 to arrive at the income in kind for that year. Spe-
cifically, the method used by Meesook was to regress the 1969
"total" income, defined as money income plus income in kind,
on the 1969 money income using the least-square regression
technique. The regression equation took the following form:

yt = a + me + u,

where Yt and Ym are the average total income and the average
money income of households, respectively, a and b are estimat-
ing parameters, and u is the stochastic disturbance. The re-
gression is carried out for each urban and rural area in each
region, and the results obtained are series of statistical
estimates of average total income in each area in each region

as a function of the average money income in that area and

region (See Appendix C).

 

F0
for 1963 i
total inco
tracting t
total inc01
classified
tain a nat:
weights are
income in k

in column 4

One
i395 should
real as wel:
Shares in C<
with the Sir.
in the Chap:
Portant, I.
hiS'her incop
tangy in bu:

 

Sec '
duty ma

   
 
 

StatistiCal
of T181; COrv

until atte»

H'v
has are S‘

79

For each linear estimate the average money income
for 1963 is then plugged in to give the estimated average
total income for a particular area and region. Simply by sub-
tracting the average money income from the estimated average
total income, what remains is the estimated income in kind
classified by size, location, and region for 1963. To ob-
tain a national income-in-kind figure, some distributive
weights are applied to the above results. The 1963 average
income in kind by income class for the whole country is shown

in column 4 of Table 2.9.

Net Corporate Retained Earnings

 

One of the reasons that net corporate retained earn-
ings should be included as income is that they constitute
real as well as potential consumption power of those who hold
shares in corporations. This reasoning is also consistent
with the Simons-Haig definition of income mentioned earlier
in the chapter. This portion of income has become quite im-
portant. In Thailand, a growing number_of people in the
higher income classes have begun to invest their surplus
money in business corporations, as is evidenced by a tremen-
dous expansion of private and public companies and increased
security market activities during the past decade. But the
statistical information on the net savings of corporations
in Thailand is still very imperfect, and the available record
of net corporate savings is certainly too low. Nevertheless,
until attempts to gauge the true extent of net corporate sav-

ings are successful, one must be content with the existing

data.

The
the national
profits afte
ifllion, Bah
zen enterpr
efprivate Cc
ntretained
Such Baht ll
tions and Bat
tions and coc

Ideal
Wild be the
C'iasses. Thi
inThailand,
53m. were us

1- Savi

Tota

(a)

(b)

 

80

data.

The source of data for corporate saving is found in
the national income accounts.37 In 1963 the net corporate
profits after taxes and dividends were estimated at Baht 819.7
million, Baht 551.1 million of which were savings of govern-
ment enterprises and Baht 264.6 million of which were savings
of private corporations and cooperatives. The total amount of
net retained earnings for 1969 was Baht 2,358.0 million, of
which Baht 1,226.2 million were the savings of public corpora-
tions and Baht 1,131.8 million the savings of private corpora-
tions and cooperatives.

Ideally, the allocation base for this kind of income
would be the distribution of investment income by income
classes. This statistical series, however, is not available

in Thailand. The following arbitrary allocation rules, there-

fore, were used:

 

 

 

 

 

 

1. Savings of government enterprises 1963 1969
Total amount (Baht million) 555 1 1226.2
(a) 25 percent allocated on a per
family basis 138.8 306.5
(b) 75 percent allocated on a
total expenditure basis 416.3 919.7
2. Savings of private corporations 1963 1969
Total amount (Baht million) 264.6 1131.8
(a) 80 percent allocated to the
top income bracket 211.7 905.4

(b) 20 percent allocated by
income distribution 52.

\D

226.4

 

The
simple- SE
theoretical
of these we
conpanies a
could be US
benefit 11011
private COmj
holds most <
being held 1
the renainir
cane distrit

The
ings by inco

and 2.9.

The .'

e. .

._.e mclusior

hated. The
A c

c?

“021 the 196

“i'TOUHted to

Cited.

But

 

inc 1
Die C1385

o‘ ' -

~ indlrect

3L4, .
usalned fr,

81

The reasoning behind the above allocations is quite
simple. Savings in the hands of public corporations should,
theoretically, be shared by all households. Thus, 25 percent
of these were allocated on a per family basis. But public
companies also operate like private firms, and their savings
could be used to expand or improve the services which would
benefit households according to their spending patterns. For
private companies, it is assumed that the top income bracket
holds most of the shares of stock; only a small proportion
being held by lower income groups. Hence the allocation of
the remaining 20 percent of private corporate savings by in-
come distributional pattern.

The distribution of the net corporate retained earn-
ings by income classes is shown in column 5 of Tables 2.8

and 2.9.

Indirect Taxes

 

The last adjustment in the income concept involves
the inclusion of indirect taxes in the income thus far esti-
mated. The amount of indirect taxes to be included are taken
from the 1963 and 1969 national income accounts. These taxes
amounted to Baht 6,461.1 million in 1963 and Baht 14,206.9
million in 1969. Thus far the matter is simple and uncompli-
cated. But the method used in allocating these taxes among
income classes is more difficult. The distributive pattern
of indirect tax burdens by size of classes of income must be

obtained from Chapter 3 of this study which deals with tax

 

 

in:

go

ii:

'13:

It,

82

incidence. Then, based on this pattern, these taxes are al-
located to different income classes. The distribution of in-
direct taxes by income classes is presented in column 6 of

Tables 2.8 and 2.9.

The Adjusted Income Concept

 

Now that all necessary adjustments have been accom-
plished, the final step is to total all the income elements.
The distribution of the adjusted income for the whole country
by income classes in 1963 and 1969 is presented in Tables 2.8
and 2.9 The Lorenz curves for the two new distributions of
income are shown in Figure 2.3.

It is clear from Tables 2.8 and 2.9 and Figure 2.3
that the relative income positions of households in lower-
income brackets has improved considerably compared to the dis—
tribution of money income as shown in Table 2.7. For example,
the income share of the lowest income bracket rose from 13.0
percent to 19.7 percent in 1963, and from 3.7 percent to 6.0
percent in 1969. Similarly the share of the top income group
in the adjusted income distribution fell from 33.5 percent
to 30.4 percent in 1963, and from 52.3 percent to 48.5 per-
cent in 1969. In terms of the degree of income inequality,
the Gini concentration ratios for the distributions of ad-
justed income are estimated at .4559 and .4822 for 1963 and
1969, respectively, an overall improvement in income equality
from the distribution of money income. This improvement is,

of course, to be expected. Income in kind alone played a

83

.uxma mom “mousom

 

 

 

 

 

 

 

o.ooa m.mve.mva m.mo~.va o.mmm.~ m.mm¢.m~ H.mmh.m o.mmm.hm o.mmo.em mommmHU Mac
m.me m.mmm.o> m.m-.m m.ome.a m.aoa.m m.anm.a m.amm.a~ h.mbn.mm um>o use ooo.ma
m.m m.omm.m m.>mo.a H.mm o.m~m m.omH H.mam.~ m.mnm.¢ mmm.>a I coo.mH
H.h N.>hm.oa n.mom.a h.mma m.an N.Hmm H.vmh.~ N.mmm.m mmm.va I ooo.~a
o.m H.mao.ma o.w>v.a h.oma m.moo.m m.mm~ m.mma.m o.mmm.m mmm.HH I ooo.m
m.HH >.mom.ma n.>m>.H n.mma m.mao.v m.mHm m.mmm.m m.m>m.m mmm.m I ooo.o
m.HH m.mem.ha m.emm.a m.mom m.mHH.m v.mhm m.>oa.m v.mmm.m mmm.m I ooo.m
o.o m.bmn.m m.o~m «.mva w.mao.v m.~HH m.>hm.a o.amv.~ ooo.m Home:
Amy Any Ame Amy Avv Amv Amy Adv
coo
W... .3 a... rage.
I 85 and 838m I886 969:
mmxme oofiflom a» E mucus 5 mucus mono: flame
menxxa” uomuHoCH mumwmeoU meoocH lymphom Imenoe. mommoao meoocH
Bumsflea
meoocH Hmcomnmm

 

 

 

 

 

 

 

Assam no 8832 5 3528
$3 .8826 9065 so .8588. 8.83 B .285 EEHE mo aofisfihflouéd flame

 

“(when

a» H U uric. V: F

\‘Ih

Fr!-
I;
”IO:

84

.88» 0mm 588m

 

 

 

 

 

 

 

o.ooa e.oma.ms H.eee.e s.mem n.mom.ee e.oos.~ m.mev.- ~.Hme.em mommmeo flea
m.om ~.e~m.m~ e.omm.a m.~mm m.amm e.~em m.emo.m H.mem.HH um>o use ooo.ma
m.HH e.mmm.m H.mms ~.me m.~me e.mem m.mve.~ s.aem.e mmm.~e I ooo.me
o.e~ m.m~s.ma o.eee.e m.mee m.mom.a m.m~s H.0mm.m e.mmo.m mmm.HH I ooo.e
m.ee m.em~.ee H.oeo.e e.eoe H.meo.~ N.emm m.mmm.~ a.~mm.e emm.m I ooo.m
m.me m.eee.me m.mee.e m.nsa e.eam.e H.H~m e.s~e.~ H.ome.e ooo.m sous:
Amy Ass rev Amy rev Amy Ame rev
9505
mmwm unease. oommmmm emphases m ”:H
mmcaFHQm Jumps:
. ease ca magma mucus. Aeneas
8“.an 853mm 5 Iugng 3 mafia awn—30 EH
985 §H§H 38088 285 . Iumzflea
Eamon? 82
9505 gonna

 

 

 

 

 

 

 

icon m0 963.3”: 5 muggy
moma .mmmmmau 9505” an .mpcgmam 285” an .9505 “humane mo coflnflfimwid manna

85

 

 

 

 

 

 

8
I".
G1963 ' '“559 f
a l
8 I — — -_ 1963 ,I
I
1969 I,
be
o
5 o a??? /
G ‘0 1i Q) I!
H do ,’
“3 6&0 I”
R Q0 ’1
g 0‘9 1’
3 3 . co /
,5 ’v" ,/
a ,/
U ’1
[I
8 i’ I/’ I
O 20 “0 60 80 100

Cululative % of Families

Figure 2.3--Distribution of Adjusted Income in Thailand: 1963 and 1969

W”

H
“in- VI

tnLI'

ist

Loki.

86

very important part in raising the relative income positions
of the lower income brackets. Various other income adjust-
ments also helped smooth out greater disparities in the dis-
tribution of cash income.38
One striking fact which must be emphasized, however,
is that despite the improvement in income equality in the
adjusted income distribution in both 1963 and 1969, the dis-
tribution of adjusted income in 1969 is more unequal than the
same distribution in 1963, as can be seen from the new set of
Gini coefficients presented above. This is also evident from
Figure 2.3, where the Lorenz curve of the 1969 adjusted in-
come distribution lies clearly outside the 1963 Lorenz curve
and farther away from the line of perfect equality. As the
distributions of money income in 1963 and 1969 are practically
the same, and with the new finding that the adjusted income
distribution in 1969 is more unequal than the adjusted in-
come distribution of 1963, one can confidently say that the
income distribution of Thai households has deteriorated rather
than improved. This is the case despite a marked increase in
the rate of growth of the gross national product over the six
years from 1963 to 1969.39
Two tentative reasons can be offered to explain this
phenomenon. First, as income in kind is accrued to rural
households relatively more than urban households and is the

most important income element that substantially raises the

income positions of the lower income brackets, the fact that

87

adjusted income distribution is worsened in 1969 compared

to 1963 could mean that lower income households are deprived
of opportunities to receive this nonmoney income. This is
possible if a large number of rural poor families imigrate
into cities or urban areas and face the usual unemployment
problems and other living hardships. At the same time,they
are foregoing their usual income in kind in the form of food
grown for own consumption. There is no study to substantiate
the exact extent of rural to urban migration in Thailand,but
the existence of such migration has been a well-known fact in
the past ten years.

Second, the urban wage and salaried earners,particular1y
the self-employed modern entrepreneurs,have become much better
off economically also in the past ten years relative to the
self-employed farmers. This could contribute to the situa—
tion where the total income positions of the upper income
brackets are rising too fast for the lower income brackets
to catch up, thus widening the existing income gap instead
of narrowing it.

As has been discussed, two income concepts are avail-
able to this study: money income and adjusted income. Which
of these should be used? Despite its extended coverage, one
must always realize that the adjusted income concept is more
or less an artificial figure incorporating various elements

(if arbirtrary adjustments. As such, it should not be taken
as the sole legitimate income concept to be used in a budget
incidence study. The best policy seems to be to use both

17WC311ey income, which is derived more or less objectively from

 

(D

"1

88

the surveys, and adjusted income, which is more artificial
but more complete, as bases for this incidence study.

The use of both income bases in this study has an-
other advantage: It gives the range within which the income
redistributional effects of the public sector can be examined.
To elaborate, suppose that the burden of the total tax sys-
tem on a representative household in income class A amounts
to 50 percent of its money income compared to a burden of 25
percent for a family in income class B. On a money income
basis alone, it would appear that households in income class
A are taxed twice as heavily as households in income class B.
However, when an adjusted income base is used to compute the
relative tax incidence, it is found that households in income
class A bear a total tax burden equal to 20 percent of their
adjusted income, whereas households in income class B bear a
total tax burden equal to 15 percent of their adjusted in-
come. In this case, then, the relative tax burden that falls
on households in income class A is only about 33 percent high-
er than the relative tax burden of households in income class
B. The "band" of differential tax incidence, therefore,
ranges from 33 to 100 percent depending upon the income con-

40

cepts used. Thus, the use of adjusted income as another in-

come base gives a more accurate picture of the incidence of

fiscal activities than does using either of these income bases

alone.

d

 

89

Summary and Conclusions

 

This chapter began with a general discussion of the
conceptual and statistical difficulties facing a researcher
attempting to estimate the distribution of income in an under-
developed country. Despite many formidable obstacles, such
an estimate is possible, and if carefully done it can be quite
reliable. In Thailand the workable sources of data are two
sample household surveys conducted in 1962-1963 and 1968-
1969. From these two surveys, and the information on the
number of households from two population censuses, the money
income distribution for each year has been estimated. The
result shows that, using the Gini concentration ratio as the
chosen index of income inequality, the money income distribu-
tion in 1969 is slightly less unequal than that in 1963, but,
overall, both income distributions were very unequal by any
standard.

To build a more complete income base, various income
adjustments have been carried out. First, the already es-
timated money income was used as the basic element of personal
income in the national income accounting sense. It was then
adjusted for the underreporting of income and for imputed
rents. Other major adjusting elements included income in
kind, net corporate retained earnings, and indirect taxes.

The final comprehensive income concept, termed "adjusted in-
come," was shown to be similar in principle to the "adjusted

family income" concept used by Joseph Pechman and Benjamin

90

Okner in their study of United States tax incidence for 1966.
The new adjusted income distribution shows considerable im-
provement in equality over the money income distribution be-
cause the income of the lower income brackets has been augment—
ed, mostly through income in kind. The distribution or ad-
justed income in 1969, however, has become more unequal since
1963. Both money income and adjusted income will be used as
bases for computing the effective rates of tax burden and ex-
penditure benefits discussed in the next two chapters.

In conclusion, it should be stated that although the
income estimates obtained in this study are the result of a
very careful investigation, they cannot claim to be the de-
finitive estimates of income distribution in Thailand. The
basic survey data used for computing the income distribution
have many defects. For example, income classifications were
inconsistent between rural and urban households, income is
considerably underreported, and nonmoney income does not re-
ceive adequate attention. While the income of the lower and
middle income classes is presumably adequately tabulated, the
income of the upper income groups probably is not. If any-
thing, the present income distribution estimate probably does
not overstate the income position of the lower income groups;
rather, the income positions of the upper income groups, es-
pecially the highest one in this study (Baht 18,000 and over),
is more than likely biased downward.

A final remark as to the nature of income distribution

thus far estimated in this study should be made. This income

91

distribution is known as a "pre-fisc" or original income dis-
tribution; it shows the relative income positions of house-
holds by income classes before the effects of taxes of all
kinds are deducted and the real benefits of government expendi-
tures converted into money terms are added. It is the distri-
bution which would have prevailed if the Thai government were
to substitute the present tax and expenditure programs with

an income-proportional tax and expenditure program.

In anticipation of what is to follow in Chapters 4
and 5, it should be noted here that the term "post-tax" in-
come distribution is reserved for the pre-fisc distribution,
where the absolute tax burdens have been deducted. By the
same token, a "post-benefit" income distribution is the re—
sulting income distribution when the monetized benefits of
government expenditures have been added to the pre—fisc in-
come distribution. Finally, the "post-fisc" income distribu-
tion is obtained when the net benefits of government fiscal
activities--that is, expenditure benefits minus tax burdens—-
are added to the pre-fisc income distribution. The extent
of income redistribution of the entire fiscal system, there-
fore, is seen through comparison of the pre-fisc and the

post-fisc income distributions.

FOOTNOTES

1This chapter deals strictly with the problems and
procedures of the estimation of income distribution; it is
not concerned with the theories of income distribution or the
explanation as to why the income is so distributed in a par-
ticular way. Neither is it concerned with the sources of
income disparities within various sectors of the economy.
These are certainly important, but they are beyond the scope
of this study. For a concise summary of the theories of
income distribution see, for example, Martin Bronfenbrenner,
Income Distribution Theory_(Chicago: Aldine Atherton, Inc.,
1971), Chapter 3; Jan Pen, Income Distribution (New York:
Praeger, 1971), Chapter 6.
- 2Henry C. Simons, Personal Income Taxation (Chicago:
University of Chicago Press, 1938), p. 50.

 

 

 

3Robert M. Haig, "The Concept of Income" in The
Federal Income Tax, ed. R. M. Haig (New York: Columbia
University Press, 1921), p. 26.

4That the sources of data are divided into four cate-
gories does not imply that each source is entirely independ—
ent of one another. Indeed a combination of techniques using
some or all of the above data sources is common and has been
used by many investigators especially in the situation where
no single data source is sufficient or is decidedly better
than the others. Research ingenuity is required here to see
which information is needed and how to get it. Two good ex-
amples of the use of these techniques are the studies by
Milton C. Taylor and associates and by Charles E. McLure, Jr.
on the Colombian income distribution in 1962 and 1970 re-
spectively. In these studies the sources of data included,
among other things, the tax rolls, the national income accounts,
the urban population surveys, surveys of urban industrial
workers, the agricultural household surveys, the estimates
of the labor force, and the population census. See Milton C.
Taylor e31§1., Fiscal Survey of Colombia, a Report Prepared
under the Direction of the Joint.Tax Program, (Baltimore:
published for the Joint Tax Program of the Organization of
American States and the Inter-American Development Bank by
the John Hopkins University, 1965), and Charles E. McLure,
Jr., "The Incidence of Colombian Taxes, 1970," Discussion
Paper No. 41, Rice University, summer, 1973 (Mimeographed).

 

 

92

93

5 . . .

Harry T. Oshima, "Income Inequality and Economic
Growth: The Postwar Experience of Asian Countries," Malayan
Economic Review, 15 (October, 1970), pp. 7-41.

6United Nations, Economic Commission for Asia and the
Far East, The 1971 Economic Survey of Asia and the Far East
(Bangkok, 1972), Chapter 3; Idem., Interregional Trade Pro-
jection, Effective Protection and Income Distribution. Vol-
ume III: Income Distribution (Bangkok, 1972), part 6.

7Udom Kerdphibule, "Income and the Distribution of
Income of Agricultural Sector," Thai Journal of Agricultural
Economics, 11 (December, 1972), pp. 1-24.

 

 

 

8For example, the data that Kerdphibule claimed to be
1970 data were actually the 1968-1969 data. The treatment of
the mean incomes of the two open-ended brackets in 1963 were
rather crude. One was also surprised at the acceptance with-
out any reservation of the distribution of income of rice
farmers in the Central Plain even when the data were gathered
from the survey of a single village of less than 300 families.

9William A. McCleary, "Sources of Change in Distribu-
tion of Income in Thailand, 1962/63 and 1968/69," Discussion
Paper No. 26, Faculty of Economics, Thammasat University,
August, 1972 (Mimeographed).

loDr. Oey Astra Meesook, who is probably the best known

authority on income distribution in Thailand today, has con-
sented to supply some of her unpublished income distribution
data for use in this study.

11The information from the reports of this Survey would
be used for the fiscal year 1963. Also for comparison pur-
poses, the town-village classification would be taken to mean
the same as urban-rural classification.

12In the 1968-1969 household expenditure survey, only
the average annual money income figures were given, and they
were consistent with the income class limit (no class jumping).
In most cases, the average income of each class was very close
to the midpoint of that class.

13There is no sacrifice in the analysis, however, be-
cause the aim of this study is to deal with fiscal incidence
at the national level only. Only the national distribution
of income would be needed.

14Specifically, the proportion of the average income
of the bottom class to that of the next higher class in 1969
is about .40, and the proportion of the average income of the
second top bracket to that of the top bracket is about .45.
However, the latter figure represented the urban income only.

94

For the country as a whole, this proportion would be less be-
cause it was likely that the average income in the rural top
bracket would be slightly less than that in the urban top
bracket. A ratio of .46, then, would be more appropriate.
When these ratios were used to multiply the midpoints of the
second bottom and the second top brackets in 1963, the re-
sults were 1800 and ll7,500--the mean incomes of the bottom
and the top income classes in 1963 respectively. Bracket by
bracket, the 1963 average income was larger than that in 1969
which should not be. But no adjustment was made for fear that
the distribution pattern in 1963, which was of interest here,
might be affected.

15Discussions on the Pareto distribution could be
found in Jan Pen; Lawrence R. Klein, An Introduction to Econo-
metrics (Englewood Cliffs, N. J.: Prentice-Hall, 1962); H.
Lydall, The Structure of Earnings (Oxford: Clarendon Press,
1968).

 

 

16To estimate the minimum income, Y in the Pareto
distribution, the quantile method was used.0 This method re-
quires that, first, two probability levels, P1 and P2 be
chosen which correspond to two levels of income, y and Y2.
P1 and P2 are simply the percentages of the total households
having income less than y, and y2 respectively. Then from a
form of the Pareto function:

b b
p. = 1- [12] and p. = I-[¥_<z]
Y1 Y2

the parameter b was first solved through:

b = log (1 - P1) - log (1 - P2)
log yz - log Y1

 

When b was substituted into one of the above equations,
yo could be solved.

Using this method for the 1963 data, b was estimated
to be 1.4672 and the minimum income was Baht 1,918. This
parameter could also be estimated through the least square
method from the linear transformation of the Pareto function,
thus:

log Ny = log A - b log y
which gave a slightly lower value of b at 1.4524.
For the mean income of the top income bracket, the

Pareto function could be used first, to compute the total in-
come of this bracket from the equation:

.lt

IBM 1

lJlH

:Ul

I1

I!

H

IOIV

flIIli

 

I‘M \ ~51

IAIN

95

a
y = I I dNy I y
to ———
p th dY
(1
l-b
_ -b - l _ _ 95 (y )
— y! ybay dy — l-b ft

ft

where ytop is the total income of the top Open-ended bracket;

dNy/dy is the number of families having income y; and yft

is the floor income of the top income bracket. When the

total income is known, a division by total families in this
bracket will yield the average income for this bracket. For
1963 such an average income of the top bracket for the whole
kingdom was estimated at Baht 248,083. This extremely large
average income of the top bracket, coupled with the also very
large average income of the bottom bracket was unacceptable

in this study. Subsequently, they were abandoned. For the
discussion on the estimation methods of the Pareto distribu-
tion, see Richard E. Quandt, "Old and New methods of Estima-
tion and the Pareto Distribution," Metrika, 10 (1966), pp. 55-
82; William R. Cline, Potential Effects of Income Distribution
on Economic Growth (New York: Praeger, 1972), pp. 58-64.

17See, for example, Mary Jean Bowman, "A Graphical
Analysis of Personal Income Distribution in the United States,"
American Economic Review, 35 (September, 1945), pp. 607-628.

18See Cline, pp. 62-63. It must be mentioned here
that perhaps the average income of the top bracket derived
from the Pareto distribution did not overestimate the true
income but the average income from the survey underestimated
the true average income of this group. This is quite under-
standable because either the income of the top group could
exist in many diverse forms which were not accounted for,
or the survey did not get a fair representation of the wealthy
families. Since there is no reason to overhaul the entire
data and build a new artificial distribution according to
the Pareto function, the survey figures would still be used
for reference.

19Thailand,National Statistical Office (NSO), Report
of the Socio-Economic Survey 1968-1969. (Bangkok, 1974).

20

 

 

 

 

 

Ibid., p. 25.

21Thailand NSO, Report of the Household Expenditure
Survey 1962/63: Whole Kingdom (Bangkok, 1966). p. 34. Here-
after, the term family will be used interchangeably with house-
hold.

 

 

96

22See Appendix B for details on how the total number
of households by income class in 1969 was obtained.

3See Oey Astra Meesook, "Income Distribution Statis-
tics for Thailand, Malaysia and Indonesia," paper prepared
for the Princeton-Brookings Income Distribution Project,
August, 1974 (Mimeographed), p. T16. One of the results of
this official adjustment would make the distribution of in-
come appear less unequal.

24M. O. Lorenz "Methods for Measuring Concentration
of Wealth" Journal of the American Statistical Association,
9 (June, 1905), pp. 209-219.

25As one may not have adequate observations to plot
as many coordinates as one wishes for the purpose of a smooth
curve, this curve was drawn using the freehand drawing tech-
nique joining the available observation points aiming at the
smoothest curve possible.

26Apart from the standard statistical measures of vari-
ations which are found in most textbooks on statistics, see
Amartya Sen, On Economic Inequality, (Oxford: Clarendon Press,
1973); O. Elteto and E. Frigyes, "New Income Inequality Meas-
ures as Efficient Tools for Causal Analysis and Planning,"
Econometrica, 36 (April, 1968), pp. 383-396; A. B. Atkinson,
"On the Measurement of Inequality," Journal of Economic Theory,
2 (1970), pp. 244-263; Harry T. Oshima, pp. 7-41, Henry Theil,
The Information Theory, (Amsterdam: North Holland, 1967),
Chapter 4.

27See C. Gini, "Measurement of Inequality in Incomes,"
Economic Journal, 31 (1921), pp. 124 - 126; also his "On the
lfleasure of Concentration with Especial Reference to Income
and Wealth," paper delivered before the Cowles Commission in
1936.

 

 

 

 

28See D. G. Champernowne, "A Comparison of Measures
of Inequality of Income Distribution," Economic Journal, 84
(December, 1974), p. 787.

29N. C. Kakwani and N. Podder, "On the Estimation of
Lorenz Curves from Grouped Observations," "Efficient Estima-
tixan of the Lorenz Curve and Associated Inequality Measures
frwom Grouped Observations," International Economic Review,
14, (June, 1973), pp. 278-291; Idems, World Bank Development
Research Center, Discussion Paper No. 10, October, 1974
Utimeographed). The method used in this study was taken al-
fMMst exclusively from their second article. It must be grate-
fully noted here that the use of Kakwani and Podder method
was suggested and assisted to the author by Drs. Byron Brown
and Daniel Saks, two of the members of his thesis committee.

 

me Iwnu p. .

u“

6.5 I7. I

out

«an .qu Isl» n\u

mm

NF! 1 fl“ 1] I.

[In III) .1 a 6:5

\c ith

FIHL

Ifils Fee

«a!

.n-\ \NH

 

97

30Hereafter, the KP method will be used to generate

not only the Lorenz curve of pre—fisc income distribution,
but also Lorenz curves of post-tax, post-benefit, and post-
fisc income distributions. The Gini concentration ratios
will likewise be estimated. The meaning of these terms will
be explained at the end of the chapter.

31In the United States and most Western economies, the
Gini coefficients usually range between .3 and .4. Any coun-
try with a Gini ratio higher than .4 is usually regarded as
having more than "acceptable" level of income disparities.
However, such simple-minded comparison of Gini coefficients
among countries is fraught with danger due to differences
in the income concepts, the size of households, the charac-
teristics of heads of households, the time periods under
study, the reliability of income estimates, and even in the
methods used to compute the Gini coefficients. Therefore,
such comparison will not be attempted here.

 

 

32Joseph A. Pechman and Benjamin A. Okner, Who Bears
the Tax Burden (Washington, D. C.: The Brookings Institution,
1974).

33

There are certainly more items of adjustments in
the money income concept such as the imputed interest, the
unshifted portion of the corporate income tax, the contribu-
tions to the social security system, but for Thailand, such
adjustments could not be made because (a) the imputed interest
data are not available; (b) the corporate income tax was re-
latively small, and since the standard incidence assumption
was a total forward shifting, it is of no concern here; and
(c) Thailand has yet to have a social security system. It

is admitted, however, that the effect from the failure of
these adjustments, if any, would be the understatement of the
upper income classes, particularly the top income group.

34Thailand, National Economic and Development Board
(NEDBL National Income of Thailand, 1968 ed. and 1972-73 ed.
(Bangkok, 1970 and 1974).

35See Thailand, NSO, Report of the 1970 Population
339 Housing Census, (Bangkok, 1974), Table 5 of Housing
Section.

36These figures were derived from multiplying the mid-
Emtints of each rent class by the total number of households
in. that class. The sum of these--Baht 78.0 million per month
for urban areas, and Baht 30.8 million per month for rural
areeas--was divided by the renter households of 296,243 families
in urban areas and 176,036 families in rural areas to arrive
at the average actual rents mentioned in the text.

 

 

 

98

37See NEDB, National Income of Thailand, 1968-69 edi-
tion, pp. 10-11.

38This finding is directly opposite from what Pechman
and Okner have found in the United States where the distribu-
tion of adjusted family income is found to have worsened over
the money family income distribution. This contrasting phe-
nomenon, however, is not difficult to explain: In a developed
economy such as the United States, the most important adjust-
ing elements in the income concept are capital gains, imputed
rents and interests which accrue mainly to households in
higher brackets, whereas in an underdeveloped economy such
as Thailand, one of the most important income adjusting ele-
ments is the income in kind which, according to the present
statistical data, enhances the income positions of the lower
relative to the upper brackets. Nevertheless, Dr. Milton C.
Taylor, the major advisor of this researcher, still believes
that with more complete data it could be shown that income
of the top income bracket is still underreported.

39The rate of growth of GNP in Thailand from 1963 to
1969 was estimated at 92.3 percent. See NEDB, The National
Income of Thailand, 1968-69 edition, p. 23.

40In several tax and expenditure incidence studies,
alternative incidence shifting assumptions were used in el-
1ocating the incidence of taxes and expenditures to various
income classes to see the range of differential incidence
as a result of the use of such assumptions. The use of more
than one income base affects the outcome of an incidence
study in the same manner as the use of more than one alloca-
tion assumption. In the fiscal incidence study in Thailand,
however, the use of alternative income bases deserves greater
emphasis because unlike other more-developed fiscal systems
the income distribution statistics are not as reliable, and
as such the use of alternative income bases is more appro-
priate.

 

 

 

CHAPTER III

THE INCIDENCE OF TAXES

Taxation may have different purposes according to the
different levels of government that collect the taxes. This,
of course, refers to the Musgravian tripartite division
of government functions.1 Taxes are collected by the alloca-
tion branch to be used as resources for the provision of
public goods; they are collected by the distributive branch
as a mean to redistribute income toward a certain "prOper"
pattern; and in the stabilization branch taxes serve as a
means to fight inflation or deflation. Musgrave's model
is a conceptually innovative device for understanding the
government's revenue functions, although in practice it is
difficult to see which tax is to serve which purpose. More
than likely, many taxes are collected to serve three purposes
simultaneously.

This chapter deals with the statistical estimation
of "tax incidence," a specialized term meaning the burden of
taxes and other governmental revenues distributed by size
classes of income. Following the example set forth by
Musgrave et 31.,2 the burden of taxation will be defined as
the difference between a taxpayer's actual money income net

of tax and what this income would be in the absence of the

99

100

tax. In other words, a tax burden is synonymous with a re-
duction in the taxpayer's real income. Several issues
involved in the process of tax incidence estimation will be
discussed below. First, certain theoretical problems inher-
ent in a study of this sort must be confronted.I Second,

a decision must be made as to which government revenue
programs will be covered and, third, how each tax is to be
allocated to which income group. The significance of the
empirical results obtained will be discussed at the end

of the chapter.

Theoretical Problems in Tax Incidence Studies

 

Despite the apparent ease of the tax incidence es-
timating procedure, that is, the apportionment and allocation
of the tax burden in money terms among income groups, a tax
incidence study creates numerous theoretical problems.
According to conventional methods used in tax incidence
studies, the tax burden borne by each income class is sub-
tracted from the income of an individual or household in
the same income class to arrive at a new pattern of income
distribution: the after-tax or post-tax income distribution.
To be able to obtain the post-tax income distribution simply
by deducting the tax burden from the original, pre-tax (or
pre-fisc) income levels, one might have to make a strong
assumption that the pre-tax income distribution would remain

unchanged in the absence of the government or the tax programs.

101

This assumption is sometimes referred to as a zero-government
or zero-tax counterfactual assumption.3

An objection to this counterfactual assumption has
been raised by Alan R. Prest, who claims that such an
assumption is so unrealistic that it renders the conven-
tional tax incidence study meaningless.4 This opinion has
been supported by Carl S. Shoup, Richard M. Bird and Luc
De Wulf.5 It is difficult to defend the assumption against
this criticism, since it is quite obvious that the existing
pattern of pre-tax income distribution already has been
influenced by the continuing presence of the government
or the previous tax policies. Furthermore, if all taxes
were abolished or repealed, the price and income effects
would send a wave of adjustments throughout the income
structures of households which only coincidentally might
remain the same as before the repeal.

Nonetheless, Musgrave, who was among the first
economists to undertake a tax incidence study in the United
States, has developed a strong defense.6 He claims that,
instead of assuming a zero-tax system, or "thinking" away
taxes, the adjustments resulting from the repeal of taxes
could be assumed to be distributionally neutral. Or the
"differential incidence" approach could be adopted whereby
the present tax system is assumed to be replaced by a
preportional tax system which yields equal revenues. Post-

tax income distribution then could be compared with that

102

income distribution which would have resulted if the sub-
stituted prOportional taxes were in effect (which,
incidentally, would be the same as the existing pre—tax
income distribution). This differential incidence argument
is very simple yet very convincing. Its use obviates the
need to be too concerned about this particular conceptual
difficulty in conventional tax incidence studies. Indeed,
most tax incidence studies conducted since that of Musgrave
and Associates in 1951, including the present study, expli-
citly or implicitly have adOpted this assumption.

Other problems remain, however. A more important
theoretical criticism maintains that the only prOper pro-
cedure in a tax incidence study is to take account of the
total effects of taxation on relative prices and income.

In other words, these critics favor the general equilibrium
model in analyzing the effects and repercussions from the
imposition of a tax on relative prices and income throughout
the economy. The pioneering general equilibrium tax
incidence model of A. C. Harberger is often cited as an
example of what a proper tax incidence study should be.7

This criticism can be answered on the ground that a
general equilibrium analysis is not empirically feasible.
Although the model offers greater realism and precision than
the partial equilibrium model, which is used in most conven-
tional tax incidence studies, general equilibrium incidence

research is very difficult if not impossible to carry out

103

in practice. The estimation problems are insurmountable.
To adOpt the qualitative method suggested by C. S. Shoup,
that is, to make a logical deduction concerning tax effects
in the general equilibrium setting, is still no better a
procedure than the judicious interpretation method of
conventional tax incidence study based on partial equili-
brium analysis.8 Furthermore, a general equilibrium inci-
dence model is based on many simplifying assumptions, the
validity of which can be seriously questioned. Therefore,
partial equilibrium analysis will continue to be used as
the standard method in calculating the incidence of

taxation.9

A General Methodology

 

The fact that this research in an intertemporal study
of tax incidence almost completely invalidates any criticism
that the zero-government counterfactual assumption might
be implicitly assumed. The main emphasis of this research
is no longer placed on the isolation of pre-tax and post-tax
distributions of income--the main objective of a conventional
tax incidence study--but on the comparison of post-tax
distributions of income between two years. Most critics
of traditional incidence studies admit that "comparisons (of
incidence) over time are conceptually easier,"10 and that

"tabulations purporting to show the distribution of burden

of a total tax system by income classes, under certain

104

assumptions about shifting and incidence, are useful only
as two or more such tabulations are compared, to show
changes in the distribution of the burden when taxes are

11 The intertemporal approach constitutes an

changed."
obvious improvement in methodology: If errors are made in
assigning tax burdens to Specific income groups, it is

quite possible that such errors are constant from one year

to another. An intertemporal study, therefore, removes most
of the conventional criticisms.

In general, the following estimation procedure will
be used here: (1) taxes and other governmental revenues in
1963 and 1969 will be selected; (2) each major tax or revenue
will be discussed as to its shifting assumptions and burden
allocation rules; and, (3) after the burden of each tax is
allocated to various income classes, the total pattern of
incidence will be assessed and the redistributive effects

of government tax policies between 1963 and 1969 will be

compared.

Selection of Taxes

 

Several views must be considered in selecting which
tax or revenue source to study. Up to this point, the term
"tax" as used in "tax incidence," for example, has connoted
the compulsory aspect of government revenue programs. This
usage conforms to the belief that only compulsory items
levied for the purpose of raising revenue for general ex-

penditure should be included, and those fees and sales

procl
rati«
acti‘
calL

bene

105

proceeds from public enterprises should be excluded. The
rationale is to consider only that part of government fiscal
activity which poses a "burden" to households, not the so-

called quid pro quo portions from which specific and direct

 

benefits are derived in exchange for family payments to the
government.

A second point of view argues that it is sometimes
difficult to determine which taxes are based on the benefit
principle and which taxes are not; the distinction between
fees and taxes is not always clear. A case in point is the

payroll tax, which could be considered either a quid pro quo

 

payment or a burden. Moreover, the alleged benefits
received from a government service often are not perfectly
related to the payments, thus creating some burdens. It
also can be argued that many charges do not represent a
payment for a service that would be sold by a private firm:
in the absence of government, for example, road fines also
would probably be absent. This view, therefore, would
include both taxes and fees in a study.

The third viewpoint advances a step farther and
maintains that if the study aims at an overall investigation
of government fiscal activities, that is, the incidence of
tax burdens as well as expenditure benefits, the selection
of taxes should be as comprehensive as possible. Not only
those taxes defined as compulsory revenue sources, but also

all other government revenue sources, such as fees, proceeds

106

from governmental sales and services, and profits from
government enterprises, should be examined. If only some
revenue items are included when the distributive impact of
both taxes and expenditures is being measured, the incidence
would be distorted.

This third position is adopted here because it is
the most apprOpriate for a total fiscal incidence study
such as this one. Consequently, what previously has been
referred to as "tax incidence" now should be interpreted
to include the incidence of taxes and other government
revenue items as well. Henceforth, "taxes" will be used as
a convenient term for actual taxes, public fees, royalties,
and other domestic receipts of the government.

In addition to the general consideration discussed
above, certain modifications must be made because the study
deals specifically with Thailand. First, this study will
be concerned only with the tax burden of Thai residents:
any tax or revenue whose burden falls on foreigners will
be excluded whenever possible. Ideally, that portion of
indirect taxes which is assumed to fall on foreign tourists
would be excluded, but there is no way of determining
which taxes, and in what amount, are borne by foreign
visitors. Therefore, no adjustment has been made. It is
assumed that such taxes are small or do not substantially
affect the distributive pattern of the resident tax burden.12

Second, only that fiscal activity over which the

central government of Thailand has full control is deemed to

107

be of interest here. Government revenues from foreign
grants or loans and other types of technical assistance
are excluded. Of course, these revenues have a redistri-
butive impact on Thai households, but this influence is
assumed to be distributionally neutral.

