MSU

LIBRARIES
m

 

 

RETURNING MATERIALS:
PIace in book drop to
remove this checkout from
your record. FINES wiII
be charged if book is
returned after the date
stamped beIow.

 

 

 

 

 

ENERGY PRICE SHOCKS, AND SHORT-RUN AGGREGATE SUPPLY:
THE EVIDENCE FROM FIVE INDUSTRIALIZED COUNTRIES

By

Alireza Nasseh

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Economics

1982

ABSTRACT

ENERGY PRICE SHOCKS, AND SHORTvRUN AGGREGATE SUPPLY:
THE EVIDENCE FROM FIVE INDUSTRIALIZED COUNTRIES

By

Alireza Nasseh

The dramatic increases in energy prices that occurred
when the Organization of Petroleum Exporting Countries
(OPEC) quadrupled the world price of oil in 1973—197”, and
again more than doubled it between the end of 1978 and early
1980, have had profound implications for the economies of
all the industrialized countries. These energy price shocks
played a dominant role, by most accounts, in bringing about
the.deep recession and high inflation of thefimidv19703.

The main purpose of this study is to evaluate the
effects of large increases in the price of energy since 1973
on the economies of the main industrialized countries. To
this end,,a model of short-run aggregate supply with explicit
treatment of energy prices was developed. We have concentra—
ted on the supply side of the economy since the important
effects of energy prices on macroeconomic activity occur
largely through aggregate supply.

The model has been estimated and simulated for five

industrial countries; the United States, Canada, the United

Kingdom, Germany, and France. Large increases in the price
of energy were found to have had substantial disruptive
effects on the economies of most of the countries we have
examined. In particular, the energy price shocks generated
a sizable inflationary pressure in the U.S., Canada, U.K.,
and Germany. Furthermore, the adverse effects of higher
energy prices on labor productivity, real wages, the capital
utilization rate, and energy consumption were found to be
significant and similar across countries, except for France.
In the German economy, the effects of higher energy prices
appear to be relatively larger than elsewhere, while the
effects in the U.S., Canada, and the U.K. are similar but
smaller than those in Germany. In France, energy price
shocks do not appear to have had a significant effect on the

economy.

TO MY PARENTS,

AND TO MY WIFE, SHAHRZAD

ii

ACKNOWLEDGMENTS

One cannot complete a program of graduate study with-
out the assistance and encouragement of a great many people
who deserve recognition for their efforts.

I must first express my sincere gratitude to Professor
Robert H. Rasche, Chairman of my dissertation committee, for
his invaluable advice and guidance throughout this study.

He was a rich source of inspiration and encouragement. I
would also like to thank Professor James M. Johannes for
many hours spent reading the various drafts of this disser-
tation, and for his countless insightful suggestions. In
addition, my thanks go to other members of my committee,
Professors Edmund Sheehey and Christine E. Amsler, who also
read the entire manuscript and made many valuable suggestions.

Finally, I shall always be grateful to my parents,
whom I owe an immense debt of gratitude for their unending
support and encouragement through the years, and to my wife,
Shahrzad, whom I owe my deepest thanks for her love and her
patience in awaiting the completion of this dissertation.

Terie Snyder typed the dissertation, and I would

like to thank her for her efforts.

iii

TABLE OF CONTENTS

Page
List of Tables. . . . . . . . . , , , vi
List of Figures . . . . . . , . , , . , , , , , , , , , viii
CHAPTER
ONE - INTRODUCTION ..‘ , . , , . ‘ , , , , , , , , , 1
TWO - A REVIEW OF LITERATURE AND RECENT ENERGY
PRICE BEHAVIOR . . . . . . . . . . . . . . . . “
2.1 - Introduction , , , , , , , , , , , , H
2. 2 - Crude Oil Prices . . . 5
2. 3 - A Review of Empirical Studies of Energy
Price Shocks . . . . . . . . . . . . . . 9
THREE - A THEORETICAL MODEL OF SHORToRUN AGGREGATE
SUPPLY o o o o o o o o 0 Q o 0 Q o o o o 22
3.1 - Introduction . , , , , , , , , 22
3. 2 - The Model , , , , , , , , , , , , 22
3. 2.1 — Price Equation . . 27
3.2.2 - Wage Rate Equation or Short— Run
Phillips Curve . , . . . . 33
3.2.3 - Short Run Demand for Labor . . , #3
3.2.u - Relationship Between Manhours
and Employment . , . , HS
3.2.5 - Supply of Labor (Labor Force
Participation) Equation , , , , #7

3.2.6 - Demand for Energy. . . , , . , , 50

3.2.7 - Supply of Energy . . , . . , , , 52

3.2.8 - Capital Utilization Rate . . . . 52

FOUR - EMPIRICAL RESULTS . . . . . . . . . . . 56
u.1 - Introduction . . . . . . . . . . . . 55
u. 2 - The Basic Estimation Technique . . . . . 56
u. 3 - Model Estimation . . . . . . . . . . 62
A. — United States of America. . . . . . 62

B. - Canada. . . . . . . . . . . . . . 66

C. - United Kingdom. . . . . . . . . . 73

D. — Germany . . . . . . . . . . . . 77

E. — France. . . . . . . . . . . . . . . 8“

F. — Japan . . . . . . . . . . . . . . . 88

iv

CHAPTER PAGE

FIVE - THE EFFECTS OF ENERGY PRICE SHOCKS. . . . . . 93

5.1 - Introduction. . . . . . . . . . 93

5.2 - The Effects of Energy Price Shocks. . . 93
A. - United States of America . . . . . 93
B. - Canada . . . . . . . . . . . . . . 99
C. - United Kingdom . . . . . . . . . . 10“
D. - Germany. . . . . . . . . . . . . . 109
E. - France . . . . 115

5.3 - The Relative Impact of. Energy Price
Shocks. . . . . . . . . . . . . 115

5.” - The Effect of Higher Energy Prices
on the Shape of Aggregate Supply. . . . 126

SIX — SUMMARY AND CONCLUSIONS . . . . . . . . . . . 128
APPENDIX - DATA DEFINITIONS AND SOURCES . . . . . . . . 131
BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . 1M3

TABLES

”.1

Results of
Results of
Results of
Results of

Results of

LIST OF TABLES

Historical
Historical
Historical
Historical

Historical

Actual Performance of
1973*19790 o o o 0

Simulation
Simulation
Simulation
Simulation

Simulation

the U.S. Economy

Historical Simulation - U.S.A

Absence of Abrupt Increases in Price of Energy

- U.S.A.

for
For
for
for

for

U.S.A.

Canada

U.K.

Germany.

France

Deviation from Base Case Forecasts - U.S.A.

Actual Performance of the Canadian Economy

1973-19.79. . o o o 0

Historical Simulation — Canada.

Absence of Abrupt Increases in Price of Energy

— Canada - . .

Deviation from Base Case Forecasts - Canada

Actual Performance of U.K. Economy - 1973-1979.

Historical Simulation - U.K.

Absence of Abrupt Increases in Price of Energy

— U.K.

Deviation from Base Case Forecasts - U.K..

Actual Performance of Germany's Economy

vi

PAGE

9a

95

97

98

100

101

102
103
105

106

107

108

110

TABLES PAGE
5.1u - Historical Simulation - Germany- - - - - . - . 111

5.15 - Absence of Abrupt Increases in Price of Energy
- Germany 0 o o o o o I 0 0 0 . ' ° ' ° ' ' ll 3

5.16 - Deviation from Base Case Forecasts - Germany - 11a

5.17 - Actual Performance of France's Economy

1973-1979 . . . . . . . . . . . . . . . . . 115
5.18 - Historical Simulation - France - - - . . - - ~ 117
5.19 - Absence of Abrupt Increases in Price of Energy

- France. . . . . . . . . . . . . . . . . . 113
5.20 - Deviation from Base Case Forecasts - France- - 119

5.21 - The Marginal Impact of a 1% Increase in the
Price Of Energy " U.S.A- o o o o o o o o o o 121

5.22 - The Marginal Impact of a 1% Increase in the
Price of Energy — Canada. . . . . . . . . . 122

5.23 — The Marginal Impact of a 1% Increase in the
Price of Energy — U.K. .. . . . . . . . . . 123

5.2” - The Marginal Impact of a 1% Increase in the
Price of Energy - Germany . . . . . ... . . 12”

5.25 - The Marginal Impact of a 1% Increase in the
Price of Energy « France. . . . . . . . . . 125

5.26 - The Estimates of Own Price Elasticity of
Aggregate Supply in Five Countries. . . . . 127

vii

LIST OF FIGURES

PAGE
Chart 2.1 — Saudi Arabian Crude Oil. . . . . . . . . . 8
Chart 2.2 - Output Growth in Six Countries . . . . . . 10
Annual Percentage Rate
Chart 2.3 - The Rate of Increase in Prices in Six
Countries. . . . . . . . . . . . . . . . . 11
Annual Percentage Rate
Figure 3.1 - A Model of Short-Run Aggregate Supply . . 2H

viii

CHAPTER ONE

INTRODUCTION

The economies of the world were so dominated in the
19708 by the effects of two oil crisis that the period has
been aptly named "the OPEC decade".1 The dramatic increases
in energy prices that occurred when the Organization of
Petroleum Exporting Countries (OPEC) quadrupled the world
price of oil in 1973-1970, and again more than doubled it
between the end of 1978 and early 1980, have had profound
implications for the economies of all the industrialized
countries. Other commodities have experienced rapid and
substantial increases in price. The prices of grain and
other agricultural products, for example, have experienced
price fluctuations on the order of 300 or ”00 percent
over years. Few people, however, would be as concerned
about these events, or expect them to have anywhere near
the impact on our standard of living that increases in
the price of energy are likely to have. Higher energy
prices have contributed to reduced economic growth in many
countries, and in the longer-run may result in basic

changes in lifestyles.

 

V‘—

IThis was the title of a special review of the
world economy in the 19703 which was published in the
Economist, 29th December 1979.

How fast energy prices rise in the future will
depend in part on the rate at which conventional energy
resources become scarcer and more difficult to find, in
part on the rate of technological change that lowers the
cost of nonconventional energy sources, in part on the
behavior of the OPEC cartel, and in part on the domestic
energy policies of various countries. Although the
dramatic increases of 1973-197N are not likely to continue,
we should probably expect to see energy prices continue to
rise in real terms, at least slowly, for the next few
decades.

This study develops a model of shortwrun aggregate
supply for major oil consuming nations with explicit
treatment of energy prices, and explores the effects of
a change in the price of energy (oil) on the aggregate
supply. The emphasis is on the supply side of the economy
since important effects of energy on macroeconomic activity
occur largely through aggregate supply. An increase in
energy prices represents an increase in the cost of a major
factor of production. Consequently, an increase in energy
prices relative to other prices should result in a decline
in the amount of goods and services supplied by the economy
at any given level of prices, that is the aggregate supply
curve should shift to the left. An attempt is made to
measure the extent to which the aggregate supply curve
shifts to the left, at any level of output, due to higher

energy prices.

An overview of the pattern of the crude oil price increases
during the last decade, and a review of the empirical
studies of the effects of energy price changes on the U.S.
and other industrialized countries' economies are included
in Chapter Two.

In Chapter Three, a model of short-run aggregate
supply with explicit treatment of energy prices for major
oil consuming nations is developed.

Chapters Four and Five present empirical results.
Specifically, in Chapter Four, the results of the statistical
estimation of the model for five industrial countries; the
United States, Canada, Germany, France, and the United
Kingdom are given. In Chapter Five, the empirical results
of the effects of energy price shock are presented, in
turn, for each country, and also a comparison of the
relative effects of energy price shock is made among the
countries included in this study.

Finally, Chapter Six presents-some overall

conclusions.

CHAPTER TWO

A REVIEW OF LITERATURE AND RECENT ENERGY PRICE BEHAVIOR

2.1 - Introduction

 

The behavior of crude oil prices in the 19705
presents a record of apparent chaos. The roughly fourfold
increase in the price of crude oil from 1971 through 1979
was partly responsible, by most accounts, for bringing
about one of the longest and deepest post—war declines in
economic activity in most industrial nations and causing
a nearly world wide inflationary epidemic. The oil
consuming nations experienced a second sharp rise in the
price of crude oil in 1979-1980, in the wake of the Iranian
Revolution. Although the increase was not as great in
percentage terms as that of 1s7a-197u, it came on top of a
very sluggishrecovery in Western Europe and Japan, and
a high underlying rate of inflation in the United States.

In this chapter we first look in detail at the
pattern of the crude oil price increases during the last
decade. Next, we survey the results of several attempts
to analyze and quantify the macroeconomic effects of the
1973-197H oil price increase on the U.S. economy, as well

as other industrial countries‘ economies.

2.2 - Crude Oil Prices

 

During the 19708 the average producer price of crude
oil rose from about $2 a barrel to about $31 a barrel
(May 1980). The increases in the price were largely
concentrated in two periods, 1973-1979 and 1979-1980.