Third, the question arises as to whether domestic
borrowing should be included in the revenue analysis. If
the study is concerned with tax incidence alone, then the
income distributive effects of internal public debts should
be included. In a total fiscal incidence study such as
this one, however, the question of including internal
public debt could be postponed until both the tax and
expenditure incidence patterns are known.

Table 3.1 presents the total taxes and other
revenues collected by the Thai government from resident
households in 1963 and 1969. Overall, the amount collected
in 1969 more than doubled that of 1963, rising from Baht
8,807.3 million to Baht 18,525.2 million. In both years,
import duties accounted for the highest share of tax
collections: 31.4 percent in 1963 and 28.6 percent in 1969.
The next largest category was business taxes, with 17.7 and
18.5 percent respectively. The revenue from selective sales
taxes jumped impressively in 1969 over 1963, but this was
Partly due to the inclusion of taxes on fuel and oil,

lxreviously classified under import duties. The share of

Ifiice premiums or ad valorem taxes on rice exports dipped

108

Table 3.l--Tota1 Adjusted Taxes and Other Government
Revenues, by Type, 1963 and 1969 (Amounts in
Millions of Baht)

 

 

 

 

 

 

 

 

 

1963 1969
Type Of Taxes Amount 22:; Amount :::;

Individual Income Taxes 468.2 5.3 1,119.5 6.0
Corporate Income Tax 312.2 3.5 851.1 4.6
Business Taxes 1,561.3 17.7 3,490.5 18.8
Selective Sales Taxes 400.5 4.5 2,521.1 13.6
Import Duties 2,764.3 31.4 5,294.0 28.6
Rice Premiums 798.4 9.1 1,235.6 6.7
Other Export Taxes 363.2 4.1 424.3 2.3
Taxes on Property 147.4 1.7 390.1 2.1
Royalties and Permits 422.0 4.8 691.9 3.7
Gov't Sales and Services 322.0 3.7 361.1 1.9
Gov't Monopolies 776.1 8.8 605.3 3.3
Gov't Enterprises 226.3 2.6 636.4 3.4
Other 245.4 2.8 949.3 5.1

Total 8,807.3 100.0 18,525.2 100.0
Sgurge: Thailand,Department of Comptroller General, Repgrts

of Receipts and Outlays of the Kingdom of Thailand,

 

1963 and 1969.

109

slightly in 1969 compared to 1963, as did other export
taxes.

One striking fact about the tax collection pattern
shown in Table 3.1 is that the individual income tax con-
tributed only 5 to 6 percent to total tax collections in
1963 and 1969. This is extremely low by world standards.
The corporate income tax was also relatively low, comprising
less than 5 percent in 1969. Taxes on prOperty, which could
partly be classified as direct taxes, were very small,
accounting for 1.7 and 2.1 percent of the total tax collec-
tions in the two years under study. The remaining collec-
tions could be grouped as nontax revenues of the government,
and together they amounted to about 20 percent of total

tax collections in each year.

Allocation of Tax Burdens

 

Once the taxes and other revenue sources are selected,
the next task is to allocate their "burdens" to household
units grouped according to size classes of income. But
first one must decide who should bear the brunt of the burden:
consumers, producers or factor owners. As a rule, if there
is any empirical evidence as to who actually bears the
burden, it should be used. If not, then some theoretical
justification must be relied upon. This problem is concerned
With what is known as the shifting of taxes or tax incidence.
In partial equilibrium analysis, whether or not the tax

burden is shifted, or in what direction, depends upon many

110

factors, such as the nature of the tax, the elasticities
of demand and supply of the taxed products and the factors
of production, and so on. This aspect will be analyzed in
greater detail when each tax is discussed.

The next step is to translate these allocations by
economic categories into allocations by households classified
into various income classes. For example, if it is decided
that the tax on tobacco is shifted entirely to tobacco
smokers, then this tax will be further allocated to house-
holds according to the pattern of their expenditures on
tobacco products. This and other distributive patterns are
normally obtained from household expenditure surveys.

In Thailand, the 1962-1963 household expenditure
survey and the 1968-1969 socio-economic survey provide the
necessary data for the allocation of taxes to households
by income classes. These surveys give the average expenditure
on each commodity group by each representative family in
different income classes. The eight commodity groups on
which household income was spent in 1963 and 1969 were:
food, dwelling (household goods and household operations),
clothing, transportation; education, reading and recreation;
medical and personal care, alcoholic beverages and tobacco,
and miscellaneous expenditures. 'When the average expenditure
is multiplied by the total number of families in that income
class, the result is the total expenditure of that income

Class on each particular commodity group. This pattern of

lll

household expenditures by income classes serves as a base
upon which most taxes are allocated. (See Appendix D for
detail.)

When certain taxes cannot be properly allocated to
households according to the available expenditure patterns,
other allocation rules must be used. For example, taxes on
rice exports, which are believed to fall on farmers, are
allocated to households according to the pattern of distribu-
tion of rural income; taxes on private passenger cars are
allocated entirely to the tOp income group; and so forth.
These and other allocation formulas or rules together form
the bases upon which the tax burdens are distributed in
this analysis.13

The foregoing methodological procedure can be stated
mathematically. Assuming that:

i = 1, . . ., 1 is the income class;
j = l, . . ., m is the type of tax;
k = 1, . . ., n is the commodity group upon
which expenditure is made and the pattern of

which is used as the allocation formula;

b = absolute tax burden;
e = household expenditure; and
t = absolute amount of tax;

then, it is true that:

112

where:
1
wki = eki / .2 eki,
1:1
and:
m
Bi = Z b.i
j=1 3'
where:
bji = the burden of tax j on a household in the

income class i;

wki = the share of income class i in the total

expenditure on commodity group k;

tjk = the absolute amount of tax j collected

from commodity group k; and

Bi = total tax burden borne by income class i.
From the above, it follows that the effective tax

rate of income class i would be represented as Bi/Yi' where

Yi is the total household income of income group i.

The Analysis of Taxes

 

The following section will discuss each major tax in
turn, taking into account the tax shifting assumptions and
the problems encountered in the actual allocation of burdens

in each year.

Individual Income Tax

 

Regarding who should bear the burden of individual

income taxes, the commonly accepted View is that the income

113

earner bears the entire burden. The reasoning is rather
obvious, since most people usually do not have the Opportu-
nity to shift such a burden directly to others, especially

in the short run. Theoretically, this is equivalent to
assuming that the factor supplies--in this case labor alone--
are perfectly inelastic, which is not at all so restrictive
an assumption as it might seem. It is often argued that
taxes on income reduce work effort, thus changing the

pattern of earning and income distribution. But, as Musgrave
has shown, the work effort could increase, decrease, or
remain unchanged depending upon the strength of the income

and substitution effects.14

If it is assumed that the change
in income distribution due to the initial change in the work
effort (before final adjustment) is neutral, then the
assumption that the individual income tax rests with the
intial payee is plausible.15
The Revenue Department of the Ministry of Finance
was responsible for collecting personal income taxes in
1963 and 1969, and statistics showing the amount of tax
received from each taxable income class were made available
in the Annual Report of that Department. By adjusting the
1969 taxable income classes to coincide with the household
income classes used in this study, the amount of tax by
desired income brackets can be estimated. Switching from

the actual tax received by the Revenue Department by taxable

income brackets to family income brackets reduces the amount

114

of tax per family considerably because only a small fraction
of households which were supposed to pay tax actually did so.
Inefficiencies in both tax administration and enforcement
were to blame for this low income tax collection.

Since no revenue data were available for 1963,
an assumption had to be made that the pattern of tax payment
was approximately the same in that year as for 1969. As
a result, approximately 97 percent of the total personal
income tax was allocated to the top income class in 1963,

and the rest to the second highest income bracket.

Corporate Income Tax

 

Perhaps the single topic in taxation that has
prompted the most extensive debate is the incidence of the
corporate income tax. Traditionally it is held that a firm
in a purely competitive market with a profit maximization
objective cannot shift the burden of the corporate profit
tax in the short run because such a tax does not affect
optimum output and production. Therefore, the tax burden
will fall entirely on profits. However, if the market is
imperfect, or if the objective of the firm is not profit
maximization but sales maximization, or some "satisficing"
objective, the traditional view loses something of its
validity. When Marian Krzyzaniak and Richard Musgrave
showed econometrically in 1961 that the firm could and did
fully shift the corporate tax forward, even in the short run,

the traditional view was seriously challenged. Furthermore.

115

they also ushered in an era of numerous studies both

supporting and disagreeing with this interpretation, and the

. 16

controversy continues.
It is not the intention of this study to discuss,

or even to review, the debate on the incidence of the

. 17
corporate income tax.

Suffice it to say that the results
are still inconclusive as to how much of this tax should

be shifted forward in terms of higher product prices or
backward in terms of lower wages and salaries. A researcher
must use his own judgment as to what will be the likeliest
shifting assumption in a particular economic setting.

In previous tax incidence studies, most economists
assumed that the burden of the corporate income tax was
shared between the consumers of the firm's products and the
firm's shareholders. But as there appears to be no compelling
rule about how to apportion the shares, this study assumes
that all of the corporate income tax in Thailand is shifted
forward to consumers and none is borne by corporate owners
or workers.

In Thailand, where the market is far from perfect
and where cost-plus or full-cost pricing is normally the
practice, the 100 percent incidence shifting assumption is
not at all unreasonable. In their study of tax incidence
in Latin America, Bird and DeWulf commented: " . . . the

absence of a full shifting assumption among the great variety

of assumptions used is curious," and " . . . it would still

116

seem as valid to assume full forward shifting as some other
arbitrary percentage on the basis of existing evidence for

"18

Latin America. The full shifting assumption also was

used in the study by Jay S. Salkin of tax progressivity in
Thailand.19 The adoption of this assumption also eliminates
the need for adjustment in the income concept for the
unshifted portion of the corporate income tax.20
The allocation of the corporate income tax in
Thailand used the following procedure. First, corporations
were broken down by types of commodities or services sold
and the amount of taxes paid. This information was available
directly from the Revenue Department in the form of tables
showing the amount of corporate tax received by types of
business activities. (See Appendix E for detail.) Each of
these taxes was then allocated to households by income
classes according to the pattern of their expenditures on
the taxed products. In cases where the product and the tax
were not exactly matched, some judgments were made as to what
prOportion of the tax should be allocated and on what
allocation bases. For instance, there was no allocation base
available for consumption expenditure on cement, so 25 per-
cent of the tax on the profits of cement companies was
arbitrarily allocated as expenditures on household operations,
25 percent on transportation (road building), and 50 percent
on the total expenditure pattern.

Although the actual corporate income tax collected

in 1969 was known, the detailed breakdown of such tax by

117

business activities was available only up to 1968. To
obtain the same breakdown for 1969, it was assumed, then,
that the pattern of tax collected in 1969 was exactly the
same as that in 1968. This assumption enables the 1968
pattern to be used for 1969.

Business Taxes and Stamp Duties, Selective
Sales Taxes, and Import Duties

 

 

The most important instances of indirect taxes
are business taxes, stamp duties, selective sales taxes and
import duties. Their burdens are assumed to be shifted
fully to the consumers of taxed products or services.
This assumption is widely accepted, and most studies have
adOpted it without questioning its theoretical justification.
But it is this indirect tax assumption in relation to the
direct tax assumption discussed earlier that, according to
A. R. Prest at least, threatens the very foundation of
conventional tax incidence studies. Prest pointed out that
for an indirect tax to be shifted fully forward, either the
supply of the taxed product must be infinitely elastic or
its demand completely inelastic. This, of course, runs
counter to the direct tax case, where factor supply is
assumed to be completely inelastic or factor demand perfectly
elastic. In a study of short-run incidence, the use of
these two assumptions would be incompatible.21

Musgrave, however, argues that the supply of the

consumer goods does not have to be infinitely elastic for the

118

full shifting assumption to apply. Imagine a situation
where two persons have a prOportional tax levied on their
income. Suppose such a prOportional income tax is substituted
by an excise tax of equal yield on one of the commodities
which is purchased and consumed by one and not the other
taxpayer. It is obvious that only the consumer of the good
being taxed bears the full burden of the excise.22 A. B.
Cartter also advances an argument supporting the full shifting
assumption, maintaining that the government could use the
indirect tax proceeds of any one commodity to buy a sufficient
quantity of that commodity to keep output unchanged. In
this case, where compensating government expenditure is
assumed, prices can be thought to rise by the exact amount
of indirect taxes without requiring either the supply to
be perfectly elastic or the demand perfectly inelastic.23
Although, as Earl Rolph has demonstrated, the burdens
of indirect taxes could shift backward to factor owners,2
the shifting magnitude is uncertain and the reasoning less
convincing than the traditional assumption. In Thailand,
as in most other underdevelOped countries, the basic wage
rates are already very low, so the possibility that these
taxes will have further depressing effects on wages is very
slim. Therefore, in this study the incidence of business
taxes, selective sales taxes, and import duties will be

assumed to be shifted fully forward to consumers of the taxed

products or services.

119

The aforementioned taxes were allocated to consumers
in various income classes according to the patterns of their
expenditures on the selected eight commodity groups, and few
other allocation bases. Some details concerning the empirical
problems involved in allocating these taxes are discussed

below. (See also Appendix E for detail.)

Business Taxes

The disaggregated data on the amount of business
taxes collected by types of business activities were available
from the Revenue Department. Business activities were
roughly classified into 13 categories and further classified
into subcategories (see Appendix E). These classifications
permit the taxes to be allocated according to the established
pattern of expenditures on commodities and services fairly
easily. As in the case of the corporate income tax, a
few categories of taxes could not be readily allocated on
the basis of any specific commodity group, and subjective
judgment was used in allocating them to the nearest
commodity group or groups. For example, taxes on oil and
fuel were allocated on the basis of expenditures on house-
hold operations, transportation, and the total expenditure

pattern.

Selective Sales Taxes
Only a few commodity groups were burdened with selec-

tive excise taxes. Major commodities included beer and other

120

alcoholic drinks, tobacco products, cement and petroleum
products. The amount of taxes collected from these products
was reported in the publications of the Excise Department,
the Department of the Comptroller General, and the Bureau of
the Budget. With the exception of cement, petroleum
products, and a few other commodity groups, the bulk of the
taxes was allocated according to the pattern of household

expenditures on alcoholic beverages and tobacco products.

Import Duties

Taxes on imports were reported in a most disaggre-
gated form by the Bureau of the Budget (from the tax
records supplied by the Customs Department). Imports were
broken down into ten categories according to international
standard classifications, which was helpful in ensuring
accurate allocation (see Appendix E). Of course, problems
arose with intermediate goods such as chemical compounds,
minerals and other raw materials. The taxes on these
intermediate goods were treated as taxes on the products

which used such raw materials as main producing agents.

Export Taxes

 

If export taxes were to be treated in the same way
as import taxes, then the burdens of export taxes would
fall on foreign consumers of the exported product. But in
a small exporting economy such as Thailand, this reasoning

would not be valid. First, each of the country's exports

121

comprises only a small portion of world exports, and supply
conditions are quite rigid. Second, the world demand for
Thai products is believed to be very elastic.25 These
conditions constitute a compelling case for the assumption
that it is the domestic producers who bear the burden of
export taxes. In other words, the taxes would be assumed
to be shifted backward, not forward.

Regarding Thai rice exports, which earn the largest
amount of foreign exchange, considerable research has been
conducted to determine who should bear the burden of rice
premiums or taxes imposed on the values of rice being
exported. After taking into account the demand and supply
conditions and various other factors, most studies have
reached almost the same conclusion that has been reached
here: Thai farmers bear most, if not all, of the burden
of rice premiums. Thus, rice premiums were allocated to
farm households according to their share in the distribu-
tion of rural income.26

Regarding other Thai exports, such as rubber, tin
and teak, the world demand and domestic supply situation does
not differ much from rice. Thus, the burdens of taxes on
these exports also were assumed to fall on domestic
producers. For rubber, taxes were allocated according to
theedistribution of income of the rubber growers, a special

distributive series derived from an agricultural census.

For tin, however, the burden of the taxes was assumed to

122

fall entirely on the highest income bracket because most

tin mines in Thailand are owned by a few wealthy families.27

Taxes on Property

 

Taxes on property as used here are not "prOperty
taxes" in the usual sense, that is, taxes on the ownership
and improvements of farm lands, residential homes, business
establishments, and so forth, but are simply taxes on two
specific kinds of transactions: automobile ownership and the
registration of real estate transfers. In Thailand, except
for automobiles, which are treated as prOperty, no other
prOperty is subject to taxes. Taxes on the registration
of new ownership of real estate are not taxes on property
in the real sense, but are fees that the government
charges for the legalization of these transfers.

To allocate automobile taxes to various income
classes, the procedure adopted requires a detail of taxes
broken down by types of automobiles. This information is
available in the annual report of the Police Department
which is entrusted with the authority to collect automobile
taxes. In this report, it was reported in both 1963 and
1969 that about 25 percent of the automobile taxes was
contributed by private passenger cars, another 25 percent
by public buses and taxis, and the remaining 50 percent by
Private trucks and other motor vehicles.

Regarding the taxes on private passenger cars, it is

assumed that the owners of this type of vehicle bear the

123

full burden of the taxes. As the tOp income group is the
only group wealthy enough to own a private passenger car

in Thailand, households in this group were assumed to bear
all 25 percent of the automobile taxes in both 1963 and 1969.
Another 25 percent of the taxes originating from public
transport vehicles are assumed to fall on all households
according to the pattern of their expenditures on transport
services, so the taxes are allocated accordingly. The
incidence of the remaining 50 percent of taxes contributed
by trucks are assumed to be spread over all commodities
transported by trucks. Therefore, the patterns of total
household expenditures in 1963 and 1969 were used in allo-
cating this portion of automobile taxes in the respective
year.

Regarding the registration of real estate, it is
assumed that the income positions of households, to a great
extent, determine the ownership of real estate: the wealthier
the family, the more immovable prOperty or real estate it
should have. Therefore, the taxes on the registration of real
estate were allocated to households in direct prOportion to
the share of each income class in the distribution of family
money income.

Royalties and Permits, Government Sales

and Services, Government Enterprises and
MQnOpolies, and Other Miscellaneous Revenues

 

 

 

Determining the incidence of various nontaxed revenues

of the government is not easy. The fact that these are more

124

benefit-based charges than taxes in the real sense casts
doubt on the existence of a "burden." Nevertheless, one can
certainly perceive the incidence of these revenues in terms
of the money that has been paid more out of necessity than
voluntarily in exchange for government goods and services.
The inclusion of these revenues is much more relevant when
the incidence of both government receipts and outlays is
considered.

The data on goods and services for which royalties
and permits are required were available from the Department
of the Comptroller General and were in sufficient detail to
permit an allocation to households on the basis of their
expenditure pattern for such goods and services. This
procedure assumed the shifting of tax burdens to consumers
rather than producers with the exception of tin royalties,
which were assumed to fall on producers. This latter
incidence shifting assumption is consistent with the facts
that the proportion of tin used within the country in relation
to the total tin output is negligible, and the world demand
for tin is elastic. The types of activity that require
permits include the sale of liquor, lottery tickets, guns
and fireworks, gambling, and alien registration. Most of the
revenues from these permits were allocated on the commodity
consumption expenditure pattern, but the alien registration
fees were allocated according to the distribution of house-
holds on the assumption that aliens (mostly Chinese) are

normally distributed throughout the income ranges.

ser
goo

to]

such

were

Bmea
house;
tobacr

accorc

125

The burden of receipts from government sales and
services was assumed to fall on those who pay for these
goods and services. Again, these revenues were allocated
to households according to the patterns of their consumption
spending. The same allocation rule was used for other
revenues from government enterprises and monOpolies. This
proved to be a simple procedure once it was determined that
the taxes charged on the products of government enterprises
and monopolies would be passed on entirely to consumers of
such products.

The largest revenue items from government monopolies
were the proceeds from the Tobacco Monopoly and the Lottery
Bureau. The proceeds of the former were allocated to
households according to the pattern of their spending on
tobacco products and alcoholic beverages, and the latter
according to the pattern of their expenditures on recreation.

The miscellaneous category of government revenues,
consisting of fines, unused funds from various government
departments, interest from loans, and other revenues, was
considered item by item and the burdens allocated accordingly.
For public fines, one-half was allocated according to the
distribution of households and the other half according to
the distribution of income. All of the left-over funds were
allocated equally to each household, as it was assumed that
these funds would be respent in the next year's general

expenditures. Interest was allocated according to the

126

distribution of income, and all other revenues according

to the pattern of total expenditures.

Empirical Results of Tax Analysis

 

Tables 3.2 and 3.3 present the distribution of
absolute tax burdens by type of taxes and by income classes.
This form of presentation does not show how these burdens
affect income distribution. There are at least two ways
in which the income redistributional results of these tax
burdens can be shown. First, one can determine the effective
rates of taxation by absolute income levels, a method almost
universally adOpted in incidence studies. Second, the
income redistributional impact can be traced through changes
in the patterns of income distribution after the amount of
taxes is subtracted from the existing income levels. In
other words, the so-called post-tax income distribution can
be compared with the "pre-fisc" income distribution to see
if there is any redistribution. In doing so, either a class
by class comparison or a measure such as the Gini concentra-
tion ratio, or both, may be used.

Joseph Pechman and Benjamin Okner, in their study
of tax incidence in the United States, have indicated a third

28 Effective tax

way in which results can be presented.
rates can be presented by decile (or percentile) groups.
Because the first method (effective tax rates by income

classes) uses an arbitrary income classification for absolute

127

 

 

m.~om.m H.4ov.~ H.mmo.e s.mmo.~ m.oam.e o.aem.a Hence
«.mem a.me m.e~ m.mm ~.me ~.ma Magoo
m.e- o.me m.a~ «.mm n.5m q.ee mmmfludumucm u.>oo
H.ees m.mee o.mm m.mo~ m.ema o.sme mmaaodocoz u.>oo
o.-m m.ms ~.mm o.es o.mm o.om mmofl>umm one modem u.>oo
o.-v e.HmHI s.oe m.os e.ee H.~e mangoes new mmsuamsom
v.5ee e.mm m.ea o.em m.me o.m~ suuoaoum :0 waxes
~.mem m.em e.em ~.moe e.ee «.me mmxme uuodxm umnuo
e.mms «.mee ~.mm ~.mm~ H.mee e.eee masseuse muse
m.eee.~ H.ome m.emm e.mee «.mme «.mme moaned usedeH
m.ooe e.ms m.em e.moa m.me m.mm mmxme mmemm m>suooeom
m.aem.e m.eem ~.eme H.emm e.mm~ e.esm mmxme mmmcamsm
~.~Hm o.~oe m.mm e.me a.mv m.ae xme oeoocH openeduoo
~.mee H.eme H.4H II II II xme msoocH eases>flecH
mommmeo um>o can mmm.na mam.ee amm.m ooo.m

sea ooo.me Iooo.~H Iooc.e Iooo.m “was:

 

 

 

 

 

 

Aunmmv mommmau mEoocH

 

mmxme we mdha

 

 

Aucmm mo mcowaaflz cwmucsoe<v
mama .mommoau mEoocH we mcoonsm xme mpsaomnm mo nodusnwuumonIm.m wanes

128

 

 

 

 

 

 

 

 

 

«.mmmda méhmm movmfn mmomrn omen...” m.ovH.~ «.mmmJ «.mmHJ H38
m.mvm Comm m.mm 9:. fimm o6: moma mmba Hmnuo
vémw H.mmm mév Emm o.mm m.om o.mm Ndw mmmwumumufiw H38
m.mow o.mom 56¢ Cam mdm 5.2. m.mm m.mm mmfinomocg #38
Team oéma CNN o.mm H.mm 5mm me m.NN mug BM modem #38
mémm v.mmv m.mm «.mm Hem 013 m2: mom made—Hum can mmfldmma
Comm méHm m.mm Ndm o.mm v.mm N.mm «.mH E95 GO mmxmu
mémv m.NmH mém 5mm m.om 0.3. «.2. Adm Manama unomxm H930
m.mmN.H mémv H23 mKHH mos. Emma «.mma «.mm mEDdEHm mug
oémmfi m.omm.m m.wmm m.mmv m.mvm ade h.m~.m Noam mega “H095”
H.Hmm.m m.mmo.H coma m.mmm Emom Adam adhm wdma mmxcm. mwdwm “5.300me
m.omv.m meow; ofiwm mmam m.mvm moo". o.o~.m N.,—”om mmxmb mmmcwgm
Himmm o.ooe Emm 0.3. H.mm 0.5m mdm va some $505” mumuonwund
m.mHH.H mimoé m.mm «.mH N..m II II. II Won. 285 nggflBH
udmmmau um>o cam mmm§H mmmJH mmmJa mmm.m mmmfi ooo.m .
H2 Iooo.wH Iooo.m._.. Iooo.NH Iooo.m Ioood loco.m “OED mmxme MO was?

 

32mm: mommmau mEoocH

 

 

 

Seam mo 9832.: 5 8588 $3 8888 285 so 885m 58. 338nm mo Sflsflfimaolma Some.

129

income levels that tends to exaggerate the significance of
data in some parts of the income distribution and to
diminish it in others, the third method is more attractive.
It is a most useful complement to the other two methods
because it can be used to compare the tax burdens of groups
representing the same number or percentage of family units
in different parts of the income scale. It must be pointed
out, however, that Pechman and Okner had at their disposal
complete observations of total households and total income
(from the so-called MERGE file) which enabled them easily to
compute the effective tax rates by decile or percentile
groups. Unfortunately, similar data are not available in
Thailand. Although it is possible to estimate a Lorenz
curve of tax burdens in this study from which effective
rates by decile or percentile groups could be computed

using numerical analysis and computer techniques, this would
be unusually complicated, and it was decided that it would
not be desirable. As a result, only two methods of tax

incidence presentation will be adopted here. The two are:

Effective Tax Rates by Income Classes

 

Effective rates for 13 different types of taxes by
income classes are presented in Tables 3.4 and 3.5. These
effective rates were based on the money income of households
in 1963 and 1969. Tables 3.6 and 3.7 present the effective
tax rates based on adjusted income of households in 1963 and

1969 respectively.

130

 

 

 

 

 

 

 

m.m~ m.om o.mm m.~m v.m~ h.~v Hmuoe
h.o v.0 o.o n.o m.o m.H “mayo
n.o v.0 o.o m.o m.o a.a mmmflumumuam u.>oo
m.~ m.H m.m m.~ m.~ ~.a mmaaomocoz u.>oo
m.o 0.0 m.o m.o H.H m.a mmow>uwm pom mmamm u.>oo
N.H n.a m.o m.o o.a m.a muafiumm pom mmfiuammom
v.0 m.o ¢.o m.o v.0 m.o muummoum co mmxme
H.H m.o H.H N.H m.H H.H mwxme uuomxm umzuo
m.m o.H o.~ m.m m.m m.m mEoHEwum mowm
o.m o.m m.s m.s H.m m.va .mmfluso anomaH
~.H 5.0 N.H N.H w.a H.m mmxme mmamm m>fluomamm
m.v o.m v.¢ m.v m.m v.m mmxma mmmCflmsm
m.o m.o m.o m.o m.o v.a xme meoooH mumuomuou
v.H m.m m.o II II II xmB mEoocH Hmspfl>HoQH
mmmmmau uw>o com mmm.ha mmm.HH amm.m ooo.m
Had ooo.ma nooo.ma uooo.m looo.m noon: mmxme mo mama

 

Aunmmv mommmao wEoocH

 

 

 

mmma .mmmmmHo mEoocH xn mEoocH wmcoz co pwmmm mmuom

Ammmmucmoummv
xma m>wuommmmunv.m magma

131

 

 

 

 

 

 

 

 

 

5.00 5.00 0.00 0.00 0.00 0.00 0.00 0.50 00000
0.0 5.0 0.0 0.0 0.0 0.0 0.0 0.5 00000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000000000 0.000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 000000000200.>00
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0000>000 000 00000 0.000
0.0 0.0 - 5.0 5.0 0.0 0.0 5.0 0.0 0005000 000 000000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000000 00 00000
5.0 0.0 0.0 5.0 0.0 0.0 0.0 0.0 00x00 000000 00000
0.0 0.0 0.0 0.0 0.0 5.0 0.0 0.0 05005000 0000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.00 000000 000050
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00x00 00000 0>0000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00x05 00000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 x09 050000 000000000
5.0 0.0 0.0 0.0 0.0 In I. .: x00 050000 00000>0000
0000000 00>0 000 000.50 000.00 000.00 000.0 000.0 000.0
000. I000.00 -000.00 -000.00 -000.0 -000.0 -000.0 00000 00005 00 0059

 

3:000 0000000 050000

 

 

 

0000000000000 0000 .0000000 050000 00 050000 00002.00 00000 00000:x00 0>0uummmmrnm.m 00000

132

 

 

 

 

 

 

0.00 0.00 0.00 0.00 0.00 0.00 00000
0.0 0.0 0.0 0.0 0.0 0.0 00000
m.o 0.o m.o m.o m.o 0.0 mmmwumumucm u.>ou
0.0 0.0 0.0 0.0 0.0 0.0 0000000002 0.>oo
0.o m.o 0.o v.o m.o m.o mm00>0mm com mmamm u.>ow
0.0 0.¢ 0.0 0.0 0.0 0.0 0000000 000 000000000
0.0 0.0 0.0 0.0 0.0 0.0 00000000 00 00000
m.o 0.o 0.0 0.o 0.o m.o mmxma uuomxm umsuo
0.0 0.0 0.0 0.0 0.0 0.0 .05000000 0000
0.0 5.0 0.0 0.0 0.0 0.0 000000 000000
0.0 0.0 0.0 0.0 0.0 0.0 00000 00000 0>0000000
0.0 0.0 0.0 0.0 0.0 0.0 00000 00000000
0.0 0.0 0.0 0.0 0.0 0.0 000 000000 0000000oo
0.0 0.0 N.o II II II x09 mEoocH 009©0>wch
0000000 00>o 0:0 000.50 000.00 000.0 000.0
000 u000.00 [000.00 I000.0 -000.0 00000 00000 00 0000

 

 

Aunmmv mommMHU mEoocH

 

 

 

mmma .mmmmMHU mEoooH an mEoocH omumsmom co ommmm mmumm

Ammmmuomoummv
x09 m>wuommmmulo.m manna

133

 

 

 

 

 

 

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00000.
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0050
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000000000 0.80
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000082 0.80
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000000 000 00000 0.80
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0000000 000 000000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.00 00000000 00 00000.
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000. 000000 00000
0.0 0.0 0.0 0.0 0.0 0.0. 0.0 0.0 0.500000 0000
0.0 0.0 0.0 0.0 0.0. 0.0 0.0 0.0 00008 000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0020.0. 00000 90000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000. 00000000
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00.0 00800 00000008
0.0 0.0 0.0 0.0 -u I I I 00.0. 00800 0000000000
.03 0,... -00 -00 -00 -00 -00 0...
000 0000.0. 00 000.0.

 

 

 

 

000000 0000000 00800

 

 

 

0000000000000 0000 .0000000 00800 00 00800 00000000 00 00000 00000 00.0 0>0000000I00 0000.0.

134

Let us first consider the effective rates based on
money income. On the average, the 1969 tax collections
appeared to have imposed heavier burdens on families as a
whole than was the case in 1963 (28.7 percent compared with
25.5 percent). On a class by class basis, the tax incidence
of both years shows a generally regressive trend from the
lowest to the highest income bracket. In 1963, the lowest
income group contributed 42.7 percent of its money income
to the government in tax payments, whereas the highest
group contributed only 20.8 percent of its money income.

The corresponding figures in 1969 were 47.0 percent and

25.7 percent for the lowest and highest income classes,
respectively. The trend of effective tax rate is distinctly
regressive for the first few income groups, but this tapers
off in the higher income brackets. In any case, the t0p

income group bore the least tax burden relative to its income

 

position in both years.

When effective tax rates were computed using adjusted
income as the denominator, the incidence picture changed
markedly. The generally regressive trend which characterized
the effective tax rates based on money income is now absent.
Instead, the new effective tax rates for both years show a
mildly fluctuating trend which can at best be described as
proportional. In 1963 the average effective tax rate for all
income classes was 11.3 percent; in 1969, the same rate was

12.7 percent. In both years, the lowest income bracket

135

still bore a higher tax burden than the national average
and the highest income bracket lower. In 1963 households
in the lowest income bracket seem to have suffered the most
from taxation relative to households in other income posi-
tions, which was also true for 1969. But since the effective
tax rates varied only about 2 percentage points among income
classes, it is perhaps apprOpriate to call the tax system
proportional in both years.

The pattern of the effective rates also can be
depicted graphically, as shown in Figures 3.1 and 3.2.
Here, a line which connects all the points corresponding to
the effective rates of each income level roughly indicates
the progressive or regressive nature of the tax system.
A line that generally slopes downward would indicate a
regressive incidence pattern, and one which generally slopes
upward a progressive pattern. For a proportional incidence
pattern, the line would be somewhat parallel to the horizontal
axis. As can be seen in Figures 3.1 and 3.2, the tax
incidence patterns in 1963 and 1969 were almost identical;
they were shown to be mildly to moderately regressive across
the income range under a money income base, and somewhat
proportional under an adjusted income base.

It is possible, in fact instructive, to show the trend
of the tax programs in 1963 and 1969 by regressing the ef-
fective tax rates upon the average money income of a repre-

sentative household in each income class. The $10pe of the

136

mom“ was new" .mouuuao onoocm an .onoouH 50:0: co cocoa nevus was o>avoouhmutdwm magmas

$28 no messages
uouuaao oaoocH

 

 

 

 

 

 

 

ma mu NH 0 o n o
lllllllllllllllllllllll : ON
1!! ii!!!)
A 0:
$2

mead lllllllllll : om
Avsoouomv

moans use

ObnvOOMhm

137

mood can mom“ .nonmuao o-oocH up .nnoosH voooshv< so conga nouum was o>aaoommanu~.n «Mamas

Apron mo acnuusonev
noon-Ho osooca

 

 

 

 

 

ma ms Nu m o m o
HHH lllll Jul llllllllllllllll I lllllllllllllllllll ,hHHIIIllflulllllll
tIIIIIIIIx
. ow
to:
moma11,

mom“ nnnnnnnnn s om
Aesoouomv
moans nae

ObHfiOOMMH

138

tax incidence regression curve is useful in indicating the
regressivity or progressivity of the tax system: a negative
SIOpe would mean a regressive tax system, a positive slope a
progressive tax system, and a zero slope a proportional tax
system. After fitting the curve, it is shown in Table 3.8
that, based on money income, the 1963 and 1969 tax programs
were quite regressive, with the earlier slightly more
regressive than the latter programs. But with adjusted
income as a base, although the tax programs were still
slightly regressive, the regressivity was reduced to almost
zero, signifying a tendency toward proportionality in the
tax systems.

As for the tax burdens by type of taxes in 1963 and
1969, most taxes were regressive throughout, with the notable
exception, of course, of the personal income tax, for which
the top income class and one or two classes below it bore all
of the burdens and the lower income classes none. But
although only a few higher income brackets bore the burden
of the income tax, this burden relative to the income of the
highest bracket is quite small: only 3.9 and 3.2 percent
for the tOp income group in 1963 and 1969, respectively.

As expected, all major indirect taxes, such as busi-
ness taxes, selective sales taxes, and import duties showed
a pronounced regressive pattern, especially in the lower
income brackets. Another tax which was particularly burden-

some to the lower income groups was the rice premium,

139

Table 3.8--Generalized Tax Incidence Functions “Effective
Tax Rates) = a + b (Annual Average Money Income) ]

 

 

Intercept SlOpe R

 

(1) On Money Income Base
1963 33.3 -.390 .9562
1969 37.1 -.375 .9638
(2) On Adjusted Income Base
1963 12.4 -.056 .9992

1969 13.4 -.003 .9922

 

 

 

 

aAverage Annual Money Income is in Thousands of Baht

140

especially in 1969. Other taxes, which did not show a
clearly regressive trend, showed a U-shaped pattern where
the lowest and the highest income brackets bore a higher
burden than the middle income brackets. These taxes
include the corporate income tax, taxes on property, royal—
ties and fees. All other government revenues had a mildly
regressive incidence pattern throughout. This is quite
understandable and, indeed, to be expected, since the bulk
of the nontax revenues came from such sources as tobacco
and distillery monOpolies and lottery operations, products
which are not consumed in proportion to income.

The Effects of Tax Programs on
the Distribution of Income

 

 

As the tax structures in 1963 and 1969 appear to be
regressive across the whole range of money income, one can
reason that income distribution after taxation would become
more unequal. This is demonstrated in Table 3.9 which
presents the results of the after-tax money income distribu-
tion after the absolute amount of taxes is deducted from
pre—tax money incane distribution.

In Table 3.9, one immediately sees that, under a
money income base, the distribution of income became more
unequal over the period for the lower income groups. In
1963, the lowest two income groups suffered relatively more
than other groups from tax incidence, while in 1969, all

but the top two income brackets experienced a deterioration

141.

 

 

 

 

 

 

 

o .2: o .2: o . 03 o .03 o .2: o .ooa 063 o 63 mommmao a?
m.m¢ m.mv m.om m.om v.¢m m.~m n.mm m.mm um>o can ooo.ma
>.m m.m m.m m.m mam.~alooo.ma
m.AH A m.HH A m.mH A A.ma
m.m A.n m.> H.m mmm.qa|ooo.ma
m.m o.m H.m m.m mmm.aancoo.m
o.vm A o.qm A m.nm A v.m~
m.HH w.AH m.oa m.oa mom.m Iooo.m
H.~A m.HH m.va m.va n.m m.m m.ma o.¢H mmm.m loco.m
o.w o.o m.ma m.ma m.~ n.m o.oa o.ma ooo.m “was:
xMBiumom omHmeum xme1umom omfimlmum xmg1umom omflmlmum xmBWHmom omHmeum
Esme
$3 $3 $3 $3 mommflu 9505

 

 

 

833“me 8.85 Bums“?

 

:Qfifixmuumwanmikxnnmucnz

 

 

 

Ammmflcwoume $3 o5 82 .3988
885 E 8.85 8pm: Pa Em ammo: xmebmom can 03QO .«6 :oflsflflma mmflfioummlm .m manna

142

in their relative income positions. Under the adjusted
income base, the changes in post—tax income distribution was
small in both years but the tOp income groups is still shown
to have improved its income position. The changes in the
distribution of income after-tax are presented in Table 3.10
in terms of Gini concentration ratios.

Although Table 3.10 reveals nothing new, it does
give concise, quantitative support to the previous conclu-
sions that: (1) being generally regressive, Thailand's
tax programs in 1963 and 1969 caused the post-tax income
distribution to become more unequal in both years, but (2)
the extent of increase in inequality however, was slightly
less severe under the 1969 than the 1963 tax programs.

Table 3.10 may be examined in the following manner:
First, using the money income base, the 1963 tax system caused
a deterioration of income equality, as measured by the Gini
concentration ratio, by 8.4 percent (that is, comparing post-
tax to pre-fisc income distribution), whereas in 1969 the
percentage of equality deterioration was 5.4 percent. If
the pre-fisc and post-tax income distributions in 1963 are
compared to the respective distributions in 1969, income
inequality in the latter year is shown to have been reduced
by 1.4 percent and 4.2 percent, respectively. This point
requires further clarification. If the existing tax programs
in 1963 and 1969 were replaced by income proportional taxes.

the improvement of money income distribution in 1969 over

143

 

 

 

 

 

 

oo.ou mAmq. mmmv. mead
mo.q+A An.m+ A
Ho.A+ moms. mmmv. moma
wmmm
msoocH omumsno< :o Ame
A z A mm.m+ 54mm. ommm. amma
mA.eu hm.A-
av.m+ Noam. nmmm. mood
ommm mEoocH.mm:oz so AHV
mflwme «wwme mauanmmmz wwwme «HMWMH
xmelumom omflmlmum ucmflowmmuou xmelumom omwmlwum
mmma cam mama cmmzumm ficfio :H A+V Ham»
mummuocH Mo
ucmflofluumoo Acflo Ans mmmmuoma mucmfioflmmmoo ficflu
CH A+v mmmmuocH no mmmucmoum
AIV ommmuomo mumucmoumm m

 

 

 

 

 

mood pom moma .huflamsvmcH mEoocH cw
momcmso mmmucmoumm paw mowumm coaumuucmocoo wcwo xmelumom can omHmIoumuloa.m manna

144

that in 1963 would have been about 1.4 percent. With the
direct effects of the existing tax systems upon the income
of households, however, the post-tax 1969 income distribution
is improved over the same distribution in 1963 by 4.2
percent. Therefore, it could be concluded that, based on
money income, the 1969 tax programs had.a smaller income
de-equalizing effect than the 1963 tax programs, and such
an effect worsened the index of income equality by about
2.8 percentage points (that is, 4.2 minus 1.4) less than
would have been the case without the present tax programs.
On the adjusted income base, however, the changes
within the year were quite small. In 1963, the tax programs
worsened the Gini coefficient by just over one percent. On
the contrary, in 1969 the tax programs seemed to have im-
proved the post-tax income distribution but only by less than
one-tenth of one percent. This change is much too small to
be given any significance; it is likely to have been caused
by a statistical error. A more credible conclusion would
be that the 1969 tax programs had neutral or zero effect
upon the post-tax adjusted income distribution of households
in Thailand. Also, as is shown in Table 3.10, the preefisc
income distribution was about 5.8 percent more unequal in
1969 compared to 1963. The fact that the post-tax income
distribution in 1969 was still 4.7 percent more unequal
than the post-tax income distribution in 1963 means that

the 1969 tax programs had reduced income inequality by

145

about 1.1 percentage points over the 1963 programs.
Obviously, the equalizing effect of the 1969 programs was
still much too weak to overwhelm the de-equalizing influences
outside the tax sector.

In summary, the tax programs in Thailand contributed
in varying degrees to the worsening of income equality within
1963 and 1969. Although over time the tax systems were less
conducive to income de-equalization, they were still far
from being an instrument for income redistribution from

the rich to the poor.

Summary and Conclusion

 

This chapter has attempted to estimate the incidence
of Thailand's total tax system in 1963 and 1969. First,
the theoretical problems confronting a tax incidence study
were discussed, and the justification for the methods used
in the present study was given. The general methodology
to be followed requires that the taxes and other revenues
be selected, and then, based on certain shifting assumptions
and allocation formulas, these taxes in absolute amounts
be allocated to households by size classes of income. Next,
these absolute burdens are divided by the income base of
each income class, resulting in effective tax rates. These
rates show the percentages of income which households are
assumed to have contributed to the government in taxes. The

income redistributional effects of the government tax

146

programs can be seen either in the pattern of these effective
rates across income classes or in the change in the post-
tax distribution of income.

Several general conclusions may be noted:

(1) Based on money income, Thailand's programs were
generally regressive across income classes, with the highest
income bracket bearing a lower burden than all other
brackets.

(2) When the income base was changed from money to
adjusted income, the tax incidence pattern changed from
regressive to almost prOportional. This was mainly because
adjusted income accounted for much of the increase in the
income of the lower brackets. As a result, the effective
tax rates of these lower income groups were significantly
reduced.

(3) Under the adjusted income base, it was clear
that the government had not caused the distribution of income
of households to become more unequal, but it also had not
made it become significantly more equal. Nevertheless, one
is inclined to suspect that the Thai tax system more likely
contributed to the worsening of household income distribution
rather than to its improvement, since the income of the
upper income brackets was likely to be understated and that
of the lower income brackets overstated on an adjusted income
basis.

(4) The above results also can be demonstrated by

changes in the pattern of income distribution after absolute

 

ta

(ii

an

le

ad

ta

UR

147

tax burdens are subtracted from the initial income
distribution. Regarding money income, the post-tax
distribution of income became more unequal for both 1963

and 1969, although the 1969 tax programs made it slightly

less unequal than did the 1963 tax programs. Concerning
adjusted income, the post-tax distribution of income remained
practically unchanged although forces other than government
tax policies caused initial income distribution to become more
unequal.

(5) Finally, what is particularly disquieting about
the general tax incidence picture is that the effective
burdens of the highest income group still are lower than the
effective burdens of the lowest income group. For this

reason alone, the tax system deserves to be reevaluated.

148

FOOTNOTES

1See Richard A. Musgrave, The Theory of Public
Finance (New York: McGraw-Hill, 1959), Chapter 1.

2Richard A. Musgrave et a1" "Distribution of Tax
Payments by Income Class: A Case Study for 1948" National
Tax Journal, 4 (March, 1951): 1-48.

3It is counterfactual because it is against the fact:
the government or the tax system cannot simply be assumed
away. The term was first used by Eugene Smolensky and his
associates. See Jean Behrens and Eugene Smolensky, "Alterna-
tive Definitions of Income Redistribution," Public Finance,
28 (1973), pp. 315-332, and Morgan Reynolds and Eugene
Smolensky, "The Post-Fisc Distribution: 1961 and 1970 Com-
pared," National Tax Journal, 27 (December, 1974), pp. 515—
530.

 

 

 

 

 

 

4Alan R. Prest, "Statistical Calculations of Tax

Burdens," Economica, 22 (August, 1955): PP. 234-45.
Prest also quoted Edwin Canaan as saying that "such an inquiry
(tax incidence study) is a'will-o-the-wisp' and absolutely
useless."

5Carl S. Shoup, Public Finance (Chicago: Aldine,
1969); Richard M. Bird and Luc De Wulf, "Taxation and Income
Distribution in Latin America: A Critical Review of
Empirical Studies," IMF Staff Papers, 23 (November, 1973),
pp. 639-81.

6Richard A. Musgrave, "Estimating the Distribution of
the Tax Burden," Chapter 2 in Problems of Tax Administration
in Latin America: Papers and Proceedings of a Conference
held in Buenos Aires, Argentina, October, 1961 (Baltimore,
1965): pp. 41-42. In fact, this differential incidence
approach had already been used in the original Musgrave gt
31. study in 1951.

7Arnold C. Harberger,"The Incidence of the Corporate
Income Tax," Journal of Political Economy, 70, (June,
1962): pp. 215-240.

 

 

 

 

 

 

8Shoup, Chapter 2.

149

9Charles E. McLure, in spite of his high praise of

Harberger model as "a breakthrough in the analysis of tax
incidence in his paper," General Equilibrium Incidence
Analysis: The Harberger Model after Ten Years," Discussion
Paper No. 37, Rice University, Fall, 1972, used conventional
partial analysis in his tax incidence studies in West
Malaysia, Colombia and Panama.

10Alan R. Prest, "The Budget and Interpersonal
Distribution," Public Finance, 23 (1968), P. 92.

11Shoup, p. 11.

 

12From the surveys of tourist expenditures carried
out by the Tourist Organization of Thailand (TOT) (Tourist
Expenditure Surveys, Bangkok: TOT, various issuesL it was
shown that the total expenditures by tourists in 1963 and
1969 were estimated at Baht 476.5 million and Baht 1406.3
million respectively. These were by no means small amounts,
but it was impossible to know how much of these expenditures
went to the government as taxes. The use of some nominal
tax rates to compute the amount of taxes that were supposed
to be paid by these tourists was hazardous; and considering
the existing laxity in tax compliance and tax enforcement,
the resultant tax figures were likely to overstate the
actual taxes paid. In all, it was not unreasonable to
assume that these taxes were small or would not greatly
affect the total incidence pattern.

13For details of these allocation formulas, see
Appendix D.

14

 

Musgrave, The Theory of Public Finance, pp. 232-40.

 

15Musgrave also argued that either the factor
supplies could be assume fixed or that the effects of changes
in relative factor prices on relative product prices could
be ignored since factor proportions were likely to be very
much the same for each combination of goods consumed at
each different point on the income scale. In other words,
the nominal burdens could be assumed to stay put. See his
"Estimating the Distribution of the Tax Burden," in Income
Redistribution and the Statistical Foundations of Economic
Policy, eds. C. Clark and G. Struvel, (London: Bowes and
Bowes, 1964). P. 201.

16See Marian Krzyzaniak and Richard A. Musgrave, The
Shifting of the Corporate Income Tax (Baltimore: John
Hopkins, 1963).

 

 

150

17For literature on this tOpic, see for example,
Marian Krzyzaniak, ed., Effects of Corporate Income Tax
(Detroit: Wayne State University, 1966); Richard S. Gordon,
"The Incidence of the Corporation Income Tax in U. S.
Manufacturing 1925-62," American Economic Review, 57
(September, 1967), pp. 731-58; John G. Cragg, Arnold C.
Harberger and Peter Mieszkowski, "Empirical Evidence on the
Incidence of the Corporation Income Tax," Journal of Political

 

 

Economy, 75 (December, 1967), pp. 811-821; Peter Mieszkowski,
"Tax Incidence Theory: The Effects of Taxes on the Distribu-
tion of Income," The Journal of Economic Literature,

7 (December, 1969), pp. 1103-1124; T. Mathew, "Sh1fting and
Incidence of Corporation Income Tax: A Review," The

Indian Economic Journal, 20 (October-December, 197277'

pp. 2574274.

18Bird and DeWulf, pp. 660-661.

 

 

 

19Jay S. Salkin, "On the Direct Measurement of Tax
and Progression in Thailand," National Tax Journal, 27
(June, 1974), pp. 301-18.

20

The results of the distribution of tax burdens
according to alternative shifting assumptions for corporate
income tax practiced in most incidence studies will not be
presented in this study because they were not substantially
different from the results using full shifting assumption.
This is not surprising, for the corporate tax collection was
very small in Thailand contributing less than 5 percent of
the total government revenues in either 1963 or 1969.

21Prest, "Statistical calculations . . ."

22Musgrave, "Estimating . . .," (1965), pp. 42-46.
See also W. Irwin Gillespie, The Incidence of Taxes and
Public Expenditures in the Canadian Economy (Ottawa:
Queen's Printer, 1966), pp. 40-51.

23Allan M. Cartter, The Redistribution of Income
in Postwar Britain: A Study of the Effects of the Central
Government Fiscal Program in 1948-49 (New Haven: Yale
University Press, 1955), pp. 19-22.

24Earl R. Rolph, "A Proposed Revision of Excise-
Tax Theory," Journal of Political Economy, 60 (April, 1952),
pp. 102-117.

25See mainly Sriprinya Ramakomud, Thailand's Foreign
Trade: Structure and Policy, 1951-60 (Bangkok: Kuruspha
Press, 1968).

 

 

 

 

 

 

 

151

26 I I O O
The researches on the 1nc1dence of r1ce prem1ums

in Thailand were quite numerous; see for example Sura
Sanittanont, Thailand’s Rice Export Tax: Its Effects on
the Rice Economy (Bangkok: National Institute of Develop-
ment Administration, 1967); Edward Van Roy, "The Pursuit of
Growth and Stability Through Taxation of Agricultural Ex-
ports: Thailand's Experience" Public Finance, 23 (1968),
pp. 294-317; Trent Bertrand, "Rural Taxation in Tahiland,"
Pacific Affairs, 42 (Summer, 1969), pp. 178-188; James C.
Ingram, Economic Change in Thailand, 1850—1970, 2nd ed.
(Stanford: Stanford University Press, 1971).

27See Wit Satyarakwit, "Tin: A Comparison of Gravel
Pump and Dredging Mining," M. A. Thesis, Faculty of Economics,
Thammasat University, August, 1971.

28Joseph Pechman and Benjamin Okner, Who Bears
the Tax Burden? (Washington, D. C.: The Brookings
Institution, 1974).

 

 

 

 

 

 

 

CHAPTER IV
THE INCIDENCE OF PUBLIC EXPENDITURES

While the previous chapter examined the revenue side
of government activities and the subsequent distribution of
tax burdens by size classes of income, this chapter will ex-
amine the other side of the fiscal coin: public expenditures
and the distribution of their benefits by income classes of
households. As mentioned in Chapter 3, one of the main pur-
poses of collecting taxes is to enable the government to pro-
vide both general and specific goods and services to taxpayers.
This process has repercussions on the income distribution of
households, since each householder's tax paid is usually not
matched by an equal amount of government expenditure for that
household. Therefore, to account for the complete income re-
distributional effects of a fiscal system, it is imperative
that the incidence of public expenditures be evaluated, in ad-
dition to the incidence of taxes. To omit this item would be
tantamount to assuming either that the benefit incidence is
neutral or that it is unimportant, neither of which view is
tenable on an a priori basis.

In estimating the income redistributional impact of
public expenditure programs, a researcher faces problems much
more severe than those encountered in estimating tax incidence.
Most difficult of all are the theoretical and operational

152

153

issues of how to quantify the expenditure "benefits" and al-
locate these to different income classes. The problems seem
almost insurmountable when dealing with such "public goods"
as national defense, public health and sanitation, public
administration, and protection. Once these matters are dis-
posed of, the rest of the procedure follows smoothly. The
methodology in this study will be to establish some rules for
allocating the quantified amounts of benefits to various in-
come classes. Many of these rules are the same as those used
earlier in the allocation of taxes. After all benefits are
allocated to each income bracket, the effective rates of ex-
penditure benefits or benefit incidence are then calculated
by dividing such benefits by the income bases of each income

bracket.

Theoretical Problems in Expenditure

 

Incidence Studies

 

Broadly speaking, the theoretical questions involved
in a benefit incidence study can be summed up as follows:
(1) Who should be regarded as the beneficiaries of govern-
ment spending: those who directly receive government money,
or those who are ultimately affected by what the government
does? (2) It is likely that there will be more than one group
of households or one segment of the population which will re-
ceive the benefits of (or be affected by) government spending.
In this case, should one be interested mainly in the primary

beneficiaries, or in the secondary and subsequent beneficiaries

154

as well? (3) How does one quantify the benefits expected
to be received by the designated groups of households in the
case of certain specific expenditures, or by the population
as a whole in the case of general public expenditures? (4)
Is it necessary to determine the true beneficiaries? Can
one discuss the benefits of government expenditures without
knowing who the actual beneficiaries are? These questions

will be discussed in detail below.

Money Flow Versus Benefit Approach

 

Two approaches may be used in analyzing the incidence
of expenditure benefits: the money flow approach and the
benefit approach.1 The first would concentrate only on the
actual recipients of direct payments made by the government
either in exchange for goods and services or as transfers,
without considering who would ultimately benefit from those
services. For example, salary payments to public teachers
would be considered benefits to the teachers themselves. ‘While
students obviously also would benefit from this educational
expenditure, such benefits are not considered in the money
flow approach. The second, or benefit, approach emphasizes
the end results of public spending; of concern is the eventual
beneficiary, not the recipient of any money payments. In the
case of the educational expenditure mentioned above, for ex-
ample, the final beneficiaries under the benefit approach
would be the students and their families, not the teachers.

In the case of defense spending, it is the entire population

155

which enjoys the benefit of protection, not the military per-
sonnel employed or the firms under contract to produce arma-
ments.

While the money flow approach is more appropriate
when dealing with governmental transfers whose recipients are
believed to be the final beneficiaries, it is inadequate in
nontransfer cases and in the situation where the government
provides substandard or no services. Moreover, as Adler has
pointed out, the money flow concept is based on an implicit
assumption that the income of government employees would be
zero if the government did not employ them, which of course
is not always true in the real world; if the government did
not employ them, they often would find work elsewhere.2 As
the foregoing indicates, it is conceptually more plausible
and useful to evaluate the effects of government services in
terms of the final beneficiary, not the recipient of govern-
ment funds spent on a certain public project. Generally,
therefore, the benefit approach is adopted in most fiscal in—

cidence studies.

Direct and Indirect Recipients of Benefits

 

When the government spends its money, it must have in
mind who will be the first to receive the benefits directly.
If the spending is on educational services, the children at-
tending school and their families would be the primary re-
cipients; if the spending is on road building, the road users

would be the direct recipients, and so on. This does not mean,

156

however, that the effects of the benefits would terminate
with the primary beneficiaries. More than likely, there will
be countless circumstances in which secondary or other in-
direct beneficiaries exist. For example, an increase in the
number of students or educational activities would also boost
book production and the income of teachers, and an increase in
road use as a result of highway expenditure would benefit pro-
ducers of transport equipment as well as the construction in-
dustry.

Closely related to the idea of indirect benefits is
the concept of externalities. A well-educated population
could result in the externalities of increased productivity
and a more propitious climate for technical advancement. The
externalities of a healthy population could be a longer and
more productive life, and of an educated population, less so-
cial delinquency and a more socially responsible electorate.
Recently Charles E. McLure has tried to emphasize the pecuni-
ary aspect of externalities in his study of how relative prod-
uct and factor prices are affected by the public purchase of
goods and the employment of factor services. His discussion
of "expenditure incidence," which involves changes in the dis-
tribution of income resulting from marginal differences in pri-
vate and public spending patterns (as purchasing power is
transferred from the private to the public sector), can be con-
ceived as another example of how and why the concept of the di-
rect or indirect benefits of externalities is important in the

study of the redistributive impact of a fiscal system.3

“an”?

157

The question now is whether it is possible, in
practice, to estimate all tangible as well as intangible ef-
fects of government spending on direct and indirect benefici-
aries. Unfortunately, given the present state of the art ofea
benefit incidence study, it is very difficult, if not impos-
sible, to go beyond the direct or primary benefits of govern-
ment expenditures. As will be shown later, even an assessment
of primary benefits is difficult. The full effects of govern-
ment spending defy theoretical estimation at this point, much
less quantitative measurement. It is possible, of course, to
attack the problem of total effect estimation through cost-
benefit analysis, but one must be willing to accept certain
simplifying assumptions and subjective judgments in dealing
with the following: the scope of possible effects; how long
the effects will last; the opportunity cost of a particular
activity; what discount rate to use; whether all possible
costs and benefits have been considered; and so on. A full-
scale cost benefit study of benefit incidence, apart from the
above shortcomings also would be very costly. All in all, it
seems more practical to evaluate only the direct, primary
benefits. In doing so, all secondary and other effects as
well as all externalities must be ruled out. This is the posi-

tion adopted by most benefit incidence studies.

The Measurement of Benefits

 

Having selected the benefit approach and determined

that only direct benefits are measurable, the next problem is

158

how to measure the benefits from public expenditures. If
these are classified into three categories, namely, transfer
expenditures, specific expenditures, and general expenditures,
valuation of the first category poses little or no problem.
If it is assumed that the benefits accruing to transfer re-
cipients are not shifted to others, then the value of the
benefits in money terms may be considered equivalent to the
amount of such transfers. It is the specific and general ex-
penditure categories that present problems.

Specific government expenditures are expected to pro-
vide benefits to readily identifiable groups. Funds spent
on education should specifically benefit students in school;
those spent on highways should benefit highway users and con-
sumers of highway-transported products; subsidies for low-
cost housing should benefit poor families; and so forth. How-
ever, those goods and services are not necessarily private
goods in their entirety, and thus valuation of benefits through
the price mechanism and exclusion principle is not always pos-
sible. For example, not all roads charge tolls, and some educa-
tional services are free. A more manageable valuation of spe-
cific benefits is possible if one is willing to make the impor-
tant assumption that the total of benefits derived from a cer-
tain expenditure program is equal to the total costs of that
program. Under this assumption, the allocation of expenditure
benefits is, in reality, the allocation of costs of public

programs to expected primary beneficiaries. This "accounting

159

approach," called the cost incurred on behalf of" method by

W. Irwin Gillespie and the "input approach" by G. S. Sahota4
implies that (l) the average cost of the provision of goods
and services is constant, and (2) such provision is efficient
in a Pareto Optimal sense: there is no overextension of ac-
rivities and no waste.5 These assumptions are required to as-
sure that marginal cost, average cost, and marginal benefit
are all equal at an efficient equilibrium.

For pure social goods provided by the government under
such general expenditures as national defense, public adminis-
tration, or diplomacy, the valuation problem is even more in-
tractable because the true preferences of households are not
known. A social good (or service) is consumed in equal amount
by all, its enjoyment cannot be made subject to price payments,
and no one can be excluded, even if he does not pay for it.
Specifically, if the provision of a social good were to be
financed through taxation, no one would reveal his true prefer-
ence for fear that he would have to pay for the good and some-
one else would get a free ride. Generally, since most of these
benefits accrue to all families free of charge, families do
not directly indicate the value they place on each good or
service. Therefore, it is impossible to determine the total
amount of benefits from each expenditure that should be added
to the income of families.

Although the provision of a social good could be de-
termined through voting systems, it has been shown that there

are no perfect systems.6 In fact, the valuation of public

160

goods is not possible unless one assumes that the consumer's
preference function, or the social welfare function, or both,
exist and are known. Even Paul A. Samuelson's celebrated
theory of public good does not help because it does not ex-
plain how public goods are determined or valued; it only
states in a specific theoretical framework the condition
under which a public good together with a private good are
to be optimally or efficiently allocated.7

To overcome these difficulties, economists again re-
sort to the accounting approach, where total benefits are as-
sumed equal to total costs. As the marginal valuation of
benefits of each household is still not known, certain utility
functions (or allocation criteria) must be assumed. Although
crude, there seems to be no better alternative, especially
when empirical manageability is taken into account. As K. V.
Green has said: "Given some plausible alternative models and
our present state of knowledge about the demands of various
groups for public services, the assumption that such benefits
and costs are equal may be no worse than any other assumption."8
De Wulf has suggested the "behavioral approach," which also
uses the valuation of benefits at production cost. However,
after such benefits are allocated to different subgroups of
the population, a further "subjective valuation" of the pre-
sumed beneficiaries is made so as to estimate the real value
of the benefits as closely as possible.9 But, unless some
theoretical or operational framework is established for this

subjective valuation, the results could be so diverse as to

161

render the expenditure incidence study totally unreliable.
Despite the many shortcomings of the accounting approach, De
Wulf admits that, in the absence of knowledge about the poli-
tics of decision making, it is useful in the evaluation of
publicly provided services to various subgroups of the popu-

lation.10

Expenditure#Beneficiaries and
the "Spendee" Concept

 

 

The preceding section discussed how the benefits re-
ceived by households from public expenditures are to be valued,
a task which is unusually difficult because preference sched-
ules are unknown. Recently a group of World Bank experts
has attempted to substitute beneficiaries with the "spendee"
concept, that is, households or groups within the population
on whom government monies are spent. Emphasis no longer is
placed on the value recipients attach to public benefits, but
only on the magnitude of government spending for different
activities and on whom it is helping.11

In a sense, this is an attractive solution to the re-
vealed preference problem faced in expenditure incidence stud-
ies. The study now becomes more manageable and less value laden
since the focus shifts to who receives how much in terms of
costs and away from benefits. For a study of the distributive
impact of public expenditures alone, the spendee concept al-
leviates many of the usual conceptual difficulties; the prob-
lem becomes one of straightforward statistical estimation.

But it does not offer a solution for a full fiscal incidence

162

study, where emphasis is placed on how much the government
has helped in the distribution of income of households through
public expenditures in exchange for what has been taken away
from households in taxes. In other words, if taxes are as-
sumed to have burdens which fall upon taxpaying households,
then public expenditures similarly are assumed to generate
benefits in varying degrees to the same households. The
change from a beneficiary to a spendee concept thus does not
completely solve the main problems of benefit incidence study.12
Practically speaking it seems that the valuation or
measurement of the benefit of public expenditures must be
carried out by the total benefits equal total costs method.
Once the extent of benefits (costs) is known, the next ques-
tion is how to allocate these benefits to households in dif-
ferent income classes. For specific expenditures the bene-
ficiary groups already have been decided upon; the remaining
problem is to determine the bases upon which to distribute
the benefits. Gillespie's study of the United States offers
an example. The benefits of an expenditure on highways were
first divided between highway users and nonusers. For non-
users, cost was allocated according to the distribution of
real property value. For users, the share of such costs was
further divided between cars and trucks; that allocated to
trucks was distributed proportionately to "consumers of trans-
ported products," and that allocated to cars was similarly
13

distributed to "consumers of passenger travel." This is

the method usually accepted in a highway benefit study.

163

For general expenditures, various studies have adopted
one of several techniques. Benefits may be assumed to be dis-
tributed either equally among households in all income classes
(per capita or per household basis); proportionately on the
basis of family income; according to certain utility functions,
such as the ones used by Henry Aaron and Martin McGuire, and

Shlomo Maital;14

or a combination of these bases may be used,
for example, one-half distributed on a per capita basis, the
other half on an income proportional basis. It must be noted
that there is no definite rule specifying who the beneficiary
group must be, or should be, for a certain expenditure cate-
gory. Criteria vary according to the nature and extent of

the given expenditure from one economic setting to another.
The beneficiaries of highway expenditures in Gillespie's U. S.
study, for example, might be different from those in Thailand.
Also, the allocation bases may differ from country to country

depending upon the availability of statistical data and other

distributional information.

A General Methodology

 

The foregoing theoretical discussion has indicated
the general direction this study on expenditure incidence in
Thailand will take. The following specific methodology has
been adopted:

(1) With the exception of the transfer category for
which the money flow method is considered more appropriate,

the benefit approach will be used for most expenditures.

164

(2) Only direct, or primary benefits will be con-
sidered in this study; all indirect benefits and externalities,
either short or long run, are assumed constant or distribu-
tionally neutral.

(3) In general the accounting approach of benefit
valuation will be used. It will be demonstrated during the
course of this investigation why this technique is as accept-
able as the more realistic behavioral method or cost-benefit
analysis. Regarding the allocation of general expenditures,
only one standard set of assumptions will be used to allocate
such expenditures as defense, public administration, law en-
forcement and justice. For reasons to be explained later, no
alternative assumptions will be used to allocate general ex-
penditures.

(4) The beneficiary, as opposed to the spendee, con-
cept will be used. A designated beneficiary may not always
actually receive the benefits allocated to him, but it is as-
sumed that under normal circumstances he would.

(5) The procedure for allocating expenditure bene-
fits to different income classes follows the pattern of tax
burden allocation described in Chapter 3. First, the expendi-
ture items are selected and classified into economic functions.
Second, within each expenditure category a judgment is made as
to which groups of households or which income classes are to
receive the benefits and whether those benefits will remain
with the recipient or be shifted to others. Third, the dis-

tribution of these benefits by income classes is accomplished

165

using some of the allocation bases described in Chapter 3,

such as the distribution of income, distribution of households,
and distribution of household expenditures by income classes.
In addition, a few other bases upon which to distribute these
benefits will be used. Fourth, after all benefits are allo-
cated to all income classes, the effective rates of these ex-
penditure benefits relative to each corresponding income group
will be calculated to determine whether the expenditure pat-
tern is progressive (pro-rich) or regressive (pro-poor). The
change in the pattern of income distribution after government

expenditures also will be examined.15

Selection of Public Expenditures

 

Public expenditures can be classified in several ways
to suit several purposes. They can be categorized as consump-
tion (current) expenditures and capital expenditures; or ex-
haustive expenditures, essentially government purchases of
goods and services and nonexhaustive expenditures, which cover
such transfers as old-age pension and unemployment benefits.
As indicated earlier, most studies classify public expendi-
tures into specific or allocable, general or nonallocable,

and transfer expenditures.

In this study, a slightly different classification
will be used. The specific/general/transfer breakdown will
be retained, but this does not imply that specific expenditures
are always allocable as pure private goods, and that general
expenditures are always nonallocable as pure social goods. It

will be seen later that some of the programs in the specific

expenc
genera
ture p
dence

expend
functi

Specif.

 

 

166

expenditure category in Thailand are more in the nature of
general than specific goods, and some of the general expendi-
ture programs can indeed be allocated with reasonable confi-
dence to a specific, identifiable income bracket. Basically,
expenditures in Thailand will be grouped according to their
functions or purposes. Six functional classifications are
specified for all types of expenditures as follows:
Group 1: Economic services
a. Agriculture
b. Power and Fuel
c. Industries
d. Transportation and Communication
e. Other Economic Services
Group 2: Educational Services
a. Primary Education
b. Secondary Education
c. Higher Education
d. Vocational Education and Other
Group 3: Health and Social Welfare Services
Group 4: General Services
a. Defense
b. Law Enforcement
c. Public Administration
d. Other General Services
Group 5: Interest Payments on Public Debt
These functional classifications conform closely with

the method used by the National Statistical Office, but differ

from

curre:
ed to
ample
crite

found

the a
were
Of t:
are c
benefi
CludE
ments

On p1

and S

and E

tram

167

from the method used by the Bank of Thailand, which adds
current and capital categories. This study initially intend-
ed to follow the Bank's practice, but after checking the ex-
amples of current and capital expenditures for the consistent
criteria that distinguished one from the other, none were
found. It seems that almost all expenditure categories under
economic and educational services are listed under capital
spending, and the rest under current spending. Therefore,
the finer separation of expenditures into current and capital
groupings did not seem to serve any useful purpose and only
the functional classifications will be used here.

Once the functional classifications were decided upon,

the actual amounts of expenditures for each government agency

 

were obtained from the Report of the Receipts and Outlays_gf

the Kingdom of Thailand, published annually by the Department

 

of the Comptroller General. The amounts used here, however,
are only that portion of expenditures which is believed to
benefit Thai residents; that benefitting nonresidents is ex-
cluded. For example, portions of interest and principal pay-
ments on foreign loans are subtracted from the total payments
on public debt because, as transfers, their benefits fall di-
rectly on foreign debt holders.

The total adjusted amounts for each expenditure group
and subgroup are shown in Table 4.1. Expenditure groups 1,
2, and 3 can generally be described as specific expenditures,
and expenditure groups 4 and 5 as general expenditures and

transfers, respectively. The table also shows that the level

Table 4

 

168

Table 4. l—Total Adjusted Public Expeaditures of Thailand by
Functimal Categariss, 1963 and 1969 (Mounts in

Millions of Baht)

 

 

 

 

 

 

 

 

 

1963 1969
Emotional Categories
Per- Per-
Mount cent Anount cent

Group 1: Economic Services 2,560.2 24.6 5,962.4 26.7

Agriculture 767.7 7.5 2,222.6 9.9

Power and Fuel 271.8 2.6 286.1 1.3

Industries 179.3 1.7 173.1 0.8

Transportation and

Cannmication 1,079.3 10.4 2,959.8 13.2

Other Econcmic Services 253.1 2.4 320.8 1.4
Grow 2: Eflucational

Services 1,777.5 17.0 3,674.3 16.4

Primary 1,024.7 . 9.8 1,997.9 8.9

Secondary 234.9 2.3 405.0 1.8

Higher 193.8 1.9 559.6 2.5

Vocational and Other

Educatim 324.0 3.1 711.8 3.2

Grow 3: Health and Social

Welfare 929.3 8.9 2,338.1 10.5
Grow 4: General Services 4,022.1 38.6 9,136.2 40.9

Defense 1,637.6 15.7 3,718.9 16.6

Law mfiorcanent 797.9 7.7 1,558.4 7.0

Public Adninistration 1,414.6 13.6 3,063.8 13.7

Other Gmeral Services 172.1 1.7 795.1 3.6
Grow 5: Payments on

Public Debt 1,139.4 10.9 1,245.5 5.6

Total 100. 0 22, 356. S 100. 0

 

 

10.428.5

 

 

 

Sources: Thailand, Department of Ccmprtroller, Re
Incomes and Outlays of the Kingdom of T

t of the Receipts

 

 

hail—and, 1963 and 1969

issues (Bangkok, 1965 and 1971) ; National Statistical Office,

Statistical Yearbook of Thailand.

 

of publ
1963 to
culture
ser but
health .
a1 serv.
reductic

other 56

Um next
CldSSes.
who the

PeCted b

0n agric.

C:Ommlll'lic:

59%

 

portatic

 

mately 7

were SP€

lOCated

ments '

169

of public spending increased by more than 100 percent from
1963 to 1969. Increased spending was most pronounced in agri-
culture, transportation and communication, and defense. Les-
ser but still substantial increased expenditures were made in
health and welfare, higher education, and miscellaneous gener-
al services. On the whole, educational services suffered some
reduction, and the public debt services were reduced, while

other services experienced some gains.

Allocation of Public Expenditures

 

Once the absolute amounts of expenditures are known,
the next step is to allocate them to households by various
classes. Each expenditure subgroup will be discussed as to
who the direct beneficiaries are and on what basis these ex-

pected benefits should be allocated.

Group 1: Economic Services

 

Expenditures for economic services include funds spent
on agriculture, power and fuel, industries, transportation and

communication, and others.

Agriculture

 

Expenditures on agriculture ranked second after trans-
portation and communication in both 1963 and 1969. Approxi-
mately 70 percent of total agricultural monies in both years
were spent on irrigation projects, while the remainder was a1-
1ocated to the Ministry of Agriculture and its various depart-

ments, such as Rice, Forestry, Livestock, Fishery, and

Agric
secto
is a1;
is all

bracke

proble.
and ca;
benefi1
Of eque
with th
be adOp
in the

beving i.

ComPlet.

 

 

170

Agricultural Extension. Since it is assumed that the farming
sector benefits directly from these expenditures, 50 percent
is allocated equally to each rural household and 50 percent
is allocated in direct proportion to the share of each income
bracket in the rural family income.

Allocating the benefits from irrigation creates some
problems. Outlays on the construction of dams, reservoirs,
and canals are capital spending which is expected to yield
benefits in future years. The normal allocation method, that
of equating the costs of capital expenditures in any one year
with their benefits in the same year, which is the method to
be adopted here, appears to violate common sense. For example,
in the year when construction of a dam is underway, costs are
being incurred, but the dam would not yield any benefit until
completion, perhaps two or three years later. Therefore, it
would seem unreasonable to match costs with benefits in the
same year. However, one could argue that it is possible to
impute the benefits for that year by discounting the stream
of future benefits to the year in question, which could very
well equal the costs of construction in that year.

A more acceptable way of imputing the probable bene-
fits from such investment expenditures as irrigation and
other water resource development is through cost-benefit analy-
sis. The use of this method in the case of water resources

projects is perhaps the most established practice in all areas

of devel
tional p
preparin
method 1
cessful .
tion. I.
and publ
Direct b!
in farm :
governme1
such as :
of farm ]
PurchaSe1
Public be
and iHVe:
mumity 1
effect 0:
the eXPe

the tOta

and indi

 

 

171

of development program evaluation.16 Virtually all irriga-
tional projects in Thailand use cost-benefit techniques in
preparing feasibility reports for external borrowing. The
method is universally used and is fashioned after its suc-
cessful application by the United States Bureau of Reclama-
tion. In Thailand, three kinds of benefits, direct, indirect
and public, are normally evaluated in an irrigation project.
Direct benefits from a new dam or reservoir include increases
in farm property or farm income, investment in farms, and
government revenues from land taxes and taxes on farm products,
such as rice premiums. Indirect benefits incorporate sales
of farm produce for local and nonlocal processing, and farm
purchases for family living and production expenses. Finally,
public benefits involve, for example, settlement, employment,
and investment opportunities in the area, increases in com-
munity living facilities and services, and a stabilization
effect on the economy. After summing all the benefits over
the expected life of the project (normally 25 to 30 years),
the total is compared to the cost of the project, both direct
and indirect. The resulting ratio usually is high enough to
erase any doubt about the project's profitability. To give
only two examples: the total benefit-cost ratio of the Kang
Krachan Project (Rathburi Province in the Central Plain) was
estimated at 13.3 to l, and the direct benefit-cost ratio of
the Lam Ta Kong Storage Dam Project, in the northeastern prov-

ince of Korat, was estimated at 9.1 to 1.17

cos1
eval
irri
the
pecte
that
probl
funct
that 1
lems,
tion,
constr
instit
leads 1
and lor
Crop in
manY do
tribUtII
Call in
EStimat
pert in
"The be
tem ar
inVOIVe

Sh0u1d

172

It would appear more accurate to use the benefit-
cost results from these feasibility reports as a basis for
evaluation in this study. But, as the actual results of most
irrigation projects have indicated, the benefit computed in
the feasibility reports tends to overstate reality. The ex-
pected benefits are computed generally on the assumption
that technical and institutional factors will not create any
problems in the actual construction of the project and in its
functioning once completed. Experience has proved, however,
that many irrigation projects suffer serious technical prob-
lems, among them a water level that is too low for distribu-
tion, unfinished or too shallow distributing canals, and poor
construction due to the use of low—grade materials. On the
institutional side, an inadequate control of most operations
leads to large-scale corruption, causing tremendous waste
and long delays. Moreover, many farmers fail to grow a second
crop in the dry season, even when the water is available, and
many do not exercise sufficient care in maintaining the dis-
tributing canals in good working order. All these problems
call into serious question the accuracy of benefit-cost ratios
estimated on the order of nine or ten or more to one. One ex-
pert in agricultural development in Thailand has commented:
"The benefits from the uses of the present irrigation sys-
tems are much lower than expected when compared to the costs
involved; the efficiency of investment is far below what it

should be."18

the v
ing c
more

strat

173

The purpose of the foregoing has been to show that
the valuation by existing cost-benefit analysis in a develop-
ing country such as Thailand is by no means superior to other
more qualitative approaches. The irrigation example demon-
strates that equating total benefits to total costs in sub-
sequent expenditure evaluation is actually quite reasonable.
Since it can be conceived that some projects generate some-
what more benefits than costs, while others generate less,
the total benefits, overall, may reasonably be assumed to be
equal to total costs.

For the allocation of benefits from irrigation expendi-
tures to various income groups, 70 percent is allocated to
farmers according to their share in the distribution of rural
family income. The reasoning for this is that the bigger the
land holding and hence the higher the income position, the
larger the benefits derived from water services. The remain-
ing 30 percent is allocated to merchants, middlemen, trans-
port operators, exporting companies, and all those who are
associated with the process of transferring farm products
from farm to market.19 As these groups are likely to reside
in urban areas, the benefits are assumed to fall on urban
households according to their shares in the distribution of
urban family income. The same allocation rules are used for

1969 and 1963.

Power a

are sma
expendi
with e1
mission
Authori
0f the «
EXpendi1
househoj
Authori1
t0 the I
tric pov
(most of
the rema
in PrOpc

to hollse

tramSPOI

 

MiniStr
and a f
miniStr

CUt has

174

Power and Fuel, and Industries

 

The subcategories of power and fuel and industries
are small in comparison to the others. Most power and fuel
expenditures are allocated to government bodies associated
with electricity generation, namely, the National Power Com-
mission, the Provincial Electricity Authority, the Lignite
Authority, the Metropolitan Electricity Authority, and two
of the government-operated oil refineries. That portion of
expenditures which clearly helps provide electricity for rural
households, such as funds spent by the Provincial Electricity
Authority, is assumed to benefit rural households according
to the pattern of their income distribution. Since most elec-
tric power expenditures aid urban more than rural households
(most of the latter still do not have electricity service),
the remaining benefits are distributed to households partly
in proportion to their shares in the total income and the
urban income.

The benefits of refinery expenditures are distributed
to households according to the patterns of expenditures on
transport services, household operations, and total expendi-
tures.

Most industrial expenditures are allocated to the
Ministry of Industries, some of its subsidiary departments,
and a few state enterprises. The allocation of benefits from
ministry services is quite difficult because there is no clear-

cut basis for judgment. The rule adopted here is to allocate

apprOI
basis
tures
benef;
tures

food,

TransE

the la
consis
tion a
are as
the C1
benefi
graph

benefi
the po
catiOn
proper;
hold b
with O

lng ma

 

pOrtiC
SerVi\

trunk

175

approximately one-half the expenditures on a per household
basis and the other half in proportion to income. Expendi-
tures allocated for government enterprises are assumed to
benefit households according to the patterns of their expendi-
tures on the products of such enterprises, whether these be

food, clothing, household goods, or others.