The price of Saudi Arabian oil, which until recently
was the standard by which most other crude oil prices were
set, was $1.30 a barrel at the beginning of the 19708,
much lower than it had been throughout the 19508, and much
of the 1960s. With moderate inflation in the industrial
countries, this would mean imported oil prices fell somewhat
more in real terms. The steady decline in the dollar price
of oil was brought to a halt in 1970, partly owning to
action by the Organization of Petroleum Exportinngountries
(OPEC). Throughout the postwwar period, the oil companies
has generally controlled production and pricing, despite
the fact that by the early 19708 the OPEC countries were the
source of about 85 percent of world exports of oil. The
OPEC countries pushed increasingly for participation in
ownership and changes to existing contracts, securing new
agreements in 1970 and 1971 which paved the way for further
substantial changes in both price and participation.

The posted price of Saudi Arabian oil, which had
been $1.80 a barrel throughout the 19608 (even though actual
discounted prices had fallen as low as $1.28), was raised
between 1971 and early 1973 in three steps to $2.59 a barrel,

and was at $3 a barrel before the Yom Kippur war in

October 1973. The war and subsequent production cutbacks
initiated a series of large price rises. In October the
posted price of Arabian light crude increased to $5.11 a
barrel, and in December 1973 the price was more than doubled
to $11.65, a 350 percent increase on the price a year before.

The immediate impact on the industrialized countries
did not come from the price rise alone, but also from
production cutbacks during the winter of 19739197“ and
embargoes until March 1978 on the United States and the
Netherlands. After this enormous surge, prices remained
largely unchanged for almost the next five years - while
declining in real terms.

The period of relative oil price stability was
brought to an end at the close of 1978 and the beginning
of 1979, sparked off by the Revolution in Iran. The new
series of price increases raised posted prices to $13.3u
a barrel and then to $lh.55 a barrel in early 1979. The
28th June meeting of OPEC in Geneva raised the Saudi
benchmark price from $1u.55 to $18.00, with a price ceiling
of $23.50 which was itself soon broken by the more militant
OPEC members. The Saudi price was later raised to $2“ and
then to $26 (in January 1980) and to $28 (in May 1980).

The average official OPEC export price of oil,
which stood at $12.87 at the end of 1978, had moved to
around $31 a barrel by May 1980, an increase of 1M0 percent.
The average of $31 a barrel covers a wide spread which for
light oils ranges from Saudi oil selling at $28 a barrel

and Iranian oil selling at around $35 a barrel.

Using the posted price of Saudi Arabian crude oil as
a rough indicator of world oil prices, and export unit values
of industrial countries as a rough index of the price of
their product8,Chart2L].8hows the movement of the nominal
price and the relative price of oil (that is, the nominal
price of oil relative to export unit values of industrial
countries) from 1955 to 1980.

The nominal price of oil which had been roughly
steady throughout the 19508 and 19608, increased almost
300 percent in 197981980. By contrast, the relative price
of oil which has been declining through most of the 19508
and 19608, after rising sharply in 1973-197u, fell steadily
from 197” to 1978, due both to inflation in the industrial
countries and to depreciation of the dollar; In 1979 the
deterioration of the relative price of oil ended and it
jumped sharply upward. Moreover, the relative price has
increased less in percentage terms in 1979-1980 than it
did in 1973-197”.

A casual review of the record for the United States
and the five industrial countries is supportive of the view
that all countries have shared a similar adverse impact due
to energy price developments in the seventies. Charts 2.2
and2. 3 show the realvoutput growth rates and the rate—of‘price
increases, respectively, for each of the six countries
from 1960 to 1980. For France and the United Kingdom,
the growth rates are for real GDP and the implicit price

deflator. For the other countries, GNP data are used to

30

25

20

x>f-f-OU
p
01

10

 

 

 

 

 

 

. “"NOHINRL PRICE
""" RERL PRICE (1972 PRICES)
7 I I I I
1955 1960 1985 1970 1975 1980

Source: International Financial Statistics, IMF

CHART 2.1

SAUDI ARABIAN CRUDE OIL

to measure output and prices. As can be seen in Chart 2-2
in 1979-1975, and then again in 1979-1980 the growth rate

of real output fell sharply in each of the six countries.

In 1979-1975, the rate of increase in prices rose sharply

in each of the six countries (Chart2.3). Subsequently the
inflation rates declined in all six countries. In 1979

the decline in the inflation pace ended and it jumped upward

again in 1979-1980.

 

2.3 - A Review of Empirical Studies of Energy Price Shocks

During the last several years considerable theoretical
and econometric effort has been made to understand and to
model the effects of oil or-—more generally-—energy price
increases. Furthermore, some analysts have attempted to
measure the extent to which these higher energy prices have
adversely affected the level of prices, or rate of inflation,
output, productivity, investment and capital utilization, and
other macroeconomic developments. There have been, however,
considerable differences in the theoretical approaches taken
' by different analysts and also, the angle from which they have
looked at the issue of energy price shocks and their effects
on the economy.

The most successful approach has been to consider
energy as a primary input in production and concentrate on
the effects of a rise in the price of energy relative to the
price of output from the goods and services sector. This
has implications for the income of non-energy factors of

the production and for aggregate demand, but many<xf

IO

 

 

 

 

\ \ \\ ..I
Y \.\ \\\.. \\
9 NE \. nfix . \\
\
. WM . mm m . \\\
s N P K RAH .
O H a o W R I q!” l
UCJU F: o 0v IV
. _. _” . s.\\
u . . . \ \\\
I.__ rixi «I,
0”; '0
‘ .H‘ l
\ ‘
\\\‘\HH\\ ’0‘ I
\‘H‘. .
is
O In. \
““ “‘
, ..I
I I II
a, to. I 'l“
0’0 s \
,0 Q o I \
4’00
1 win
0‘.‘ ‘ Q l \ I
1111.\. \\\’.oooo~ I I
~\ \\ a‘\\ ll
\ \\ ‘\‘ I
\\ A blil \Q‘ \‘
/ :70 \\
fl /LD'CO.NO I“
l ‘0 ‘\“ ” \
IIOIO “8089‘ - \\\\
’0'0’01. ’ \ ~~
I‘8I‘QII.“‘ {\5 \\\ I
‘Q‘a‘IIIIII n \\ \\\
.Ii .v
[0"] \"” j
I ’
1L1 \\ v
8‘.‘ ’0". \\
I“‘!‘ l O l ‘\
1“.‘ O.\ O
/ \\ \ "
\A\ L\ .
q 4 . I4 4 4| a q
‘ 2 o a a 4 2 o 2
1 I. 1. .

1982.8 1986 1987.8 1970 1972-5 1975 1977.5 1980

1980

International Financial Statistics, IMF

Source:

CHART 2 . 2

OUTPUT GROWTH IN SIX COUNTRIES

ANNUAL PERCENTAGE CHANGE

ll

30

 

_ U .8 .
----- CRNnoR I
26 .J --- JRPRN

"'" U .K .
-- oERNRNv
---- FRANCE

 

 

f I

I I I I
1980 1982.5 1985 1987.5

1970 1972.5

 

 

T T
1975 1977.5
International Financial Statistics, IMF

Source:
CHART 2.3

THE RATE OF INCREASE IN PRICES IN SIX COUNTRIES
ANNUAL PERCENTAGE CHANGE

1980 '

12

the effects come through induced changes in production
techniques and costs, which are the determinants of the
aggregate supply. For this reason, energy price increases
and similar events have commonly been referred to as
"supply shocks" in recent macroeconomic literature.

Traditionally, macroeconomists have limited their
analysis of supply side behavior to simple models of wage
and price dynamics. These models could be apprOpriate
for the analysis of some types of macroeconomic problems,
but they cannot answer all the questions.

More comprehensive analysis of the effects of
higher energy prices on the economy, for example, on the
rate of inflation, unemployment rate, productivity and
other macroeconomic issues, have led to use more
disaggregated aggregate supply models.

One of the major theoretical works - in this area -
has been done by Dale W. Jorgenson and Edward A. Hudson
(1979). Their model is basically an integration of a
Walrasian "Input-Output" model and a Keynesian aggregate
demand model. This model has been used in a number of
empirical studies regarding the macroeconomic effects of
higher energy prices.

D. Jorgenson and E. Hudson (1978) analyzed the
impact of higher energy prices by using their "dynamic
general equilibrium model of the U.S. economy". A feature
of the model is its reliance on a close relationship
between the quantity of capital and energy use a that is,

energy and capital are considered complements with a low

13

elasticity of substitution between them. But a higher
degree of substitution is thought to exist between energy
and capital, on the one hand, and labor, on the other. The
model was used to simulate two economic growth paths over
the 1972-1976 period. In the first simulation, actual
values of the exogenous variables, including world oil
prices, were employed as the basis for model solution. In
the second simulation, 1972 energy prices were employed
over the whole 1972—1976 period. Since all other exogenous
variables were the same, the differences in simulated
economic activity can be attributed to the impact of the
oil price increase.

Their model results show real GNP in 1976 was
reduced by 3.2 percent by the increase in energy prices
from 1972 to 1976. This implies that oil price increase
accounted for part, but not all, of the recession of the
mid 19708. Also, total energy consumption was 8.8 percent
lower in 1976 with the energy price increase than without.
An analysis of the sources of this energy saving suggested
a combined reduction of final demand away from energy
and energy-intensive goods and services, a restructuring
of patterns of input into production away from energy, and
a reduced rate of economic growth. The model also showed
labor input lower by 0.5 million jobs or just over 0.5
percent with the energy price increase than without it,
and GNP per unit of labor 2.57 percent lower. Also, the

level of capital stock was reduced by 3.0 percent by the

 

CE

7516

is
is
ma:
prc
shc

and

19

increase in energy prices.

Their overall conclusion for the U.S. economy was
that higher energy prices have resulted in a decrease in
the level of real GNP and a slowdown of economic growth,

a decline in the growth of productivity, reduction in
capital stock and investment and using more labor intensive
methods of production.

R.G. Anderson (1979), after critically examining
the theoretical basis of the specification of the factor
demand functions of each sector in the Hudson-Jorgenson
model, argues that, "the Hudson-Jorgenson type model, which
is a dynamic general equilibrium model of the U.S. economy,
is probably best suited to the analysis of longer~run
macroeconomic issues, such as the slowdown in the
productivity growth. As Sargent, Sim8,and others have
shown, theoretical analysis of "business cycle" fluctuations
and stabilization policy is possibly best conducted within
very simple supply-side models."

K.A. Mork and R.E. Hall (1980a), also have found
substantial disruptive effects on the United States economy
due to large and unanticipated changes in the price of
energy in 1973-1979. They have constructed a medium sized
macroeconomic model of the U.S. economy with explicit
treatment of energy. Their model attempts to consider both
substitution away from energy as an input to production
that occurs because of energy price shock and also traces

the monetary and general macroeconomic effects of it. The

15

important features of their model are a technology
constraint with the three inputs, capital, labor, and
energy, and with a flexible functional form; a money demand
function with gross output as the transaction variable;

a permanent income consumption function; rational expecta-
tions; and some short-run rigidities, notably in wage and
price determination and in the investment process.

Mork and Hall simulated the effects of the energy
price increase experienced in 1973 and 1979, then compared
this simulation with a stable growth case with steady 5.0
percent inflation. According to their analysis, energy
price shocks generated a sizable recession in the model,
i.e., "a severe recession combined with an extraordinary
increase in inflation". Their finding suggests that the
energy price shock depressed real output by 2.9 percent in
197a and by 5.0 percent in 1975, or about threeoquarters
of 197H-1975 recession. The rate of inflation was increased
by N.0 percent in 197u and about 2.0 percent in 1975.
Consumption declined by about 3.2 percent in response to
the decline in real income associated with the increase in
the price of energy, and investment decreased by $7 billion
”(1972 dollars). The predicted effect on employment was
an increase in the unemployment rate of about 1.0 percent
in 1979, rising to 1.9 percent in 1975. Their overall
conclusion is that, the energy price shock was clearly
a major cause of the 197u-1975 recession and inflation in

the United States.

 

—7—_—"

16

In another study, Mork and Hall (1980b) examined the
effect of the second energy price shock of 1979-1980. The
results of this study indicate a serious adverse impact on
the U.S. economy because of this shock, but a good deal
less so than in 197u-1975. Almost 2.1 percentage points
are added to the inflation rate in 1979 and about 1.3
percentage points in 1980. Also, the energy price shock
has resulted in a decrease in real output of $15 billion
(1979 dollars) or 1.0 percent in 1979 and as high as $55
billion or almost R.G.percent in 1980. COnsumption has
reduced by 2.5 to 3.0 percent, and unemployment has
increased, not so much in 1979 but by a little more than
1.0 percentage points in 1980, because of energy price
shock.