Transportation and Communication

 

Transportation and communication expenditures are
the largest in the Expenditure Group 1. As expected, they
consist of budgetary allocations to the Ministry of Communica-
tion and its various departments. The services of the ministry
are assumed to benefit every household equally, while those of
the Civil Aviation Department (in 1969 only) are assumed to

20 Of the Post and Tele-

benefit only the top income class.
graph Department spending, 25 percent is believed to have
benefited only the top income class, as this figure is roughly
the portion of outlays on international postal and telecommuni-
cation systems. Another 25 percent is distributed on an income
proportional basis, and the remaining 50 percent on a per house-
hold basis. Expenditures of the Port Authority, dealing mainly
with overseas shipping, and of the Department of Harbour, deal-
ing mainly with domestic water transport, are distributed pro-
portionately according to household income shares and the pat-
terns of household transport expenditures respectively. The

services of the state railways are divided into passenger and

trunk categories. Since about 53 percent of earnings in both

196
fro:
utec
ter:

has:

the
asic
give
allc
in 1
that
Year

be b

176

1963 and 1969 came from passenger services, and 47 percent
from trunk services,21 53 percent of the benefits are distrib-
uted to households on the household transport expenditure pat-
tern basis and 47 percent on the total expenditure pattern
basis.

The expenditures of the Departmemt of Highways and
the Department of Land Transport have specifically been set
aside for discussion here so that special attention may be
given to them. They constituted the single largest budgetary
allocation of all other governmental departments in Thailand
in 1969 and the second largest in 1963, obviously indicating
that great emphasis was placed on highway construction in both
years. It is necessary, therefore, that any benefit allocation
be based on a careful study of the highway situation.

Similar to expenditures on irrigation projects, funds
for highway construction are investment expenditures, the bene-
fits of which in any given year are difficult to measure. How-
ever, using more or less the same methodology employed in deal-
ing with agricultural expenditures, it is assumed that thelxnur-
fits of highway spending in 1963 and 1969 are equal to its
costs in the same period. This is not a major problem; more
difficult is the distribution of these benefits among households
of different income brackets.

In his studies on highway expenditures in the United
States and Canada, Gillespie assumes that expenditures are in-
22

curred on behalf of both users and nonusers of highways.

Users are defined mainly as those who own and operate automobiles

and
user
PIOE

near

use]
and

pros
roac
veys
dice
Way

trax

177

and those who consume transported commodities, whereas non-
users are those who benefit from the access facilities to their
property and whose property increases in value as a result of
nearby roads.

Gillespie uses three methods to divide costs between
users and nonusers and the cost share between passenger cars
and trucks. For the former, he employs the relative use ap-
proach: the distance traveled by vehicles on through traffic
roads and local or access roads is compared through road sur-
veys. The percentage share of through traffic road travel in-
dicates the portion of highway expenditures that benefits high-
way users, and the percentage share of local and access road
travel the portion that benefits nonusers.

Another method for separating user from nonuser bene-
fits is the net earning credit approach. It is assumed that
the frequency of use of primary and local roads in relation to
their respective construction costs determines the benefits ac-
cruing to road users and nonusers. Although the costs of con-
struction and maintenance of through traffic roads are much
higher than for access roads, the cost per vehicle mile travel-
led is usually much lower for primary than for secondary roads.
Compared to the relative use method, the earning credit method
tends to yield higher benefits for nonusers. However, the re-
sults of U. S. road studies, as cited by Gillespie, have shown
that both methods give remarkably similar results.

For the division between the benefits to'cars and.trucks,

Gillespie uses the incremental cost method. This divides the

cost 0:
the f 0:
vehicle
vide a
imagine
cars a1
quire c
are nee
inches

COStS c

benefit
States
little
much le
Users i
in Thai
benefit
muSt be
coneent

riVe Va

rOad UQ

 

 

the to‘

and 19

178

cost of road construction into "basic" and "specific" costs,
the former being those necessary to provide a road for light
vehicles, the latter being additional costs necessary to pro-
vide a road for successively heavier vehicles. For example,
imagine a road with a one-foot thick surface for use by both
cars and trucks. It could be established that light cars re-
quire only a four-inch surface; the additional eight inches
are needed for the heavier trucks. The cost of the first four
inches would therefore be spread over all vehicles, and the
costs of the next eight inches over trucks of varying sizes.

This brief summary indicates how the problem of highway
benefits is dealt with in systems such as those in the United
States and Canada, but unfortunately, all these methods have
little applicability in Thailand. There, the road system is
much less developed, the separation between road users and non-
users is not practicable, and, most important there is no study
in Thailand comparable to the one used by Gillespie toaallocate
benefits to private cars and trucks. An entirely new approach
must be adopted which eliminates the user/nonuser dichotomyanui
concentrates mainly on the assumption that all households de-
rive varying benefits from public highway expenditures.

In Thailand it is possible to classify the number of
road using vehicles into three main types: passenger cars,
buses and taxies, and trucks. The prOportion of each type to
the total number of vehicles is assumed to be the same in 1963
and 1969, that is, 60 percent trucks, 25 percent buses and

23

taxies, and 15 percent private passenger cars. It is then

assumed that 15 percent of total highway expenditures in

1963 ;
income
sengei
locate
expend

taxies

of tru
that

beginn
voliune
Departr
ducted
senger
Checkpc
Data a:

fur tru

The Cla
that or
hOuSehC
the fol
eXampi¢

other

 

road tl
fQOd .

cloth

1

179

1963 and 1969 are incurred on behalf of households in the top
income bracket, since only this group can afford to own pas-
senger cars. Twenty-five percent of highway spending is al-
1ocated to all households according to the pattern of their
expenditures on transport services (those using buses and
taxies). '

The remaining 60 percent is allocated to consumers
of truck transported products. Information concerning goods
that are transported by trucks nationwide is available only
beginning in 1966, the first year a preliminary survey on the
volume of road transport in Thailand was conducted by the
Department of Land Transport. These national surveys, con-
ducted each year since 1969, check all vehicles except pas-
senger cars that pass through a number of randomly selected
checkpoints during the two annual week-long survey periods.
Data are gathered on points of departure, destination, and,
for trucks, the commodity content of their hauls.

It is this last information that is of interest here.
The classification of transported commodities differs from
that of commodity groups used in the 1962-1963 and 1968-1969
household expenditure surveys, but it is possible to rearrange
the former classification to coincide with the latter. For
example, rice, maize, vegetables) groceries, meat, fish, and
other such products,which are classified separately in the
road transport surveys,can be grouped under the category of
food. Similarly, weaving materials can be grouped in the

clothing category, and construction materials can be partly

180

assigned to the household operations group and partly to the
transport and total expenditures categories. The next step

is to determine the share of each commodity in the total trans-
ported volume so that the portion of highway expenditures al-
located to trucked transport may be similarly allocated. The
best indicator as to the share of each transported commodity

in total benefits is its total weight, which is equal to the
number of trips multiplied by the weight of each trip. But
since the incremental cost method does not enter into con-
sideration in this study, the different weights of different
trucks does not matter here. Instead, it is assumed that more
benefits from highway expenditures should be imputed to the
commodities which use the roads more often. In other words,
only the number of trips or the number of trucks with different
commodities passing through checkpoints will be used as a guide-
line for allocating highway benefits.

Table 4.2 shows the distribution of highway benefits
assigned to trucked transport by allocation bases. According
to the results of the above mentioned road survey, the largest
percentage of commodities transported by trucks (21.7 percent)
was of mixed types which could not be clearly assigned to any
of the eight commodity groups. So, the benefits were allocated
to households by the pattern of their total expenditures.
Still, a large portion of truck-transported commodities was in
the food and housing categories (20.3 and 19.1 percent reSpec-
tively) of which the benefits were allocated to households ac-

cording to the patterns of their expenditures on food and

181

Table 4 .2—Distribution of Highway Benefits Assigned to Trucked

Transport, by Allocation Bases, 1963a and 1969

 

 

 

 

m . . 1...... Share teams...
ocation Bases of Benefitsby (Million Baht)
Each Allocatim
Base 1963 1969
Food Expenditures 20.3 103.0 312.4
Housing Expenditures 19.1 96.9 294.1
Clothing Expenditures 1.2 6.4 19.3
Transport Ecpenditures 11.7 59.5 180.4
'Ibbacoo quaenditures 3.7 18.8 56.9
Total Expenditures 21.7 110.1 334.0
Total Non-Food Expenditures 7.2 36.7 111.2
Rural Income Distribution 12.2 62.0 188.1
Income of Rubber Growers 3.0 15.0 45.6
Total 100.0 508.4C 1,542.1d

 

 

 

 

Source: Computed fran the 1969 results of the volune of land transport
3 urvey conducted annually by the Department of Land Transport.

aDue to lack of data, the distribution of benefits in 1963 is
assumed to have the same pattern as in 1969.
bSee Appendix D for (htails and explanation of allocation bases.
C'd'Ihese figures are the total anounts of highway benefits assigned
to trucked transport in 1963 and 1969, respectively.

182

household operations, respectively. Other major commodity
groups consisted of rice and other farm products, and fuel.
Based on Table 4.2, a substantial amount of highway benefits

could be distributed fairly easily and objectively.

Other Economic Services

 

Major expenditure items assigned to the "other" eco-
nomic services group include expenditures of the office of
the Under-Secretary of the Ministry of National Development,
the Department of Soil Development, the Department of Land
Co-operatives, and the Bank for Agriculture and Agricultural
Co-operatives (for 1969 only). As most of these expenditures
tend to benefit the farming sector, they are allocated mainly
according to the share of rural households in the rural in-

come distribution.

Group 2: Educational Services

 

As a group, educational expenditures accounted for
less than 20 percent of total expenditures in 1963 and 1969.
Among the four subgroups--primary, secondary, higher, voca-
tional and other-~primary education received the highest share,
with vocational and other education a distant second. In Thai-
land, primary education is compulsory for seven years (Prathom
l to Prathom 7). Secondary education occupies the next five
years of a student's life (Mor Sor l to Mor Sor 5). The stu-
dent then may go on to college, where he normally spends four
years to obtain a bachelor's degree. The above is what gener-

ally is referred to as an "academic" line of schooling. There

183

is also "vocational" schooling from the secondary through
college levels.

Most public primary schools are administered by the
Provincial Administration Organization in each province under
the academic supervision of the Ministry of Education and with
financial support from the Ministry of Interior. Some local
municipalities also operate their own public primary schools
separately. Public secondary and vocational education are
generally operated and administered completely by the Ministry
of Education. All universities were under the control of the
Office of the Prime Minister before 1972, at which time a
special Bureau of State Universities was established.

The functioning of the educational system in Thailand
is not of interest here. What is of concern is to whose bene-
fit the educational expenditures should be allocated. Accord-
ing to the benefit principle adopted in this study, children
attending school and their parents, not the teaching profes-
sion, are the proper beneficiaries of public educational
spending. In his study of educational benefits in the United
States and Canada, where statistical sources usually are among
the best, Gillespie was able to develop two allocation formu-
las: the distribution of school children under 16 years of
age by parental income class, and the distribution of univer-
sity students by parental income class. The first of these
formulas is obviously inadequate because it includes pre-
school children and excludes older students still attending

secondary schools. However, Gillespie pointed out that no

184

better allocation alternative was available.24

Primary Education

 

In Thailand statistics similar to the above do not
exist. Until there is a household survey to determine the
number of children in school by class of parental income, the
only way to allocate the educational benefits of primary and
secondary education is to make reasonable but arbitrary as-
sumptions. Since primary education is compulsory and the
government must provide schooling for every child, it would
appear that primary schooling is somewhat similar to a public
good, and every family benefits equally, and no family can be
excluded. But to allocate actual primary education expendi-
tures equally on a per household basis would be incorrect for
two reasons: (1) expenditure per pupil is greater for schools
in urban than in rural areas, and (2) households in lower in-
come brackets often cannot afford to permit children to at-
tend school all day, or even to attend at all, since they can-
not afford books and clothing and/or the children are needed
for farm work at a very early age. It is evident therefore,
that the income position of households influences who receives
educational benefits. With this in mind, one-half of primary
education expenditures is arbitrarily allocated on a per house-
hold basis for both 1963 and 1969, and the other half in pro-

portion to income.

185

Secondary Education

 

Relatively few Thai students who finish primary school
continue on to the secondary level. In 1963, of all students
attending public primary and secondary schools throughout the
country, 96 percent were enrolled in primary schools and only
4 percent in secondary schools.25 In 1969, these percentage
shares remained the same.26 The problem may lie either on
the demand or the supply side, that is, either parents cannot
afford to send their children to secondary schools, or sec-
ondary and higher educational facilities are lacking or inade-
quate, or both. It is believed, however, that the demand out-
weighs the supply factor. If so, then a family's income posi-
tion plays a significant role in determining the educational
future of children beyond the primary stage. It is assumed,
then, that the benefits of secondary education expenditures
should be distributed to families according to their share in
the distribution of family income. Admittedly, this is not
the best formula, but neither is it unreasonable nor implausi-

ble.

Higher Education

 

In 1968 the National Educational Council of Thailand
conducted a survey of the cost of living of university stu-
dents.27 It originally was intended that at least one ques-
tion would inquire about the income of the student's parents.
For some unexplained reason, this question was excluded from

the actual survey.

186

Fortunately, there is another source of information,
albeit a limited one, on the income distribution of parents
whose children are enrolled in universities. These data,
classified into 11 income brackets, can be found in a publica-

tion of Thammasat University, General Facts About First-Year

 

Students at Thammasat University, 1972-1973. Since no similar

 

information exists for other institutions, the pattern des-
cribed for Thammasat University is used as representative of
all university students. There are strong arguments in sup-
port of this decision. First, the data for both 1972 and 1973
are quite similar (a correlation coefficient of .9907), imply-
ing that students from the same income classes entered Tham—
masat University in these two years. Second, Thammasat Uni-
versity may be considered typical of other universities; there
is no apparent reason to think that the university's policy is
biased in favor of or against any classes of households or stu-
dents, or that any specific groups of students tend to enroll
there more than at any other institution. Therefore, the aver-
age income distribution of parents of first-year Thammasat Uni-
versity students in 1972 and 1973 is used as the allocation

formula for higher education expenditures in both 1963 and 1969.

Vocational and Other Educational Services

 

Most vocational education expenditures are allocated
to households in the same manner as secondary school expendi-
tures. The notable exception is expenditures by the Department

of Teacher Training, which are allocated on the same basis as

primary education benefits (since most prospective teachers

187

will be employed in primary schools). For other educational
services, 50 percent of expenditures on educational administra-
tion, educational techniques, and fine arts are arbitrarily al-
located to households on a per household basis, and the other
50 percent proportional to income. Expenditures on national

research are allocated to all households equally.

Group 3: Health and Social Welfare Services

Expenditures in the health and social welfare category
consist mainly of budgetary allocations to the Ministry of
Public Health, Department of Public Welfare, Department of
Public Works, Department of Community Development, Accelerated
Rural Development Programs, and the Government Pension Plan.
The Ministry of Public Health and its two largest departments,
Health and Medicine, are responsible nationwide for the opera-
tion of most public hospitals and preventative medical meas-
ures such as vaccinations and health education. Thailand has
no health insurance system, but the costs of public hospitali-
zation and other medical services are low, and poor families
often receive various treatments free of charge. The main
problem, however, lies in insufficient medical facilities,
especially the low number of hospitals, health clinics and
doctors. _

An ideal way to account for the benefits received by
various households from public health expenditures would be
to keep records of the income of those who come for medical

services and the estimated share of the costs of such services

188

for each income class. Unfortunately, such data are not
available, and it is necessary to make an arbitrary assump-
tion. Since households in lower income brackets often re-
ceive free or less expensive medical services than do those
in higher income brackets, it is reasonable to assume thattflua
benefits are distributed according to the share of the pp-

ciprocal of household income. In other words, expenditures

 

are allocated in such a way that the poorer the household,
the greater the benefits received.

On the public welfare side, the Department of Public
Welfare helps find jobs for the unemployed, raises orphans,
cares for the elderly, gives temporary relief to households
during natural disasters, and so forth. Similar to health
expenses, these public welfare expenditures are allocated to
families according to the distribution of the reciprocals of
their income. Government pensions, also included in this
expenditure category, are allocated to government employees
using the distribution of civilian officials in 1963 and 1969.
Oddly enough, these pensions are the only explicit public
transfers in the entire range of government spending.

Services of the Department of Public Works in the
area of public utility construction are assumed to benefit
mainly urban households and are distributed according to the
distribution of urban income. On the other hand, services
of the Department of Community Development and the Office of
the Accelerated Rural Development Programs benefit rural house-

holds almost exclusively. Their expenditures are thus

189

allocated to those households according to the share of each
income class in the distribution of rural income.

In all, public expenditures in the health and social
welfare category are relatively small compared to others.
It is somewhat odd that, although most Thais are generally
poor, the government is least active in the area of public
transfers, from which poor households in need of assistance

could benefit most directly.

Group 4: General Services

 

Expenditures in the general services group-—defense,
public administration, justice, and law enforcement--have
been variously called general, nonallocable, and public goods
expenditures. Having established that all individuals or
households derive equal benefit from such expenditures, and
that consumption by one person has no effect on the consump-
tion of another, most researchers allocate benefits from these
general expenditures equally to every individual or household.
However, it increasingly has become necessary to question the
wisdom of such a practice; what is theoretically sound may
not be plausible or valid in reality.

The conventional expenditure incidence study frequently
lumps all general expenditures into one group and then allocates
them to individuals or households using different criteria.
Since there is no definite set of guidelines, several alterna-
tive assumptions are necessary. However, if one looks closely

into the disaggregated details of a general budget, it can be

190

seen that some of the generally conceived expenditures provide
positive benefits to identifiable groups. At least they are
not entirely "unallocable" as usually assumed.28 Because few
funds fall into this category, and because the majority of
general expenditures may be truly general or nonallocable, a
compromised approach is adOpted in this study. Rather than
lump all such expenditures and use alternative allocative as-
sumptions, attempts are made to disaggregate them to the ex-
tent practicable so that a reasonable rule may be applied to
each subgroup. The following section will discuss, in turn,
the allocation rules of expenditures on defense, law enforce-

ment, public administration, and other miscellaneous public

activities.

Defense
Being predominantly a military-controlled country,
Thailand Spends a great deal on defense. The budgetary alloca-
tions to the Ministry of Defense accounted for almost 16
percent of the total budget in 1963 and almost 17 percent in
1969. Unlike expenditures by other government organizations,
defense spending is shrouded in secrecy; no details are pub-
lished, especially concerning a quite substantial portion
set aside for "secret services.f A detailed examination of
defense spending is thus impossible and one must rely only on
the conceptual characteristics of defense benefits.
Theoretically, national defense is often cited as

a good example of a pure public good; it is purported to

191

protect equally the life and property of every citizen. This
is a correct theoretical conclusion, but it also could be ar-
gued that the higher the income or wealth status of an indi-
vidual or family the greater is his desire to protect it. In
short, the wealthier a person is, the more he stands to lose
in the event of invasion and conquest, and consequently the
more he appreciates protection from the defense system.29
Therefore, rather than allocate benefits from defense spending
on a per capita or per household basis, one could allocatetflmmi
according to the share of each individual or household in the
distribution of income (or wealth if the data are available).
Such is the allocation formula adopted in this study. In con-
trast to the conventional per capita allocation, the propor-
tional allocation rule is gaining wide acceptability. A. C.
Harberger, for example, has said:

There are a great many government outlays

for general purposes (administration, the

courts, national defense) whose assignments

as benefits to particular income groups is

necessarily quite arbitrary. The notion

that these benefits are roughly proportional

to income and/or wealth, however, seems at
least to be a plausible approximation.30

Law Enforcement

 

Expenditures in the law enforcement category can be
divided into two subgroups, justice and police. For the jus-
tice system (Ministry of Justice, Department of Public Prose-
cution, and Department of Correction), it is reasoned that,
as every person or household should be equally served by the

due process of the law, benefits should be distributed equally

192

to all individuals or households. This allocation rule is
rather straightforward and quite obvious.

The same cannot be said for police expenditures, which
rank very high among major government outlays. It is true
that the strength of police forces is proportionately much
greater in urban and metropolitan than in rural areas. It is
also true that special branches, such as traffic, fire, and
highway police are oriented toward serving individuals or house-
holds in central cities or towns which are mainly in the higher
income brackets. Consequently, it would be incorrect to assume
that every household in the nation enjoys the same service from
police operations. Rather, it would be more appropriate to di-
vide the police budget in half and distribute the first half
to households on a per household basis and the other half to
.households on an income proportional basis. This formula is

used here.

Public Administration

 

Expenditures of government organizations which have
not been classified previously are here classified as public
administration expenditures. These are quite large, amounting
to almost 14 percent of total expenditures in Thailand in 1963
and 1969. Two types of government organizations are included
in this category: those constituting general bureaucratic ad-
ministration, such as the Department of Local Government,
Ministry of Finance, Department of Land, and Ministry of Eco-

nomic Affairs, and those that deal with other government bodies

193

or the general affairs of the state, such as the Ministry of
Foreign Affairs, Department of Technical and Economic Coopera-
tion, State Service Commission, and Office of Government Audit.
Normally one would allocate public administration ex-
penditures on an equal per capita or per household basis be-
cause they are aimed at satisfying social wants. Neverthe-
less, it could be argued strongly that in Thailand, where the
administrative system is very centralized, the effects of
general government are more likely to be felt in urban than
in rural areas. It is more correct, therefore, to allocate
benefits according to an arbitrary 75—25 percent rule; 75 per-
cent of the bulk of public administration expenditures is dis-
tributed to each household equally, and the other 25 percent
on an income proportional basis. A portion of these expendi-
tures primarily benefit government officials, for example, the
Government Official Assistance Fund and government scholarships.
This portion is distributed to official households according
to their shares in the income distribution of government of-

ficials.

Other Miscellaneous Expenditures

 

Some expenditures are set aside for the government's
Reserve Fund, which is used during natural disasters or for
augmenting certain regional development projects which unex-
pectedly run short of funds. Since no details are given as
to how the Reserve Fund is spent, it is allocated equally among

households.

194

Group 5: Interest Payments on the Public Debt.31

The long-standing controversy as to who should benefit
from interest payments on the public debt has yet to be de-
cided. There are various ways to look at the nature of public
debt and the effects of its interest payments.32

First, in a classical setting, where full employment
is assumed, interest payments can be construed as the cost
or factor return incurred by the government in providing cer-
tain goods or services through loan finance. If this view is
adopted, then the benefits of interest payments should be al-
located to families in proportion to their consumption of
debt-financed goods and services. But since government loans
are not always tied to provision of specific goods and ser-
vices, it is quite impossible to determine who benefits. Even
when a project is known to be specifically financed by a cer-
tain loan, it still may be very difficult to estimate the pro-
portion of goods and services consumed by families in each in-
come class.

Second, in a less than full employment setting, the
creation and liquidation of public debt is seen as a fiscal
compensatory action of the government. Assuming the government
could monetize the debt if it so wished, its decision not to
do so could be considered the desire of the government to sup-
press an inflationary increase in private expenditures. In
other words, interest payments can be seen as the price paid
for the purchase of nonspending or illiquidity. If this view

is adopted, the beneficiaries of such payments would be those

195

affected in one way or another by liquidity control, something
very hard to pinpoint, of course. As McLure has pointed out,
allocating the true benefits of buying illiquidity and reduc-
ing aggregate demand is a complex question, and the answer is
far from clear.33
Third, interest payments on public debt can be con-
ceived simply as transfer payments which serve to redistribute
income in the course of providing loan-financed goods and ser-
vices. It is reasoned that, since taxes are levied to pay for
interest payments on public debt, and if the distribution of
tax payments by income classes is not identical with the dis-
tribution of interest payments, there will be a redistribution
of income among families. To treat these payments as transfers
is also consistent with the general practice of national income
accounting. If this View is adopted, then interest charges
should be allocated to families that receive the payments.
Despite McLure's insistence that allocating benefits
of interest payments to holders of the public debt makes no
sense,34 this approach offers the most practical way to trace
the beneficiaries, and it has been adopted by most researchers

35 For lack of a better alternative,

on expenditure incidence.
this approach also is adopted here.

The next immediate question is: On what basis should
these interest payments of the Thai government be allocated
to which recipients? There are four main groups of domestic

debt holders: the Bank of Thailand, Government Savings Bank,

commercial banks, and other financial institutions and domestic

196

sectors. Of these, the commercial banks and other financial
institutions combined held about 38.0 percent and 38.8 percent
of the domestic debt in 1963 and 1969, respectively: the
share of the Bank of Thailand was 41.4 percent and 31.4 per-
cent, and that of the Government Savings Bank 18.5 percent

and 29.5 percent.36 A very small share was also held by the
national Exchange Equalization Fund. Therefore, about 40 per-
cent of the benefits from interest payments accrued to share-
holders of commercial banks and other financial institutions,
and they are likely to represent only from the top income
bracket. Small savers in the lower income groups who normally
save with the Government Savings Bank are quite numerous, but
their share of benefits is small compared to major corporate
savers. As a result, and due to a lack of information on the
distribution of interest payments by income classes, one-half
of interest payments on public debt for 1963 and 1969 are al-
located entirely to the top income group, and the other half
is allocated to all families in proportion to their share in

total income.

Empirical Results of Expenditure Analysis

 

The procedure for estimating expenditure incidence is
identical to that used in estimating tax incidence in Chapter
3. After deciding upon criteria for allocating all public ex-
penditures disaggregated down to departmental levels, the ex-
penditures are distributed to various income brackets based

on the chosen criteria. The basic results, as shown in Tables

197

4.3 and 4.4, are the absolute amounts of expenditure benefits
by income classes, by five main groups of expenditures and
their related subgroups. As in Chapter 3, these basic results
may be presented in two ways: effective expenditure rates by
absolute income levels using both money income and adjusted
income as bases, and the change in the pattern of income dis-
tribution as a result of public expenditures (the so-called
post-benefit income distribution) and its related change in
terms of Gini concentration ratios.

Before examining the empirical results, the reader
must be cautioned about the nature and procedure of this ex-
penditure incidence study. First, the estimates of absolute
benefits by income classes are exactly that: estimates. The
fact that they are quantitatively stated should not mislead
one into thinking that they are the exact "benefits" accruing
to different income classes. The beneficiaries are families
on whose behalf the costs of public expenditures are incurred,
and some benefit allocation criteria are quite arbitrary owing
to various conceptual and statistical difficulties. Moreover,
these families have only one characteristic in common, their
income level; the estimates say nothing about such interfamily
differences as the size of the household, age and education of
heads of households, their occupations, and so on.

Second, when the average expenditure benefits per family
are obtained by dividing the class benefits by the number of
families in that income class, the result is the benefits of a

typical or representative family only. In reality, each family

198

 

 

 

 

 

 

 

 

mp2}: «583 ASS; immim 0.8m; fimmms A38.
:2; 98A. mi TR: was 0.3. noon 33.3 no Bow—son
as: 93 93 8mm hmm mam 850
33; ~63 and «SH ~§~ has. Eggnog ofifio
has A32 88 1162 on: with uogmonomfi 33
p.53 mém NJ: as? imam hm: 8.8an
Name; mama :18 as? The ads; 8088 A898
mama Emma home as». 32: 3.2 magnum momma: a £891
oémm 9% 8mm 3K 1.4m was go pom 83388
114.3 1:3 3: «.1 To as 88E
38 was so... C.3 Tam mom @888
3.8; TEN mom ~68 do: 92... go
met; 3% as: seam oém «a: moonsom 83888
dmmm ado 98 4.8 TS «.2. 850
mas; has he: SEN an: ofim ooflfluonmfih.
m . as m .2 A. 3 o .3 A Am 98 moehmoofi
EAR 93 man 9: Tom mam Hon...— 8... .858
to: some m .21 can he: 92A 83883
Teams 1152. TEN 31$ imam spam «pogo oeaSom
8820 .88 Ba 118.2 an: 8a.... 8?».
s I s I s I. s
A2 8o 3 8o 3 So a 8a m .88: noenooouno 8353a
Bone moonSo 8.85

 

 

 

Ann—om mo msonflnz 5 3595.
mom." 63630 9505 an .mousuflofimxm mo momma >9 .muflmoswm gagfillmé manna

199

 

 

m.omm.- A.~mo.m ~.oH¢.A m.mm~.a A.~me.~ ~.~am.~ a.mmo.~ m.m~o.~ Among
888A ~80. ~.~a 88 8B 88 83 88 88 838 no 328
A68 843 82 83. 82. ~82 88A 88A 850
888m 88s 88: 88~ 88m 8:; 88m 5.8m. 881...me 833
888A ~.~$ 88 8m: ~84 83~ ~.:.~ ~.$~ ufihonflé 33
m.me~.~ m.msm.a m.am~ m.oom m.mqm ~.mmm ~.asm o.mme omooumo
~.ome.m A.~mm.m m.smm A.o~s m.oaa m.AmH.H ~.omm.H m.m~o.a nooeonom Hoooooo
A88~ ~48 88~ 88~ 88~ ~48 82H 88~ 8088 2882 a 5A8:
8A: 82... 83 8mm 88 88 ~.$ m8 ammo o8 88382,
8.8 ~88 8- 8- 88 88 8~ m.~ 888
88v ~48 8R ~.~m 85 82 8R #2 88.88
888A 89% En: ~83 8A3 88~ 88m 88~ Esp
8453 888; H.8~ 8mm~ 8on 83. 85. 83m 888m A8388
:8 8H: 8- 88 8% 88 85 A.m~ go
8am8~ 888A 88A 88~ 88~ 88~ 8B~ 848 8888889
35 88 8m 82 82 83 83 8: 8883
A88 82: 88 8m~ 88 8mm 8:. 83 Anon o8 838

o . -~.~ 888 a . 3H m .2: m .o- A. A8 a .8~ 8 «S 83838
888m 888~ 8.8 88 :8 ~58 ~48 88m 8338 8888
momnflo 826 o8 885 88: 883 88m 88m 88m

A2 888 .888 -882 -88m $86 -88». ~85

 

 

 

 

 

 

 

 

Gnome mommflo 8.85

 

mowuomoumu maven—59$

 

 

38mm m0 9338. 5 Boonie
mood momma-HO mEoQoH an. 683%? NO woman. an .Bflmocwm managinvé manna.

200

within the same income bracket receives different benefits.
For example, if one family has children in public schools or
universities and another family has none, the benefits from
educational expenditures clearly would fall on the first fami-
ly at the expense of the second. Furthermore, one family may
pay taxes, while another family successfully manages to evade
them. These discrepancies must be assumed negligible or to
cancel each other out for manageability in research.

Third, the ostensibly precise nature of the effective
expenditure rates must not be overemphasized. Such rates are
merely the result of mathematical division and should be taken
as a general approximation of how much is added to original

family income as a result of public expenditures.

Effective Expenditure Rates by Income Classes

When the total income of each income bracket is used
to divide each corresponding amount of expenditure benefits,
as shown in Tables 4.3 and 4.4, the effective expenditure
rates are obtained. Tables 4.5 to 4.8 show these effective
expenditure rates based on money and adjusted income, respec-
tively. For all income classes, the benefits are classified
by functional categories of expenditures.

As can be seen from Tables 4.3 and 4.4, the absolute
amounts of expenditure benefits accruing to each income class
are not very informative, apart from the noticeable fact that
the top income group also receives the highest share of bene-

fits.

201

 

 

 

 

 

 

 

 

 

~.8 88 88 8- ~4~ 88 H38.
8m 80 8A 8A 8A 8A 88 33:3 :0 magma
8o 80 ~.o 80 8o 8A .850
3. 8A 8~ 8m 8m 83 893% 338
8~ AA 8H 8H 8m 8... ”8.8885 33
8s 8s 8v 83. 8s 81‘ 88mg
8: Es 8m 8: 82 8s~ noonnom 8.88m
~.~ 8A 8... a m 8~ in 833.38 8883 o 588
80 8o 80 8o A; 5H 880 o8 888803
8a «4 8o ~.o u u 888
ho ho to To To To E88
8... EA H.~ 8~ A m 81. goo
~.m 8s 8s m a 8m «8 833.38 H8888
To 8o To 8o 80 A; 850
on 8m 8~ 8~ 8m 81. 5388888.
8o 8° 8° 8° 8a A; 88835
8a 8o 80 8o 8o 8o A8...— 98 838
m.~ e4 8A m.~ 8~ 8a 85883
8A To .8 8o ms 8: 80858 3.588
888 .36 pa 88S 88: 88m 88m
s s s s
A2 .8 3 $8 ~A 18 o 18 m n85 88888 magmas
Aunmmv mommmao 9:85

 

 

mos: so 033 .mommmau 8505 an .mousuflosmmxm mo monks an .8umm ousuwocomonm 938mmmlmé 83mm.

A8omoooono$ $2 .885

Table 4.6—Effective Ebcperditure Rates, by types of Ebcpenditures, by Inccme Classes, Based on
Adjusted Income, 1963 (Percmtages)

Incone Classes (Baht)

202

 

 

 

 

 

 

Hg m OVNVM m MMNV N H I-INwN m m
o o o o o o o o c o o o o o o o o a o
HUI m HCCHO N H000 H L“ Nv-Io-IO H 00
4:6 H
Fl
0
ch”!
88 H I‘Q‘Hgm N CDMFM m I" mxom' N O
o o g o o o o o o o
m‘ m OOOHO N 0000 O m NCO m 2
Hg
I
88 N OVNMM O HMNQ‘ m \D VQMI—I m m
o o o o o o o o o o o o o o 0
cm "I HOOP-IO N u—looo H V Net-IO o f:
~§
NI‘
I-II-l
I
83: m I—IQ‘NNQ‘ v-I WWII-1Q: m m ”@O‘N w m
0 O O O Q o O o o 0
90‘ m v-IOCI—IO N H000 H m Nev—Io O 2
Ԥ
EDI—I
r-l
I
om Ln MMMMQ- m \Dm In C CD OMNM I‘ V‘
°¢ o o o o o o o o o' o o o o o o o o 0
cm W HOOHO N HO 0 H In NI—INO O 2
s 5
mm
C
8 Q‘ Nmmem I‘ ON'LO ON 0‘ Vle-n ln Q‘
0 0 I o o O o O o 0 o o o o O o o 0
(r; m HOOP-Io N No C O \O HHMO o :

 

 

 

Ebcpefliuire Categories

Economic Services

 

Power and Fuel

Agriculture
IndLstries

Transportation
Other

Educational Services

Vocational and Other

Secondary
Higher

Health & Welfare Services

General Service

Public ministration

Law Ehforcenent
Other

Defense

Payments on Prblic Debt

‘Ibtal

 

Table 4. 7-Effective Expenditure Rates, by Types of Expenditures, by Income Classes, Based on mney
Insane, 1969 (Percentages)

Income Classes (Baht)

203

 

 

 

 

 

 

 

 

m
38 N Q'Q'Mkoln I‘ Hva-l \D H mfi‘hN m ‘0
0 o o o o o o o o o o o o o o o o o o
g 0‘ MOCQ‘O In ”OCH "3 V mNQ‘l-I H a
PI H
U
0
8% In l‘fl'maflm 0‘ O‘Dva‘ ln 0 qu'Q‘ a) I‘
o o o o o o o o o o o o o c o o o o 0
§ w NOOQ‘O v NOo-lo N O Inf-INC N m
mg H N
H
I
00‘
8% m mmNQLn (n WWI-no In 0‘ met-Io C N
o o o o o o o o o o o o o o o o o o o
‘\ m moomo Q‘ Noel—I Ln N LnNQ'O-I H N
In" H ('5
HF!
I
ca
88 m mle-Im o mWQ'v-I H I‘ ml-‘IQC o W
o o o o o o o o o o o o o g o o o o o
\s m MCOQ'O tn NCOv-I In M vao-l PI 00
NV H m
.4.-I
I
00‘
00‘ N mmNI-‘IW M NKDMF‘I O Q‘ mmmm o m
cos 0 o o o o o o o o o o o o o o o o o o
~~ OI mcoqo l-n MOOH Ln Ln LannI-I H In
G‘F‘I H m
._.
ég“ o NmNMF o O‘WMN I.“ m ml-Ilnafl o I‘
o o o o o o o o o o o o o o o o I o o
00‘ o Q'oovo \D MCOH m I‘ lnMIOI-I PI I‘
~~ PI H M
W”
I
CO‘ 0‘ ommmm ‘0 mm In C W mwv—If-‘I O 0
CG 0 o o o o o o 0 0| 0 o 0 o o o o o o
oa‘ o moofi'o I‘ \06 H m N mfi‘mm H I.“
~ ~ H N Q
mm
50
o o wawo In mm m o m ml-‘IO‘I‘ o O‘
o o o o o o o u 0 0| I g o o o o o o o
g~ In Noemi-I Q‘ Ho N H Q‘ “30°F H In
m H H H H V‘ F'IN m

 

 

 

Expenditure Categories

 

ervices

Emmmic S

Agriculture

PowerandFuel

Industries

Transportation
Other

Edmational Services

Higher
vocational am! Other

Health & Welfare Services

ervices

General S

Defense

ation

law Enforcement
Ptblic Adnin '

Payments on Ptblic Debt

Total

 

Table 4.8-Effective Expenditure Rates, by ‘Iypes of Expenditures, by Income Classes, Based on
Adjusted Incane, 1969 (Percentages)

Income Classes (Baht)

204

 

 

 

 

 

 

 

 

g H Ian—ION In VMQ‘In \0 v \Dl—Il—IID 0‘ V
o o o o o o o o o o o o o o o o o o c
H
U
o’d
88 H MNHMN Q‘ mMI‘Q' N co mhr-IN M I‘
o o o o o o o o o o o o o o o I o o 0
~ V‘ HOONO N COCO H V NOu—(O H m
m'E v—l
._.
(U
I
00‘
O . . O . O . . . C O O O C . O . O .
~‘ fi' l—IOONO N H000 N \D MHNC 0 ID
In“ H
F'IF-I
I
00‘
88 In mNu-h-lm In VMNID ID as O‘l-IQ'In In C
o o o o o o o o o o o o o o o o o o o
~~ Q‘ HOONO N Hooo N IO NHNO o I‘
NV v-I
HI—I
I
88 N I‘Nv-IO‘M Q. mMNIn m 0 \DNIOID Q q
a o o o o o o o o o o o o o o o g o o
00‘ Q‘ HOOP-IO N Hooo N I‘ NHNO O S
s s
¢H
v-I
ég“ I-I [\NHmm In WWII-(In Q‘ f—I Q'mmm v Q'
o o o o o o o o o o o o o o o o o 0
cm Q‘ HOOP-Io N HOOD r-| I‘ (Vt-INC 0 I3
s s
ID“
I
00‘ m I‘Nl—Imm IO CNN In C m CIDF'If-I m In
cm 0 o o o o o o 0 0| 0 o o o I o o o 0
cm (’3 HOOHO N HO o H I‘ Nl—IMH O a
~ ~
MID
o
o H ONNan 0 mm ID 0 m \DmmF-I m I‘
o o o o o o o o 0 0' o o o o o o o o 0
~ Q' NOOF‘IO Q mo 0 m N I-INInN O m
m H N

 

 

 

Expenditure Categories

 

Economic Services

Agriculture

PowerandFuel

Industries

Transportation

Educational Services

Health & Welfare Services

General Services

Law Enforcement
Public Administration

Defense

Payments on Public Debt

Total

 

205

When expressed in terms of effective rates, the impact of pub-
lic expenditures upon the income distribution of households in
1963 and 1969 becomes clearer. Using money income as the base,
the lowest income bracket clearly received the highest share
of benefits from public expenditures in both years. The ef-
fective expenditure rates for the lowest,income group were

49.8 percent and 85.9 percent in 1963 and 1969, respectively,
compared to rates of 26.8 percent and 28.7 percent for the
highest income group in the same years.

The effective rates were much higher in 1969 than in
1963, meaning, in part, that the benefits accruing to the low-
est income group were much greater in 1969 than in 1963. This
pattern also applies to the top income class. As a matter of
fact, the effective expenditure rates across the entire in-
come range improved in 1969 over 1963, meaning essentially that
during these six years the expenditure benefits for all income
groups increased faster than their money income, and all clas-
ses received relatively higher gains from public spending in
1969 than in 1963.

When the effective expenditure rates are computed using
adjusted income as the base, the results are understandably
lower than the rates based on money income. The change, of
course, reflects the pattern of adjusted income distribution
vis-a-vis money income distribution in that the effective rates
of the lower income groups, particularly the two lowest ones,
are now only slightly larger than those of the higher income

groups. As indicated in Chapter 2, this is because, in the

206

adjusted income as compared to the money income distribution,
the income positions of the lower brackets increase relatively
more than those of the higher brackets. This change seems to
have "flattened out" the distribution of effective expenditure
rates across income classes from the lowest to the highest.
The effective expenditure rates from the bottom income bracket
is now shown to have been 14.4 percent and 23.7 percent in.l963
and 1969 respectively, whereas the corresponding effective:rates
for the top income bracket are 28.7 percent and 13.0 percent.

The distribution of effective expenditure rates by in-
come classes in 1963 and 1969 using both money income and ad—
justed income as bases can be depicted graphically. This is
shown in Figures 4.1 and 4.2. It can be seen that, on a money
income base, the income distributional effects of public ex-
penditures were generally regressive or pro-poor in both years
(and more so in 1969, as noted above). But the degree of re-
gressivity falls off rapidly after the first income bracket
and remains only slightly regressive for the rest of the in-
come classes. The exception is 1963, when the effective rates
turn slightly progressive at the highest income bracket. On
an adjusted income base, the pattern of effective rates seems
to have levelled off considerably, as noted earlier. The gen-
eral pattern could not now be called clearly regressive, be-
cause there is some fluctuation in the middle income range in
both years.

Similar to the treatment on the tax side, if one re-

gresses these effective expenditure rates on their respective

207

area on. new“ .nonnaao o-oocH an .vnoocm 50:0: :0 con-m uovum ounaauconxm o>aaoommmuua.: ouswuh

3.3 no 8:323
mooaqao ouoocH

a 3 2 a s n 0

ON

om

 

 

ll

Aacoouomv moons
Saigon—nu

u>uuooumm

208

mow" use mead .uomuqao onoocm an .o-oocH coauzfiu< no conga noun: ounpavconxm o>uaoommmno~.a ousmuh

Aandm mo nusuusonav
nounaao osoocH

| l“||'ll|l‘||'ll'l'l I
-'|lil'|'|'l- I _I"'-""""'

 

$2
8.: .: ........

fom

vow

 

A

v

Avcouuomv moan:
ounaucnonxm
o>uvoouum

209

average annual income levels using the ordinary least-square
method, the resulting slope of the regression curve could be
used to roughly indicate the degree of regressivity of the
expenditure system.

As shown in Table 4.9, the slopes of the regression
curves that indicated the pattern of the.l963 effective expen-
diture rates based on money and adjusted income are estimated
at -.369 and -.026, respectively. For 1969, the corresponding
slopes are -.935 and -.076. This generally confirms the con-
clusion that on a money income base the expenditure rate pat-
tern is slightly regressive, but on an adjusted income base
it is barely regressive, so that for all practical purposes
it may be regarded as proportional (particularly for 1963).

The detail of effective expenditure rates by type of
expenditures also may be found in Tables 4.5 ‘U3 4.8. On the
average, across income classes, in 1963 and 1969, the effec-
tive rates of expenditures on general services were highest
followed in second and third places by expenditures on eco-
nomic and educational services. In 1963 the average effec-
tive rate for public debt services was higher than that of
health and welfare services, but in 1969 the situation was re-
versed. Among budgetary allocations for general services, de-
fense always generated the highest average rate of benefits
to households. General public administration, primary educa-
tion, and transportation and communication were among the ma-
jor expenditure sub-categories that generated high average

effective expenditure rates. Effective rates of expenditures

210

Table 4.9--Generalized Expenditure Incidence Functions
KEffective Expenditure Rates) = a + b (Annual
Average Money Income)a]

 

 

 

Intercept SIOpe R2
(1) On Money Income Base
1963 37.0 -.369 .9546
1969 54.9 -.935 .8982
(2) On Adjusted Income Base
1963 13.7 -.026 .9983
1969 19.0 -.076 .9809

 

 

 

 

aAverage Annual Money Income is in Thousands of Baht

211

that were presumed to benefit poor households relatively more
than the rich, such as agriculture and health and welfare ser-
vices, even taken together, were quite small, accounting for
only 4 percent of household money income in 1963 and 7 penxnu:
in 1969. Obviously, the lower income brackets gained when the
two types of expenditures increased in 1969 compared to 1963--
notably in agricultural and irrigation spending--but the in-
crease was still quite small on the whole. A fairly high rise
in benefit rates also came from spending on general services,
namely, defense and public administration.

Comparing the effective rate of each type of expendi-
ture among income classes, the lowest bracket generally re-
ceived relatively higher expenditure benefits than any other
except on such items as higher education and debt payments.

In these two expenditures, the top bracket was the greatest
beneficiary. That the total effective expenditure rate for
the lowest income group based on money income was very high
must be noted carefully. The bulk of these benefits (about 50
percent) was generated by such expenditures as defense, law
enforcement and public administration spending. It may be ar-
gued that from the point of View of the poorest household
these benefits might be ranked lower in utility than the same
amount of benefits from such expenditures as agricultural,
educational, health and welfare spending. However, this point
is value-loaded and, as such, should be left to the judgment

and evaluation of each individual.

212

The Effects of Expenditure Programs
on the Distribution of Income

 

 

The previous section dealt with the usual way in which
the income distributional effects of public expenditure pro-
grams may be demonstrated by effective expenditure rates by
income classes. This section discusses the effects of public
expenditure programs upon the pattern of income distribution.
In other words, the focus is directed to the change in the
pattern of househOld income distribution after expenditure
benefits are computed and assigned to each income class. Once
the absolute amounts of estimated benefits are added to the
income of each corresponding income bracket, a new distribu-
tion should emerge. Table 4.10 shows these new distributions
based on money and adjusted income concepts.

Inspecting Table 4.10, one notes the small degree of
income redistribution as a result of public expenditure pro-
grams. On the money income base, the degree of redistribution
is quite apparent, particularly in the lowest and highest
brackets, but on the adjusted income base, the pre-fisc and
the post-benefit income distributions are shown to be almost
identical in 1963 and 1969. In other words, the expenditure
programs of the Thai Government in 1963 and 1969, considered
alone, effected very little change in the pre-fisc distribution
of household income. This conclusion is supported by the change
in the Gini concentration ratios.

As shown in Table 4.11, within each individual year

there is an improvement in income equality as indicated by the

213

 

 

 

 

 

 

 

 

 

o.ooa o.ooa o.ooa o.oon o.oo~ o.oo~ o.oon o.oon mmmmmao Haa
m.~v m.m¢ v.om m.om o.om m.~m s.~m m.mm um>o_u:m ooo.wH
m.m m.m m.o m.m mmm.sauooo.ma
~.HH m.HH i m.~a I H.ma A
~.s H.s » o.w H.m maa.¢~-¢oo.~a
9m 9m ma 3.. mmmfiuoooa
o.s~ o.v~ I m.m~ _H «.mm A
m.HH o.HH m.on m.oa I amm.m -ooo.o
m.HH m.HH m.vH m.¢H o.on m.m H.¢H o.¢a amm.m -ooo.m
m.m o.m a.m~ m.mH ~.m k.m m.va o.m~ ooo.m “was:
udwfimbmom 08E unusambmom omEbE uflmcmmbmofiomflufim £088.38 03me
mean mama moms moms Insane

mmmmMHU . 9:85

 

aflfiflfiflo 283 Bums DE

 

GOUBHHEAQ 9805 >982

 

 

 

$3 can mom.” .mmmmmfiu 0.505”
an 9605 @3ng can mwcflz ufimmcmmuumom new omwmnmum mo cQwubnfi—umflo mmmusmoummlloaé manna

214

 

 

 

 

mm.mu mmmv. mmmv. mmmm
mm.m+ I mm.m+ A .
mm.ou mmmv. mmmm. momm
$8 9505 839.9% 5 .m
. mm.m- mmmm. ommm. momm
mm.mu i mm.mn A
mm.mu momm. mmmm. mmmm
mmmm_maao:H mucazjco .m
coflsflfiflo 838:5 ad .89 Smusflfimd cogs—Emma
9605 93H m5 5:33 «BUSH 90ch
ummmmamaumom ommmumum 858088 uflmfimuumom omEbum

 

83 can mama 503qu
”Emmomumwoo go
5 at 38905 no
TV mmmmuooo mmflfioumm

 

H50 cum

TI mmmwuocH no

TV mmmouooo
magnum

 

 

mucwwowmmmoo H50

 

HMO?

 

 

mmmm new mmmm .mummmsmmcm
mEoocH cum 80298 mmficwamm new moflmm Signoocoo H50 ummmcwmlumom paw Damaging . v 0309

215

reduction in Gini concentration ratios from the pre-fisc in-
come distribution to the post—benefit income distribution. In
1963 there was an improvement of 5.7 percent in income equality
(or a 5.7 percent reduction in income inequality) when the
post-benefit income distribution was compared to the pre-fisc
income distribution based on household money income. The com-
parable improvement in 1969 was 7.2 percent. As expected,

the percentage improvement is greater with distribution based
on money than on adjusted income. Although when each indivie
dual year is viewed, the 1969 expenditure programs are shown
to have more income redistributive power than the 1963 pro-
grams. The effect is not sufficiently powerful or redistribu-
tive to offset the extraneous factors that caused the deterio-
ration of the adjusted income distribution from 1963 to 1969.
This is evidenced by the fact that the 1969 post-benefit in-
come distribution was about 4.6 percent more unequal than the
1963 post-benefit income distribution. As this is still smaller
than 5.8 percent, which is the rate by which the pre-fisc in-
come distribution had worsened from 1963 to 1969, one could
conclude that, on an adjusted income base, government expendi-
tures reduced income inequality by 1.2 percentage points from

1963 to 1969, which is not a large amount.

Summary and Conclusion

 

The statistical estimation of public expenditure bene-
fits or expenditure incidence--the subject of this chapter--

is the obverse side of a fiscal incidence study, the reverse

216

being the tax incidence estimate. The chapter began with a
discussion of the framework in which the study is to be car-
ried out: the definition of a benefit, the identification
of beneficiaries, and the selection of proper beneficiaries.
Also discussed were various theoretical difficulties associ-
ated with the quantification and allocation of the benefits
of specific expenditures when beneficiaries are identifiable,
as well as expenditures on such public goods as national de-
fense, public health and law enforcement.

In general, the conventional total benefits equal to-
tal costs approach, or the approach which equates benefits to
"costs incurred on behalf of" expected beneficiaries, has been
adopted in this study. It offers the most consistent and
practical way of dealing with the empirical estimates of ex—
penditure incidence.

After the 1963 and 1969 expenditures were classified
into functional categories,and each type or category was ex-
plained as to its nature and the bases upon which the intended
benefits could be allocated, all expenditures were systemati-
cally distributed to households in different income brackets.
It was noted that the five major groups of public expendi-
tures in Thailand, classified according to their function,
are: economic, educational, health and welfare, general and
debt payments. Of these, general service expenditures, en-
compassing defense and general public administration, are the
largest.' To determine the absolute amounts of expenditure

benefits by income classes, money income and adjusted income

217

bases were used as denominators to arrive at the effective
expenditure rates by type of expenditures as well as by total
expenditure. These calculations form the basic estimates of
expenditure incidence.

Expenditure incidence in Thailand in 1963 and 1969
ranged between slightly regressive to mainly proportional de-
pending upon the income bases used. When money income is used
as a base, the incidence pattern is generally regressive be-
cause the distribution of money income had a much wider dis-
parity than the distribution of adjusted income. When adjusted
income is used as the base, the incidence pattern becomes
practically proportional. The lowest income bracket usually
received the highest benefits per income than all other in-
come brackets except for such expenditures as higher education
and debt payments.

Agricultural, primary education, and health and welfare
expenditures proved to be the major items benefiting the lower
income households relatively more than the upper income house-
holds. The fact that transportation expenditures benefited
the lowest and the highest income classes relatively more than
others may be because the lowest group, in the farming sector,
gained through easier access to markets to sell their produce,
and the highest group benefited through shorter travel time
for business and pleasure. Other expenditures were small and

tended to benefit all income classes equally.

218

The change in the pattern of income distribution after
government expenditures were allocated was quite small on the
money income base and even more so on the adjusted income base.
The computed rates of improvement in income equality due to
government expenditures range from only 1.5 to 3.0 percent,
depending upon which income concept is used.

A general conclusion which could be drawn from this
chapter is that public expenditures in Thailand as represented
by the 1963 and 1969 budgets have done very little to redis-
tribute income between rich and poor households. The im-
provement in income equality as a result of these expenditure
programs in the situation where they could be most redistribu-
tive (that is, when money income is used as a base) was less
than 8 percent in 1969 and 6 percent in 1963. Otherwise, the
effects of government expenditures left Thai households in
very much the same condition as before the inception of the
programs.

Although it is true that the 1969 expenditure programs
were more income redistributive than those in 1963, it is
doubtful that this can be taken as an indication of the gov-
ernment's intention or deSire to redistribute income or re-
duce income inequality through expenditure policies. In 1969
there was a noteworthy increase in relative benefits going to
lower groups at the expense of higher income groups through
expenditures in such areas as agriculture, primary education,

transportation and health, but such increases were relatively

219

small on the whole.

In all, the role of the government between 1963 and
1969 as an income redistributor through its expenditure poli-
cies was rather limited. If the lower income brackets are
to benefit more from government services, emphasis should be
placed more on funding programs more attuned to their needs,
such as various agricultural services, basic educational
facilities, improved health and welfare services, and so on,
than on such general services as defense, law enforcement and

public administration.

FOOTNOTES

1John H. Adler was the first to use this categoriza-
tion of approaches. See his study, "The Fiscal System, the
Distribution of Income and Public Welfarefl in Fiscal Policies
and the American Economy,ed. Kenyon E. Poole, (New York:
Prentice-Hall, Inc., 1951), pp. 359—421.‘

 

21bid., p.362.

3See Charles E. McLure, Jr., "On the Theory and
Methodology of Estimating Benefit and Expenditure Incidence,"
paper presented at the Workshop on Income Distribution and
its Role in Development, Programs of Development Studies,
Rice University, April 26, 1974, (mimeographed). McLure
distinguishes between the "expenditure incidence" and the
"benefit incidence" with the former conforming to the study
in the change in the distribution of income resulting from
changes in relative product and factor prices, and the latter
the change in household or individual income position after
receiving real and imputed benefits from public expenditures.
Since McLure's expenditure incidence study exists only in
theory, the special meaning given to expenditure incidence
which normally is interpreted as benefit incidence only serves
to confuse the matter, at least in the context of the present
study. Therefore, no such special meaning is followed here,
and expenditure incidence and benefit incidence are still re-
garded as synonymous.

4W. Irwin Gillespie, "Effect of Public Expenditure
on the Distribution of Income" in Essays in Fiscal Federalism
ed. Richard A. Musgrave, (Washington, D. C., The Brookings
Institution, 1965), pp. 122-186. (Gillespie's U. 8. Study
from now on); Idem., The Incidence of Taxes and Public Ex-
penditures in the Canadian Economy (Ottawa: Queen's Printer,
1966). (Gillespie's Canadian Study from now on); Gian S.
Sahota, "Public Expenditures and Income Distribution in Panama,"
United States Agency for International Development (USAID),
Panama City, August, 1972, (mimeographed).

5

 

 

 

These points were made by McLure, pp. 49-59.

6Referring to the famous Impossibility Theorem by
Kenneth J. Arrow. See his Social Choice and Individual Values,
2nd ed. (New York: Wiley, 1963).

7See the two oft-quoted papers by Paul A. Samuelson
"The Pure Theory of Public Expenditure," Review of Economics
and Statistics, 36 (November, 1954), 387-89, and "Diagrammatic

220

 

 

 

221

Exposition of a Theory of Public Expenditure," Review of
Economics and Statistics, 37 (November, 1955), 350-56.

 

8Kenneth V. Greene, "Collective Decision—Making Models
and the Measurement of Benefits in Fiscal Incidence Studies,"
National Tax Journal, 26 (June, 1973), p. 184.

9See Luc De Wulf, "Fiscal Incidence Studies in Develop-
ing Countries: Survey and Critique," IMF Staff Papers, 22
(March, 1975). PP. 61-131.

 

101bid., p. 108.
11Development Economics Department, Public and Private
Finance Division, World Bank. "Distributive Impact of Public

Expenditures: A Research Proposal," September, 1973
(mimeographedL

12To put it more obtusely, a researcher in expenditure
incidence study should not feel uneasy about using the word
"beneficiaries" particularly when another, more detached word
is to be used which in fact has the same meaning.

l3See Gillespie's U. S. Study, pp. 140—5.

l4See Henry Aaron and Martin McGuire, "Public Goods and
Income Distribution," Econometrica, 38 (November, 1970), pp.
907-20; Shlomo Maital, "Public Goods and Income Distribution:
Some Further Results" Econometrica, 41 (May, 1973), pp. 561-
68. The benefit allocation criteria used by Aaron, McGuire
and Maital are quite unique but somehow their principal use
in this study is rather inappropriate. See Appendix H in this
study for the fiscal incidence result in Thailand based on
Aaron-McGuire-Maital methods and the reasons why they are
not used in the main study.

15Admittedly, the general methodology chosen here does
not differ a great deal from what has been adopted in most
other expenditure incidence studies. Nevertheless, this present
study is an intertemporal study where the changes in the pat-
terns of income distribution as a result of government expendi-
tures between two reference years are compared. On the belief
that the conceptual difficulties inherent in an incidence study
are much reduced with an intertemporal investigating method,
this is an important improvement over the conventional one-
period benefit incidence study where the original or "pre-fisc"
and "post-benefit" income distributions within a single period
are compared.

16See, for example, Otto Eckstein, Water Resource Develop-
ment (Cambridge, Mass.: Harvard University Press, 1958). John
V. Krutilla and Otto Eckstein, Multiple Purpose River Develop-
ment (Baltimore: Resources for the Future, Inc., 1954).

 

 

222

17Thailand, Royal Irrigation Department "Feasibility
Report on Lam Ta Kong Storage Dam Project, Korat Province,"
Vol. 1: Project Evaluation, Bangkok, April, 1963 (mimeo-
graphed): Idem., "Report on Kang Krachan Project," Bangkok,
February, 1961 (mimeographed).

18Ammar Siamwala, "Rice," Thammasat Universitprournal
3 (May, 1974), p. 22. Leslie E. Small, however, argues that
at least the irrigation expenditures in the Central Plain
areas produced the positive internal rate of return on the
investment in the order of 6 to 9 percent. The problem of
lack of farmers' initiative in response to the opportunities
presented by the system (such as their non—production in dry
season despite the availability of water) is still recognized
and regarded as real and serious, and deserves special atten-
tion. See Leslie E. Small, "Water Control and Development in
the Central Plain of Thailand," Southeast Asia, (Spring, 1974),
pp. 679-697.

19The idea that the allocation rules be so establish-
ed originated from the studies on marketing channels of vari-
ous agricultural products in the Central and Northeastern
parts of Thailand. From these studies, it was known how a
certain agricultural product passed through from a farmer to
a middleman to a transport operator and so on until it reach-
ed the final consumer within each farming season. The per-
centage share of the profit margin to each of these agents
would be a good indicator of how benefits from increased
yields as a result of irrigation spending should be allocated.
The 70-30 breakdown seems to be an appropriate criterion after
various studies have been consulted. See Siamwala, pp. 65-74
for a summary of such marketing channel studies.

20The Department of Civil Aviation operates all air-
ports in Thailand including Bangkok International Airport.
Since only the really well-to-do travel by plane domestically
and internationally, it is logical to assume that only the
top income group benefits from this type of public expendi-
ture.

 

 

21This 53/47 breakdown was obtained from Thailand,
National Statistical Office, Statistical Yearbook of Thailand,
No. 29, p. 281.

22See Gillespie's U. S. Study, pp. 140-45; see also
Gillespie's Canadian study, pp. 97-105.

23Information obtained from the Department of Land
Transport and the Police Department.

24See Gillespie's U. S. Study, pp. 146-147 and
Gillespie's Canadian Study, pp. 105-110.

 

223

25Thailand, Ministry of Education, Education in
Thailand: 1963 (Bangkok, no date).

26Thailand, National Statistical Office, Quarterly
Bulletin of Statistics, 20 (March, 1972), p. 17.

 

 

 

 

27Thailand, National Educational Council, Private
Expenditures of University Students, 1968-1969, (Bangkok,
1973).

 

28A rather odd, not so important but quite interesting
example is seen in the part of defense appropriation which
was spent on building and operating some golf courses in
Thailand or the creation of a governmental administrative
body supervising the operation of security market. Only
higher income groups could benefit from such "general" ex-
penditures.

29See Jacob Meerman, "Fiscal Incidence in Empirical
Studies of Income Distribution in Poor Countries," AID Dis-
cussion Paper No. 25, December, 1972 (mimeographedh p. 6.
Perhaps a good test regarding the benefits of a defense sys-
tem is to see who is most affected when such defense system
collapses. It would be interesting to know the income dis-
tribution of those individuals or families who managed to
flee a country after its defense system collapsed under ex-
ternal attack such as in South Vietnam or Cambodia in 1975.
If most of these refugees were in higher income brackets,
then the assumption adOpted here is not entirely false.

30Arnold C. Harberger, "Fiscal Policy and Income Dis-
tribution," paper prepared for the Workshop on Income Dis-
tribution in Less Developed Countries," Princeton University,
June 12-13, 1974, p. 11.

31In the actual estimation, small principal payments
were also included in these interest payments; the same
analysis is assumed to apply for both these payments.

32See Gillespie's U. S. Study, pp. 157-158 and
Gillespie's Canadian Study, pp. 128-137.

33McLure, p. 47.

34Ibid.

35Among them Gillespie; Sahota; Richard A. Musgrave
and Peggy B. Musgrave, Public Finance in Theory and Practice
(New York: McGraw-Hill, 1973), Edita Tan, "Taxation, Govern-
ment Spending, and Income Distribution in the Philippines,"
paper presented for the Joint JERC-CAM Seminar on income

224

distribution, employment, and economic develOpment in South-
east and East Asia, Tokyo, December 16-20, 1974.

36Thailand, Bank of Thailand, Monthly Bulletin.
12 (June, 1973): p. 36.

 

CHAPTER V

THE NET FISCAL INCIDENCE

Unlike the previous two chapters, this one involves
no major theoretical difficulty; it deals with the straight-
forward subtraction of the tax incidence of Chapter 3 from
the expenditure incidence of Chapter 4 to arrive at the net
fiscal incidence of the government tax and expenditure ac-
tivities. More specifically, this chapter will account for
the income distributional effects of government taxes and ex-
penditures in two customary ways. First, the effective ex-
penditure rate of each income bracket will be subtracted by
the effective tax rate of the same corresponding income class.
The result may be called the effective fiscal rate or the net
fiscal incidence of the Thai fiscal system. Second, the net
effect of expenditure benefits minus tax burdens of each in-
come class will be added to the "pre-fisc" or the original
before-tax and before-benefit income distribution to see
whether or to what extent the "post-fisc" or the after-tax
and after-benefit income distribution changes. Alterations
in the Gini concentration ratios associated with changes in

income distributions also will be estimated.

225

226

Effective Fiscal Rates or Net Fiscal
Incidence by Income Classes

 

 

As the estimates of tax and expenditure incidence in
Chapters 3 and 4 were carried out on the same five income
brackets in 1963, and seven income brackets in 1969, the ab—
solute fiscal incidence and the effective fiscal rates for
those years can be obtained simply by subtracting the tax
burden (effective tax rate) pertaining to each particular in-
come bracket from its expenditure benefit (effective expendi-
ture rate). For example, the absolute net fiscal incidence
of the lowest income bracket (under Baht 3,000) in 1963 would
be Baht 2,229.7 million minus Baht 1,911.0 million, or Baht
318.7 million.

Table 5.1 shows the absolute net fiscal incidence by
income classes in 1963 and 1969. The effective fiscal rates
can be estimated by dividing these absolute incidence figures
by the relevant money and adjusted income levels. The result-
ing effective fiscal rates by income classes are shown in
Table 5.2.

Essentially, an effective fiscal rate shows the rela-
tive net benefit (or net loss if the effective fiscal rate
has a negative sign) which the households in different income
brackets have received from the tax and expenditure policies
of the government as a percentage of their money or adjusted
income. The use of income as a weight is important in this
incidence study because these rates across income classes could

reveal whether the fiscal structure is regressive (pro-poor),

227

Table 5.1—-Absolute Net Fiscal Incidence, 1963 and 1969
(Amountsin.Millions of Baht)

 

 

 

Income Classes
(Baht 1963 1969
Under 3,000 '318.7 941.5
3,000- 5,999 138.5 697.2
6,000- 8,999 452.2
} 402.7
9,000-11,999 375.1
12,000-14,999 190.8
} 67.1
15,000-17,999 163.7
18,000 and over 694.1 1,010.8
All Classes 1,621.2 3,831.3

 

 

 

Source: Tables 3.2, 3.3, 4.3 and 4.4.

228

Table 5.2-~Effective Fiscal Rates Based on Money and a
Adjusted Income, 1963 and 1969: Assumption I

 

 

 

 

 

Based on Based on
Income Classes Money Income Adjusted Income
(Baht)
1963 1969 1963 1969
Under 3,000 7.1 38.9 2.1 10.7
3,000- 5,999 2.8 11.6 1.2 4.0
6,000- 8,999 6.6 2.7
}4.4 } 2.1
9,000-11,999 6.3 2.9
12,000-14,999 3.7 1.8
}1.5 } 0.7
15,000-17,999 3.7 1.9
18,000 and over 6.0 3.0 2.9 1.4
All Classes I 4.7 5.9 2.1 2.6

 

 

 

 

aThese are effective fiscal rates under Assumption I
where the budget deficit is assumed to add directly to the
income of households without reduction in real income through
possible price increase.

229

progressive (pro-rich), or proportional. In general, a regres-
sive fiscal system would generate relatively greater net bene-
fits to lower than to higher income classes, so that the ef—
fective fiscal rates will fall as the level of income rises.
The opposite is true with a progressive fiscal system. A
proportional or neutral fiscal system would be indicated by
a constant effective fiscal rate across income brackets.

Before proceeding to discuss the effective fiscal
rates illustrated in Table 5.2, the reader should be informed
about one important point about these rates. Both the 1963
and 1969 budgets were unbalanced: The expenditures to resi-
dents exceeded the tax receipts from residents by Baht 1,621.2
million in 1963, and by Baht 3,831.0 million in 1969. These
deficits are assumed to raise or add to the income of house-
holds without any reduction in the income of the same house-
holds either through taxes or through price increase. Table
5.2 was calculated on the assumption that the budget deficit
has a "zero repercussive effect" upon any price adjustments
that might occur due to the existence of such a deficit. In
other words, the deficit merely adds more income to house-
holds without them having to sacrifice or suffer in any other
way.1 Hereafter this assumption will be referred to as As-
sumption I.

With the preceding qualification in mind, Table 5.2

may be examined. Three points are apparent:

230

(1) The average effective fiscal rates for all in-
come classes for 1963 and 1969 were 4.7 and 5.9, respectively,
under the money income base, and 2.1 and 2.6, respectively,
under an adjusted income base.

(2) All effective rates are positive across all in-
come classes, meaning that, under Assumption I, all income
classes gained absolutely from government fiscal activities.

(3) The overall net incidence pattern in 1969 could
be described as moderately regressive (pro-poor) under the
money income base to slightly regressive under the adjusted
income base. In 1963 no clear incidence pattern is discerni-
ble. As the effective fiscal rates fluctuate slightly through-
out the income range, the net fiscal incidence pattern could
appropriately be termed a wandering proportional one. It is
noted, however, that the effective fiscal rates for the top
income class are higher than the average fiscal rates for all
income classes. This indicates that, indeed, the effects of
tax and expenditure policies was to redistribute income from
some lower income brackets to this top income bracket, cer-
tainly a redistribution in the wrong direction if one hypoth-
esizes that a more egalitarian distribution is desirable.

The pattern of net fiscal incidence perhaps may be
seen more easily in the graphical presentation in Figures 5.1
and 5.2. In these figures the effective fiscal rates are
plotted along the vertical axis and the average money income

of households along the horizontal axis.

231

mood can mood .uumuuao usoocH mp .oaoonH mono: no woman now-m agenda o>uuoommuru«.m ousmmh

35m mo uccaaozs
avocado onoocH

 

"'
‘ '
‘Il ‘ '-
‘I"
|
I,"
1“
l‘-..
‘
"Q'I‘E
“'
'I‘

 

moms
moms ..........

can

On

on

 

é

Annuouomv
omega Hannah
o>uuoomum

232

mom“ was mwmd .mommnau oaoonH hp .osoosH vouusnv< no woman «ovum Hannah o>daoommmnnm.m ouawnk

Scam mo 33.323
mouuuao osoocH

A
(

J

1

J

 

 

 

moms
moms ..........

o“:

om

 

t

Avcuouomv
nuaam Hannah
o>uvoomma

 

I..IIIIIII II I11

233

From Figures 5.1 and 5.2 it is clear that the 1969 fiscal sys-
tems were more regressive than those in 1963 both in the level
and the pattern of net incidence. Under a money income base,
the 1963 net fiscal incidence has a somewhat U-shaped pat-
tern, although a very shallow one, whereas the 1969 net fis-
cal incidence has a monotonically downward sloping pattern
throughout. Under an adjusted income base, the net fiscal
incidence patterns were practically prOportional in both years,
with the situation in 1969 being slightly more favorable to
the lowest two income classes.

Following the practice adopted in the tax and expendi—
ture analysis, it is possible to generalize about the net fis-
cal incidence pattern in the same way as the tax and expendi-
ture incidence patterns, that is, by regressing the effective
fiscal rates upon the average household money income using the
ordinary least-squares method. The functional form for the
net fiscal incidence pattern, however, differs slightly from
the original functional form for the tax and expenditure in-
cidence patterns. Instead of expressing effective fiscal
rates as a function of the annual average money income of
households, the new functional form expresses effective fiscal

rates as a function of the logarithms of the annual average

 

money income of households. The main reason for the change
is because the new functional form gives a much better good-
ness of fit than the original form.2

The results of the generalized regression functions

of net fiscal incidence are shown in Table 5.3, and these

234

Table 5.3—-Genera1ized Net Fiscal Incidence Functions:
Assumption I KEffective Fiscal Rate) = a + b log
(Average Annual Money Income)a]

 

 

 

Intercept Slope R2
(1) On Money Income Base
1963 8.14 -l.044 .8085
1969 109.51 -25.066 .8420
(2) On Adjusted Income Base
1963 0.08 +0.40? .8388
1969 29.19 -6.472 .9014

 

 

 

 

aAverage Annual Money Income in Thousands of Baht

235

give general support to the results in Table 5.2 and Figures
5.1 and 5.2. The net fiscal incidence pattern in 1963 was
barely regressive under the money income base and actually
turned progressive under the adjusted income base, the same
pattern in 1969 was clearly regressive throughout under both
income bases. Comparing fiscal activities in 1963 and those
in 1969, the latter were definitely more income equalizing
than the former. The more exact extent to which the two fis-
cal programs helped change the equality in the distribution
of income will be discussed in the next section.

The Effects of Fiscal Activities
on the Distribution of Income

 

 

Of paramount concern in a fiscal incidence study is
the following question: What is the impact of government
fiscal activities upon the distribution of income of house-
holds? The effective fiscal rates discussed earlier showed
mainly the level of fiscal incidence or net benefits weighted
by two income concepts, and the general pattern of progres-
sivity or regressivity of the fiscal programs. This section
will examine the effects of such fiscal programs from an-
other aspect, that is, the relative income position of each
income class after net benefits from the government sector
were added to the original, pre-fisc income levels.

Table 5.4 presents the distribution of post-fisc money
and adjusted income of households in 1963 and 1969. Let us
first consider the money income portion of that table. Re-

gardless how one views it, the degree of change in the pattern

236

 

 

 

 

 

 

 

 

o.oom o.oom o.oom o.oom o.oom o.oom o.oom o.oom mpmmmmo mam
m.mm m.mm m.mm m.mm m.mm m.mm . m.mm m.mm pm>o_pcm coo.mm
m.m m.m m.m m.m .mmm.mm.ooo.mm
m.mm i m.mH l m.mm i m.mm i
m.m m.m . m.m m.m mmm.mmnooo.mm
m.m m.m m.m m.m mmm.mmuooo.m
m.mm I m.mm .1 m.mm i m.mm i
m.mm m.mm m.mm m.mm . mmm.m -ooo.m
m.mm m.mm m.mm m.mm m.m m.m m.mm m.mm mmm.m -ooo.m
m.m m.m m.mm m.mm m.m m.m m.mm m.mm moo.m puns:
ommmcumom ommmlmum ommmlumom ommmlwum ommmuumom ommmnmum ommmlumom ommmlmum
mmmm mmmm mmmm lusmmv

 

pogflbmd .8805 8...ng

 

 

comm—53.339 mcduaH 5:9.“

 

8330 9.85

 

 

H coin—Ema

momma cam mood .mmmmmfiu

9505 an 9505 8pm: mg 98 moan: ommmluwom paw omfimlmum mo 2033.“ng mmfifioummllv .m magma.

237

of income distribution from the pre-fisc to the post-fisc
situation can only be described as insubstantial. In 1963
the lowest income bracket gained less than one-half of a
percentage point in its relative income position, whereas

the positions of the middle three brackets slightly worsened,
and the top income group gained by about 0.5 percentage
points. In 1969, the degree of change was more distinct,

yet still small. Of seven income brackets, the first four
appear to be gainers and the last three losers, with the low-
est income group gaining the highest and the top income group
losing the most.

When adjusted income is used as the base, the redis—
tributional results become even smaller. The pre-fisc and
the post-fisc income distributions in both 1963 and 1969
look almost identical, attesting to the fact that the fiscal
activities of the government had very small ifiCome redistribu-
tional effects indeed upon Thai households. That there was
more income redistribution in 1967 than in 1963 is obvious
from Table 5.4. Still, it would be interesting to see ex-
actly how much the fiscal policies of each year affected the
inequality of income as measured by Gini concentration ratios.

Table 5.5 presents the estimates of Gini concentra-
tion ratios before and after government fiscal activities
took effect, and the percentage changes in these ratios both
between the pre-fisc and the post-fisc situations of a single

year and between the post-fisc situations of 1963 and 1969.

238

 

 

 

 

ao.mu mama. mmmv. mama
av.m+ a mm.m+ a
ma.o+ aame. . mama. mama
mmmm
msoocm pmumsmpa co Ame
om.vu mama. ommm. mama
mm.m| a mm.a| a
Na.ou aaam. mmaa. mama
mmmm mEooca mmcoz so may
mam”. mam”. 8:: am”. am”.
mcu canvas
umamnumom omamlmum ucmaommmmou omamuumom omamuwum
. . uwmw

 

mama new mama cmm3umm
ucmaoammmoo acau
ca A+V mmmmuocH no
Ala mmmmuomo wmwucmoumm

 

acau ca A+v
wmmmuoca Ho
Ala mammuomo
womucmoumm

 

mucmaoamwmoo acaw

 

 

 

 

a ceaumfismmd ”mama mam mama .muaamsvmca OEOOCH
Ca mmmcmco momucmoumw Ucm moaumm Geaumuucmocou acao omamlumom can omamumumlum.m magma

239

On a money income base, the 1963 post—fisc income distribution
improved from .5627, the pre-fisc Gini ratio, to .5581, re-
ducing the inequality area by 0.8 percent. The similar change
for 1969 was from .5550 to .5289, or a rate of improvement in
income equality of 4.7 percent. Comparing the post-fisc dis-
tributions based on money income in both years, the effect
of the 1969 fiscal policies accounted for a reduction of al—
most 4 percentage points in the inequality index from 1963.
Such an effect is much reduced when adjusted income
is used as a base. The pre-fisc Gini ratios were .4559 and
.4822, respectively, for 1963 and 1969. As a result of taxes
and expenditures, they changed to .4566 and .4725, indicating
a very slight worsening in income equality in 1963 and a
slight improvement of the same in 1969. Nevertheless, the
overall effect in comparing the 1963 to the 1969 post-fisc
adjusted income distribution is seen to be a general worsen-
ing in the degree of income equality by about 3.5 percent.
This certainly does not mean that fiscal activities in 1969
made households worse off than fiscal activities in 1963.
On the contrary, 1969 policies did redistribute income from
higher to lower income classes, thus improving the Gini:ratios
of income inequality. However, as the pre-fisc income dis-
tribution based on the adjusted income concept became more
unequal in 1969 compared to 1963, the tax and expenditure
policies in 1969, which were only slightly regressive to pro-
portional, were not strong or redistributive enough to off-

set the deterioration in income equality due to governmental

240

and other undetermined external influences.

Since the 1969 pre-fisc distribution shows an approxi-
mate 5.8 percent deterioration in income equality from 1963
on the distribution of income based on adjusted income, and
since the increase in the Gini ratio in the 1969 post-fisc
income distribution was only 3.5 percent, then roughly 2.3
percentage points should be regarded as the extent to which
1969 fiscal policies reduced income inequality from 1963
based on adjusted income. Similarly, on a money income base,
the post-fisc distribution in 1969 was 5.2 percent more equal
than the similar distribution in 1963. Compared to 1.4 per-
cent which is the extent to which the 1969 pre-fisc distribu-
tion was more equal than the 1963 pre-fisc distribution, it
could be said that the 1969 fiscal activities had about 3.8
percentage point (that is, 5.2 percent minus 1.4 percent)
more income equalizing power than the 1963 fiscal activities.
These two figures (3.8 percentage points and 2.3 percentage
points) should indeed be identical; the small divergency is
probably due to errors in statistical estimation.

It could be concluded, therefore, that between 1963
and 1969 the government, through its taxes and expenditures,
effected between a 2.3 and 3.8 percentage point reduction in
income inequality as measured by-Gini coefficients, a feat
which could easily have been accomplished by imposing only
a 2.5 percent extra tax on the income of the highest income
group in 1969 and transferring this amount of tax revenue to

the lowest income households.5 Also, the reader is reminded

241

that the 2.3 to 3.8 percentage point reduction mentioned
above is true only under Assumption I, where the budget defi-
cit is assumed to have no detrimental effect through price
adjustments but to add directly to the income of households.
Had this assumption been proven untenable or false, the pat-
tern of post-fisc income distribution could have changed,

and this, of course, could have resulted in a greater in-
equality index. It is possible, as will be shown later, that
depending on alternative assumptions about the income distri-
butional effects of a budget deficit, the government could
indeed worsen, not improve, the income equality of households

through its fiscal activities.

Budget Deficit in Fiscal Incidence Studies

 

The significance of the existence of a deficit in the
public sector is generally recognized in almost all fiscal in-
cidence studies. Gillespie devoted a substantial part of his
Canadian study to a discussion of how the problems should be
dealt with in an empirical investigation.3 Similar efforts
also have been made by J. A. Johnson and G. S. Sahota.4 The
results, however, have been less than gratifying, for there
is still no single most satisfactory way of dealing with the
empirical estimation of the possible distributional impact
of the budget deficit and its financing.

A budget deficit is said to exist when the govern-
ment's expenditures exceed its revenues in any fiscal year.