Mork and Hall also examined different monetary and
fiscal policy options open to policy makers to prevent
the adverse effects of energy price shocks on the economy.
They found that monetary policy could accommodate the shock
and maintain private investment, but at the cost of higher
inflation in the longerwrun. On the other hand, a policy
intended to limit inflation «.specifically, through monetary
contraction, would worsen the adverse effects of the energy
shock on real economic activity. They found policies with
favorable effects on investment and capital formation 2 i.e.,
a little monetary accommodation of the energy shock, have
a significant advantage in preventing the adverse impact
of energy shock on real economic activity, without too

high a price in added inflation. Specifically, they

17

suggest cuts in the payroll tax as one of the most promising
tools of macroeconomic policy to provide substantial
stimulus to the real economy without adding to inflation.

There have been several attempts to modify the major
econometric models of the U.S. economy to include energy
prices and to estimate the effects that the sharp rise in
world oil prices had on the United States economy in 1979
and 1975.

Pierce and Enzler (1974) analyzed the oil price
rise in the M.I.T.-PennvSSRC (MPC) model. Their estimates
showed that the rise in imported oil price in 1973-197u
and the partial rise in domestic oil prices raised the
consumption price deflator by 2.1 percent in 1979. Also,
the real output was decreased by'l.2 percent in 1979 and
2.8 percent in 1975 due to oil price increases.

G. Perry (1975) employed two large econometric
models of the U.S. economy a the Michigan quarterly
econometric model and the quarterly econometric model of
the Federal Reserve Board (FRB). He concluded that higher
oil prices, coupled with the failure of fiscal and monetary
policies to offset their depressing effect on the total
demand, had been the primary cause of the l97uw1975
recession. His estimates showed that, by the end of 197A,
the depressing effect of higher oil prices had reduced
real GNP (1972 dollars) by about $35 billion and added 1.0
percentage points to the unemployment rate; Also, higher

oil prices had directly added 3.5 percent to the consumption

18

price deflator.

Eckstein (1978) analyzed the effects of the energy
price increases in the Data Resources, Inc. (DRI)
econometric model. The base of comparison is the simulation
of the model with actual values of the variables. The
effect of the oil price increase is estimated by simulating
an alternative scenario where the wholesale price for fuel
and power raises by 6.0 percent per year from early 1973,
instead of its actual average of 27.u percent annual rate
of increase in 1973—1975 period.

Eckstein's estimates indicate that real output
growth rate was lower by 1.0 percent in 197a and 2.6 percent
in 1975, inflation rate was higher by 3.0 percent in 197a
and 1.“ percent in 1975, and output per hour was lower by
2.1 percent and 0.H percent in 1978 and 1975 respectively.
Eckstein concluded that, "the energy crisis was not the
only factor pushing the economy off its equilibrium path
in 1973-197M. However, without it, the economy would have
suffered no worse than a year of a small GNP decline in
1979, and would have seen 1975 as the first year of
recovery".

Comparison of the above results present some
problems. First, each author assumes that fiscal and
monetary policies are unchanged in simulating the model,
but there are some differences in the exact specification
of this assumption. Second, inflation is measured by

different indice8.in the various model. Furthermore,

19

different base years had been used for deflation. The

model also differs in the way the energy price is entered,
and the extent of the energy price change. Models without

an energy sector (Pierce and Enzler, Perry) raise consumption
price relative to output prices. Models with energy sector
(Eckstein, Mork and Hall) raise or lower an energy price
directly. However, all models predicted a significant
recession as a result of the energy price shock.

There have also been several studies by R.H. Rasche
and J. Tatom (1977a, 1977b, 1981) regarding the adverse
effects of sharp increases in the price of energy in
1973-1979 and then in 1979-1980 on the U.S. economy as well
as other industrialized countries' economies. Rasche and
Tatom (1977b) estimated a CobbeDouglas type production
function for the output of private business sector in terms
of inputs of capital, labor and energy. The Cobb-Douglas
production function assumption implies relatively high
price and output elasticities of demand for energy and
unit partial elasticities of substitution between energy
and capital labor. However, they have cited a number of
empirical evidence supporting the notion that these
properties apply at least to the U.S. economy, Japan, and
Western Europe. Their conclusion is that the rise in the
price of energy resources relative to that of output have
reduced economic capacity of the economy, have raised the
.general level of prices, and have reduced potential output
and productivity of existing capital and labor resources.

They estimated that four to five percentage points of

20

both the higher price level and reduction in national
output in 1979 were due to the sharp increases in the price
of energy in 1973-l97u. Furthermore, the increase in
energy costs in 197a alone reduced labor and capital
productivity by about 9.0 percentage points, and for the
1973-1978 period, higher energy prices reduced the growth
rate of labor productivity by 1.3 percent per year.

Other analysts have argued that the effects of
increased energy prices on the economy have been much
smaller. For example, Denison (1979) rejects the rise in
energy prices as an explanation for reversal in productivity
growth, based largely upon a study by Perry (1978). He
concludes that the results obtained by Rasche and Tatom
establishing the importance of energy in the productivity
decline are overstated. Perry analyzed industrial energy-
consumption data, and concluded that the own price
elasticity of energy demand was much lower than unity, which
is the assumed value in the Rasche-Tatom analysis. Perry
estimated that higher energy prices have caused not more
than 0.2 percent 1088 of labor productivity and potential
output between l973«1976. Perry concluded that Rasche and
Tatom overstated the importance of energy because of the
upward bias in their estimates of the own price elasticity.
However, as demonstrated in an appendix of one of their own:
papers, Rasche and Tatom (see Rasche and Tatom, 1977 and
Tatom, 1979) show that a lower estimate of the own price

elasticity of demand for energy than the value assumed in

21

their analysis increases rather than decreases the impact

of energy price changes on real aggregate output. Thus,

Denison's dismissal of energy as the main factor responsible

for the major declines in productivity growth is incorrect.
In a recent study, Rasche and Tatom (1981) found

the same adverse effects of higher energy prices on output

and productivity for five other industrial countries; Canada,

Japan, U.K., Germany, and France, as they had estimated for

the United States. Furthermore, they found the effects of

the 1979-1980 energy price shock on the economy to be

comparable to the 1973—1979 experience.

CHAPTER THREE

A THEORETICAL MODEL or SHORT-RUN AGGREGATE SUPPLY

3.1 - Introduction

 

The purpose of this chapter is to develop a model of
short-run aggregate supply in major oil consuming countries
with explicit treatment of energy prices, and to study the
influences of a change in the price of energy (oil) on
economic variables such as inflation and unemployment. We
shall concentrate on the supply side of the economy since
important effects of energy on macroeconomic activity occur
largely through aggregate supply. In contrast to a demand-
oriented theory of inflation, an energy price increase
directly touches the supply side by exerting a cost push on

the price level while reducing output.

3.2 - The Model

 

A typical short—run aggregate supply model with an
energy sector added would consist of:

(I) A price determination equation which relates
the price level to the wage rate and other

input prices via some sort of mark-up.

(II) A wage rate equation or shortvrun Phillips

curve .

(III) A short—run demand for labor.

22‘

23
(IV) A labor supply or labor force participation
equation. '
(V) A demand function for energy.
(VI) A supply of energy equation.

(VII) A capital utilization rate equation, (Demand

for capital).

Given the quantity of output and a fixed stock of
capital goods in the short-run, (I) - (VII) will determine
the price level and unemployment.

The effect of a higher relative price of energy can
be shown graphically. In Figure 3.1, the system of equations
representing the aggregate supply model are shown in panels
(a) to (g). In panel (a), the aggregate production function
is presented, tmuademand for and supply of labor are shown
in panel (b). In panel (d), the wage determination equation
is depicted and in panel (8), the (markeup) price equation
is presented. The “5° line in panel (f) simply transforms
the rates of inflation (p) from panel (e) to (g), which
shows the aggregate supply curve, presented in terms of a
relation between the inflation rates and level of output.
The aggregate supply curve represents the amount of output
that producers are willing to supply at various price levels
(inflation rates). The amount producers are willing to
supply at different price levels in the shortvrun (a period
short enough that the capital stock can be assumed fixed)
depends on such factors as the amount of the capital stock,
the state of technology, the size of labor force, and the .

prices of two variable factors of production; labor andnenergy.

 

24

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE 3.l

A MODEL OF SHORT-RUN AGGREGATE SUPPLY

25

Let us start with the initial equilibrium condition.
Points A to G in panels (a) to (g) indicate the initial
equilibrium state where the economy fully utilizes capital
and labor, producing output Yp at its corresponding rate
of inflation pd, and the unemployment rate is 0*.

In a purely nee-classical economy where wages clear
the labor market instantaneously, an unexpected increase in
the relative price of energy, assuming the economy is a
price-taker in the world energy market, would reduce energy
use, output, and the relative price of labor and capital
services in the short—run. This is shown by shift in
aggregate supply curve fromS S to S 81 , and the reduction

0 0 1

in full-employment output YP to YP'.

reduction in energy usage as a third factor of production

Essentially, a

lowers the marginal productivity of existing capital and
labor in the short—run. Over a longer period of time, the
capital stock is reduced since the supply price of capital
goods (assumed to be given in the long-run) does not decline.
The result is that the longer-run output loss and real wage
decline are larger than the shorterun.

In our model, on the contrary, wages (and prices)
respond slowly to unexpected changes in energy prices (or to
all other surprises in the economy). Due to wage rigidity
because of long-term labor contracts among other reasons,
adjustment to full-employment equilibrium in the labor market
is not instantaneous, but is delayed for a considerable period

of time. Also, prices are sluggish in their response to

26

changes in production costs. During the period following

an energy price increase, but before the accomodating change
in the wages, labor is priced too high for full employment.
During this period, the demand function for labor determines
the level of employment, which may then be well below the
supply of labor. This can be interpreted as a characteriza4
tion of the Keynesian hypothesis of wage rigidity and is an
attempt to embody the view that the labor market achieves
equality of supply and demand in the long—run but the process
takes time.

An increase in the relative price of energy reduces
energy usage, which in turn decreases the marginal productivity
of existing capital and labor. In Figure 1, this will shift
the aggregate production function down, and the demand for
labor to the left, DL'. Assuming the real wage is fixed in
the short-run at (%%), the level of employment is determined
on the short-side of the labor market at Ll, which represents
an increase in the unemployment rate above the rate of
unemployment, U*. The higher rate of unemployment U1,

corresponds to the rate of change of wages w in panel (d).

1
In panel (e) the increase in relative price of energy shifts
the price line upward to p', corresponding to the rate of

change in wages wl. Also the lower level of employment, L1,
will result in a lower level of output, Y1, as it is shown
in panel (a). Therefore, the effect of a rise in the relative

price of energy, in the short-run, is to reduce the output

below its full employment level and increase the rate of

27

unemployment and the inflation rate.

3.2.1 - Price Equation

 

The price equation of the model determines an
aggregate price index for the economy. Most specifications
of the price equation now found in the literature are
derived from one of the two alternative theories about price
formation; the mark—up theory and the marginal cost pricing
theory. It is interesting to observe that the mark-up and
the marginal cost theories need not be inconsistent with
each other, at least as applied to aggregate price level of
an economy. It is possible to derive the same price level
equation, beginning with either view.

One of the recent expositions of the markwup theory
as it pertains to olig0polistic price behavior is presented
by deMenil (1979), whose empirical work agrees with Gordon's
conclusion (1971, 1975) that price formation is principally
a matter of cost factors, with direct influence of demand
pressure entering secondarily or not at all. Wage costs
are generally very important in the determination of the
price of final goods and services. In addition the costs
of other factors of production also appear to have prominent
effects.

Let us assume that the representative firm in the
economy has some monopoly power and sets its prices as high
as it can so as to maximize profits while still preventing

the entry of an additional firm into the market (this

28

assumption is valid for industries ranging from nearly
perfect monopoly on the one hand, to nearly perfect competi-
tion on the other). The firm does this by determining price
as a mark-up over minimum long-run average cost, where the
mark-up itself is a function of the shape of the industry
demand curve and the scale of production at the point of
minimum long—run average cost.

Minimum long-run average cost depends on the shape
of the "exante" putty production function, which will be
interpreted as an aggregate production function describing
the summed output of many firms, all of which are operating
at the bottom of a U-shaped cost curve, and the supply of
factors of production. (The inputs are assumed to be
available in perfectly elastic supply at the going price.)

The production function is taken to be of the Cobb-

Douglas form:
Y = e”tL°‘1<BEY (3.1)

where Y is output, L is labor input measured in manahours,

K is capital input, E is the flow of energy resources, t is
time, u is the trend rate of growth of output due to
technological change, and a, B, y are the output elasticities
of the respective inputs. The production function is assumed
to exhibit constant returns to scale, that is<:+-Bi-y = l.

The total cost function is taken to be as:

TC = wL + rK + pE E (3.2)

29

where w is wage rate, r is price of capital input per unit,
pE is price of energy input per unit.

With inputs priced competitively, efficient production
requires to minimize total cost (3.2) subject to production
function (3.1); The necessary (or first order) conditions

for cost minimization are:

EE = NY
a.
fig = AY
(3.3)
EE = AY
-—-.,
Y = eutLaKBEY

‘ Using the first—order conditions, the total cost function,

can be written as:

TC = wL + £»wL + l~wL
(I a
TC = § (8 + e + y) wL (3.u)
TC = l wL
a

and the conditional demand for labor is:

- _ -B - -8 'Y
L = Y e “tr-9) (l) (31) c—"—) (3.5)
a a r pE

From (3.”) and (3.5) the minimum total cost function will be:
TC : e'ut Y (fate-.87“Y waerEY (3.6)

and the minimum long—run average cost function is:

AC = e-ut a-aB-By.ywar8pBY (3.7)

30

if we let m (m 2 1) be the mark—up factor, a markeup on

average cost means:
p = m (AC) = met-uta-QB.BY-Ywar8p£Y (3.8)

taking logarithm from relationship (3.8), then the total

differential of the relationship (3.8) can be written as

follow:

1n p = A + 1n m + a 1n w + 8 1n r + y 1n pE - ut
or

d ln p = d ln m + a d 1n w + B d ln r

+ y d In pE — d(ut)

or

p = ms+aw+ Br +YpE-A(ut) (3.9)
where;

A 1n (a’as'ey’y)

The price equation given by (3.9) represents the
equilibrium or desired price p*. Recent studies by Taylor
(1979, 1980) and Hall (1979) emphasize the fact that all
prices are not set in unison across the economy but are
generally staggered, and that a primary determination of the
price decision by firms is the prevailing price outstanding
in the market. At any point in time some, but not all, firms

will be adjusting their prices and must therefore take into

account the most recent price decisions of other firms.
Moreover prices are maintained for more than one period,

long-term profit considerations and transaction costs which

31

involve in frequent price setting can be an explanation
for these short-term price rigidities. Therefore, given
this sluggish adjustment of actual to equilibrium or

desired price, we adopt the following simple adjustment

mechanism:
. = 0* - O .
Ap n(p Pt—l) (3 10)
or
o = 0* - o '
P up + <1 n) Pt_1 (3.10)

where n is the adjustment coefficient (0 s n S l).
Combining equation (3.9) and (3.10)’, we obtain the

basic price equation as:
P = (lvn)Pt_1*~nm+-naRI-ner+-nyPE-na(ut) (3.11)

This basic price equation is modified in the following
way, to take into account for short-run factors: the mark-up
factor is assumed to vary with demand pressure; that is,
short-run variation in the markvup factor is determined by
the variation of elasticity of demand for output over the
cyclical movement of the economy.

It seems reasonable to assume that, when the demand
is very high and the capacity for output is close to being
utilized fully, the elasticity of demand for output with
respect to price is relatively low. Thus, the supplier may
feel that he can afford to increase his markaup somewhat
without significant loss of sales of his output. Thus one

would expect that the markvup factor tends to be higher than

 

32

usual when the demand for output is larger relative to
the capacity.

When the demand for output is low relative to the
existing capacity, firms in the industry are likely to find
that the elasticity of demand with respect to price is
relatively high. In this situation we expect the mark—up
factor to be smaller than usual.

It is likely that these tendencies of the markvup
factor to fluctuate with the demand condition will be
reinforced if the current demand condition is expected to
continue for some time into the future, and therefore not
only the current demand conditions, but also expected demand
conditions would be relevant in the determination of the
markoup factor.

An appropriate measure of the demand pressure
affecting the mark-up factor is the ratio of industrial
production to its expoential trend, (as was used by J. Ball
and M. Duffy, 1970). We expect this ratio to be large
and increasing when demand is high relative to capacity,
and small and decreasing when demand is low relative to
capacity.

For empirical convenience, we shall adopt the form:

1n m = b0 + bl 1n q + b2 1n qt“1 (bl > 0, b2 < 0)
(3.12)
or
O = O O ‘
m blq + b2 qt—l (3.12)

where q is the ratio of industrial production to its

 

33'

exponential trend.
With this modification, the equation of aggregate

price index can be written as follow:

P = (I-n)Pt_li' nbléi'nbzqt_14-nawi-n8r
. (3.13)
+ nApB - nA(ut)
For estimation purposes, our price equation will be as:
P : ao+alw+a2r+ “BEBE... unfit-1+ (1561+ (165115.11
(3.1a)
+ a7t

3.2.2 - Wage Rate Equation or ShortuRun Phillips Curve

 

The Phillips curve began as the result of an empirical
investigation of U.K. wage behavior by Phillips (1958),
was extended and put into a theoretical disequilibrium
context by Lipsey (1960), and was applied to the U.S. and
set in a policy context by Samuelson and Solow (1960).

The basic hypothesis of the Phillips curve is an
application to the labor market of the familiar idea that
price increases in response to excess demand and decreases
in response to excess supply. For the sake of convenience,
the rate of unemployment is used as a proxy for excess
demand (or excess supply) in the labor market. The relation—
ship between the rate of change of money wages and unemp10y«
ment rate is generally thought to be non«linear; when the
unemployment is near the frictional level, small variations

in the unemployment rate might be expected to have a big

 

3N

impact on the resulting wage change, on the other hand,

if the rate of unemployment is already large, a small
variation in this rate might be expected to have little
impact on the rate of change of money wages. However, recent
empirical studies (Cagan, 1977) and Papademos (1977) have
demonstrated that non-linear and linear specifications seem
to do about as well over the sample period through the

mid 1970's;

In addition, Phillips argued that the change in the
unemployment rate and the percentage change in consumer
prices would also influence the percentage change in money
wages. Phillips' justification for including the change in
the unemployment rate in the wage equation was that the
change in unemployment rate is a plausible indicator of
future labor market conditions. When the rate of unemployo
ment is decreasing, employers would be expected to offer
higher wages for the service of labor than when the rate of
unemployment is constant. Conversely, when the rate of
unemployment is increasing, employers will be less willing
to grant wage increases to workers. Thus, the relationship
between the‘rate of change of money wages and the change
in the unemployment rate will be inverse.

The effect of the rate of price increase on wages
is clear. Phillips, and Lipsey (1960) pr0posed that workers
would, at least partially, protect the real value of their
earnings by bargaining for compensation for increases in

the general level of prices.

 

 

 

 

curve, i
importa
employe
union 1
c) Iabc

accordi

but do
trade-T
(1966,
that tj
trade-
Pested
result
labor

that b
0n the
momina
PPoduC.
aVePage
iDCpeaE
the act
actual :
highEF E
‘I’Ol‘keps )

that ‘

35

In the subsequent extensions of the basic Phillips
curve, new explanatory variables were included, the more
important of which were: a) profits as an indicator of
employers' ability to concede to higher wage demands; b) trade
union membership as a proxy for worker's aggressiveness;

c) labor productivity as a basic influence on wage rates
according to neo-classical economic theory.

These modifications to the Phillips curve qualify,
but do not challenge, the conclusion that there is a longvrun
trade—off between inflation and unemployment. Friedman
(1966, 1968) and Phelps (1968) were the first who argued
that there was a temporary, but not a longvterm stable
trade-off between unemployment and inflation. Their argument
rested on the idea that the short—run tradeaoff was the
result of expectational error by economic agents. Friedman's
labor market analysis (1968) made the explicit assumption
that both the demand for labor and supply of labor depend
on the real wage rather than on the nominal wage. Since the
nominal wage was evaluated in terms of the current actual
product price by employers and in terms of the expected
average consumer price level by workers, employment could
increase as long as the expected price level lagged behind
the actual level (thus, simultaneously allowing a lower
actual real wage to induce increased hiring by firms, and a
higher expected real wage to induce a higher labor supply by
workers). However, the expectational error can not persist

so that employment returns to its equilibrium level. In

36

equilibrium the expected and actual price level are equal,
and so in equilibrium only one level of unemployment and
output is possible.

Friedman named the associated unemployment rate
(given population, technology, and labor participation) as
the Natural Rate of Unemployment, NRU. The natural rate
hypothesis also implies that attempts to hold the unemploy-
ment rate below (above) the natural rate will cause an
upward (downward) acceleration of prices indefinitely.

The initial reaction of U.S. economists to the NRH
was that the policy implications of the NRH could be safely
ignored, on the empirical grounds that U.S. prices and
unemployment were inconsistent with it.

In its simplest form, the expectation hypothesis as

applied to wage equation can be written as:

wt = f(x) + apte (f' > 0)

where w is the percentage change of money wage, pte is

t
the expected rate of inflation at time t, and X represents

a measure of excess demand for labor (in turn related to the
unemployment rate). Empirical confirmation of the NRH
requires that the estimated coefficient on Ste be unity, -
assuming the conditions that characterize the long run;

a) be = p, that is expected price changes equal actual price
changes; b) p = w, that is wage changes are fully reflected

in the final prices.

Initial empirical evidence on the NRH suggested that

 

 

th
fo
th
Pe
to
th.

thT

ex;

stu

the

i.e,

(196

the

In 5917,

37

the hypotheses can be rejected for the U.S. and U.K. but not
for Canada. More specifically, empirical wage studies for
the U.S. by Turnovsky and Wachter (1972), Gordon (1971),
Perry (1970), among others have found that the wage response
to expected price change is significant but is less than one;
that is, the coefficient a is approximately 0.3 to 0.5. On
the other hand, studies for Canada by Turnovsky (1972) and
J. Vanderkamp (1972) have found that the wage response to
expected price change is not significantly different from one.
However, more recent empirical evidence such as the
studies by Gordon (1972) and McCallum (1976), suggest that
the coefficient a is equal to one, implying the absence of
a long-run tradeuoff between inflation and unemployment,
i.e.; a long-run vertical Phillips curve.
Ever since the publication of studies by Friedman
(1968) and Phelps (1968), it has been customary to include
the expected rate of inflation as an important factor in
the explanation of shortorun wage and price changes. Even
if there is no longsrun tradevoff between inflation and
unemployment, as implied by the NRH, the short-run tradeooff
may be highly important. The nature of this short‘run
trade-off is strongly influenced by what determines inflation
expectation. A usual assumption about this expectation
formation is that the expected rate of inflation is determined
by historical rate of inflation. In particular, "adaptive
expectation", as introduced by Cagen (1956), have been used

in several theoretical and empirical studies of inflation

38

by Nordhaus (1975) and Tobin (1975), among others. Indeed
the widespread use of adaptive expectations suggested that
an ever-accelerating rate of inflation could maintain an
unemployment rate below the natural rate of unemployment.

More recently, this type of assumption about
expectations has been challenged by the notion of "rational
expectations", first introduced by Muth (1961) and incorpor-
ated in the complete model of inflation by Lucas (1972) and
Sargent and Wallace (1975). Expectations are defined as
rational when the expectation error is unrelated to all
the information available when expectations are formed, i.e.,
expectation of inflation is an unbiased prediction of actual
inflation, given all the information available just before
the period begins. Sargent and Wallace demonstrate that
macro economic stabilization policy, even in the short run,
cannot influence unemployment if expectations are formed
"rationally", unless the policy makers act in an unpredictable
way.

The use of rational expectations in the theory of
inflation and unemployment has been criticized by FiScher
(1977), Gordon (1976 and 1977), Phelps and Taylor (1977),
Taylor (1979) among others. It is not so much.the assumption
of rational expectations in itself as_the implicit assumption
about complete shorteterm flexible prices and wages that
these economists criticize. Longsterm profit considerations
and transaction costs involved in frequent price setting

on the part of firms and multiperiod wage contracts have

AL

“1

A$k

39

often been mentioned as the underlying reason for sluggish
short-run wage and price adjustment. These short-run price
and/or wage rigidities can result in a short-run Phillips
curve, so that the policy implications of the usual rational
expectations model mentioned above are incorrect. Fischer
(1977) constructed a rational expectations model with long—
term labor contracts. He demonstrated that short-run wage
rigidities due to the long-term contracts enables stabilize-
tion policies to affect the behavior of real output in the
short-run. According to Gordon (1977): "In light of wide«
Spread evidence that prices adjust sluggishly to demand and
supply shocks, it is hard to avoid the conclusion that for
short-run analysis the application of rational expectations
to economic policy should be regulated to the same scrap
heap of discarded ideas where lie the earlier classical
models of perfect market clearing laid to rest by Keynes
forty years ago".

"It is important to recognize that sluggish shorteterm
price adjustment is not "irrational" and does not in any way
contradict the idea that expectations should be formed
rationally;'Gordon (1976).

The specification of the wage adjustment relationship
- the short-run Phillips curve e includes the conventional
variables which have been found to be significant in previous
researchJ The rate of unemployment, which serves as a proxy
for the excess demand in the labor market, will appear in

three differnt functional forms, namely linearly and inverted

NO

to the first and second power. Though theoretical consiv
derations suggest a nonlinear form due to the downward
rigidity of money wages and to the presence of bottlenecks
at low levels of unemployment, some empirical research has
discovered that the linear form is more suitable. Rather
than make an arbitrary choice at the outset, we will
experiment with all three forms in order to determine the
one which is appr0priate for each individual nation.