There are several avenues Open for the government to finance

242

such a deficit. It may print more money, use up past reserves,
borrow from domestic as well as foreign sources, or receive
financial help from overseas. Normally, the government will
use a combination of these methods. What is most pertinent
here is the government policy of increasing the money supply,
since this will normally raise the price level in most econo-
mies unless they are characterized by high unemployment or
underutilization of resources. The increase in the price
level could either partially reduce the income generating
power of the deficit (that is, through the income multiplier),
or eliminate it completely depending upon other factors.

The problem is further complicated by the fact that
even if the overall extent of the price increase is known, it
is still difficult to know how different income brackets are
differently affected by such an increase. Various studies
have been directed to answering the question of who would be
affected by price increases, and by how much, but the results
obtained so far are not conclusive. Generally, however, it
is believed that lower income groups suffer more under infla-
tion compared to higher income groups, for several reasons:

A higher proportion of the income of the poor is spent on
necessary consumption, such as food and clothing, which usu-
ally is subject to price increases more than other items, or
the earnings of the poor do not increase as fast as the price
level, or other factors.6 Also, if the traditional view is
correct that inflation harms wage-earners and other fixed-

income groups but benefits property owners and business firms,

243

then this would imply that the lower brackets will suffer
more than the upper brackets since wage and salary earners
generally tend to be bunched in the lower to middle income
groups, and the business operators and property owners in
the higher income groups.

But the foregoing discusses only one dimension of the
problem; many other questions regarding a budget deficit
could be considered, among them: What if deficit financing
causes not only a price increase but also unemployment? How
does one isolate the portion of the deficit that affects only
nonresidents? How does one account for the distributional
effects of the deficit financed by foreign assistance? These
problems cannot be solved unless certain assumptions are made
as to how the government's deficit should be treated. Most
fiscal incidence studies seem to ignore the income distribu-
tional effects of the public deficit and assume that it only
adds to existing income, just as it has been treated thus far
in this study. -A few alternative assumptions designed to
"eliminate" the deficit also have been used, resulting mainly
in the change in the level of net incidence but not its pat-
tern.

In Thailand the "functional" budget deficit amounted
to Baht 1,621.2 million in 1963 and Baht 3,831.9 million in
1969. By "functional" is meant the deficit which affects
only the resident households considered in this study, not

the actual accounting deficit, nor the total amount of public

‘244

borrowing to finance the deficit in that year. In other words,
the functional deficit is derived simply from subtracting the
total adjusted tax revenues from the total adjusted expendi-
tures in each year.

Three assumptions concerning the treatment of budget
deficits are considered in the case of Thailand. Assumption
I has already been discussed: The deficit is assumed to have
a zero or no repercussive effect on price adjustments due to
the existence of the deficit and its financing. This has been
the standard assumption so far. According to Assumption II,
the price effect is assumed to cancel out completely the in-
come generating effect of the deficit. The deficit is then
eliminated, and the budget is hypothetically balanced. How-
ever, only the level of net benefit is changed, not its pat—
tern. Assumption III is somewhat similar to Assumption II,
that is, the price effect is assumed equal to the income ef—
fect, but the distribution of the effect by income classes is
assumed to be more burdensome to the lower than to the upper
income classes. As the net fiscal incidence according to
Assumption I has already been considered, only Assumption II
(deficit with "neutral" repercussive effect) and Assumption
III (deficit with "de-equalizing" repercussive effect) will be
discussed below.

When the deficit is eliminated and the effect on ex-
isting income distribution is neutral (Assumption II), the
change in the effective fiscal rates is simply a matter of

subtracting the average net benefit rate for all income

245

classes from the effective fiscal rate derived under Assump-
tion I of each income bracket. The new effective fiscal rates
would not be all positive, meaning that the deficit finance
benefited some and not others. However, post-fisc income
distribution would remain the same, since the neutral effect
touches only the level of effective fiscal rates, not the en-
suing distribution of income.

Under Assumption III, the deficit is said to cause a
relatively heavier burden to the lower than to the upper in-
come classes, due to price adjustments. As there is no basis
upon which such a deficit could be deducted from each income
class, it is arbitrarily assumed that each household is equal-
ly affected by it. In other words, financing the deficit has
the effect of a poll tax upon each head of household in Thai-
land. Since most households are found in the lower spectrum
of the income distribution scale, the elimination of a deficit
in this way appears to "de-equalize" income equality.

The results of the hypothetical elimination of a func-
tional budget deficit using the above two assumptions are
shown in Tables 5.6 and 5.7. Table 5.6 shows the effective
fiscal rates under Assumption II and Table 5.7 shows the re-
sulting percentage distribution of income under Assumption
III.

Although the pattern of effective fiscal rates in
Table 5.6 is the same as that shown in Table 5.2, use of As-
sumption II gives a clearer view of whether and in which

direction the redistribution has occurred. In 1963, on both

 

 

246

Table 5.6--Effective Fiscal Rates Based on Money and
Adjusted Income, 1963 and 1969: Assumption II

 

 

 

 

Based on Based on
Income Classes Money Income Adjusted Income
(Baht)
1963 1969 1963 1969
Under 3,000 +2.4 +33.0 +0.1 +8.1
3,000- 5,999 -1.8 +5.7 -0.8 +1.4
6,000- 8,999 +0.6 +0.1
- -0.3 +0.1
9,000-11,999 +0.4 +0.3
12,000-14,999 -2.3 -0.8
-3.2 -1.3
15,000-17,999 -2.2 -0.7
18,000 and over +1.3 -2.9 +0.8 -l.2
All Classes - 0 0 0 0

 

 

 

 

 

 

 

 

Fli' III‘II Illl‘l

247

 

 

 

 

 

 

 

 

 

 

GOauafiaHumao 968a moan: n2

 

cognaaumao mEooca mung

 

o.ooa odoa o.ooa mama odoa mama o.ooa 0.8a «dummao a2
a.mv a.am m.am m.mm m.ma m.ma . m.mm m.mm um>o_a:m oom.aa
m.m m.m a.a a.a mmm.mauooo.aa
m.aa a m.aa a m.ma a a.ma a
a.m a.m m.m a.m mmm.aauooo.~a
m.m m.m m.m m.m mmm.aanooo.m
a.mm a o.am a a.am a a.a~ a
m.aa a.aa a.oa a.oa mmm.a uooo.a
m.aa m.aa ~.aa m.ma m.m m.m m.ma m.ma mmm.m uooo.m
o.a o.a m.ma m.ma m.m m.m m.aa m.ma coo.m amass
omamauumom 8315 Baabmom oma aumam omapuumom ammumam 8amau8m omamuoam
38am:
mama mama mama mama mammmao mg

 

 

aaa coaumasmma

umama pan mama .mmmmmau 9:85
man .9593 pmumflnaad paw Name—oz omamlumom Em Omahnmam mo companahamao mmflsmoummllm .a magma.

248

money and adjusted income bases, the middle three income
brackets seemed to be the net losers or redistributors of
income to the top and bottom income brackets. In 1969, it
was also clear that the top three income groups under both
income bases were the net losers. Again, it must be mentioned
that the extent of redistribution is quite small in both years.
A new kind of picture emerges when the deficit is
eliminated under Assumption III. As expected, the lower in-
come classes now appear to be the heavy losers (in 1963) or
small gainers (in 1969); in all cases, the top income class
received positive gains. The net fiscal incidence pattern
no longer could be called slightly regressive to proportional,
as before, but rather proportional to slightly progressive.
The above assertion is supported by the change in the
post-fisc income distribution under Assumption III shown in
Table 5.7. That the lower income classes are worse off and
the upper income classes better off after the effects of gov-
ernment fiscal activities than before is clearly evident in
both years. In 1963, for example, the post-fisc income share
of the lowest bracket fell from 13.0 percent to 11.8 percent
under the money income base, and from 19.8 percent to 19.2
percent under the adjusted income base. In 1969, the income
position of the lowest income class was actually improved by
a very small percentage, but the positions of the three low-
est brackets taken together deteriorated. Improvement in the
top income group's position ranges from as low as 0.3 to as

high as 1.7 percentage points.

 

an
an
In
su
co:
co.

fr:

ce]
AS:
lee
thu
110‘
uni
fi
th

991

is
0n
tr
ha
to

Cr

249

The pre-fisc and post—fisc Gini concentration ratios
and percentage changes in these ratios in, and between, 1963
and 1969 under Assumption III are presented in Table 5.8.

In 1963, the post-fisc Gini concentration ratios under As-
sumption III increased from .5627 to .5838 based on money in-
come, and from .4559 to .4660 based on adjusted income. The
corresponding increase in 1969 was from .5550 to .5614, and
from .4822 to .4853. The percentage increase in income in-
equality from the pre-fisc to the post-fisc situation is es-
timated at between 2.2 to 3.8 percent in 1963 and between

0.6 to 1.2 percent in 1969, depending upon the income con-
cepts used. This is a startling change from the results under
Assumptions I and II because, under those assumptions, at
least the post-fisc mgngy income distribution improved over
the pre-fisc mgngy income distribution in both 1963 and 1969;
now, under Assumption III, the post-fisc income distribution
‘ under both income bases becomes more unequal than the pre-
fisc income distribution. In short, under Assumption III,
the government aggravated instead of mitigated income in-
equality in Thailand.

The question then arises as to whether Assumption III
is too extreme. It does not appear to be. In fact, in the
only study on the effects of price increases upon income dis-
tribution of households in Thailand, Dr. Oey Astra Meesook
has commented that it is the lower income groups relative
to the higher who are more adversely affected by general in-

creases in the consumer price index. She reasons as follows:

250

 

 

 

 

 

aa.o+ mama. «mam. mama
aa.m+ a. mm.a+ a
mm.~+ oaam. mama. mama
mmmm
msooca pmumsaam co Ame
aa.a+ aaaa. oaaa. mama
ma.m- a mm.a- a mm.m+ amam. mmaa. mama
mmmm mEoocH mono: co aav
.umao .umao .umao .umaa
mEooca mEooca woaawm wEoocH mEooca
Omahlumom. omamlwum may canvas omamuumom omamumum
unmaoamwmoo

 

mama paw mama cmmzumm
unmaoammmou acao
ca a+v mammHUCH no
any mmwmuomo mmmucmoumm

 

acao ca a+v
mmmmHOCH Ho
Ala mmmmaomo
momucwoamm

 

mucmaoamwmoo acau

 

and»

 

 

aaa coaumeammd

”mama can mama

.muaamsvmsa mEooca

ca momcmnu mmmucwoamm 0cm moaumm coaumaucmocou acaw omamlumom 0cm omamlmamula.m manna

,_—‘.
.n—‘L

”

.—

-
_

fl
-Q

. u-I‘

“

w

A“...-
a
‘—

‘

dif
inc«
dif:

per:

eSti;
mateE
ernme
Come
these
Sente(
inCOm

Pre-f

251

The poor suffer more first of all because
total expenditures form a large percentage
of their incomes so that there is less of

an income margin, if any, to enable them

to keep real consumption at a constant level.
Moreover, when real consumption declines for
the poor it is a serious problem if they are
near or actually below subsistence level.
Our calculations show that further aggrava-
tion of the problem is brought on by the un-
favorably larger overall price increase fac-
ing the poor during recent years.7

Although Dr. Meesook has quantitatively compared the
different distributional effects of price increases across
income classes, such a scheme cannot be applied here due to
differences in income class structures, income concepts, and
periods under study. It is sufficient, nonetheless, to say
that price increases due to deficit financing could have a
de—equalizing effect upon the existing income distribution
if it is assumed that each household equally bears the burden

of such price increases, as Table 5.8 has shown.

Summary and Conclusions

 

This chapter has brought together the tax incidence
estimation of Chapters 3 and the expenditure incidence esti-
mates of Chapter 4, netting out the residual effects of gov-
ernment taxes and expenditures upon the distribution of in-
come of households in Thailand in 1963 and 1969 and between
these two years. As in Chapters 3 and 4, the results are pre-
sented in two ways, according to effective fiscal rates and
income classes, and according to the changes in the original,

pre-fisc income distribution after the net benefits have been

‘—‘

adde

tior

had
est
the
est
the
both
to t
in 1

siVe

1963
mEdSI
Dre-1
Undel
tribL
tribL

equal

252

added and showing the related changes in the Gini concentra—
tion ratios.

Under a money income base, the net fiscal incidence
had a U-shaped pattern in 1963 with the lowest and the high-
est income groups enjoying the highest net benefits; in 1969
the pattern was monotonically downward sloping from the low-
est to the top income group. Under an adjusted income base,
the net fiscal incidence became more or less proportional in
both 1963 and 1969, as expected. Because the net benefits
to the lowest income classes were much larger in 1969 than
in 1963, the 1969 fiscal programs were definitely more regres-
sive or pro-poor than the 1963 programs.

The post-fisc distribution of money income both in
1963 and 1969 showed an improvement in income equality as
measured by Gini concentration ratios when compared to the
pre-fisc money income distribution in the respective years.
Under the adjusted income base, however, the post-fisc dis-
tribution became slightly more unequal than the pre-fisc dis—
tribution in 1963, while in 1969 the improvement in income
equality was still evident in the post-fisc distribution.

The foregoing conclusion is based on the assumption
that the budget deficit which existed both in 1963 and 1969
only adds directly to the income'of households. If the
budget deficit is assumed to raise the price level, and if
each household is assumed to suffer the price increase equally,
then it can be shown that the post-fisc income distribution in

both 1963 and 1969 would be more unequal than the corresponding

-”-—— __———-———‘———a

pre-
assu
trib
ture
to h
inco

come

gene
tiOH.

Of V,

253

pre-fisc income distribution. The purpose of this latter
assumption is to demonstrate that the extent of income redis-
tribution of the Thai government through its tax and expendi-
ture policies is so small that, by changing the assumption as
to how the budget deficit differentially would affect each
income class, the government fiscal programs could become in-
come de-equalizing instead of income equalizing.

Although the year 1969 showed a promising trend toward
greater income equality (or less income inequality) over the
year 1963 as a result of government fiscal policies, apparently
this trend was too weak and too slow to be meaningful. This
was mainly because, on the tax side, too much revenue was col-
lected from indirect taxes and too little from direct taxes,
and, on the expenditure side, too much benefit to the lower
income classes was derived from such general expenditures as
defense, law enforcement and public administration. It may
be questionable to assume that one Baht's benefit from these
general expenditures has exactly the same value as one Baht's
benefit from such specific or transfer expenditures as educa-
tional spending and child allowances, at least from the point
of View of the poorest section of the population.

If the more urgent objective of the government is to
help the poorest section of the population, it would be most
desirable for fiscal policies to be reoriented toward fewer
indirect taxes and general expenditures, and toward more di-
rect taxes and more specific and transfer expenditures, es-

pecially the categories which contribute most to the direct

 

 

ar

hc

in
si
1e‘
the
dis
Chi
nua
in .
the
cruc
inco

come

State

QIOWt

254

and indirect raising of the income positions of the poor

households.

One final note about the significance of income level
in relation to income inequality should be mentioned. By
simply looking at the considerable increase in the income
level in 1969 compared to 1963, one could be led to believe
that the growing economy was simultaneously easing the income
distribution problem. The study by this researcher and
Chintana Chernsiri pointed out that even when the average an-
nual income of a family in the lowest income class was tripled
in 1969 it was still not sufficient to raise that family above
the poverty level as defined for that year.8 What is more
crucial, of course, is the inequality in the distribution of
income, a factor that easily could be overshadowed by the in-
come rise.

In conclusion, it seems appropriate to quote a cogent
statement made by Oey Astra Meesook on the issue of income
growth and income inequality:

A major question...is whether it is in fact
the case that rapid overall growth of income
makes it easier to bring about changes in the
way that income is distributed. The argument
usually put forward is that a rapid rise in
income makes it unnecessary to actually re-
distribute income. By simply ensuring that

a larger proportional increase accrues to the
poor than to the rich, it is possible to im-
prove the distribution of income in a way
which is not disagreeable to the rich. Al-
though this is a notion which is being bandied
about a great deal, our analysis suggests that
rapid growth of income may simply dull the
pains of inequality: in a situation of high

growth rates, it becomes less noticeable that
there is still a great deal of inequality

255

around. People are
that their position
thus less likely to
some.9

fooled into thinking
has improved and are
be politically trouble-

FOOTNOTES

1The most obvious case where this is possible is when
the government finances the deficit through some windfall
revenues, through past savings, or through foreign grants
in a less than fully employed economy.

2The other reason is that the original functional
form when used on the 1963 money income data gives a general-
ized net fiscal incidence curve with a positive slope, sig-
nigying a progressive net fiscal incidence pattern. This
contradicts the result of the change in the post-fisc money
income distribution as measured by Gini concentration ratios
(to be discussed later in the chapter) which showed that under
money income base the post-fisc income distribution in 1963
was still more equal than the pre-fisc income distribution
of the same year. After an experiment with two other func—
tional forms-~the semi-log and the double-log--which showed
the net fiscal incidence curve with a negative slope as ex-
pected, it was decided to use the semi—log functional form
which also yielded the best overall goodness of fit. There
was no contradiction regarding the net fiscal incidence slope
under adjusted income in 1963, however; the positive slope was
not unexpected.

 

 

3W. Irwin Gillespie, The Incidence of Taxes and Public
Expenditures in the Canadian Economy. (Ottawa: Queen's
Printer, 1966), pp. 164-179.

4

James A. Johnson, The Incidence of Government Reve-
nues and Expenditures (Ottawa: Queen's Printer, n.d.), pp.
73-6; Gian S. Sahota, "Public Expenditures and Income Dis-
tribution in Panama," United States Agency for International
Development, Panama City, August, 1972, (mimeographed).

 

5The experiment was carried out by deducting 2.5 per-
cent from the total money income of the highest bracket and
adding this to the income of the lowest bracket. The new
Gini coefficient was then computed from this new distribution
of income the result of which was a reduction of about 4 per-
cent in income inequality compared to the original income
distribution.

6Oswald Brownlee and Alfred Conrad, for example, claim
Specifically that inflation not only hurts the lower income
classes but also benefits the upper income classes. George
L. Bach and Albert Ando, on the other hand, maintain that for

256

 

 

on
Uni

men
Jou

Fac

257

a moderate inflation the redistributional effects are likely
to be small, and more or less neutral. On a general context
of incurring a public debt, however, Hugh Dalton and Alvin
Hansen, among others, believe that internal debt creates
higher relative real burden upon the lower income classes
and is thus likely to intensify inequality of wealth and in-
come rather than to mitigate it. See Oswald H. Brownlee and
Alfred Conrad, "Effects upon the Distribution of Income of a
Tight Money Policy," American Economic Review, 51 (May, 1961),
pp. 74-85; George L. Bach and Albert Ando, "The Redistribu-
tional Effects of Inflation," Review of Economics and Sta-
tistics, 39 (February, 1957), pp. 1-13; Hugh Dalton, Princi-
ples of Public Finance (London, 1923), p. 192; Alvin H.
Hansen, Fiscal Policy and Business Cycles (New York, 1941),
p. 179.

7Oey Astra Meesook, "The Impact of Price Increases
on Different Income Groups," Faculty of Economics, Thammasat
University, August, 1975, p. 11 (mimeographed).

8Medhi Krongkaew and Chintana Chernsiri, "The Measure-
ment of the Poverty Level in Thailand," Thammasat University
Journal, 5 (June-September, 1975) (in Thai).

 

 

 

 

9Meesook, Income Distribution in Thailand (Bangkok:
Faculty of Economics, Thammasat University, 1975), pp. 83-84.

 

 

(II I t [lull I I I ‘

CHAPTER VI

SUMMARY, CONCLUSIONS, AND POLICY IMPLICATIONS

Summaryof Purpose and Procedures

The purpose of this study was to measure empirically
the income redistributional impact of taxes and public ex-
penditures in Thailand in, and between, 1963 and 1969. In
the language of public finance this is known as a "fiscal
incidence study.” Unlike the more conventional type, however,
this study emphasized the intertemporal effects of fiscal ac-
tivities upon income distribution. In other words, it was
primarily interested in comparing the change in income dis-
tribution as a result of fiscal activities in 1969 with 1963.
To use more precise terminology, this research estimated and
compared "post-fisc" income distribution in 1969 with "post-
fisc" income distribution in 1963. Knowledge of the redis-
tributional impact would be useful not only as a means of
measuring the government's performance in terms of income
distribution, but also as a policy guideline or as an indica-
tion of the direction in which the government might move if
it wished to make some concrete changes in the existing pat-
tern of income distribution.

The persistent and severe inequality of income in

Thailand was one of the major factors prompting this study.

258

 

 

l I [Ii I I{ .II I .(I. .I I

259

Since it is undoubtedly the government, more than any other
element in the economy, which is capable of changing the pat-
tern of income distribution of individuals or households, the
investigation of the government's role in the distribution of
income as reflected through its tax and expenditure policies,
was a reasonable beginning. Such information would be a
logical first step toward an efficient economic policy aimed
at reducing the present income inequality. Admittedly, the
ideological bias of this study is toward greater income and
economic equality within the present democratic form of
government.

The scope of the research was limited to the tax and
expenditure impact on household income at the national level.
Lack of data, time, and resources did not permit a more de-
tailed investigation at the regional, provincial, or sectoral
level. Nevertheless, the information gathered more than com-
pensated for the above limitation. Within this framework,
the study attempted to estimate income distribution in 1963
and 1969 as a basis for further estimation of tax and ex—
penditure incidence.

Special attention was given to estimating the income dis-
tribution in 1963 and 1969 as accurately as possible. This
was necessary because an incorrect or inaccurate income base
could distort the true pattern of tax and expenditure inci-
dence. The basic data for income distribution estimates for
1963 and 1969 were taken from the household expenditure survey

of 1962—1963 and the socio-economic survey of 1968-1969

260

respectively. Both were conducted by the National Statistical
Office of Thailand. However, the income distribution obtained
was the distribution of money income, which is only part of
real income. This latter is defined in the Simons-Haig tra-
dition as consumption plus the change in net wealth. Many ad-
justments were made to money income in an effort to obtain a
more comprehensive income base. Among the income elements
added to money income were adjustments for the underreporting
of incOme in the surveys and for under-estimated imputed rents
of owner-occupied homes, income in kind, net corporate savings,
and indirect taxes. The new income concept, called "adjusted
income," was used as an alternative income base to the money
income concept.

Once income bases were known, the burden of taxation
by income classes was estimated. The procedures used did not
differ substantially from other tax incidence_studies. First,
it was assumed that the burden of taxation should be measured
in terms of a reduction in real income of taxpaying households.
Then the question of incidence shifting had to be answered:
Who ultimately bears the burden of each tax?

A large part of Chapter 3 was devoted to resolving
this problem. When the incidence problems were resolved, each
major tax and government revenue in Thailand was allocated to
various households based on different allocation rules. These
rules generally followed the patterns of household expendi-
tures on certain groups of commodities and other distributive

series, the information being available from the household

261

expenditure surveys. The incidence of the total tax programs
falling on a particular income class was then obtained by sum—
ming up the absolute burden of each tax pertaining to that
income class.

When these absolute burdens were expressed as percent-
ages of income levels of each class, the result obtained was
the effective tax rate by income class. Subtracting these
amounts from the original or "pre-fisc" income distribution
yielded the "post-tax" income distribution. To account for
the distributional impact of the tax system by the simple sub-
traction of the absolute tax burdens from original income
levels requires an assumption that the original or pre—fisc
income distribution was indeed what would have resulted if
the present tax system had been replaced by a proportional
income tax of equal yield. This is the so-called differential
incidence approach. Thus, the pre-fisc and post-tax income
distributions within the same year could be compared, as
could post-tax income distributions of two different years--
in this case 1963 and 1969--for evidence of a change in in-
.come distribution.

In a similar fashion, the incidence of public ex-
penditures was estimated by first assuming that public ex-
penditures generate benefits that add to the real income of
households. Conceptual difficulties inherent in the quantifi-
cation of expenditure benefits, especially those from such
social goods as national defense and public health, necessi-

tated the adoption of the accounting approach to benefit

262

allocation, that is, the assumption that total benefits are
equal to total costs. Thus, actual government spending is
allocated to expected first-round beneficiaries, known as a
"cost-incurred-on-behalf-of approach." Once this decision
was made, public expenditures were classified into five func-
tional categories, and consideration was given as to who the
prospective beneficiaries were and how the benefits should be
allocated. Then all benefits of each income class were summed
and divided by corresponding income levels to arrive at the
effective expenditure rates. These same benefits were added
to the existing, or pre-fisc, income distribution to arrive
at the new, post-benefit, income distribution. The reasoning
underlying this procedure was that the patterns of the effective
expenditure rates and the changes registered in the post-bene-
fit income distribution should reveal the income redistribu-
tional nature of government expenditure programs.

Determining the net fiscal incidence, that is, the
incidence of the entire fiscal program, was simply a matter
of finding the net benefits (expenditure benefits minus tax
burdens) of each income class and expressing them in relation
to the income level of that income class. This process yield-
ed the effective fiscal rates. When the net benefits were
added to the existing income distribution, the new income dis-
tribution obtained is the "post-fisc" distribution. This, of
course, was the final result and the one of prime interest in

this study.

263

Conclusions

 

Since the main study is divided into three parts--
tax incidence, expenditure incidence, and net fiscal incidence——

the conclusions will be presented in the same way.

Tax incidence

 

The four conclusions concerning tax incidence are:

First, under a money income base, the tax programs
in Thailand in 1963 and 1969 were generally--almost monotomi-
cally-regressive across income classes. The effective tax
rates ranged from 20.8 to 42.7 percent, and from 25.7 to 47.0
percent, of household money income in 1963 and 1969, respec-
tively. The pattern of tax regressiveness in these two years
was practically identical.

Second, under the adjusted income base, the regres-
sivity was almost eliminated; the tax systems exhibited a
generally proportional pattern of incidence. The use of the
adjusted income base therefore provided the lower bound of
the effects of the tax programs, with the money income base
providing the upper bound. If this interpretation is accepted,
then the degree of tax burden in Thailand in 1963 and 1969
ranged from regressive to proportional at best; it could in
no way be considered progressive.

Third, regarding specific taxes, only the personal
income tax and taxes on property showed a progressive incidence

pattern; the others showed either a clearly regressive pattern

264

(business taxes, selective sales taxes, and import duties),
or a U-shaped pattern (government receipts from state monopo-
lies). But very few taxes were progressive. Taxes on in-
come and property combined constituted only about 8 percent
of total government revenues in both 1963 and 1969.

Fourth, it is obvious that when a tax is regressive,
the after-tax income distribution will be made more unequal.
This was true in Thailand in 1963 and 1969. The degree of
income inequality in the after-tax income distribution as
measured by Gini concentration ratios, increased from 1.0 to
8.4 percent in 1963, and from 0 to 5.3 percent in 1969, depend-
ing upon the income concepts used. Although the Gini ratios
revealed that the tax programs of 1969 were less income de-
equalizing (or more income equalizing) than the 1963 programs,
the reduction in the Gini coefficient was very small, amount-

ing only to about one or two percentage points.

Expenditure Incidence

 

The four conclusions concerning expenditure incidence
are:

First, based on the money income concept, the expendi-
ture programs of the Thai Government in 1963 and 1969 showed
a generally regressive or pro-poor pattern from the lowest to
the highest income brackets. In 1969 the incidence pattern
was clearly and steadily downward sloping, whereas in 1963
it was also downward sloping, but it turned slightly upward

at the top income bracket. This pattern is almost identical

265

to that of the tax programs considered earlier, the only
major difference being that the expenditure incidence of the
lowest income class was much higher than the tax incidence
of the same income class. This indicates that the poorest
households reaped relatively greater net real benefits from
government fiscal activities, at least from a money income
point of view.

Second, when the adjusted income base was used, ex-
penditure incidence flattened out considerably. The 1963
pattern, when show diagrammatically, became almost a straight
line, while the 1969 pattern still showed a very high level
of expenditure incidence in the lowest income class. But
for the other classes, the incidence pattern was more or
less proportional, with a slight increase in incidence in the
middle portion of the household range. On the whole, except
for the lowest income group in 1969, all income classes seem
to have received benefits from public expenditures in amounts
proportional to their adjusted income.

Third, hardly any expenditure item was shown to be
progressive or pro-rich, with the exception, perhaps, of in-
terest payments on the public debt, which is assumed to bene-
fit upper income bracket debt holders. The expenditures most
beneficial to lower income groups included those for agricul-
ture, primary education, and health and social welfare. The
benefits from general service expenditures, including mainly
defense, law enforcement, and public administration, were

also very regressive.

266

Fourth, if expenditure benefits are assumed to add di-
rectly to the income of households, the income distribution
after expenditure benefits are allocated, or the "post-benefit"
income distribution, must be more equal than the original,
pre-fisc income distribution. In this study, the Gini concen-
tration ratios for post-benefit income distribution in 1963
improved by about 0.7 to almost 5.7 percent over pre-fisc in-
come distribution, depending upon the income concepts used.
The corresponding rate of improvement in 1969 was between al-
most 1.7 percent to over 7.2 percent. The 1969 expenditure
programs were more income equalizing than the 1963 programs,
although both could be criticized for insufficient expendi-
tures in the areas benefiting the poor most directly, such
as primary education in rural areas and health and social wel-
fare benefits. For maximum redistribution, too much was spent
in both years in the areas of dubious benefit to the poor,

such as in defense and general bureaucracy.

Net Fiscal Incidence

 

Concerning net fiscal incidence, five conclusions may
be noted:

First, the fact that government fiscal programs in
1963 and 1969 provided net gains to all income classes is well
supported by the results of this study. Net benefits for the
lowest income class were quite pronounced in 1969 under the

money income base, much more so than was the case in 1963.

267

As for the other income classes, the net fiscal in-
cidence pattern was still continuously regressive in 1969
under the money income base. But for 1963, under that income
base, the trend turned slightly upward at the highest income
bracket, thus making the entire fiscal incidence picture ap-
pear somewhat U-shaped rather than generally downward sloping,
as in 1969. Under an adjusted income base, however, the net
fiscal incidence trends in both years were flattened out, as
was true of the tax and expenditure incidence patterns. Al-
though the regressivity remained under an adjusted income base
(at least in 1969), this was again barely apparent. (For
1963 the slope of the generalized net fiscal incidence func-
tion even turned slightly positive, meaning that the fiscal
system was progressive.) This being the case, it would be
best to call the fiscal programs of 1963 and 1969 proportional.
But of the two, the 1969 fiscal programs offered greater net
benefits to all households relative to their income than did
the 1963 programs.

Second, the foregoing conclusion that every income
class gained absolutely from government tax and expenditure
programs is valid only insofar as the budget imbalance, with
expenditures exceeding taxes by quite a large amount, is as—
sumed to add to the income of households without any cor-
responding reduction of real income through possible price
increases due to budget deficits. This is known as the post—
fisc income distribution under Assumption I. Using Assumption

II, that is, if the deficit is eliminated by a hypothetical

268

collection of a proportional income tax, then there were net
gainers and net losers. The pattern of post-fisc income dis-
tribution was the same as before the elimination of the bud-
get deficit, only the level of net fiscal incidence was re-
duced. The pattern of income redistribution became easier
to see; any income class with a negative effective fiscal
rate was said to have redistributed its income to other in-
come classes with positive effective fiscal rates. In 1963
the top and the bottom income classes were clearly the net
gainers, with the three income groups in between the net
losers. In 1969, the redistribution was quite clear-cut,
with the first four income brackets being the net gainers,
and the last three brackets the net losers.

Third, it is possible, and in fact more likely, that
the existence of a budget deficit would drive prices up in
Thailand. If it is assumed that prices rise to the amount
of the deficit, and that the effect is an equal reduction in
real income of each household (Assumption III,), then the
fiscal programs in 1963 and 1969 would turn mildly progres-
sive. In other words, a budget deficit in an overall fiscal
program is very important, and different assumptions about
its income distributional impact could yield different net
fiscal incidence results.

Fourth, under Assumptions I and II regarding the bud-
get deficit, the 1963 post-fisc income distribution improved
over the pre-fisc income distribution of the same year by

about 0.8 percent using a money income base, but worsened

269

by about 0.2 percent under an adjusted income base. In 1969,
the post-fisc income distribution improved over the pre-fisc
income distribution under both income bases, and the rate of
improvement ranged from 2.0 to 4.7 percent. But under Assump-
tion III, the post-fisc income distribution actually became
more unequal than the pre-fisc income distribution in both
1963 and 1969. In 1963 the post-fisc distribution worsened

by about 2.3 to 3.8 percent, compared to about 0.6 to 1.2
percent in 1969.

Fifth, and most important as a conclusion, the fiscal
programs of the government caused very little improvement in
the existing income distribution of Thai households. The ef-
fects were so minimal that by slightly changing the assumption
as to how to allocate the burden of the budget deficit reversed
the overall effects from equalizing to de—equalizing income
distribution. One brighter note, however, is that the fiscal
programs of 1969 were definitely and consistently more income
equalizing than those of 1963. This means that the govern-
ment's impact on income distribution is becoming more favor-
able as time passes. Yet, the fact remains that the rate of
improvement over the six years between 1963 and 1969 was still
woefully slow and the magnitude of improvement was small. Also,
the income redistributional effects of the government taxes
and expenditures in 1969 were still much too inadequate to off-
set the increased inequality in adjusted income distribution

over the previous six years.

270

Policy Implications

 

There are many ways in which the government could
change the distribution of income of individuals or house-
holds. Taxes and public expenditures are merely one instru-
ment of income redistribution, but since this study is con-
cerned only with tax and expenditure incidence, it is ap—
propriate to consider policy implications only in those areas.
No doubt the present findings could have implications in
other areas, but a discussion of these issues is outside the

scope of this research.

On the Tax Side

 

At least five recommendations may be made concerning
changes in taxation policy:

First, the entire Thai tax system is much too regres-
sive. This study has clearly shown how the lower income
households have suffered the burdens of regressive taxes re-
latively more than upper income households. The main aim of
any tax reform should be to reduce the regressivity of the
tax and revenue structures as much as possible.

Second, and more specifically, reform should be fo~
cused on strengthening income or direct taxation in Thailand.
An income tax that contributes only 5 to 6 percent of total
government revenues is much too low. The main problem is not
so much with the tax rate (although present structures could

be improved to make the effective rate more progressive),

271

but with compliance and enforcement. The present tax adminis-
tration system must be thoroughly streamlined before tax col-
lections can increase.

Third, a system of property and wealth taxation must
be instituted. It is difficult to believe that in the face
of continuing concentration of wealth among a small but grow-
ing number of families, Thailand has no estate, gift, or in-
heritance taxes. An effective wealth tax, probably both on
net wealth as well as transfer tax, would be a most commend-
able income redistributing agent in addition to an improved
income tax system.

Fourth, with increased revenues from direct taxation,
the dependence on such indirect taxes as business taxes, se-
lective sales taxes, and certain import duties would diminish.
Since these indirect taxes burden the poorer income classes
relatively more than the wealthier, less dependence on them
would certainly improve income equality.

Finally, certain taxes are singularly burdensome to
the poor, such as the rice premiums and state lotteries. The
government favors these revenue sources for their convenience
in collection, but they should be abolished if the government
is truly concerned with improving the present income distribu-

tion.

On the Expenditure Side

 

Four recommendations concerning expenditures seem most

important:

272

First, expenditure programs as a whole are regressive
or pro-poor, and this is desirable. But since a budget defi—
cit is probably detrimental to a more equal distribution of
income because of inflationary pressure, attempts should be
made to have a balanced or near-balanced budget through in-
creased direct taxes and a curtailment of certain expenditure
items.

Second, those expenditures which need to be reduced
include defense and public administration. The imputed
benefits from these general services constituting 40 to 50
percent of all benefits accruing to the poorest households
has a dubious meaning at best. The benefits from a large
defense and general administrative spending may be overesti-
mated, particularly from the point of View of the poorest
households.

Third, for income distribution purposes, there is a
need for additional spending in primary education and health
and public welfare services. Primary schooling in rural
Thailand is very underdeveloped. Considering the ever-present
budget constraints, emphasis should be placed more on primary
than on secondary and highly subsidized university education.
Also, an adequate social welfare program should be instituted.
At the very least, the already planned but long delayed so-
cial security program should be put into operation.

Fourth, agricultural expenditures should be increased.
The government should be more actively involved in agricultural

credit and in provisions for guaranteed prices and in devedoping

273

market outlets. Thai farmers are probably the segment of the
population that deserve to be assisted the most in redistrib-
uted efforts.

The foregoing are some policy implications and recom-
mendations stemming from the results of this fiscal incidence
study. It is noted again that these policies are confined to
the areas of taxes and public expenditures only; there are
many other nonbudgetary policies that could conceivably gen-
erate a much greater income redistributional impact. Land
reform is one example. If a massive land reform were suc-
cessfully carried out, it would be certain to improve the
distribution of income within the country. Thai farmers con-
stitute more than three-quarters of Thailand's population
and the majority are in debt.1 Their indebtedness ranges from
a fraction of a norman annual crop yield to many times that
amount, and within a few years they will lose-the ownership
of the land to creditor landlords or money lenders. Many of
these farmers are forced into debt through no fault of their
own, but because of adverse production and marketing condi-
tions, credit conditions, and the general neglect of the

government.

Suggestions for Further Research

This intertemporal study of fiscal incidence in Thai-
land is by no means satisfactory in every respect. There are
undoubtedly several shortcomings in analysis, statistical data,

and in misplaced emphasis. Future research should be conducted

274

with the following possible improvements in mind:

(1) Income and expenditure surveys should be carried
out frequently, preferably on an annual basis.

(2) It is to be hoped that household surveys will
become much more reliable as the system is regularized and
the researchers become more experienced. In these surveys,
an emphasis should be placed on perfecting the income elements,
including nonmoney income, on consistency in the methods of
eStimation, and in the presentation of the results.

(3) Tax and expenditure records should be collected
with as much disaggregated detail as possible. The analysis
should be expanded from a national level to a regional, sec-
toral and provincial level.

(4) The practice is becoming widespread of studying
in detail, the income distributional effects of specific
public expenditures such as those for irrigation and public
health. This involves much more than simply assuming who the
beneficiaries of such expenditures are, and then allocating
those expenditures; actual field studies are necessary in
which sample surveys are conducted to determine who receives
the benefits and how much, using some index of benefit valua-
tion. For example, regarding health expenditures for public
hospital operations, income records could be kept of the in-
come groups who come for medical service and the amount of
services provided. This method of study is very expensive,
of course, but it would provide a much clearer idea of the

areas to which the government should reorient its policy for

275

equity purposes.

(5) If possible, researchers should attempt to in-
corporate in their studies the income redistributive in—
fluences of the non-budgetary policies and external influences
such as the effects of foreign aid and loans. Certainly, this
is not the work for one or even a few researchers. Perhaps a
government body could be established to deal specifically
with this problem of evaluating the income redistributional

impact of the entire economic structure.

FOOTNOTE S

lSee Uthis Narksaward, Report on the Indebtedness
and Marketing Practices of Rice Farmers in the Central Plain
of Thailand, B. B. 2510-2511. (Bangkok: National Research
Council and United States Operation Mission, 1970).

 

 

 

276

APPENDICES

APPENDIX A

THE ESTIMATION OF THE NUMBER OF HOUSEHOLDS IN 1963

APPENDIX A

THE ESTIMATION OF THE NUMBER OF HOUSEHOLDS IN 1963

l. The pOpulation figures by regions taken from
the 1960 population census are compared to the population
figures by regions taken from the 1970 population and hous-
ing census. From the total rate of increase of population
in each region over 10 years, the average annual rate of
growth is estimated.

2. The population in 1963 by regions and locations
is then estimated by raising the 1960 pOpulation by three
times the average annual growth rate of population. The
result is shown in Table A.l.

3. The average numbers per family by regions (also
shown in Table A.1) are obtained from the 1962-1963 House-
hold Expenditure Survey, and they were used to divide the
population in each corresponding location and region in
1963 to arrive at the number of families by regions.

4. Multiplying these families by the percentage
distribution of families by income classes (Table 2.1) and
summing up all urban and rural households in all regions
resulted in the total number of households, as shown in

Table A.2.

277

278

Table A.l--Estimated Population in 1963 (Numbers in Thousands)

 

 

 

Regions Urban Rural Total
North 855.1 (5.1) 5,397.7 (5.3) 6,252.8
Central-East 1,126.4 (5.3) 5,731.5 (5.6) 6,857.9
Northeast 776.8 (5.4) 9,124.9 (5.9) 9,901.7
South 595.0 (5.6) 2,976.8 (5.2) 3,571.8
Bangkok 2,115.6 (5.5) -- 2,115.6

Whole Kingdom 5,468.9 23,230.9 ‘ 28,699.8

 

 

 

 

Note: Figure in parenthesis is the size of household in that
region and location taken from the Reports of the
1962-1963 Household Expenditure Survey publISHed by
the National Statistical Office.

 

279

Table A.2—-Estimated Number of Households in 1963 (Numbers
in Thousands of Households)

 

 

 

Incomgagiasses Urban Rural Whole Kingdom

%
Under 3,000 129.7 2,359.3 2,489.0 48.1
3,000- 5,999 175.2 898.8 1,074.0 20.8
6,000-11,999 336.8 674.1 1,010.9 19.5
12,000-17,999 159.3 141.5 300.8 5.8
18,000-23,999 76.1 49.9 126.0 2.4
24,000-29,999 38.1 12.5 50.6 1.0
30,000-35,999 29.2 8.3 ’37.5 0.7
36,000-47,999 25.3 8.3 33.6 0.7
48,000-59,999 14.1 4.2 18.3 0.4
60,000 and over 29.2 4.2 33.3 0.6
All Classes 1,013.0 4,161.0 5,174.0 100.0

 

 

 

 

APPENDIX B

BASIC INFORMATION FOR THE ESTIMATION OF THE

DISTRIBUTION OF MONEY INCOME IN 1969

APPENDIX B

BASIC INFORMATION FOR THE ESTIMATION OF THE

DISTRIBUTION OF MONEY INCOME IN 1969

Since the basic information for the estimation of
the 1969 money income distribution in this study is obtained
from the retabulation of the original 1968-1969 socio-econo-
mic survey results by Dr. Oey Astra Meesook, the outcome of
which differed slightly from the officially published
figures, it is necessary that such basic data be provided
for the interested reader.

1. Table B.1 shows the number of households in 1969

taken directly from the Repert.gf Population and Housing

 

Census, 1970. Since the census was conducted primarily

 

during 1969, no further adjustment was made to these figures.

2. The distributions of urban and rural households
by regions are shown in Table B.2. They provide the weights
by which the urban and rural households are subdivided into
regions.

3. The distributions of urban and rural households
by income classes in each region are available from the 1968-
1969 socio-economic survey (Tables B.3 and 3.4). With these
distributions, the number of urban and rural households by

income classes in each region can be estimated.

280

281

4. Given the average annual money income by income
classes in each region (Tables B.5 and B.6), also available
from the 1968-1969 socio-economic survey, the total money
income of households by locations, regions, and income

classes easily can be computed.

Table B.l—-Number of Households in Urban and Rural Areas, 1969

 

 

 

Location figgifggi?s Percentage
Urban 762.1 12.9
Rural 5,146.4 87.1
Whole Kingdom 5,908.5 100.0

 

 

 

Source: Thailand, National Statistical Office, Report of
Population and Housing Census, 1970 (Bangkok, 1974),
Table 2.

 

 

Table B.2--Percentage Distribution of Urban and Rural
Households by Regions, 1969

 

 

 

Region Urban Rural
North 11.0 24.9
Central 16.3 23.6
Northeast 10.7 36.4
South 10.8 13.6
Bangkok 51.2 1.5

Total 100.0 100.0

 

 

 

Source: Thailand, National Statistical Office, Report of
Socio—Economic Survey, 1968-1969 (Bangkok, 1974),
p. 36.

 

 

282

Table B.3--Percentage Distribution of Urban Households by

 

 

 

 

Income Classes, 1969
Income Classes Regions
(Baht) North Central Northeast South Bangkok

Under 3,000 1.5 1.2 1.4 1.2 0.5
3,000- 5,999 12.5 5.0 3.7 6.4 1.2
6,000- 8,999 12.5 7.4 9.6 13.9 4.4
9,000-11,999 16.5 10.5 13.2 14.5 8.9
12,000-14,999 10.6 14.5 11.5 11.9 10.2
15,000-17,999 8.2 11.1 9.4 10.1 9.8
18,000—23,999 14.8 18.2 19.5 14.0 19.8
24,000-29,999 5.4 12.3 10.3 8.8 12.4
30,000-35,999 5.3 5.3 6.1 . 5.4 7.9
36,000-47,999 7.1 4.8 4.2 5.1 9.0
48,000-59,999 1.1 3.4 5.0 3.8 5.6
60,000 and over 4.8 6.3 6.2 5.0 10.4
All Classes 100.0 100.0 100.0 100.0 100.0

 

 

 

 

 

 

Source: Results of the retabulation of the original data
tapes of the 1968-1969 socio-economic survey by
Dr. Oey Astra Meesook of Thammasat University.

283

Table B.4--Percentage Distribution of Rural Households by

 

 

 

 

Income Classes, 1969
Income Classes Regions
(Baht) North Central Northeast South Bangkok
Under 3,000 18.4 7.0 52.1 17.2 1.1
3,000- 4,499 19.4 9.7 15.6 19.5 3.1
4,500- 5,999 14.9 9.4 7.3 19.2 5.8
6,000- 7,499 12.3 11.2 5.8 14.2 6.0
7,500— 8,999 8.9 10.6 4.2 8.7 3.2
9,000-10,499 4.9 10.3 4.7 6.2 7.3
10,500-11,999 4.0 6.1 1.1 3.6 6.6
12,000-14,999 7.2 10.2 3.1 3.8 15.3
15,000-17,999 3.8 7.4 1.6 3.0 11.3
18,000-32,999 5.5 12.6 3.4 4.2 24.8
33,000 and over 0.8 5.6 1.1 0.4 15.6
All Classes 100.0 100.0 100.0 100.0 100.0

 

 

 

 

 

 

Source:

Results of the retabulation of the original data

tapes of the 1968-1969 socio-economic survey by
Dr. Oey Astra Meesook of Thammasat University.

284

Table B.5--Average Urban Annual Money Income, by Income

Classes,

1969 (Amounts in Baht)

 

 

 

 

 

 

 

 

 

 

Income Classes Regions
(Baht) North Central Northeast South Bangkok
Under 3,000 2,326 1,578 1,617 1,302 1,747
3,000- 5,999 4,503 4,171 4,265 4,522 4,545
6,000- 8,999 7,530 7,526 7,128 7,192 7,429
9,000-11,999 10,144 10,394 10,302 10,302 10,278
12,000-14,999 13,112 13,244 12,899 13,166 13,216
15,000-17,999 16,182 16,066 16,070 16,324 16,186
18,000-23,999 20,196 20,208 20,117 20,313 20,256
24,000-29,999 25,838 25,650 25,639 25,965 26,303
30,000-35,999 32,286 32,327 32,786 31,871 32,169
36,000-47,999 40,295 40,628 40,664 40,033 40,670
48,000-59,999 53,703 52,035 52,865 53,614 52,705
60,000 and over 115,354 101,226 130,291 126,847 117,123
Average 20,955 24,188 26,213 23,233 32,844
Source: Results of the retabulation of the original data

tapes of the 1968-1969 socio-economic survey by
Dr. Oey Astra Meesook of Thammasat University.

285

Table B.6-~Average Rural Annual Money Income, by Income

Classes,

1969 (Amounts in Baht)

 

 

 

 

 

 

 

 

 

 

Income Classes Regions
(Baht) North Central Northeast South Bangkok
Under 3,000 2,078 1,832 1,546 2,125 1,732
3,000- 4,499 3,638 3,674 3,622 3,589 3,694
4,500- 5,999 5,125 5,145 5,160 5,082 5,266
6,000- 7,499 6,491 6,665 6,556 6,682 6,743
7,500- 8,999 8,028 8,188 8,164 8,226 8,013
9,000-10,499 9,699 9,618 9,542 9,688 9,708
10,500-11,999 11,245 11,096 11,081 11,044 11,094
12,000-14,999 13,117 13,087 13,104 13,104 13,403
15,000-17,999 16,083 15,881 16,148 16,097 16,215
18,000-32,999 23,470 22,861 23,058 22,386 23,158
33,000 and over 45,947 60,287 131,372a 51,946 59,824
Average 7,488 13,009 5,963 6,831 21,488
Source: Results of the retabulation of the original data

tapes of the 1968-1969 socio-economic survey by
Dr. Oey Astra Meesook of Thammasat University.

a . '
Computationally correct figure

APPENDIX C

THE ESTIMATION OF INCOME IN KIND IN 1963

APPENDIX C

THE ESTIMATION OF INCOME IN KIND IN 1963

1. The average income in kind by income classes for
1969 is supplied by Dr. Oey Astra Meesook and is presented
in Table C.l.

2. The relationship between "total" income, defined
as money income, plus income in kind is established through
the following linear regression equation:

Y = a + me = u,

t

where Yt is total income, Ym is money income, a and b are
parameters to be estimated, and u is the statistical dis-
turbance. The resulting estimates of parameters in the income
in kind functions are shown in Table C.2.

3. After substituting the 1963 average monthly
money income of each income class from each region and
location into the estimating equation, the average income
in kind by income classes, locations, and regions is obtained.
This estimation uses the average money incomes supplied by
Dr. Meesook, not the midpoint incomes, because it is
believed that Dr. Meesook's data provide better income
variations among regions. It should be mentioned, however,
that both estimated total incomes (that is, money income

plus income in kind) using both methods are very close to

one another.

286

287

4. When the average income in kind is multiplied by
the number of households in 1963, the total income in kind
for 1963 is obtained,as presented in Table C;3.

Table C.l--The Average Income in Kind, by Locations, by
Income Classes, 1969 (Amounts in Baht)

 

 

 

Incomgaglasses Urban Rural
Under 3,000 1,398 2,826
3,000— 5,999 1,221 4,418
6,000- 8,999 742 4,462
9,000-11,999 829 3,851

12,000-14,999 957 2,030

15,000-17,999 1,193 2,202

18,000 and over 8,927 5,140

 

 

 

Source: Basic data supplied by Dr. Oey Astra Meesook of

Thammasat University.

I.ll.‘..l'l.l' III!!!

 

I! iuiil Is! Iii!!!

.I .I. l.§il II I

288

Table C.2--The Estimated Income in Kind Regression Functions

(Yt = a + me + u)

 

 

 

 

 

 

 

 

Locations and Regions Intercept Slope R2
Urban
North 952 .9929 .9995
(6.65) (161.59)
Central 927 1.0130 .9995
(4.64) (137.55)
Northeast 913 1.0247 .9998
(2.65) (93.34)
South 777 1.0234 .9997
(6.80) (234.13)
Bangkok 911 1.0063 1.0000
(14.56) (693.43)
Rural
North 1,776 1.0248 .9907
(12.96) (49.65)
Central 1.876 1.0146 .9999
(34.35) (300.10)
Northeast 3,477 0.9719 .9988
(63.47) (67.70)
South 1,271 1.0520 .9983
(17.74) (86.51)
Bangkok 657 1.0286 .9999
(4.53) (214.73)
Source: Results supplied by Dr. Oey Astra Meesook of
Thammasat University.
Note: Figures in parentheses are t-statistics.

289

 

 

 

 

 

 

 

a.mo~.aa a.amm m.mma a.mma.a a.mao.~ a.mam.a _spamcax mama:
m.mam a.mm m.mm a.amm a.mmm a.mmm apnea
m.mm~.m a.am m.ma m.aaa m.aaa m.maa.m pmmmnupoz
m.aaa m.ma a.am a.mma m.mma a.maa papa
a.mma.a a.mm m.aaa a.mam a.mom m.aaa appacmo
o.mom.a m.ma a.am m.mmm m.mam m.maa.a 5.82
.mmmmm
m.aaa m.aaa a.am m.mma m.ma a.ma xoxmamm
m.maa m.mm a.mm m.mm m.aa m.a napoa
m.mma m.am m.mm m.ma a.am a.mm ammonuuoz
a.mm m.ma m.a a.am m.m m.a pmam
m.aaa m.ma a.mm a.mm a.am m.ma amuucmo
m.aaa m.ma a.ma a.mm a.am m.ma appoz
ammmmm
pm>o_a:a mmm.ma mmm.aa mmm.a moo.m
8830 momma -ooo.ma name .a .68 .m 88: 898m a8 8883
aaa

 

35mm: monumau m.a—08a

 

 

 

328 a0 appaaaaz ca 8885 mama 65.2 :a

9:85 HBOBIIM .U 0....nt

 

 

 

[I'llII'llli ll“

APPENDIX D

DISTRIBUTIVE SERIES FOR USE AS TAX AND

EXPENDITURE ALLOCATION BASES

APPENDIX D

DISTRIBUTIVE SERIES FOR USE AS TAX AND

EXPENDITURE ALLOCATION BASES

This appendix describes how each allocation base from
the distributive series table was obtained. This information
will enable the interested reader to retrace the steps taken
in estimating the final tax and expenditure incidence in

this study.

Tables D.l and D.2

 

Base Bl-B8: For 1963 these distributions of house-

 

hold expenditures are computed by multiplying the average
monthly expenditures by families on eight commodity groups
(shown in Tables 4.3 and 4.4 in the 1962-1963 Household
Expenditure Survey) by the number of households in the
corresponding income brackets (see Appendix A). The eight
commodity groups are listed in greater detail as follows:
(1) food purchased and prepared at home as well as food
consumed away from home; (2) housing, furnishing, and house-
hold operations; (3) clothing and materials; (4) transporta-
tion; (5) reading, recreation, and education; (6) medical
and personal care; (7) tobacco and alcoholic drinks; and
(8) gifts and contributions, taxes, and miscellaneous house—
hold expenses.

The percentage share of expenditures of each income

class in each commodity group is then computed from the total

290

“mocmucooummv mwmfl .mmfiumvm m>flusfiuuwwflllwd wanna.

291

 

 

 

 

 

 

 

0.006 0.3 0.0 0.0 0.0 0.0 x2. 885 666385 mm
0.006 0.3 0.2 0.3 0.0 0.0 macfigm 698598 mm
0.03 0.2 0.3 0.00 04 0.0 3.60606 60.885 6
0.03 m.mm m.m 0K m0 0.0 38036 56.0.6965 0N
0.006 0.00 0.3 0.00 0.3 0.0 36.80 6896 no 9586 2
0.006 0.2 a.mm 0.2 0.0m 0.00 6805 mo 9.8.66.6 3
0.03 0.3 0.: 0.: 6.6 a.mm 2.096 9.66 S
0.03 0.5 0.2 0.2 m.m m.a .885 :35 m:
0.03 m.mm 6.2 0.00 0.3 0.2 9805 638. 2
0.03 0.00 0.0 0.0 0.0 0.0 anew 9.85 08. S
0.02 6.0 0.0 0.3 0.60 0.00 39688 6.56 3
0.006 0.0m 0.3 m.mm 0.: m.ma 328:8 635 NH
002 m.m m.m 0.3 0.00 6.00 Batman 38. d
0.03 m.mm 566 0.00 0.3 0.0m .96 086.82 638 S
0.02 m.ma 0.: 0.0m 0.: 0.00 .96. 38. 0
0.006 0.00 0.: 0.6 0.3 a.mm .96 BBSSmBE 0
0.006 0.06 0.2 0.0m 0.: 0.0m .96 088.8. a
0.03 0.2 0.: m.mm 0.2 0.00 .96 88 68:82 0
0.03 0.: 0.3 m.mm 0.2 0.2 .96 838686 0
0.006 m.mm m.ma 0.0m m.ma 0.3 .96 cflumfi&mcg 0
0.006 0.2 0.: 0.0m 0.5 0.00 .96 054.68 m
0.03 m.mm 0.2 0.00 0.3 0.00 .96 96.0.8: 0
0.006 0.3 . 0.3 0.00 0.3 «:00 .96 88 H
“68.6% 000.2 000”: 000mm 000.0
mmMmU 000 2 -000 NH -000 0 -000 m 365 mcoflsflfifla mmmm
same .6688 285

 

 

 

 

 

36935869 83 63.86 ggflmfloia .o magma

Table D.2-Distributive Series, 1969 (Percentages)

292

 

 

 

 

 

 

 

 

g ooooooooooooooooooooooo
I I I I I I I I I I I I I I I I I I I I I I I
H ooooooooooooooooooooooo
21:3 ooooooooooooooooooooooo
a HHHHHHHHHHHHHHHHHHHHHHH
33:!
80 HOhNNhomvoocooonomxoONomxoh-zr
I I I I I I I I I I I I I I I I I I I I I I I
~ mmgamvaHv-u'mcoONNthMNNm
3g mq mmmmmvwam gunmen Ncommm
I
cm
83 onmoommxomvxococooooxovoom—ivwa
~~ hhmhmmmhhhvmmowmhhvaHN
ml‘ H 01.4
.4.—a
I
cm
:3 83: LONQ'HmNmFVQ‘mMHOHNmNKDu-immrv
I I I I I I I I I I I I I I I I I I I I I I I
.. mmmmmmmmmmmamommmehvhha
Nv H H o-n-I
.4.-4
V
m
0 I
m cm
:3 8g manhmmwmmomomONommHhNNm
I I I I I I I I I I I I I I I I I I I I I I I
8 .~ :mdmhaomoom—amomvHNwavo
as: .4.-4 .4.-1 .4.—4H NN
8 u
'5 83 moomuoooooNNxobommoxomoooxohNNo
I I I I I I I I I I I I I I I I I I I I I I I
om mdmahwmmNHmhhoov-«moomvvo
.. .-c .4.-4 .4.-1 Hr-JI-l .-I Hv-IN .4.-4
moo
$3: \omomNNmNoohoomooomxoxoq-vmooo
I I I I I I I I I I I I I I I I I I I I I I I
can NOMFWMHQHOMMkDOO‘OMNMOOOO
~~ .4.-4H .4.-4 l—‘I—IN N Pit-FIN
mm
o
8 HammmhommoammOhr-uooocoq—«oo
I I o I 0 O 0 0 O o o I I I I a I I I I I I I I
. (\RDBMNNBIDKDKDQ'OI‘OMOMOQOOOO
m N N m

 

Distributions

Recreation

musing Exp.

Clothing Exp.
'Ibtal Non-food Ekp.

Transportation Exp.
Medical Care Exp.
Tobacco Exp.
Miscellaneous Ebcp.
Total Household
Urban Household

‘Ibtal Exp.
Rural Household

Food Exp.
'Ibp Inoane Group
Total Income

Urban Imone

Rural Income

Ihciprocal of Incrme

Incane of Rubber Growers

University Students

'13

1c1a

Government Off

Government Pensmners

Individual Income Tax

 

 

 

 

1
2
3
4
5
6
7
8
9
0
12
13

ll

14
15
16

17

18

19

20
21
22
23

 

293

 

 

 

 

 

 

 

0.000.60 0000.0 0.0000 0.00006 0000.0 0.006.66 006360696 6596.
0000.0 0000.6 0.060 0.000 0.0.00 0000.6 00866668002
0.0006 0.060 0.060 0.000 0.060 0.000 8930.6.
0000.0 0.060 0.000 0.000 0.000 0.000 88 680066
0.0006 0.000 0.600 0.000 0.000 0.000 8068.680
0.0006 0.000 0 .000 0. 600 0.060 0.000 . 86636960566.
0000.0 0.0066 0.000 0.0006 0.0066 0.0006 055060
0000.0 0060.6 0.000 0.0006 0.000 0.0006 0:608:
0.000.06 0000.0 0000.6 0006.0 0000.0 0000.0 6.86
608 0% 000.06 000.66 000.0 000.0
8mm6m6o 000.06 000.06 000.0 000.0 6.00:0 00:80 06.602660

 

3.6000 8000.600 058:6

 

 

 

3:06 00 0866662 :6 065050

0006 .86660896 069606566 0396.10.06 06060.6.

294

 

 

 

 

 

 

 

 

 

m.amv.mm m.ovN.vN mNomJV 0.53:.“ m.mamfi 0.0.3.0. a.ammb 06.3.0” mmflflfig H638.
H.th.v v.mmv.N 0.60m H.090 m.mmm mNmm . 60.va N.NmN gaqmaawomfiz
v.mmH.m N.NMN.H otnmN Whom H.Nvm 5.5.3 m.abm qtnNN 00038
a.mooJ a.ammra waN m.mmm m.amv H..Hmm mKom o.o....m 6.6.600 H6003
«.mmmN mwaJ m.vON a.amN mNmH m.aaa Emma 03mm 8388a
m.mmvfi .70th .H.NvN 0.3.0” m.mmm H.No¢ mJKN m.ONH . 833.609.008.69
m.mmm6 m.mmvN N630 0.0mm vinNh Ndvm 9th a.mov 9.6.0.508
0.935 m.mme m.amw ._...wvm w.omm N.mo~. m.mvw mNmM 96.09603
m.mmH.mN m.mmn.oa v.mma.~ 0.2.9N m.mmN.m mNmfim v.mmm.m m.NHo.N 08h
.695 Em mmmKH mmmJH mmmJa mmm.m mmmfi ooo.m
mmmmdau ooo.ma Iooofia loco.NH I255 I256 A508 .6865 @9566 03.69500
H2 Augmmv mmmmMHU mag

 

 

 

 

35606 no 665.0302 660 3030.56; $00 566503 365090: 33:05 6360.6.

'—
‘
_

_
‘

.-—

‘—

A

_

I A

(ann‘
exaC'
hous‘
in T.
Surv.
Tabl¢

in l!

exper
exper
in Ta
hold

numbe

is ob

total

rural

fOr l

the ix

highe:

frOm T

295

(annual) expenditures, as shown in Table D.3. For 1969,
exactly the same procedure is used. The average monthly
household expenditures for eight commodity groups are given

in Table 4 in the Report of the 1968-1969 Socio-Bconomic

 

Survey, and the number of households in 1969 in given in
Table 2.6 in the text. Table D.4 shows total expenditures

in 1969.

Base B9: The distribution of total household
expenditures is estimated by multiplying the average total
expenditures, by locations, by income classes (also shown

in Tables 4.3 and 4.4 in the 3gport of the 1962-1963 House-

 

hold Expenditure Survey: Whole Kingdom)turthe corresponding

 

numbers of households (Table A.2).

Base 310: The distribution of nonfood expenditures

 

is obtained from the total household expenditures after the

total food expenditures by income classes are subtracted.

Base 811-313: For 1963, the distribution of urban,

 

rural, and total households is obtained from Appendix A:

for 1969, it is obtained from Tables 2.4 to 2.6 in the text.

Base B14: This distribution is used in cases where

 

the incidence is assumed to fall entirely on the top or the

highest income group.

Base 815-817: For both 1963 and 1969, the distribu-

 

tions of urban, rural, and total money income are obtained

from Tables 2.3, 2.4, and 2.5 in the text.

for
recig

this

the I
prod'
nunum
enabi
by s:
then

The a
into

is as
remai

Patte

inconm
entitl
Univer
Thamma:
Student
their P

fOr 197

296

Base 818: The total money income in each income class

 

for 1963 and 1969 .ns reciprocated. The share of the
recuxocated sum by each income class is then estimated to give

this distribution.

Base 819: The 1963 Agricultural Census conducted by

 

the National Statistical Office contains information on the
production of rubber by size of holdings and also the total
number of rubber growing households by size of holdings. This
enables one to compute the total volume of rubber production
by size of holdings for 1963. The average rubber price is
then used to convert it to total income by size of holdings.
The average income of rubber growers is then adjusted to fit
into the income class classification for 1963. For 1969, it
is assumed that the distribution of income of rubber growers
remains the same as in 1963, so the 1963 distributional

pattern is used.

Base B20: The distribution of university students by

 

income classes for 1963 and 1969 :hs obtained from the report

entitled General Facts about First-Year Students at

 

Thammasat University! 1972-1973, published by Thammasat

 

University. It is assumed, first, that the distribution of
Thammasat students is representative of all other university
students with respect to the overall income positions of

their parents and, second, that the pattern of distribution

for 1973-1974 is the same for 1963 and 1969. This second

fc

by
Se:
F01
she
the
to
mil.
the
Sim;
lowe
mOVe
in t
in t:
It i:
Pensj
the E

deCid

perCeJ

297

assumption tends to bias in favor of the lower income
classes because the current pattern is for these lower income
classes to be able to attend universities which they could
not attend ten years ago. But this bias should not unduly
affect the pattern of benefits received since expenditures

for higher education are quite small.

Base 821-822: The distribution of government officials

 

by size of salaries was obtained directly from the State
Service Commission and the Department of Comptroller General.
For distribution in 1969, one might notice a small percentage
share in the lowest income class. This was the result of

the program set up by the Department of Local Government

to hire some villagers to perform the duty of village
militiamen, for which each was paid a small fee. Thus

the distribution of government pensioners was estimated
simply by deleting the number of official households in the
lowest salary bracket. It is recognized that officials do
move up the income scale when they retire, but most remain

in the middle income brackets because the promotion system

in the Thai bureaucracy is, by and large, extremely slow.

It is possible to obtain the actual distribution of goverment
pensioners for 1963 and 1969 by reconstructing data from

the Department of the Comptroller General, but it was

decided that this was not worth the time and effort.

Base B23: This distribution in fact shows the

 

percentage of income tax that has been paid by households in

various
mnlin
in the .
These t4
into the
For 1961
that the

approxir

 

298

various income brackets. For 1969 the statistics on indivi—
dual income tax returns by size of taxable incomeannapresented
in the 1969 annual report of the Department of Revenue.

These taxable income brackets have been rearranged to fit

into the income class classification used in this study.

For 1963, there are no similar statistics, so it is assumed
that the pattern of income tax burden for 1963 is

approximately the same as for 1969.

APPENDIX E

DISAGGREGATED TAX DATA

Tab}

APPENDIX E

DISAGGREGATED TAX DATA

Table E.1--Disaggregated Taxes and Other Revenues, 1963 and
1969 (Amounts in Millions of Baht)

 

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories
1963 1969
1. Individual income tax 468.2 1,119.5
2. Corporate income tax classified by
the following types of business
activities: 312.2 851.1
2.1 Agriculture, forestry, fishery 2.4 2.7
2.2 Mining 40.6 17.2
2.3 Construction and repair 5.7 20.5
2.4 Manufacturing 81.7 291.7
Food 1.6 47.4
Softdrink and beer 11.2 --
Alcoholic drinks 7.3 14.8
Cement and cement products 36.2 59.9
Wood products '-- 21.5
Cosmetic and leather products -- 53.7
Glass 7.0 10.7
Other 18.4 83.6
2.5 Public Utilities and Transport 17.7 38.9
Public transport 16.5 38.1
Water and light 1.2 0.8
2.6 Commerce 97.4 306.9
Export-import 31.2 80.1
Hardware and dep't stores -- 80.6
Other 50.7 146.2
2.7 Banking, insurance, real estate 31.6 79.7

299

300

Table E.1--continued

 

 

 

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories
1963 1969
2.8 Other services 32.2 33.3
Service stations 19.6 --
Other 12.6 --
2.9 Other miscellaneous 2.7 60.2
3. Selective sales taxes 400.5 2,521.1
Entertainment taxes 47.6 85.6
Liquor 194.4 299.7
Beer 41.5 190.7
Non-alcoholic beverages 42.5 126.5
Matches 20.8 24.2
Tobacco 30.0 738.5
Cement 16.5 45.2
Snuff 6.6 10.8
Oil and fuel -- 999.7
Other 0.5 0.1
4. Business taxes 1,561.3 3,490.5
4.1 Sales type A ’928.3a 1,622.3
List 1 1,393.4
Food, drinks, and tobacco 473.0
Clothing 105.1
Cosmetics and medicine 115.2
Building materials and
stationery 221.6
Household tools and utensils 40.8
Motor vehicles 61.1
Fuel, oi1,and petroleum
products 235.3
Sundry goods 84.8
Miscellaneous 56.5
List 2 134.0
Food 2.4
Household goods 22.0
Machinery 5.2
Other 104.4

301

Table E.1--continued

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories

1963 1969
List 3 94.9
Upland crOp export 94.9
4.2 Sales type B 265.0 663.9

Refrigerators, air condi-
tioners 18.6 8.0
Electrical appliances 0.6 36.0
Passenger cars 15.1 144.9
Liquor 79.2 308.2
Painted material 1.9 3.8
Gold and jewelry 7.2 14.0
Tin and rubber exports 126.6 135.4
Other (food) 15.8 13.6
4.3 Rice and saw milling 124.9 223.7
4.4 Contractor and services 89.0 439.1
Printing 4.2 14.0
Photographic shOps 2.5 4.2
Civil construction 33.4 247.5
Advertising ' 1.9 14.7
Barbershops, beauty salons 1.7 4.5
Other 45.3 154.3
4.5 Rental services 3.0 17.1
4.6 Storage services 0.8 3.1
4.7 Hotels and restaurants 49.6 135.4
Nightclubs 5.2 12.6
Hotels 13.6 57.9
Restaurants 30.8 64.9
4.8 Transport services 14.4 48.9
4.9 Pawnshops 3.0 8.0
4.10 Underwriters, auctioneers 37.5 68.8
4.11 Real estate agencies 0.9 32.1

302

Table E.1--continued

 

 

Tax and Revenue Categories

 

 

 

 

 

1963 1969
4.12 Banking and insurance 39.0 204.5
4.13 Net detail given -- 22.3
4.14 Other 5.9 1.3
5. Import duties 2,764.3 5,294.0
5.1 Food 247.1 369.5
5.2 Drinks and tobacco 116.3 280.2
5.3 Raw materials 19.0 57.3
Wood fiber 0.7 5.6
Clothing fiber 4.8 20.9
Fertilizer 5.7 13.7
Metal and other raw

materials 7.7 17.1

5.4 Fuel, oi1,and petroleum
products 661.4 434.3
5.5 Animal and vegetable fats 3.5 8.3
5.6 Chemical products 236.9 ‘570.4
Chemical compounds 49.1 100.4
Dye 35.0 74.3
Pharmaseutical products 42.0 68.1
Perfumes 51.9 101.1
Other 58.9 226.4
5.7 Manufactured goods 732.0 1,512.8
Clothes 396.7 625.1
Leather products 1.6 6.3
Wood and cork products 2.0 6.2
Rubber products 77.6 115.3
Paper - 40.1 186.3
Non-metallic products 38.0 82.9
Silver and alloys 3.8 9.0
Other metal products 172.2 481.6

303

Table E.1--continued

 

 

 

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories
1963 1969
5.8 Machinery and transport
equipment 597.0 1,675.4
Heavy machinery 182.5 450.8
Electrical generators 130.4 376.9
Transport equipment 284.1 847.8
5.9 Miscellaneous manufactured
goods 139.6 362 2
Household fixtures 11.8 23.6
Furniture 3.4 8.7
Travel goods 1.9 5.9
Clothing 28.8 55.9
Shoes 2.7 7.6
Scientific equipment, films,
and watches 32.5 88.2
Other 58.6 172.4
5.10 Others not included elsewhere 11.5 23.6
6. Rice premiums 798 4 1,235.6
7. Other export taxes 363.2 424.3
Rice 154.5 157.4
Rubber 187.0 227.9
Wood 9.4 19.9
Rawhide 4.7 1.1
Other 7.6 18.0
8. Taxes on property 147 4 390.1
Motor vehicles and automobiles 88.6 214.7
Immovable property registration 58.8 175.5
9. Royalties and permits 422.0 691.9
9.1 Royalties 173.3 354.1
Bird's nest 1.1 2.3
Fishery 6.0 3.5
Wood 5.0 6.8
Other forest products 39.0 46.5
Tin 122.3 294.9

304

Table E.1--continued

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories

1963 1969

9.2 Permits 248.7 337.7
Radio 0.5 0.5

Liquor sale 3.1 2.0
Other excise 3.4 8.1
Forest 1.6 7.3

Rubber 20 7 8.4
Alien registration 83.2 63.5
Lottery sale 75.0 118.1
Gambling 48.5 106.9

Gun and fireworks 8.6 14.1

Civil 2.9 4.2
Health 1.3 4.8

10. Government sales and services 322.0 316.1
10.1 Estate sale 0.8 0.3
10.2 National products sale 6.9 5.5
Fuel and oil 5.5 0.4

Wood products 0.7 3.5
Agricultural products '0.6 1.7

10.3 Public utilities 28.1 --
10.4 Books and documents 0.3 0.6
10.5 Other sales 13.1 43.0
10.6 Services (fees) 221.3 204.8
Customs 6.0 5.7
Animal epidemic control 6.8 5.6
Watergate 8.7 10.8

Land and water transport 7.1 24.7
Fishing 0.6 0.4
Forest conservation 39.0 57.5
Airport tax 6.1 14.3
Vehicles 5.0 11.0

Land fees 3.3 6.2
Mineral fees 0.5 1.1
Weighting fees 7.8 16.3

Others 130.4 51.3

305

Table E.1--continued

 

 

 

 

 

 

 

 

 

 

Year
Tax and Revenue Categories
1963 1969
10.7 Rents 51.6 61.6
11. Government monOpolies 776.1 605.3
Contributions from tobacco

monopoly 733.8 450.0

Lottery Bureau 40.8 149.7

Other 1.5 5.5

12. State enterprises 226.3 636.4

13. Other Revenues 245.4 949.3

13.1 Fines and surcharges 110.7 152.9

Surcharges 32.5 51.2

Liquor fines 42.2 32.1

Tax fines 6.2 6.1

Other fines 29.8 63.4

13.2 Returned funds 39.4 603.2

13.3 Miscellaneous 99.2 193.2

Interest 5.2 23.7

Mint receipts 47.8 61.0

Other 46.2 108.5

TOTAL 8,807.3 18,525.2
Sources: Compiled from the following publications: Department

of Revenue: (1) Annual Report, 1969, (2) Corporate
Tax Statistics, 1963 and 1968, (3) Business Taxes
Statistics, 1962 and 1969: Budget Bureau: (1)

Annual Budget, 1965 and 1971, (2) Government Receipts,
1963 and 1969; Department of Comptroller General,
Report of Receipts and Outlays of the Kingdom of
Thailand, 1963 and 1969.

 

 

 

 

 

 

 

 

 

Note: Disaggregated figures for corporate and business
taxes are adjusted from the only partially complete

data available from the Department of Revenue.

a .
Details are too fragmented to be presented here.

APPENDIX F

DISAGGREGATED PUBLIC EXPENDITURE DATA

APPENDIX F

DISAGGREGATED PUBLIC EXPENDITURE DATA

Table F.1--Disaggregated Public Expenditures, 1963 and 1969
(Amounts in Millions of Baht)

 

 

 

 

 

 

10C

 

Year
Expenditure Categories
1963 1969
1. Economic Services 2,560.1 5,962.4
1.1 Agriculture 776.7 2,222.6
General irrigation services 64.9 234.2
Canals and maintenance 88.5 257.7
State irrigation
construction 63.7 664.0
General irrigation
construction 55.3 245.6
Chao Phya DevelOpment Scheme 65.1 --
Big dam construction 213.6 152.2
OUS: M. Agricultural 6.5 35.2
D. Agricultural 41.3 78.6
D. Rice 33.9 46.0
D. Fishery 26.6 66.2
D. Livestock 36.9 74.4
D. Forestry 72 6 184.7
D. Agricultural Extension -- 80.8
Rubber Organization 7.7 8.0
Rubber Growing Support Fund -- 15.0
Farmers' Assistance Fund -- 80.0
Power and Fuels 271 8 286.1
Atomic Power for Peace
Office 3.0 6.3
National Power Commission 29.7 103.5
Provincial Electricity
Authority 49.6 45.1
Northeast Electricity
Authority - 22.5
Lignite Authority 36.6 11.3
Yankee Electricity Authority 67.8 --
Refinery I (Fang) 19.1 11.7
Refinery II (Bangchak) 3.0 85.6
Metropolitan Electricity
Authority 63.0 --

Table F.1--continued

307

 

 

Expenditure Categories

 

 

1.3

1.4

1.5 Other Economic Services

Industries

OUS: M.

D. Industrial Promotion

Industries

D. Industrial Works

Frozen Food Organization
Tannery Organization
Textile Organization

Glass Organization
Thai Paper Factory

D. Mineral Resources

Board of Investment
Jute Mills

Tin Buffer Stock
Sugar Office

Battery Organization
Preserved Food Organization

Ceramic Factory
Thai Rubber Company

Sugar Industry Company

cation

OUS: M.

Transportation and Communi-

Communication

D. Land Transport
D. Civil Aviation
D. Harbor

D. Post and Telegraph

D. Highways

Port Authority
State Railways
Telephone Authority

Express Transport Organi—
zation

Thai Shipping Company

Thai Television Company

Bangkok Dock Company

Thai Airways

OUS: M. National Develop—

ment

 

Year
1963 1969
179.3 173.1
2.4 7.2
3.4 17.3
1.3 14.0
7.0 6.9
0.6 --
41.2 --
7.9 --
0.5 --
38.2 73.3
-- 3.1
-- 12.0
-- 37.5
-- 1 5
2.5 --
41.0 --
0.3 --
3.0 --
30.0 --

 

26.1 24.3
88.7 19.2
-- 198.8
6.7 36.5
95.0 59.0
758.6 2,550.9
24.6 21.9
32.3 46.4
34.0 --
9.0 --
2.6 --
1.0 --
-- 2.6
0.7 --
253.1 320.8

 

308

Table F.1-—continued

 

 

Year

 

Expenditure Categories

D. Soil Development --
D. Credit and Marketing
Co-operatives 6
D. Co—operative Auditing 5.
Board of Export Promotion 0
Tourist Organization 6
Fish Marketing Organization --
D. Land Co-operatives 50.0 6

Bank of Agricultural and

 

 

 

 

 

Agricultural Co-operatives -- 140 0
National Economic Promotion
Company 146.0 --
Union Hotel Company 13.6 --
2. Educational Services 1,777.5 3,674.3
2.1 Primary Education 1,024.7 1,997.9
D. Primary Education 1,024.7 208.6
Grants to Provincial
Administrations -- 1,789.3
2.2 Secondary Education 234.9 405.0
D. Secondary Education 234.9 405.0
2.3 Higher Education 193.8 559.6
Chulalongkorn 42.8 108.0
Thammasat 14.2 25.8
Kasetsant 18.7 66.1
Mahidol 112.8 170.4
Chiengmai -- 71.0
Silpakorn 5.4 23.0
Khonkaen -- 29.0
NIDA -- 15.4
Songkhla -- 51.0
2.4 Vocational Education and Other 324.0 711.6
D. Vocational Education 115.1 406.7
D. Physical Education 21.8 12.5
OUS: M. Education 76.2 68.5

309

Table F.1--continued

 

 

Expenditure Categories

Year

 

1969

 

3.

4.