The change in the rate of unemployment is also
included as a measure of expectations of the future of
labor market conditions.

In addition, the expected rate of inflation has been
found to be an important factor in the explanation of
short—run wage changes.

Finally, the percentage change in labor productivity
is also included. According to neoclassical economic theory,
_this variable should exert a significant influence on the
rate of wage changes. The theory postulates that labor
services will be purchased until the marginal value product
of labor just equals the money wage. Holding product prices
constant, a rise in the marginal physical productivity of
labor will shift the demand for labor curve outwards and,
as equilibrium is reestablished, nominal wages will increase.

In sum the wage equation can be written as;

o - 0e 0
w — bo + blU + 1320 + b3 p + bu Lp (3.15)

where w is the rate of change of money wages, U is the rate

of unemployment which will take on one of the three forms

NI

outlined above, 0 is the (absolute) change in the rate of
unemployment, be is the expected rate of inflation
(unobservable) and LP is the rate of change in average labor
productivity.

The wage adjustment relationship given by (3.15)
represents the equilibrium or desired wage changes, 3*.
The sluggish short-term adjustment of wages to prices,

(and other factors) can be incorporated into the wage

equation by adopting the following simple adjustment

mechanism:
. = .* - .
Aw 1(w wt_1) (3.16)
or
w = Aw* + (1-1) wt”1 (3.16')

where A is the adjustment coefficient, (0 s A s 1).
Combining equation (3.15) and (3.16)‘, we obtain the

basic wage adjustment equation as:

1.+ AblU + 1b2

{a U + Ab3pe + 1b,, LP
(3.17)

The wage adjustment equation given by (3.17) can not

Ibo + (.1--).) wt_

be directly estimated because data on pe do not exist. Thus,
an expectation generating function has to be postulated. I
The use of rational expectations appears to have become quite
conventional in recent years, and, hence it will be adOpted
for present purposes.

The idea of rational expectations approach isto

base expectations on the model itself. In particular,

1:2

expectations of pte are based on the information available
at time t. Assuming that the information set is based on
past values of exogenous variables in the model, then REH,

due to Muth (1961) specifies that:

e - 0 0 o 0
pt - E<pt|pt'1...’ Int-1..., th-1040, 000) (3018)

Now, the wage adjustment relationship can be
consistently estimated by the instrumental variable technique
with the instrument for pte being the "predicted" value Ste
obtained from the least-squares regression of pt on lagged
values of exogenous variables in the model.

In sum then, the wage equation to be estimated will

be of the form:

A

o -‘ o O o e o
w - 80 + 31 wt_1 + 320 + 83U + 8,, pt + le-IP (3.19)

Expectations of inflation can be modeled in various
forms. One simple approach is to assume "partly rational"
. (expectations, which consider the possibility that economic
agents actually utilize only a portion of the relevant
inffiormnation available to them, at time t-l, in forming their
e“‘Pecrl;ations of inflations, in other words, it might be
desirable in practice to limit information set to those
varPiables that are "most likely" to be considered by economic

agents.

The partly rational expectations hypothesis utilizes

ODIY'lflTe information contained in past inflation rates, that

is:

143

e (3.20)

P, E (3t|§t_l, Pt-2:""
An alternative form of partly rational expectations
hypothesis utilizes the information contained in past
inflation rates as well as past "money growth" information.
The justification for including money supply growth informa-
tion in an equation explaining inflation expectations is,
of course, the notion that "actual" inflation in influenced
by the recent pattern of money growth rate. A recent
empirical study by D.J. Mullineaux (1980) demonstrated that
inflation expectations are systematically influenced by
past inflation rates and past rates of money growth, but
not by fiscal-policy-related variables such as growth in
government spending or the size of the federal deficit.

Therefore, the expected inflation can be measured as:

p.t = E (ptlpt'l..., Mt'l...) (3021)

where M is the growth rate-of money supply.
3.2 . 3 - Short-Run Demand For Labor:

The simplest approach to generating an employment
fUDCtion begins with that demand for labor theory which
Specifies a desired level of employment determined by the
Planned level of output, with various other casual variables
such as factor prices and the state of technology, incorporo
ated in the function according to the particular assumptions
Characterizing the behavior of the firm, such as cost

m. . . . . . . . - o
lanIllzation, or profit max1mizat10n. It 18 then supposed

an

that in the short-run the actual employment may differ
from its desired level, since as output changes
actual employment will take time to adjust to its desired
level, there being significant costs involved in changing
this level. so, typically, a partial adjustment mechanism
is specified and when this is incorporated into the desired
function a single equation is obtained for estimation
purposes.

Let the economy be characterized by the CobbeDouglas

production function:
Y = e“tL*°‘I<BEY (3.22)

where Y is output, L* is desired labor input in manuhours,
(which is defined as the product of the number of workers
employed, M, and hours per worker, Hr - L = M.Hr), K is
capital stock and E is energy input. The assumption of
constant returns to scale is also postulated, i.e.,

a + 3 + y : 1,

The central assumption for cost—minimization is that
the optimal (desired) level of employment is set so as to
minimize a specified total cost function subject to the
constraints of the given production function. Assuming the
Cobb-Douglas production function given by equation (3.22),~
the general costaminimization procedure requires that firms
equate the ratio of marginal products of the inputs to the
(exogeneously determined) ratio of their factor prices. 80

the necessary (or first order) condition for cost minimization

us

becomes:
P
M L = aY/L* : g
MPK 877K r
(3.23~
MP )
L = . Y/L* = IL
or
- E E a
K - (a) (r) L
Y w, (3.23)‘
- _ __ a
E - (a) (PE) L
substituting (3.23)‘ into (3.22)
B
Y = eutL*°‘+B+I(§.)3 (1)7 (21) (1"...)Y (3.21:)
a a r pB
where a + B + Y = 1. Now solving (3.28) for L, we have
1n L* = A+ lnY -B.ln (g) -11n (ll—1-..: (3-25)

PE
- fl 8 1 Y . .

where A - In (a) (a) , and r and pE are pr1ces of cap1tal
and energy inputs respectively. Equation (3.25) determines
the optimal labor employment of the economy. For a variety
of reasons, the firms may not choose to carry out actions
implied by equation (3.25). In the short-run, the adjustment
implied may nOt be fully realized because: (1) expectations
with respect to output and wages are clouded with uncertainty;
or (2) because there are costs of increasing employment which
the firms may not wish to incur fully in view of uncertainty
of future employment needs; or (3) because of difficulties
involved in carrying«out a concurrent adjustment of capital

stock. For these, and perhaps other, reasons we may postulate

the following adjustment mechanism:

96

: * -
ln Lt ln Lt + (l 6) ln Ltol (3.26)

where 6 is the speed of adjustment, (0 s 6 5 1).

Equation (3.26) gives a logarithmic version of the
usual adjustment process in which only part of a gap between
current desired and previous position values is eliminated
in a given period.

Now substituting equation (3.25) into (3.26), the
labor demand function becomes:

ln Lt = 6A + elnY --9 31M? - eyln(i¥E)

(3.27)

+ (1—9) ln Lt_1 - e (ut)

For estimation purposes, equation (3.27) may be written as:

t 0 1 2 H

lnL = a+a lnY+aln(¥)+a3

1n (35-) + a
P

lnI.
E t“

1

 

3.2.8 - Relationship Between Manhours and Employmenp

The labor demand functions express the demand for
labor in terms of "manhours". In order to generate the level
of "employment" we need to divide manhours between level of
employment and hours of work. By definition, the following
relationship exists between average hour per worker, Hr,

and number of employees in the economy, M.

: i:
Lt Hrt Mt (3.29)

The hypothesis of what determines the average hours

of work per man is as follows; When output increases sharply,

H7

there is a tendency for firms to increase the average hours
of work, because it is often difficult to increase the
number of employees, especially if increase in output is to
be temporary. Given this hypothesis, we may write an

equation for Hr as:

1n Hr = b + b

Y
0 1 1n (7—) + b t (3.30)

-1 2

where Y is the output and t is time.

3.2.5 - Sppply of Labor (Labor ForceiParticipation)

 

The main purpose of the labor supply equation in the
model is to allow us to determine the unemployment rate,
given the prediction of employment, from the employment
equation. The cyclical behavior of the aggregate labor force
is best explained by labor participation function. These
functions relate the labor force participation rates,
defined as the ratio of labor force to population of various
age—sex subsets of the working population, to measures of
employment opportunities. However, in dealing with labor
force participation rates, the problem is that the dynamics
of short-run changes in labor force participation are simply
not well understood. That is, there is a lack of satisfactory
theoretical model of the factors affecting participation
decisions in the shortvrun.

The supply of workers is a function of population
and its characteristics in the longerun and of employment

opportunities in the shortsrun. However, A. Tella‘s work

 

(19E
grox
The
by I
cycl
aged
char

highe

is ta

to be

WhereI
25-54,
the ti
least

brough-

is 88115
cYCliCa
Péhtch

pm‘UCt:

WePall

Eyel Of

98

(1965) has shown clearly that there are two major age—sex
groups with different cyclical behavior (apart from trend):
The primary labor force of males aged 25-59, characterized
by high average participation rates and little response to
cyclical changes, and the secondary labor force of women
aged 15 and over plus males aged 15-29 and 55- and over,
characterized by lower average participation rates and
higher responsiveness to cyclical changes.

Therefore, the participation rate of primary workers
is taken to be a simple function of time, and the equation

to be estimated is:

= e + e t (3.31)

 

where, LP is total civilian primary labor force aged

1t
25-59, POPl is population corresponding to LFl, and t is

the time trend. Time trend is included to take account, at
least in part, of secular changes in participation rates
brought about by economic, demographic, and taste factors

not explicitly controlled for in the regression.

The participation rate of secondary workers, however,
is sensitive to labor market conditions, and to other
cyclical changes, and it is expected to be effected, in
particular, by variables such as; unemployment rate, average
productivity, and real wages.

The first variable, unemployment rate: is a measure of

overall labor market conditions and represents the general

level of business activity. An inverse correlation between

99

the participation rate of secondary workers and the rate of
unemployment is expected to exist, due to "discouraged worker
effect". Also, an inverse correlation between the partici—
pation rate of secondary worker and average labor productivity
is expected to exist. According to D.B. Suits, et al., (1975)
as average productivity of labor increases, families‘ income
will rise and this presumably will discourage the participation
of secondary workers in the labor force. Finally, the
correlation between participation rate of secondary workers
and real wage is expected to be positive, but several
empirical studies such as deMenil and Enzler (1970) have
failed to estimate a positive elasticity to the real wage.

Therefore, the equation for participation rate of
secondary workers may be written as follows:

LP“ =f+fU +f LP +fIw)+ft(332)
POP2t o 1 t-j 2 t 3 -‘p" u " -

 

where LF2 is total civilian secondary labor force, POP2 is

population corresponding to LF2, U is the unemployment rate,
LP is average labor productivity, (g) is real wage, and t
is time trend.

Having determined employment, M, from labor demand

equation, and taking POP and POP2 to be ex0genous, LFl and

1
LF2 can be determined from equation (3.31) and (3.32).

Then, by definition, the civilian labor force is equal to
LF1t + Lth
obtained as:

and so, the civilian unemployment rate can be-

50

M
U = (1 - ) * 100 (3.33)

t
t LFlt‘l’LFzt

 

3.2.6 - Demand For Energy

 

Demand for energy as a factor of production by firms
is determined by the planned level of output, and with
various other variables such as factor prices, level of
capital stock and the state of technology, incorporated in
the function according to the particular assumptions
characterizing the behavior of the firm, such as cost
minimization, or profit maximization.

Let the economy be characterized, as before, by the

Cobb-Douglas production function:

Y = ept L°K3E*Y (3.3a)

where, E* is the optimal (desired) level of energy input.
The assumption of constant returns to scale is also
postulated, a + 6 + r = l.

The key assumption for cost minimization is that the
optimal (desired) level of energy input is set so as to
minimize a specified total cost function subject to the
constraints of the given production function. Assuming the
Cobb-Douglas production function given by equation (3.39), the
general cost minimization procedure requires that firms
equate the ratio of marginal products of the inputs to the
(exogenously determined) ratio of their factor prices. So

the first order conditions for cost minimization becomes:

( )

L
(J

R}

(D

51

."fé = .YY/E" - 3.12
MPL aY7L - w
MP (3.35)
___E = vY/E* = if;
MPK BY7K r
or
P
L — (3) (13-) E*
y w
(3.35)'
8 PE
K - (~)(——) E*
y r
substituting (3.35)' into (3.39):
a B P a P B
Y = e“ £89+3+Y(9) (E) {—3) (ii) (3.36)
Y Y W 1”
where a + B + Y = 1. Now solving (3.36) for E*, we have:
pE ‘PB
In E" = A + 1n Y valn (15—) -8111 (‘5‘) " lit (3.37)

where A = 1n ($)a(%)8, and p3, r and w are prices of energy,
capital and labor inputs respectively.

Equation (3.37) determines the optimal (desired)
level of energy input for the economy. For a variety of
reasons, the firms may not choose to carrywout actions
implied by equation (3.37). In the shortvrun, the adjustment
implied may not be fully realized because of uncertainty
with respect to expected future output, and difficulties
involved in carrying out a concurrent adjustment of capital
stock. For these, and perhaps other, reasons, we may

postulate the following adjustment mechanism:

lnE ‘= TlnE

t t (3.38)

* V
+ (-1 T) ln E'tv-l

where T is the speed of adjustment of actual to desired

52

(optimal) level of energy input, (0 i 1 i 1). Now,
substituting equation (3.37) into (3.38), we have the

following equation as a demand for energy function:

lnFZ

A + Y 1 (DB)
t T T ln t -Ta 11 -;—

p .
r3 ln (7:5) + (1-.) ln Et_ -r (ut) (3.39)

1

For estimation purposes, equation (3.39) may be written as:

- + + p3 + p“5)
1n Et '-g0 gllny,c g21n(—W--)‘t g31n (—r—t

+ 81.111194 " 85“ (3.90)

3.2.7 - Supply of Energy

 

Energy prices are, since late 1973, determined on
a world market, so it can be assumed that each country faces
a perfectly elastic supply curve for this resource, and the
economy is a price-taker in the world energy market.

The relevant price which OPEC is able to dictate as
the dominant energy producer is the price of energy relative
to that output. This characterization of OPEC pricing
practices over the past seven years is probably due to their
attempt to maintain the value of their oil in terms of goods

and services that they purchase.

 

3.2.8 - Capital Utilization Rate _

The supply of capital goods, i.e., capital stock, is
taken as given in the Shortvrun. In the longvrun, however,

the supply of capital is variable as firms can add to or

53

subtract from the stock of capital depending on their
incentives. The relative price of capital goods (measured
relative to the price of output), determined by supply
conditions, is assumed to be given in the long-run (see

J. Tatom, 1979). The actual amount of capital for a firm
to employ can then be determined using this price.

The optimal employment of capital services is
determined by the planned level of output, with several
other variables such as factor prices and the state of
technology, incorporated in the function according to the
particular assumptions characterizing the behavior of the
firms, such as cost minimization or profit maximization.

Let the economy be characterized by the Cobb-Douglas

aggregate production function:

where K* is the Optimal level of capital goods.

The central assumption for cost minimization is that
the optimal (desired) level of capital services will be set
so as to minimize a specified total cost function subject
to the constraints of the given production function.

Assuming the Cobb—Douglas aggregate production function given
by equation (3.91), the general costvminimization procedure
requires that firms equate the ratio of marginal products

of the inputs to the (exogenously determined) ratio of their
factor prices. So the necessary (or first order) condition

for cost minimization becomes:

59

K = BY/K* _ 3
NFL aY7L w
(3.92)
MPK = 8Y/K* = 1;
“PE yY7E pE
or
L - (3) (3)10c
8 w
(3.92)‘
E = (I) (L) K*
8 PE
substituting (3.92)’ into (3.91):
(I a
Y = e”tI<*°‘+B*Y (9-) (I)Y(£) (31-)Y (3.93)
. B B: W PE
where a + 8 + y = 1. Now, solving (3.93) for K*, we have:
_ r r
ln K* - A + 1n Y -a1n (—) — Y 1n (—) - pt (3.99)
w pE
a. (I Y
where A = 1n (8) (%) , and r, pE and w are price of capital,

energy and labor inputs respectively. Equation (3.99)
determines the Optimal employment of capital goods. For a
variety of reasons, the firms may not choose to carry-out
actions implied by equation (3,99). In the short-run, the
adjustment implied may not be fully realized because of
uncertainty with respect to expected future output, and
difficulties involved in adjustment of capital stock itself.
For these, and perhaps other reasons, we may postulate the

following adjustment mechanism:

i:
1n K = ¢ 1n Kt + (1—¢) 1n Kte (3,95)

t l

where ¢ is the speed of adjustment of actual to desired

55

(optimal) level of capital services, (0 s o s 1).
Now, substituting equation (3.99) into (3.95), we

have:

¢A + a 1n Y - ¢a 1n (g) — ¢y ln (31)

ln K
PE

...

<1-¢) ln Kt-l - ¢(ut) (3.96)

for estimation purposes, equation (3.96) may be written as:

r
3 ln (5;)

In K h + h In Yt + h

I‘
t 0 1 1n (a)+h

2

4.

h” 1n Kt_1 + hst (3.97)

Equation (3.97) represents a short—run demand for capital
services, in other words, it is a demand for utilized
capital. Having determined the level of utilized capital,
ln Kg], from the demand equation for capital services, the

utilization rate of capital can then be obtained as:

_ U
KUR ‘ 1“ Kt ‘ 1“ Ktvl (3.93)

where KUR is the capital utilization rate.

CHAPTER FOUR
EMPIRICAL RESULTS

9.1 - Introduction

 

In the last chapter, we deveIOped a theoretical
model of short-run aggregate supply with explicit treatment
of energy prices. In this chapter, we present the statistical
estimation of the model for each country in turn. We
discuss the basic estimation technique and the potential
problem in using alternative estimation methods. The
basic ordinary least squares estimations are reported first,
followed by the results of the historical (ex-post)

simulation of the model.

9.2 - The Basic Estimation Technique

 

We will employ Ordinary Least Squares in estimating
the coefficients of the equations of the model. However,
in our system of equations, some of the equations contain
some amount of simulatanity among endogenous variables.
Given a truly simultaneous system of equations, we know that
ordinary least squares estimation is inconsistent as well
as biased, and therefore, a statistically consistent method
of estimation, such as 2SLS or other systems estimation

techniques, should be used for equation systems. However,

56

57

obtaining consistent estimates for our system of equations
is a formidable computing problem. A real problem in the
present case is the paucity of degrees of freedom. An
essential first step in computation by the well-known
consistent methods, such as 2SLS or 3SLS involves the
regression of each endogenous variable on the whole set of
predetermined variables, lagged endogenous and exogenous
variables, in the entire system. In a system.of equations
like ours, with at most twenty nine observations available,
there will not be enough degrees of freedom to estimate a
single reduced form regression for the whole system.
Another difficulty with all systems estimation techniques
is that individual parameter estimates are sensitive to

the specification of the entire system. A serious specifi-
cation error in one equation can affect the parameter
estimated in all equations of the model. Furthermore,
several writers in their empirical work on price and wage
equations, i.e., Phillips Curve studies, have pointed to
the potential inconsistency and biased estimates arising

in Least Squares estimates of an equation, when the equation
in fact is part of a simultaneous system. In some works,
such.as Perry (1969) and Eckstein and Brinner (1972),

these equations were used as part of a multiequation model
in which.price and wages were explained, with all equations
estimated by 2SLS. The estimated coefficients from 2SLS
were shown to differ negligibly from OLS, implying that the

simultaneOus equation bias was not too serious. Based on

58

this empirical evidence, and in our case the practical
problems in using a statistically consistent system
estimation method, we will employ the ordinary least
square technique in deriving the various coefficient
estimates.

A major feature that is likely to characterize an
aggregate time series model, such as the one developed in
this study is serial correlation of the error terms in the
equations. Serial correlation is likely in the present
model because of the relatively simple specification of the
equations. Many relevant variables have doubtlessly been
excluded from the equations and if these variables are
autocorrelated, it is likely that the error terms in the
equations will be autocorrelated as well. The autoregressive
correlation that we will employ to estimate our model makes
the implicit assumption that the additive error terms are

of the form:

: +
Ut p U s

where at is the uncorrelated part of the Ut'

Finally, for purpose of estimation, it is necessary
to use a proxy for inflation expectations in the wage
equation of the model. We will experiment with three
alternative approaches, described in the last chapter.
First, we suppOse that inflation expectations are formed

rationally, (see Muth, 1961). The basic idea is that

market agents form their expectations of inflation at time

59

t on the basis of information at that time. We follow the
empirical procedure suggested by McCallum (1975). This
involves regressing the actual inflation rate at time t

on lagged values of exogenous variables in the model and
using the predicted inflation rate as the expected
inflation rate.

Second, we obtain the prediction of inflation by
regressing the actual inflation rates on its own lagged
values.

Third, we obtain the prediction of inflation by
regressing the actual inflation rate on past inflation
rates and past growth rates of money supply. For each
country, several variants of each of the above alternative
approaches were estimated, differing in the number of lagged
periods used. The equation, with the smallest standard
error and most significant coefficient was chosen to obtain
a proxy for inflation expectations.

In order to evaluate the model's performance and its
ability to replicate the actual data, we will examine the
historical (ex-post) simulation of the model for each
country, over the estimation period. All simulations are
dynamic, in the sense that simulated (rather than actual)
values for the endogenous variables in a given period are
used as inputs when the model is solved in future periods.
One would expect the results of the historical simulation
(i.e., simulation through the e8timation period) for each.

endogenous variable to match its corresponding historical

60

data series rather closely. A measure that is most often
used to evaluate the performance of a system of equation is
the RMS (root-mean square) simulation error. The RMS

simulation error for the variable Yt is defined as:

 

2

n
RMSE = g 2 (Y: - Yi)
t=l

where Y: is simulated value of Yt’ Y: is actual value, and

n is the number of periods in the simulation. The RMSE is

a measure of the deviation of the simulated variables from
its actual time path. Of course, the magnitude of this
error can be evaluated only by comparing it with the average
size of the variable in question. A low RMSE for the
endogenous variables in the model would be a desirable
measure of the model's performance and its ability to
replicate the actual data.

Low RMS simulation error are only one desirable
measure of simulation fit. Another important criteria is
how well the model simulates turning points or rapid changes
in the actual data.

A series of useful simulation statistics related to

the RMS simulation error and applied to the evaluation of

historical simulations are Theil's statistics, defined as:

 

 

..s ..a 2
M _ (Y -Y )'
U - (1)2 (Ys - Ya)2
n - t- t
2
US - . (as -.- Ga») -.
' ‘ ‘2

l s a
(a); (Yt .- Yt)

 

V6

3(

61

and
2(1-o)osoa

1 8._ a72
HZ (Y,c Yt)

 

where Y8, Ya, as, and 0a are the means and standard

deviations of the series Y: and Yi, respectively, and p is
their correlation coefficient. UM, Us, and UC are called
the "bias", the "variance", and the "covariance" proportions,
respectively, and they are useful as a means of breaking the
simulation error down into its characteristic sources.

The bias proportion UM is an indication of systematic
error, since it measures the extent to which the average
values of the simulated and actual series deviate from each
other. One would hOpe that UM would be close to zero. A
large value of UM (above .1 or .2) would mean that a systema—
tic bias is present.

The variance proportion US indicates the ability of
the model to replicate the degree of variability in the
variable of interest. If U8 is large, it means that the
actual series has fluctuated considerably while the
simulated series shows little fluctuation, or vise versa.

Finally, the covariance proportion UC represents the
remaining error after deviations from average values and
average variabilities have been accounted for. The ideal

values for Theil's statistics are UM = U8 = 0 and UC = l.

 

wh
th

83'

62

9.3 - Model Estimation

 

A..- United States of America

All the equations of the model for the U.S.A. were
estimated using annual data for the private business
sector of the economy over the period 1973 to 1979. In
all preliminary experiments, the following equation
for inflation expectations, which is based on the
empirical procedure suggested by D.J. Mullineaux (1980),
performed best in terms of tastatistics and (adjusted)
coefficient of multiple correlation for the overall
regression (R2), and therefore was used to obtain a proxy

for inflation expectations in the U.S.A.

Pt = -.007 + .58pt“1 - .95ptm2 + .90pt._3 + .29M

(1.13) (3.50) (2.62)‘ (3.30) (1.91)

tvl

+ .096Mt_2 + .27Mt-3

(.56) (1.75) (9.A1)

761
ll

.69_ F(6,22) = 11.20 SER = .015 DW = 2.12

where the numbers in parentheses are t—statistics and M is
the percentage rate of growth of money supply.2

All the behavioral equations of the model were
estimated using ordinary least squares (OLS) with a Cochrane‘

Orcutt adjustment for those equations for which the presence

 

2The hypothesis that the sum of the coefficients on
lagged p, plus the sum of the coefficients on lagged M is
equal to one, cannot be statistically rejected. Given the
insignificancy of the constant term, this would imply that
long-run inflation has been a monetary phenomenon.

63

of autocorrelation proved to be significant.