Royal Institute

National Education Council

D. Teacher Training '

D. Educational Techniques

D. Fine Arts

National Research Council

National Sport Promotion
Organization

Health and Social Welfare
3.1 Public Health

OUS: M. Public Health
D. Medicine

D. Health

D. Medical Science

3.2 Public Welfare

D. Public Welfare

D. Religious Affairs
D. Labor

Government Pawn ShOps
Government Pensions

3.3 Other Social Services

D. Public Works

MetrOpolitan Water Work

D. Community Development

D. Science

Zoo Organization

Town Planning Office

Accelerated Rural DeveloP-
ment Office

General Services
4.1 Defense

M. Defense
National Security Council

com-heave
combat-alum

HM

929.3 2,338.0

 

 

 

 

305.3 619.2
24.1 27.7
153.4 317.0
120.1 266.5
7.6 8.1
443.6 747.8
73.3 170.8
15.4 24.8
'-- 16.8
8.0 --
347.0 535.3
180.4 971.0
143.8 432.3
-- 165.4
25.1 84.4
6.9 11.1
1.0 1.1
3.7 12.1
-- 264.6

 

1,637.6 3,718.9

 

1,637.2 3,716.8

0.4

2.1

310

Table F.1--continued

 

 

 

 

 

 

Year
Expenditure Categories
1963 1969
4.2 Law Enforcement 797.9 1,558.4
M. Justice 66.3 82.2
D. Correction 95.8 153.0
D. Public Prosecution 24.4 33.6
D. Police 611.3 1,289.6
Public Administration 1,414.6 3,063.8
OUS: Prime Minister's
Office 3.8 5.4
Office of the Prime Minister 44.4 28.7
Budget Bureau 3.2 10.8
Office of the Cabinet

Secretariat 3.5 5.4
Tax Supervision Office 5.5 10.2

National Economic Develop-
ment Board 116.2 7.8
Office of Judicial Council 2.0 2.8
State Service Commission 25.4 44.7
Audit Office 10.9 16.4
National Statistical Office 10.7 32.9
D. Public Relations 35.9 47.6
D. Central Intelligence 7.9 16.5
D. Meteorology 6.3 18.1
OUS: M. Economic Affairs 7.6 16.0
D. Internal Trade 2.8 6.0
D. Foreign Trade 3.2 5.0
D. Business Intelligence 2.5 5.4
D. Trade Registration 3.7 5.3
D. Economic Relations 13.5 9.7
OUS: M. Interior 46.5 47.4
D. Land 42.3 85.8
D. Local Government 429.6 917.6
OUS: M. Finance 47.9 340.0
D. Treasury . 57.7 54.8
D. Comptroller-General 22.7 33.9
D. Customs 38.6 69.5
D. Excise 49.1 63.6
D. Revenue 62.2 103.9
M. Foreign Affairs 86.9 139.3

D. Technical and Economic
Co-operation 53.9 445.9
Royal Household Office 26.9 50.1

311

Table F.1--continued

 

 

 

 

 

 

 

Year
Expenditure Categories
1963 1969

Royal Privy Council 0.6 1.2
Parliament Secretariat 14.6 24.9
Postal Charges 3.5 8.0
Foreign Training 3.0 0.5
Foreign Conferences 14.4 17.3

Government Official
Assistance Fund 101.8 297.2
Phones and Television Charges 3. 11.0
Office of Fiscal Policy -- 57.3
4.4 Other 172.1 795.1
Reserve Fund 172.1 762.7
Counterpart Fund -- 32.5
5. Payments on Public Debt 1,139.4 1,245.5
5.1 Repayments 763.7 217.6
5.2 Interest and Management Fees 375.8 1,027.9
TOTAL 10,502.5 22,904.3
Sources: Compiled from the following publications: National

Note:

Statistical Office,
1965 and 1971;

Statistical Yearbook of Thailand,

 

Department of Comptroller General,

Report of Receipts and Outlays of the Kingdom of

 

Thailand, 1963 and 1969.

 

OU
M
D

S

= Office of the Under-Secretary
Ministry
Department

APPENDIX G

THE LORENZ FUNCTIONS OF VARIOUS

DISTRIBUTIONS OF INCOME

APPENDIX G

THE LORENZ FUNCTIONS OF VARIOUS

DISTRIBUTIONS OF INCOME

As noted in Chapter 2, a part of the main procedure
in this study was to estimate the Lorenz (curve) functions
that would fit the pre-fisc, post-tax, post-benefit, post-fisc,
and a few other variations of income distributions for 1963
and 1969 so that the exact Gini concentration ratios for
both years could be estimated. This method was adopted
from N. P. Kakwani and N. Podder as presented in "Efficiency
Estimation of the Lorenz Curve and Associated Inequality
Measures from Grouped Observations," World Bank Development
Research Center Discussion Paper No. 10, October 1974
(mimeographed). The function of the desired Lorenz curve

is a double-log regression equation of the following form:

log [2:3] = a' +alog[m] +Blog [/2-m]+u,
.5 5 .5

where p and q are the cumulative prOportions of households
and income by income classes, respectively: a',a , and g are
the parameters to be estimated; and u is the stochastic
disturbance term. The ordinary least—squares method is used
to obtain various Lorenz functions. The parameter values and
the resulting Gini coefficients (G) are presented in Tables

6.1 and G.2.

312

313

Table G.l--Lorenz Functions of various Distributions of Income, 1963

 

 

 

Distributions . —2
of Income a a B R’ G

Pre—Fisc: Money -.2238 1.0281 '.9817 1.0000 .5627
(-57.95) (58.16) (477.71)

PreéFisc: Adjusted -.2692 1.2677 .9793 1.0000 .4559
(-38.30) (34.60) (250.86)

Post—Tax: Money -.2021 .9538 .9842 1.0000 .6102
(-51.98) (56.70) (484.73)

Post-Tax: Adjusted -.2650 1.2654 .9803 1.0000 .4605
(-37.90) (34.87) (252.95)

Post-Benefit: Money -.2359 1.0888 .9809 1.0000 .5304
(-50.00) (48.55) (385.79)

Post-Benefit: Adjusted -.2709 1.2747 .9792 1.0000 .4529
(-37.61) (33.83) (244.45)

PostéFisc: Money -.2218 1.0465 .9828 1.0000 .5581
(Assumptions I & II) (-47.90) (49.07) (398.40)

PostéFisc: Adjusted —.2674 1.2734 .9800 1.0000 .4566
(Assumptions I & II) (-37.15) (33.95) (245.31)

PostrFisc: Money -.2092 1.0110 .9846 1.0000 .5838
(Assumption III) (-48.86) (52.80) (436.42)

Post-Fisc: Adjusted -.2610 1.2579 .9810 1.0000 .4660
(Assumption III) (-37.96) (35.47) (258.09)

 

 

 

 

 

 

Note:

Figures in parentheses are t—values.

 

314

Table G.2--Iorenz FUnctions of Various Distributions of Income, 1969

 

 

 

Distributions . —2
of Income a a B R G

Pre-Fisc: Money -.2400 .9775 .9760 1.0000 .5550
(-125.69) (836.38) (849.34)

Pre-Fisc: Adjusted -.3082 .9616 .9581 (.9998 .4822
(-28.90) (147.62) (148.49)

Post-Tax: Money -.2147 .9834 .9825 1.0000 .5847
(-58.41) (436.66) (445.03)

Post—Tax: Adjusted -.3085 .9616 .9579 .9998 .4819
(-26.99) (137.74) (138.54)

Post-Benefit: aney -.2617 1.0156 .9720 1.0000 .5151
(-77.33) (102.27) (552.29)

Post-Benefit: Adjusted -.2972 1.0456 .9644 .9998 .4738
(-26.92) (31.09) (165.17)

PosteFisc: Money -.2536 1.0144 .9740 1.0000 .5289
(Assumptions I & II) (~72.11) (98.97) (534.58)

Post-Fisc: Adjusted -.2977 1.0487 .9644 .9998 .4725

(Assumptions I & II)

PostéFisc: Mbney -.2328 .9860 .9780 1.0000 .5614
(Assumption III) (-92.95) (138.99) (762.41)

PostrFisc: Adjusted -.2881 1.0368 .9663 .9999 .4853
(Assumption III) (-24.55) (29.28) (156.30)

 

 

 

 

 

 

Note:

Figures in parentheses are t-values.

 

APPENDIX H

THE NET FISCAL INCIDENCE BASED ON

AARON-MCGUIRE-MAITAL METHOD

 

APPENDIX H

THE NET FISCAL INCIDENCE BASED ON

AARON-MCGUIRE-MAITAL METHOD

Dissatisfied with the arbitrary ways in which public
expenditures were allocated among income classes in most
fiscal incidence studies, Henry Aaron and Martin McGuire
prOposed a theory indicating the proper way in which such
public good expenditures should be allocated.1 Essentially,
they contend that the imputed benefits of public goods
provided by the public sector accruing to each family should
be determined by some value of private good benefits
received by that particular household. A brief description
of the Aaron-McGuire theory is given below.

In a neoclassical framework of efficient consumption,
Aaron and McGuire assume that a utility function for each
family is known or can be determined and is a function of the

value of private goods and pure public goods in the form
vi = f[(Yi + yi) p]

D S I I
where U1 is the utility of family i: (Y; + Y;) is the summa-
tion of disposable income of and the benefits from specific
good expenditures received by family i (together, it is the
estimated value of private goods consumed by family 1 alone):
and P is the amount of public goods. A public good, P, has all

the characteristics well known in the literature, for example, it

315

316

is consumed in equal amounts byall, consumption by one has no
effect on consumption by others, and no one can be excluded
from enjoying it. Aaron and McGuire also assume that the
utilities of public goods and other goods are independent.

For any two families a and b, the optimal or efficient
condition for consumption is attained when the marginal rate
of substitution of total private goods and public goods of
family a is equal to the similar marginal rate of substitution

of family b, that is, when

 

a a a b

MRSa = dY /dP = fp/f = EX
MRSb dYb/dP fb/fb fa
P Y Y

where f2, the first partial derivative of Ua with respect to P,

is assumed equal maf:,or when MRSaf; = MRSb b = . . . =

Y
MRSif; = K is a constant.

Then assuming that each family's marginal rate of
substitution between public goods and other goods is known
and that the total and marginal cost of public and specific
goods nsknown for all relevant outputs of these goods, for a

bundle of public goods to be efficiently shared by all

families, it is required that

 

XMRSl

where Yp is the total public good benefits (which are equal

to the actual public good expenditures), and Y; is the imputed

317

public good benfits accruing unfamily i. As only the ratios

of MRS's matter, ZMRSl can be set equal to 1 so that

Y1 = Y MRSl = Y —E— = Y —i— ,
p p E> f1 E> f1
y Y

What this equation has shown is that "to each household
should be imputed a function of the total value of the public
good proportional to the reciprocal of its marginal utility
of private good expenditures."2
Aaron and McGuire then specify f; on the basis of dis-

posable cash income (or the original income net of tax

burdens) plus the estimated value of specific goods received,

or Y: + Y; = Y1. Then two utility functions are assumed
U1 = A log Y1 + B (1)
u1 = E - J%7 (2)
Y

where A, B, C, and D are arbitrary constants. Differentiating
these utility functions with respect to Y1, equations (1) and

(2) become

dUl/le

n
H1
k: w-

? A/Y ; (3)

P.

dUi/in C/Y. . (4)

ll
H1
L<
I

318

If utility function (2) is used,the imputed public good

benfits to family i would be

that is, the public good expenditures should be allocated

to family i according to its proportionate share in

the total money income net of tax plus the estimated value of

specific good benefits. The above condition implies that the

higher the income position of a family, the higher the public

good benefits it will receive, other things being equal.
Aaron and McGuire believe that the marginal utility of

income (for the United States) lies between -1 and -2.

Shlomo Maital in a subsequent study further claimed that

from at least three independent studies on the marginal

utility of income there is valid reason to believe that the

value of the "true" marginal utility is -1.5.3
So, in order to determine the net fiscal inci-

dence in Thailand using time Aaron—McGuire-Maital (A-M-M)

method, the 1969 fiscal activities based on money

income distribution are selected as an example.

First, the relevant tax burdens are deducted from each income

class to arrive at the distribution of disposable income in

the Aaron—McGuire sense. Then, the benefits of specific

expenditures by income classes (rearranged from Chapter 4 by

exchanging such public goods as defense, public administration,

and other similar categories) are added to disposable

319

income to arrive at the total private good income distribution,
which will be the base for public good benefit allocation.

The marginal utility of -1.5 claimed to be true by
Maital for the United States is also assumed to be true for
Thailand. This value is used to find the share of each
income class in the total distribution of private good income,
after which the total public good expenditures are allocated
according to the pattern of the shares obtained. The results
of this allocation and the new post-fisc income distribution
using the A-M-M method are presented in Table H.1 below.

It can be seen from Table H.1 that the resulting
post-fisc income distribution based on the A-M—M method has
become much more unequal than the same distribution based on
Assumption III in this study, which is regarded here as the
most de-equalizing assumption. The lowest income class
in the distribution under the A-M-M approach shares only
2.9 percent of the total money income after the effects of
government fiscal activities are taken into account, whereas
under Assumption III and Assumption I of this study the same
income class receives 3.7 percent and 4.9 percent of the
total income, respectively.

At the same time the highest income group, under the
A-M-M assumption, always receives the highest income share.
The Gini concentration ratios for the post-fisc income distri-
bution are estimated at .6041, .5614, and .5289 under the A-M—M

method, Assumption III, and Assumption I, respectively.

320

 

 

 

 

 

$3. 38. $8. ufiflammmoo go
o.ooa o.oeH o.oeH m.mmv.mo e.mmm.HH m.mmo.em a.mma.me mmmmmHo Ham
a.am m.mm a.am m.mem.mm m.mmm.m m.mee.om m.moa.m~ um>o new ooo.ma
m.m m.m m.m o.~e~.e a.mwm o.m>m.m e.m~a.m mam.n~uooo.ma
e.e o.m e.e m.eeo.m «.mev o.~em.v u.eom.m mem.eauooo.ma
m.a m.m m.m m.mmm.m a.mom m.m¢~.m m.mo~.e mme.aaueeo.m
e.ea v.oa e.a o.e¢m.m m.ma» o.mmm.m m.ame.v mmm.m ueoe.m
m.m m.m m.e m.ame.m m.aem m.amm.v ~.Amm.m mmm.m loco.m
m.v m.m m.m m.aflo.~ m.mma m.mmm.a m.~m~.H ooo.m gene:
A a w zrzua
H .udmma. HHH .ammma muammewm meooaH msooeH an v
@006 eooo gem
aoafiflbfle 83:3me 332a 393a 28888 888 889m
9605 omflhlumom 9.005 omwmuumonfi

 

 

 

 

 

 

 

Beam no 8932 5 88:: 143m men. 5 8,5 805% one 693m:

HSHmzumuHmUzlcng so @093 9505 m0 20343330 omflmlumom on» no Smwuazoolai 3an

321

This indicates that the A-M-M method, if used, would cause
at least a 7.6 percent increase in income inequality in
Thailand under the money income based in 1969 over the most
de-equalizing assumption used in this study.

The use of this A-M-M method would provide an ideal
case in support of the contention that the government of
Thailand has effected very little change in income distribu-
tion. Indeed, instead of improving the present distribution,
the public sector actually worsens it. However, the A-M-M
method is not used as the main method here because it is
believed to be inconsistent or incongruous with the approach
adopted in this study regarding the allocation of public good
benefits (which is to examine the disaggregated details of
public good expenditures and then allocate the benefits to
either general or specific beneficiaries according to the
chosen objective or subjective criteria). Oddly enough, the
present demonstration does not refute the main findings in
this study; rather, it adds more weight to those findings,
namely, that the Thai government is not merely indifferent

to the issue of income redistribution, but is Opposed to it.

 

lHenry Aaron and Martin McGuire, "Public Goods and
Income Distribution," Econometrica,38 (November, 1920),
pp. 907-20.

2

 

Ibid., p. 911.

3Shlomo Maital, "Public Goods and Income Distribution,"
Econometrica, 41 (May, 1973), pp. 561-68. '

 

SELECTED BIBLIOGRAPHY

SELECTED BI BLIOGRAPHY

Aaron, Henry and McGuire, Martin. "Public Goods and Income
Distribution." Econometrica, 38 (November, 1970),
pp. 907-920.

 

Adelman, Irma and Morris, Cynthia T. "An Anatomy of Patterns
of Income Distribution in Developing Nations." U.S.
Department of State, Agency for International Develop-
ment (AID), 1971, (Mimeographed).

. Economic Growth and Social Equity in Developing
Countries. Stanford: Stanford UniversIty Press, 1973.

 

Adler, John H. "Development and Income Distribution."
Weltwirtschaftliche Archiv, Bond 108, Heft 3 (1972),
pp. 329-343.

 

. "The Fiscal System, the Distribution of Income,
and Public Welfare." In Kengon E. Poole, ed.

Fiscal Policies and the American Economy. Prentice-
Hall, 1951. Pp- 359-421.

 

Ahluwalia, Montek S. "Income Inequality: Some Dimensions
of the Problem." Finance and Development, 11
(September, 1974), pp. 2-9.

 

Aitchison, J. and Brown, J. A. C. The Lognormal Distribution.
Cambridge: Cambridge University Press, 1957.

 

Arrow, Kenneth J. Social Choice and Individual Values.
2nd ed. New York: Wiley, 1963.1

 

Atkinson, A. B. "On the Measurement of Inequality."
Journal of Economic Theory, 2 (1970), pp. 244-263.

 

., ed. Wealth, Income, and Inequality. Hammonds-
worth: Penguin Books, Inc., 1973.

 

 

Ayal, Eliezer B. "Public Policies in Thailand under the
Constitutional Regime: A Case Study of an Under-
developed Country." Ph.D. Dissertation, Cornell
University, 1962.

Each, George, L. and Ardo, Albert. "The Redistributional
Effects of Inflation." Review of Economics and
Statistics, 39 (February, 1957), pp. 1-13.

 

 

322

323

Barna, Tibor. Redistribution of Incomes through Public
Finance in 1937. Oxford: Clarendon Press, 1945.

 

 

Behrens, Jean and Smolensky, Eugene. "Alternative Defini-
tions of Income Redistribution." Public Finance,
28 (1973). PP. 315-332.

 

Bertrand, Trent J. "Rural Taxation in Thailand." Pacific
Affairs, 42 (Summer, 1969), pp. 178-188.

Bhagwati, Jagdish N., ed. Economics and World Order.
London: MacMillan, 1972.

 

Bird, Richard M. "Income Distribution and Tax Policy in
Columbia." Economic Development and Cultural Change,
18 (1970): pp. 519-535.

. Taxation and Development: Lessons from Columbian
Experience. Cambridge: Harvard UnIVersIty Press,
1970.

 

 

and DeWulf, Luc. "Taxation and Income Distribution
in Latin America: A Critical Review of Empirical
Studies." International Monetary Fund Staff Papers,
23 (November, 1973), pp. 639-681.

 

Bonnen, James T. "The Absence of Knowledge of Distribution
Impacts: An Obstacle to Effective Public Program
Analysis and Decisions." In U.S., Congress, Joint
Economic Committee. The Analysis and Evaluation of
Public Expenditures: The PPB System. Washington,
D.C.: U.S. Government Printing Office, 1969,
pp. 419-449.

. "The Distribution of Benefits from Cotton Price
Supports." In Samuel B. Chase, Jr., ed. Problems in
Public Expenditure Analysis. Washington, D. C.:
Brookings Institution, 1968, pp. 223-254.

 

 

 

Boulding, Kenneth E. and Pfaff, Martin, eds. Redistribution
39 the Rich and the Poor. Belmont, Calif.: Wadsworth,
1972.

 

Breit, William. "Income Redistribution and Efficiency Norms.‘
In Harold M. Hochman and George E. Peterson, eds.
Redistribution Through Public Choice. New York:
Columbia University Press, 1974, pp. 3-21.

Bronfenbrenner, Martin. "Equality and Equity." The Annals
of the American Academy of Political and Social
Science, 409 (September, 1973). PP. 9-23.

324

. Income Distribution Theory. Chicago: Aldine,
1971.

 

 

Brownlee, O. H. and Conrad, Alfred. "Effects upon the
Distribution of Income of a Tight Money Policy."
American Economic Review, 51 (May, 1961), pp. 74-85.

 

Cartter, Allan M. The Redistribution of Income in Postwar
Britain: A Study of the Effects of the Central
Government Fiscal Program in 1948-1949. New Haven:
Yale University Press, 1955.

 

 

 

Champernowne, D. G. "A Comparison of Measures of Inequality
of Income Distribution." Economic Journal, 84
(December, 1974), pp. 787-816.

 

Chenery, H.B. e: 31. Redistribution with Growth. London:
Oxford University Press, published for the World
Bank and the Institute of Development Studies,
University of Sussex, 1974.

 

Clark, Colin and Struwel, Geer, eds. Income and Wealth,
Series X, Income Redistribution and the Statistical
Foundation of Economic Policy. New Haven: Interna-
tional Association for Research in Income and Wealth,
1964.

 

 

 

Cline, William R. "Distribution and Development." Journal
of Development Economics, 1 (February, 1975),
pp. 359—400.

 

. Potential Effects of Income Redistribution on
Economic Growth: Latin American Cases. New York:
Praeger, 1972.

 

 

 

Dalton, Hugh. Some Aspects of the Inequalitypof Incomes
in Modern Communities. London: George Routledge
and Sons, Ltd., 1920.

 

 

Development Economics Department, Public and Private Finance
Division, World Bank. "Distributive Impact of Public
Expenditures: A Research Proposal." September, 1973
(Mimeographed).

 

DeWulf, Luc. "Do Public Expenditures Reduce Inequality?"
Finance and Development, 11 (September, 1974),
pp. 20—23.

. "Fiscal Incidence Studies in Developing Countries:
Survey and Critique." International Monetary Fund
Staff Papers, 22 (March, 1975), pp. 61-131.

 

 

 

325

Dodge, David A. "Impact of Tax, Transfer, and Expenditure
Policies of Government on the Distribution of Personal
Income in Canada." The Review of Income and Wealth,

 

21 (March, 1975), pp. 1-52.

Eckstein, Otto. Water Resource Development. Cambridge:
Harvard University Press, 1958.

 

Edwards, Richard C.: Reich, Michael; and Weisskopf, Thomas E.
The Capitalist System: A Radical Analysis of American
Society. Englewood, N.J.: Prentice-Hall, 1972.

Franzen, Thomas; Lougren, Kerstin: and Rosenberg, Irma.
"Redistributional Effects of Taxes and Public Expen-
ditures in Sweden." The Swedish Journal of Economics,
77 (1975), pp. 31-55.

 

Friedman, Milton. Capital and Freedom. Chicago: University
of Chicago Press, 1962.

 

. "Choice, Chance, and the Personal Distribution
of Income." Journal of Political Economy, 61 (August,
1953), PP. 277-290.

 

 

Galbraith, John Kenneth. The New Industrial State. 2nd ed.
Boston: Houghton-Mifflin, 1971.

 

. "Power and the Useful Economist." American
Economic Review, 63 (March, 1973), pp. 1-11.

 

 

 

 

Gaswirth, Joseph L. "The Estimation of the Lorenz Curve
and Gini Index." Review of Economics and Statistics,
54 ( , 1972), PP. 306-316.

Gillespie, W. Irwin. The Incidence of Taxes and Public
Expenditures in the Canadian Economy. Studies of
the Royal Commission on Taxation, Number 2. Ottawa:
Queen's Printer, 1966.

 

 

. "Effect of Public Expenditure on the Distribution
of Income." In Richard A. Masgrave, ed. Essays in
Fiscal Federalism. Washington, D. C.: The Brookings
Institution, 1965, pp. 122-186.

 

 

 

Gordon, David M. "Taxation of the Poor and the Normative
Theory of Tax Incidence." American Economic Review,
62 (May, 1972): pp. 319-328.

 

Greene, Kenneth V. "Collective Decision-Making Models and
the Measurement of Benefits in Fiscal Incidence
Studies." National Tax Journal, 26 (June, 1973),
pp. 177-185.

 

326

Haig, Robert M., ed. The Federal Income Tax. New York:
Columbia University Press, 1921.

 

Hansen, W. Lee and Weisbrod, Burton A. "Distributional
Effects of Public Expenditure Programs." Public
Finance, 27 (1972), pp. 414-420.

. "Who Pays for a Public Expenditure Program?"
National Tax Journal, 24 (December, 1971): pp. 515-517.

 

 

Haq, Mahbub ul. "Employment in the 19705: A New Perspective."

A Statement Prepared for the Society for International
Development 12th World Conference, Geneva, 16-19 May,
1971 (Mimeographed).

Harberger, Arnold C. "The Incidence of the Corporate
Income Tax." Journal of Political Economy, 70 (June,
1962), PP. 215-240.

 

. "Fiscal Policy and Income Distribution." Paper
prepared for the WorkshOp on Income Distribution in
Less Developed Countries, Princeton University,
June 12-13, 1974.

 

Harcourt, George C. Some Cambridge Controversies in the
Theory of Capital. Cambridge: Cambridge University
Press, 1972.

 

 

Hirschman, Albert 0. "The Changing Tolerance for Income
Inequality in the Course of Economic Development."
Quarterly Journal of Economics, 87 (November, 1973),
pp. 544-566. '

 

Hochman, Harold M. and Peterson, George E. Redistribution
through Public Choice. New York: Columbia Univer-
sity Press, 1974.

 

 

Hunt, Shane. "Distribution, Growth, and Government Economic
Behavior in Peru." In Gustav Ranis. ed. Government
and Economic Development. New Haven: Yale University
Press, 1971, pp. 374-416.

 

 

Ingram, James C. Economic Change in Thailand, 1850-1970.
2nd ed. Stanford: Stanford University Press, 1971.

 

Johnson, Harry G. "Some Micro-Economic Reflections on
Income and Wealth Inequalities." The Annals of the
American Academy of Political and Social Science, 109
(September, 1973), pp. 53-60.

 

 

Johnson, James A. The Incidence of Government Revenues and
Expenditures. A Study Prepared for the Ontario
Committee on Taxation. Frank Fogg, Queen's Printer,
n.d.

 

 

327

Kuznets, Simon. "Economic Growth and Income Inequality."
American Economic Review, 45 (March, 1955). PP- 1-28.

 

. "Quantitative Aspects of the Economic Growth of
Nations: VIII, Distribution of Income by Size."
Economic Development and Cultural Change, 11 (January,
1963), pp. 1-80.

 

 

Lewchalermwong, Anan. Taxation and Tax Reform in Thailand.
Bangkok: Kurusapha Ladprao Press, 1971.

 

Lydall, H. The Structure of Earnings. Oxford: Clarendon
Press, 1968.

 

McCleary, William A. "Sources of Change in Distribution of
Income in Thailand, 1962/63 and 1968/69." Discussion
Paper no. 26, Faculty of Economics, Thammasat
University, Bangkok, 1972 (Mimeographed).

McLure, Charles E., Jr. "The Incidence of Taxation in
West Malaysia." The Malayan Economic Review, 17
(October, 1972), pp. 66-98.

 

"The Distribution of Income and Tax Incidence
in Panama, 1969." Discussion Paper no. 36. Program
of Development Studies, Rice University, 1972
(Mimeographed).

 

. "The Incidence of Colombian Taxes, 1970."
Discussion Paper no. 41, Program of Development
Studies, Rice University, Summer, 1973.(Mimeographed).

 

. "On the Theory and Methodology of Estimating
Benefit and Expenditure Incidence." Paper presented
at the WorkshOp on Income Distribution and Its Role
in Development, Rice University, April 26, 1974
(Mimeographed).

 

Maital, Shlomo. "Public Goods and Income Distribution: Some
Further Results." Econometrica, 41 (May, 1973),
pp. 561-568.

 

Mann, Arthur J. "The Income Redistributional Effects of the
Peurto Rican Fiscal System." Ph.D. Dissertation,
University of Florida, 1971.

Mathew, T. "Shifting and Incidence of Corporate Income Tax:
A Review." The Indian Economic Journal, 20 (October-
December, 1972), pp. 257-274.

 

328

Kakwani, N. C. and Podder, N. "On the Estimation of Lorenz
Curves from Grouped Observations." International
Economic Review, 14 (June, 1973): pp. 278-291.

 

 

. "Efficient Estimation of the Lorenz Curve and
Associated Inequality Measures from Grouped Observa-
tions." World Bank Development Research Center
Discussion Papers, no. 10, October, 1974 (Mimeographed).

 

Kambhu, Xujati. "Water Resources DevelOpment in Thailand
and Its Policy." Ministry of National Development,
Bangkok, December 27, 1963 (Mimeoqraphed).

Karageorgas, D. "The Distribution of Tax Burden by Income
Groups in Greece." Economic Journal, 73 (June, 1973),
pp. 436-448.

 

Kerdphibul, Udom. "Income and the Distribution of Income
of the Agricultural Sector." Thai Journal of Agricul-
tural Economics, 11 (December, 1972), pp. 1-24.

 

Klein, Lawrence R. An Introduction to Econometrics.
Englewood Cliffs, N. J.: Prentice-Hall, 1962.

Kravis, Irving B. "International Differences in the Distri-
bution of Income." Review of Economics and Statistics,
42 (November, 1960), pp. 408-416.

. "A World of Unequal Incomes." The Annals of the
American Academy of Political and Social Science, 409
(September, 1973), pp. 61-80.

 

Krongkaew, Medhi and Chernsiri, Chintana. "The Measurement
of the Poverty Level in Thailand." Thammasat
University Journal, 5 (June-September, 1975).

 

 

Krutilla, John V. and Eckstein, Otto. Multiple Purpose
River Development: Studies in Applied Economic
Analysis. Baltimore: Resources for the Future, Inc.,
1958. '

 

 

Krzyzaniak, Marian, ed. Effects of the Corporate Income Tax.
Detroit: Wayne State University Press, 1966.

and Musgrave, Richard A} The Shifting of the
Corporate Income Tax. Baltimore: John HOpkins,
1963.

 

 

and Ozmucur, Sfileyman. "The Distribution of Income
and the Short-run Burden of Taxes in Turkey, 1968."
Finanzarchiv, 32 (1973), pp. 69—97.

 

 

329

Meerman, Jacob P. "Fiscal Incidence in Empirical Studies of
Income Distribution in Poor Countries." Department of
State, Agency for International Development Discussion
Paper no. 25, December, 1972 (Mimeographed).

Meesook, Oey Astra. "Income Distribution Statistics for
Thailand, Malaysia, and Indonesia." Paper prepared
for the workshop on Income Distribution in Less
Developed Countries, Princeton University, June 12-13,

1974.
. "The Impact of Price Increases on Different Income
Groups." Faculty of Economics, Thammasat University,

August, 1975 (Mimeographed).

. Income Distribution in Thailand. Bangkok:
Faculty of Economics, Thammasat University, 1975.

 

Mieszkowski, Peter. "Tax Incidence Theory: The Effects of

 

 

Taxes on the Distribution of Income." The Journal
of Economic Literature, 7 (December, 1969), pp. 1103-
1124.

Miliband, Ralph. The State in Capitalist Sociepy. London:
Wiedenfeld and Nicholson, 1969.

Musgrave, Richard A. The Theory of Public Finance. New
York: McGraw-Hill, 1959.

. "Estimating the Distribution of the Tax Burden."
In Problems of Tax Administration in Latin America.
Papers and proceedings in a conference held in
Buesnos Aires, Argentina, October, 1961. Baltimore:
John Hopkins, 1965.

 

., ed. Essays on Fiscal Federalism. Washington,
D. C.: The Brookings Institution, 1965.

 

 

.; Carrol, J. J.: Cook, L. D.; and Frame, L.
"Distribution of Tax Payments by Income Class: A
Case Study for 1948." National Tax Journal, 1 (March,
1951). PP. 1-48.

., and Daicoff, Darwin, W. "Who Pays Michigan Taxes?"
Michigan Tax Study Staff Papers, (1958), pp. 131-182.

., and Musgrave, Peggy B. Public Finance in Theory
and Practice. New York: McGraw-Hill, 1973.

 

Nell, Edward J. "The Fall of the House of Efficiency." The
Annals of the American Academy of Political and Soc1a1

Science, 409 (September, 1973): PP. 102-111.

330

O'Connor, James. The Fiscal Crisis of the State. New York:
St. Martin's Press, 1973.

 

Oshima, Harry T. "The International Comparison of Size
Distribution of Family Incomes with Special Reference
to Asia." Review of Economics and Statistics, 44

 

(November, 1962), pp. 439-445.

. "Income Inequality and Economic Growth: The
Postwar Experience of Asian Countries." The Malayan
Economic Review, 15 (October, 1970), PP. 7-41.

 

 

Paukert, Felix. "Income Distribution at Different Levels
of Development: A Survey of Evidence." International

 

Labour Review, 108 (August-September, 1973), pp. 97-
125.

 

Peacock, A. T., ed. Income Redistribution and Social Polipy:

 

A Set of Studies. London: Cape, 1954.

 

Pechman, Joseph A. "The Distributional Effects of Public
Higher Education in California." Journal of Human
Resources, 5 (Summer, 1970), pp. 361-370.

 

 

., and Okner, Benjamin A. Who Bears the Tax Burden?
Washington, D. C.: The Brookings Institution, 1974.

 

Pen, Jan. Income Distribution: Facts, Theories, and
Policies. New York: Praeger, 1971.

 

 

Phelps, Edmond 8., ed. Economic Justice: Selected Readings.

 

Harmondsworth, Middlesex: Penguin Books, Inc., 1973.

Poole, Kenyon E., ed. Fiscal Policies and the American
Economy. New York: Prentice-Hall, 1951.

 

Prest, Alan R. "Statistical Calculations of Tax Burdens."
Economica, 87 (August, 1955), pp. 234-245.

 

. "The Budget and Interpersonal Distribution."
Public Finance, 23 (1968). PP. 80-98.

 

 

Quandt, Richard E. "Old and New Methods of Estimating the
Pareto Distribution." Metrika, 10 (1966), pp. 55-82.

Rawls, John. A Theory of Justice. Combridge: Harvard
University Press, 1971.

 

Recktenwald, Horst-Claus. Tax Incidence and Income
Redistribution: An Introduction. Detroit: Wayne
State University Press, 1971.

 

 

 

331

Reynolds, Morgan and Smolensky, Eugene. "The Post Fisc
Distribution: 1961 and 1970 Compared." National Tax
Journal, 27 (December, 1974), PP. 515-530.

 

Riggs, Fred W. Thailand: The Modernization of a Bureaucratic

 

Polity. Honolulu: East-West Center Press, 1966.
Sahota, Gian S. "Public Expenditures and Income Distribution

in Panama." USAID, Panama City, August, 1972
(Mimeographed).

Sakuntalaksna, Prachark. "An Economic Analysis of the Thai
Direct Tax System: Evaluation and Suggestions for
Reform." M. A. Thesis, Faculty of Economics,

Thammasat University, 1972.

Salkin, Jay S. "On the Direct Measurement of Tax
Progressivity in Thailand." National Tax Journal,
27 (June, 1974), PP. 301-318.

 

and Chalayonvatn, Darunee. "Changes in Tax Incidence
in Thailand from 1963 to 1968." Discussion Paper
no. 39, Faculty of Economics, Thammasat University,
March 20, 1974 (Mimeographed).

 

Samudavanija, Chai Anan. "The Politics and Administration
of the Thai Budgetary Process." Ph.D. Dissertation,
University of Wisconsin, 1971.

Samuelson, Paul A. "The Pure Theory of Public Expenditure."
Review of Economics and Statistics, 36 (November,
1954): pp. 387-389.

 

. "Diagrammatic Exposition of a Theory of Public
Expenditure." Review of Economics and Statistics,
37 (November, 1955)

 

Sanguanruang, Saeng. "Net Resource Flow Between Agriculture
and Non-Agriculture in Thailand in 1953 and 1963."
Ph.D. Dissertation, Indiana University, 1971.

Sanittanont, Sura. Thailand's Rice Equrt Tax: Its Effects
on the Rice Economy. Bangkok: National Institute
of Development Administration, 1967.

 

 

Sawyers, Larry and Wachtel, Howard M. "Theory of the State,
Government Tax and Purchasing Policy, and Income
Distribution." The Review of Income and Wealth,

21 (March, 1975), pp. 111-124.

 

Sen, Amartya. On Economic Inequality. Oxford: Clarendon
Press, 1973.

 

332

Shoup, Carl 8. Public Finance. Chicago: Aldine, 1969.

 

Siamwala, Ammar. "Rice." Thammasat University Journal,
3 (May, 1974), pp. 5-91.

 

Siffin, William J. The Thai Bureaucracy: Institutional
Change and Development. Honolulu: East-West
Center Press, 1966.

 

 

Simons, Henry C. Personal Income Taxation: The Definition
of Income as a Problem of Fiscal Poligy. Chicago:
University of Chicago Press, 1938.

 

 

Singer, Neil M. "Income Redistribution through the
Federal Budget in 1959." Ph.D. Dissertation, Stanford
University, 1965.

Small, Leslie E. "Water Control and Development in the
Central Plain of Thailand." Southeast Asia, 3
(Spring, 1974), pp. 679-697.

 

Snodgrass, Donald R. "The Fiscal System as an Income
Redistributor in West Malaysia." Public Finance,
29 (1974), PP. 56-76.

 

Stark, Thomas. The Distribution of Personal Income in the
U. K., 1949-1963. Combridge: Cambridge University
Press, 1972.

 

 

Tan, Edita A. "Taxation, Government Spending, and Income
Distribution in the Philippines." Paper prepared
for the Joint JERC-CAMS Seminar on Income
Distribution, Employment, and Economic Development
in Southeast and East Asia, Tokyo, December 16-20,
1974.

Tawney, Ralph H. Equalipy. London: Urwin Books, 1964.

 

Tax Foundation, Inc. Tax Burdens and Benefits of Government
Expenditures by Income Class, 1961 and 1965. New
York: Tax Foundation, 1967.

 

 

Taylor, Milton C. and others. Fiscal Survey of Colombia.
A Report Prepared under the Direction of the Joint
Tax Program of the OrganiZation of American States
and the Inter-American Development Bank. Baltimore:
John HOpkins, 1965.

 

Thailand. Budget Bureau. 'Annual Budget. 1963, 1965, 1969
and 1971.

 

Va

We

Wei

333

. Government Revenue. 1963 and 1969.
. Records of Government Activities, 1969.

Thailand. Department of Comptroller General. Report of

Receipts and Outlays of the Kingdom of Thailand,
1963 and 1969.

 

Thailand. National Economic and Development Board. National

Income of Thailand, 1968-1969 edition, and 1972-1973
edition. -

 

Thailand. National Statistical Office. Reports of the
Household Expenditures Survey, 1962-1963, 7 volumes.

 

. Report of the Socio-Economic Survey, 1968-1969.

Report of the Population and Housing Census, 1970.

 

. Statistical Yearbook of Thailand, no. 28,
(1968-1969) and no. 29 (1970-1971).

United Nations.» Economic Commission for Asia and the Far

East. Economic Survey of Asia and the Far East, 1971.
Bangkok, I972.

. International Trade Projections, Effective Protec-
tion and Income Distribution, Vol. III: Income Dis-
tribution. Bangkok, 1972.

 

 

Van Roy, Edward. "The Persuit of Growth and Stability through
Taxation of Agricultural Exports: Thailand's Ex-
perience." Public Finance, 23 (1968), pp. 294-317.

Webb, Richard C. "The Distribution of Income in Peru."
Discussion Paper no. 26, Research Proqram in Economic
Development, Woodrow Wilson School, Princeton Univer-
sity, September, 1972 (Mimeographed).

. "Tax Policy and the Incidence of Taxation in
Peru." Discussion Paper no. 27, Research Program in
Economic Development, Woodrow Wilson School, Princeton
University, September, 1972 (Mimeographed).

Weisbrod, Burton A. "Income Redistribution Effects and
Benefit-Cost Analysis." In Samuel B. Chase, Jr;, ed.
Problems in Public Expenditure Analysis. Washington,
D. C.: The Brookings Institution, 1968, pp. 177-222.

, and Hansen, W. Lee. Benefits, Costs, and Finance
of Public Higher Education. Chicago: Markham, 1969.

334

Weisskoff, Richard. "Income Distribution and Economic
Growth in Puerto Rico, Argentina, and Mexico."
Review of Income and Wealth, 16 (December, 1970),
pp. 303-332.

 

Weisskopf, Thomas E. "Capitalism, Underdevelopment, and the
Future of the Poor Countries." In Jagdish Bhagwati,
ed., Economics and World Order. New York: Mac-
Millan, 1972, pp. 1-49.