(9.A2)
p = -.019+-.220pt_1*.58'2w--.022r+.089pE
(1.33) (3.33) (2.79) (3.69)
- o o + .
.020q + .011qt_1 0003t
(.55) (.23) (.99)
R2 = .90 F(7,21) = 37.23 Durbin h = 1.53
SER = .0033 p = 0
39.713)
w = .026 + .929w_l+ .OOOSU—.008U- .ll9LP+ .913pe
(2.96) (.39) (2.53) (.76) (3.09)
R2 = .68 F(5,22) = 12.97 0w = 1.96
SER = .011 p = -.935
(2.93)
(9.A9)
In L = 12.332 + .681 1n Y - .026 ln (3) - .109 1n €11)
(9.33) (1.37) (6.51) E
+ .079 1n124 - .012t
(.33) (9.93)
32 = .99 F(5,22) = 562.36 SER = .0037 0w = 1.32
o = .903
(2.30)
(9.A5)
1n Hr = 3.712 + .131Y - .009t
(5.09) (7.36)
32 - .93 F(2,25) = 695.92 SER = .0099 DW = 1.73
p = .800

69

(9.A6)
LPRl = .995 - .002t
(10.72)
32 = .99 F(l,26) = 2001.6 SER = .0013 0w = .36
p = .801
(7.09)
(9.A7)
LPR2 = .962 - .002U - .023LP + .039(%) + .008t
(2.96) (2.39) (.33) (9.99)
32 = .99 F(9,23) = 953.26 SER = .003 0w = 1.57
p = .860
(3.30)
(9.A8)
pE
ln E = -1.231 + .527 1n Y - .066 In (37)
(9.02) (1.61)
pE
- .003 ln (3:) + .997 In E__1 - .001t
(.21) (3.62) (.25)
32.: .99 F(5,23) = 1217.13 SER = .019
Durbin h = .91 p = 0
(9.A9)
ln K = -.969 + .269 In Y + .001 ln (1;) — .006111(£)
PB W
(5.56) (.13) (.91)
+ .397 1n K_1 - .006t
(11.15) (1.97)
32 = .999 F(5,22) = 19926 SER = .0053 DW = 1.59
p=.61

(9.05)

In

65

(9.A10)
: * h .
L Hrl Mt (365/7)

(9.A11)
- -— *
U - (1 IF) 100

LFl + LF2 - AF is total civilian labor force,

and EM = M + GM is total civilian employment. Also, AF

where, LP

is armed forces and GM is government employment.

The estimated coefficients of the model generally
have the theoretically expected sign, except for the
percentage change in price of capital (r) and the percentage
change in the ratio of industrial production index to its
exponential trend (q) in the price equation, and unemploy—
ment rate (U) and the percentage change in labor productivity
(HP) in the wage equation. Also, the coefficients are
significantly different from zero in most of the cases.

‘It is interesting to know that in the wage equation
(9.A3), the hypothesis that the sum of the coefficients on
be and lagged w is equal to one and cannot be rejected at
the level of significance 0 = 5% (i.e., t-statistic = -l.78).
This would imply that the long—run Phillips Curve in the
United States is vertical.

An analysis of the dynamic behavior of the model,
by examining the implied longvrun elasticities of factor
demands with respect to output and factor prices, proved

the dynamic stability of the system of equations.

The mOdel can now be simulated a8 a complete system.

66

Despite the simplicity of this model, its simulation
performance is rather good. To see this, we can examine
the historical simulation of the model over the estimation
period 1952-1979. In performing this simulation the
historical values are used for the exogenous variables in
the model. The results of this simulation are presented
in Table 9.1, which shows the RMSE and the Theil's
statistics for each endogenous variable in the model along
with the mean value of each variable.

Low values for RMSE's, relative to the mean value
of the corresponding endogenous variables, indicate the
rather satisfactory performance of the model and its ability

to replicate the actual data series.

B. - Canada

All the equations of the model for Canada were
estimated using annual data for commercial indUstries over
the period 1959 to 1979..

The following equation for inflation-expectations,
which is based on the empirical procedure suggested by
McCallum (1976), performed be8t in terms of tastatistics

and adjusted R2, and therefore was used to obtain a proxy

for inflation expectations in Canada.

(9.31)
pt = .009-t1.19pt_1~—.60ptfizi-.099rt-1i*.0008rt_2
(3.62) (1.69) (1.67) (.02)
+ .088p + .067p
Bt—l Et—2

(.67) (.57)

: ..H.hfln<r—L

67

 

 

mmbb. mmmm. ommm. ommm. mmbm. :omm. bumm. AODV o AHmmB

HNNN. «soc. coda. mmmm. mmha. :Nmo. ommo. AmDu m AHmmB

Nmoo. :Nho. Huom. mama. when. muse. mmoo. AXDV 2 AHmmB
H30. Hue. Hao. 3w. mmo. OH.H om. mwzm
Hm.w mm.m :m.mH m~.m :m.w :o.m mm.m z<mz
XGH mcH Add 3.. _ mg. 3 Q

 

.<.m.D mom ZOHB<ADsz A¢OHMOBmHm mo mBADmmm

H.: mgm<8

68

32 = .63 F(6,l7) = 9.29 SER = .022 0w = 1.36

where the numbers in parentheses are t-statistics, and r
and pB are the percentage change in the price of capital
and energy, respectively.

All the behavioral equations of the model were
estimated using ordinary least squares (OLS) with a
Cochrane-Orcutt adjustment for serial correlation of the

error term.

(9.32)
6 = -.002 + .3296.1 + .0005w — .022} + .2311;E
(1.20) (.002) (1.03) (2.95)
+ .0356 + .1053“1 + .001t
(.35) (1.00) (1.52)
32 = .32 F(7,16) = 16.22 SER = .016' 0w = 1.73
p = 0
(9.33)
9 = .056+-.529w_l-.008U-.00504-.271EP+-.522pe
(3.56) (3.82) (1.05) (1.28) (3.31)
'2

R = .80 F(5,18) = 19.89 SER = .013
Durbin h = 1.36 p = 0

. (9.39)
ln L = 9.305+ .633 '1nY- .013 ln(¥;) — .109 1n (Psi)
. E
(3.69) (1.56) (6.73)
+ .2701nL_1 - .020t
(2.67) (6.19)

R2 = .99 F(5,18) = 593.3 SER = .0092

Durbin h = .63 p = 0
(9.85)
In Hr = 3.77 + .130Y - .008t
(3.23) (36.01)
32 = .99 F(2,20) = 1513.09 SER = .0097
0w = 1.93 p = .352
(1.31)
(9.36)
LPRl = .939 - .001t
(190.92) (5.37)
32 = .95 F(l,21) = 919.6 0w = 1.99 SER = .0023
p = .733
(5.17)
(9.87)
LPR2 = .358-.0020-1-.008LP-.867(%)4-.Ollt
(1.31) (.95) (.53) (9.90)
32 = .99 F(9,18) = 568.69 SER = .0096 0w = 1.91
o - .785
(6.07)
M (9.33)
In E = -3.939+-1.032 1n Y-.205111(£§)+-.003 1n (E
(5.59) (9.90) w (.95) r
+ .115 1n E_1 - .017t
(.35) (2.92)
32 = .99 F(5,18) = 1916.93 SER = .020

Durbin h.= -.11 p 0

70

(9.89)
an = -.103+.1591nY+.0151n(Br—‘-)—.0291n(§)
E
(3.03) (1.29) (1.73)
+ .676131K_1 + ,008t
“(5.39) (1.09)
R2 = .99 F(5,17) = 33919 SER = .0092 0w = 1.02
p = .227
(1.12)
(9.310)
= * *
Lt Hrt Mt (365/7)
(9.811)
- - E! a
U - (1 LP 100

where, LF LFl + LF2 is the total civilian labor force, and
EM = M + GM is the total civilian employment, and GM is
"non-commercial" employment.

The estimated coefficients of the model have generally
the theoretically expected sign, except for the percentage
change in the price of capital (r) in the price equation,
and the ratio of price of energy to the price of capital,
iln (g?) in the demand for energy equation. Also, the
estimated coefficients cf the model are statistically
significant in most cases.

Autocorrelation is not a significant problem in
the price equation, wage equation, demand for labor
equation, and demand for energy equation, while other

equations of the model were corrected for the presence of

serial correlation in error terms.

9

de

Ca
Si)
192
pre
sta

wit

 

71

In the case of Canada also, for the wage equation
(9.83), the hypothesis that the sum of the coefficients on
fie and lagged w is equal to one cannot be rejected at the
level of significance 8 = 5% (i.e., t-statistic = .92).
This would imply that the long—run Phillips Curve in
Candad is vertical.

An analysis of the dynamic behavior of the model
by examining the implied long-run elasticities of factor
demands with respect to output and factor prices proved the
dynamic stability of the system of equations.

Again, the simulation performance of the model for
Canada is surprisingly good. An examination of historical
simulation of the model over the estimation period 1957 to
1979 would show this. The results of the simulation are
presented in Table 9.2, which shows the RMSE and the Theil's
statistics for each endogenous variable in the model along
with the mean value of each variable.

Low values for RMSE's, relative to the mean value
of the corresponding endogenous variables, would indicate

the sati8factory performance of the model and its ability

to replicate the actual data series.

 

 

N o J M.meVrH.

 

72

 

 

momm. mmHm. memo. mmmm. Hmwm. mwom. umHm. AUDV o AHmmB

mmom. muem. mmbu. ammo. «mmm. mzmH. Homo. AmDv m AHmmH

mooo. :mmo. ommo. mmHo. mHso. mmmo. mmuo. AZDV z AHmmB
boo. omo. «Ho. mm. 50. m:.H mm.H mmzm
mm.: m:.m bm.mH hm.m 0H.m m:.n 5H.: z<mz
xcH mcH JGH D ma 3 m

 

<a<z<u mom 20Ha<aasz n<UHmosmHm so mansmmm

«.3 mam<B

73

C. - United Kingdom

All the equations of the model for the United Kingdom
were estimated using annual data over the period 1958 to
1979. In all preliminary experiments, the following
equation for inflation expectations, which is based on the
empirical procedure suggested by McCallum (1976), performed
best in terms of t—statistics and adjusted R2, and therefore
was used to obtain a proxy for inflation expectations in

the United Kingdom.

(9.01)
p = -.009 + .6051)".l + 1.009p-2 + .0353“l + .026r_2
(1.86) (2.69) (2.67) (.56)
+ .293p — .761p
E-l E-2
(1.26) (3.69)
32 = .73 F(6,15) = 13.36 SER = .026 0w = 2.09

where the numbers in parentheses are t-statistics, and r
and pE are the percentage rate of change in the price of
capital and energy, respectively. All the behavioral
equations of the model were estimated using ordinary least
squares (OLS), with a Cochrane-Orcutt adjustment for those

equations for which the presence of autocorrelation proved

to be significant.

79

(9.C2)
6 = -.0006+-.121p_l+-.5159-.003£+-.2lspE-.306§
(.92) (3.89) (.39) (2.23) (2.25)
-.017c'1_l+.0002t
(.10) (.18)
32 = .99 F(7,l2) = 99.19 SER = 0.19
Durbin h = .28 p = 0
(91C3)
9 = —.007-+.336w_1-.002U-.0580+-.99lEP+-l.l23pe
(1.92) (.95) (2.73) (1.32) (9.55)
32 = .73 F(5,13) = 10.65 SER = .023 0w = 2.33
p = -.539
(2.75)
(9.C9)
1n L = 11.79+ .931 1nY- .0031n(¥) - .109ln(-;—)
(1.79) (1.19) (.99) B
+ .0711nL_l - .012t
(.26) (2.33)
32 = .31 F(5,13) = 16.55 SER = .019 0w = 1.72
p = .588
(3.17)
(9.C5)
1n Hr = 3.302 + .253Yg- .005t
(2.97) (13.12)
32 = .92 F(2,l7) = 117.9 SER = .009 EN = 1.93

75

(9.C6)
LRPl = .303 - .006t
(9.99)
32 = .93 F(l,l7) = 937.73 SER = .009 0w = 1.00
p = .732 '
(9.68)
(9.C7)
LPR2 = .916- .003U-.093LP-.037(§) + .008t
(1.89) (1.56) (.59) (3.85)
R2 = .97 F(9,l9) = 169.10 SER = .003 0w = 1.02
p - .323
(6.32)
(9.03)
‘PE pE
1n E = 2.653+-.2971nY-.0161n(;r)-.0191n(7:)
(.73) (.08) (1.18)
+ .3391nE__1 - .002t
(1.16) (.22)
R2 = .73 F(5,13) = 13.97 SER = .029 0w = 1.53
p = .956
(2.29)
(9.C9)
ln K = -.793+-.0631nY - .0101n(§l)+-.0111n(§)
E
(2.32) (.63) (.69)
+ 1.0031n1<_l - .003t
(9.. 75) (..31)
32 = .99 F(5,13) = 90253 SER = .002 0w = 2.26
p = .820

(6.25)

 

III

76

(9.C10)

= n *
Lt Hrt Mt (365/7)

(9.011)

c:
I

M t
(1 if) 100

where, LF LFl + LF2 is total civilian labor force. The
estimated coefficients of the model generally have the

"a priori" expected sign, except for the percentage change in
the price of capital (r), and the ratio of industrial
production index to its exponential trend (q) in the price
equation, and the ratio of price of capital to wage rate in
the demand for capital equation. Also, the estimated
coefficients are not statistically significant in some cases.
However, the F-statistics for overall regression are highly
significant for all equations.

Autocorrelation is not a significant problem in the
price equation, and the equation for average hours of work
per worker, while the other equations of the model were
corrected for serial correlation in the error terms, using
Cochrane-Orcutt technique.

In the case of the United Kingdom, for the wage
equation (9.C3), the hypothesis that the sum of the
coefficients on lagged w and 3 is equal to one can be
rejected at the level of significance 8 = 5%, (i.e.,
t-statistic = 2.033). This would imply that the longvrun
Phillips Curve in the U.K. is not vertical, and a negative
long-run trade-off between inflation and unemployment does

exist.

77

An analysis of the dynamic behavior of the model,
by examining the implied long-run elasticities of factor
demands with respect to output and factor prices, proved the
dynamic stability of the system of equations for U.K.

The simu1ation performance of the model for U.K. seems
very reasonable. An examination of historical simulation of
the model over the estimation period 1961 to 1979 would
support this claim. The results of this simulation are
summarized in Table 9.3, which shows the RMS error and the
Theil's statistics for each endogenous variable in the model
along with the mean value for each variable.

Low values for RMSE's, relative to the mean value
of the corresponding endogenous variables, indicate the
satisfactory performance of the model and its ability to

replicate the historical data series.

D. « Germany

All the equations of the model for Germany were
estimated using annual data over the period 1959—1979.
In all preliminary experiments, the following equations for
inflation expectation, which is based on the empirical
procedure suggested by D.J. Mullineaux (1980), performed
best in terms of t—statistics and adjusted R2, and therefore

was used to obtain a proxy for inflation expectations in

Germany.

78

 

 

mmoH. :Nmm. :zmm. omom. ommm. mmmm. «mom. H039 o AHmmB

mmmm. mmzo. mmnH. mmmm. :ozo. :ooo. momo. AmDu m AHmmB

mmmm. Hmoo. ooo. mmmH. mHoo. HHoo. Hooo. AZDV x AHmmB
OHo. bmo. mHo. mam. mo. :b.m mm.H mmxm
sm.m mo.m mh.>H mH.m ou.H mm.m hm.s z<mz
XCH mcH ASH D m4 3 a

.x.D mom 20HH<ADZHm A<UHMOBMHE mmB mo mBADmmm

m.: mqm<8

P

R2

DW

where the

79

(9.Dl)

= .015+1.326p_1--1.929p__2+1.369p__3--.501p__l4
(10.10) (6.09) (9.02) (2.93)

+ .079M_ + .096M_ - .002M_ - .176M_

1 2 3 9

(1.05) (.69) (.03) (2.38)

.39 F(8,11) = 19.23 SER = .010 N = 20

2.68

numbers in the parentheses are t—statistics and M

is the percentage rate of growth of money supply.

All the behavioral equations of the model were

estimated

using ordinary Least Squares (OLS), with a

Cochrane-Orcutt adjustment for serial correlation of the

error terms.

P

DUI

DW

DW

(9.D2)

= .0013 + .539p_l + .091w — .019r + .036pE
(3.03) (.93) (2.07) (2.12)

+ .083q + .179q_1 + .0005t
(1.12) (2.25) (1.30)

.89 F(7,10) = 20.79 SER = .006 N = 18

= 2.13 p = -.627
(3.91)
(9.03) .
= .067 — .1959)“1 . .0093U — .0032U — .139LP + 1.27758
(.79) (1.73) (.73) (.92) (2.91)

1.52 F(5,12) = 9.70' SER = .013 N = 13

1.83‘ p = .373
(1.71)

80

(9.D9)
1n L = -.921 + 1.2561nY - .6711nY_l - .0031n(§)
(6.92) (1.99) (.68)
- .0991n(3L) + .598lnL 1 — .021t
. pE -
(1.53) (2.32) (1.39)
R2 = .95 F(6,12) = 59.37 SER = .0125 N = 19
DW = 2.52 p = 0
(9.05)

In Hr = 3.790 + .976Y - .009lt

(9.00), (8.19)

32 = .35 F(2,16) = 53.13 SER = .0115 N = 19
0w = 1.53 p = 0
(9.D6)
LPRl = .366 — .0087t
(3.32)
32 = .99 F(l,16) = 1799.69 SER = .0093 N = 13
DW = .37 p = .300
(5.66)
(9.D7)
LPR2 = .366 — .0012U - .0071LP + .079(%) + .002t
(9.26) -(.90) (1.15) (1.69) (.99)
32 = .97 F(9,13) = 127.09 SER = .002 N = 13
0w = 1.23 p = .363

(7.23)

81

(9.D8)
pE pE
ln E = -3.999 + 1.9631nY - .1011n(;r) + .017ln(;r)
(3.35) (1.01) (.79)
- .1971nE_l - .020t
(.68) (1.76)

R = .98 F(5,13) = 233.95 SER = .0239 N = 19
h

(9.09)
1n K = .317 + .032lnY + .0161n(g:) — .0131n(§)
(.67) (1.00) (1.10)
+ .9031n1<_1 + .0019t
(10.79) (.31)

R = .99 F(5,12) = 9511.20 SER = .003 N = 18

DW 3 1.31 p = .523
(2.60)
,(9.D10)
: * *
Lt Mt Hrt (355/7)
(9.D11)
- EL 9
U - (.1 - LP) 100

where LF = LFl + LF2 is the total civilian labor force. The
estiInated coefficients of the model generally have the

"a priori" expected sign, except for the percentage change

in the price of capital (1}) in the price equation, the
perclentage change in labor productivity (LP) in the wage
equation, and the ratio of price 'of energy to price of capital

go pE ~ . .
- Ode ln(—‘r-;-) 1n the demand for energy equation. Also, the

82

estimated coefficients of the model are statistically
significant in most cases.

Autocorrelation is not a significant problem in the
demand for man-hour equation, average hours of work per
worker equation, and demand for energy equation, while the
other equations of the model have been corrected for serial
correlation in the error terms.

In the wage equation (9.D3), the hypothesis that
the sum of the coefficients on pe and lagged w is equal to
one cannot be rejected at the level of significance 0.: 5%,
(i.e., t-statistic = .095). This would imply that the
long-run Phillips Curve is vertical.

An analysis of the dynamic behavior of the model, by
examining the implied long-run elasticities of factor demands
with respect to output and factor prices, proved the dynamic
stability of the system of equations for Germany.

The simulation performance of the model for Germany
is rather good. An examination of historical simulation of
the model over the estimation period 1962-1979 would show
this. The results of this simulation are presented in
Table 9.9, which shows the RMSE and Theil's statistics for
each endogenous variable in the model along with the mean
value for each variable.

Low values for RMSE's, relative to the mean values of
the corresponding endogenous variables, would indicate the
satisfactory perfOrmance of the model and its ability to

replicate the actUal data series.

83

 

 

:amm. maaa. maam. Heme. maaa. asaa. mmma. mass m Amuse

mama. mmma. mmma. maam. mama. maaa. mama. mass m Amuse

Hamm. msma. maaa. amma. aama. aaaa. maaa. Azaa z nmmms
aaa. ama. mHa. 9a. am. mm.m mma. maze
mm.m Hm.a ma.mm mm.m ma.mm am.m mm.s z<mz
xcm mam sum 3 an 3 a

 

Mz<zmmw mom ZOHB<ADXHw A<UHmOBmHm ho mBADmmm

3.3 mam<fi

89

E. - France
All the equations of the model for France were

estimated using annual data over the period 1959-1979. The
following equation for inflation expectations, which is .
based on the empirical procedure suggested by McCallum (1976),
was selected in the preliminary tests, based on t-statistics
and adjusted R2, and therefore was used to obtain a proxy
for inflation expectations in France.

’ (9.E1)

p = .022 + 1.29p._1 - .70p._2 - .21p-3 + .07r__l

(3.78) (1.55) (.99) (1.16)

a .0001:%_2 - .01r_3 + .099pE— + .023pB- + .19PE_

1 2
(.002) (.19) (1.53) (.25) (2.01)

3

32 = .79 F(9,9) = 3.35 SER = .013 CW = 2.23

where the numbers in parentheses are tbStatistics, and r and
pH are the percentage change in the price of capital and
energy, respectively.

All the behavioral equations of the model were
estimated using Ordinary Least Squares (OLS), with a Cochrane—

Orcutt adjustment for serial correlation of the error terms.

(9.82)

’U
I:

.0005 + .989.p._l + .265w H .015r + .052pE v .099q
(3.08) (2.35) (.60) (1.27) (1.91)

9 .053q_l+.0002t
(.33) (.19)
R = .86 F(7,ll)

ll6.96 SER = .011 N

19
Durbin h.= .18 p

II
C

85

(9.33;
w = .190 - .063w_1 - .012U + .0020 - .771LP + .392pe
(.29) (1.79) (.10) (1.91) (.69)
’2

R = .59 F(5,12) = 5.09 SER = .026 -N = 13

0w = 1.79 p = .356
(7.03)

(9.E9)

ln L = 10.60 + .9961nY — .0351n(¥) + .00021n(gL)
E

(3.95) (3.39) (.006)
+ .0391nL_l - .022t
(.26) (9.29)
‘2

R = .91 F(5,13) ? 36.27 SER = .007 N = 19

Durbin h 8_1.01 p = 0

(9.E5)
1n Hr = 3.909 + .299Y - .010t
(3.13) (9.66)
R2 = .99 F(2,15) = 595.99 SER = .005 N = 13
EN = 2.09 p = .757
(9.92)
(9.E6)
LPRl = .797 - .006t
(l25.99)(19.98)
32 = .93 F(1,16) = 781.6 SER = .005 N = 13
0w = 1.79 p = .500

(2.95)

 

86

(9.87)

LPR2 = .332 - .003U - .003LP - .050(§) + .012t

(.96) (.36) (1.15) (2.26)

32 = .93 F(9,l3) = 137.36 SER = .009 N = 13
0w = 1.36 p = .663
(3.30)
(9.E8)
ln'E = -13.792 + 3.697lnY - 2.101nY_1 — .0091n(:§)
(5.95) (3.57) (.06)
.0131n(3E) + .2161nE - .090t
r —l
(.50) (.91) (2.02)
32 = .99 F(6,11) = 209.3 SER = .029 N = 13
EN = 2.15 p = -.599
(3.13)
(9.E9)
In K =A -.559 + .1081nY + .1511nY_1 + .0111n(§:)
(1.73) (2.68) (.93)
- .0321n(§) + .7621n1<_1 - .009t
(2.27) (1.91)
32 = .99 F(6,11) = 92320.0 SER = .003 N = 13
EN = 1.53 p = .363 .
(7.23)
(9.310)
Lt =, Mt * Hrt * (365/7)
. (9.311)
U - (1 4i§%) * 100

87

where LP = LFl + LP2 is the total civilian labor force. The
estimated coefficients of the model generally have the expected
theoretical sign, except for the percentage rate of change

of the price of capital (r), and the percentage rate of change
in the ratio of industrial production index to its exponential
trend (q) in the price equation. Also, the percentage rate

of change in labor productivity (b?) in the wage equation,

and the ratio of wage rate to the price of energy, 15(5?) in
labor demand equation, and the ratio of the price of energy

to price of capital ln(§§) in the demand for energy equation
do not correspond to the expected "a priori" sign. Also, the
estimated coefficients are not statistically significant in
some cases. However, the P-statistics for the overall
regression are highly significant for all equations.

Autocorrelation is not a significant problem in the
price equation anui the demand for labor equation, while
the other equations of the model were corrected for serial
correlation in the error terms, using Cochrane-Orcutt
technique.

In the wage equation (9.E3), the hypothesis that the
sum of the coefficients on $8 and lagged w is equal to one
can be rejeCted at the level of significance a = 5%,Ci.e.,
t-statistics = 2.37). This would imply that the longvrun
Phillips Curve is not vertical.

An analysis of the dynamic behavior of the model, by
examining the implied long—run elasticities of factor demands

with respect to output and factor prices, proved the dynamic

stability of the system of equations for France.

88

The simulation performance of the model for France
seems very reasonable. An examination of historical
simulation of the model over the estimation period 1961—1979
would support this claim. The results of the simulation are
summarized in Table 0.5 which shows the RMSE and the Theil's
statistics for each endogenous variable in the model along
with the mean value for each variable.

,Low values for RMSE's, relative to the mean value of
the corresponding endogenous variable, would indicate the
satisfactory performance of the model and its ability to

replicate the actual data series.

F. - Japan

All the equations of the model for Japan were
estimated using annual data over the period 1350 to 1979.
In all preliminary experiments, the following equation for
inflation expectations performed best in terms of tcstatistics
and adjusted 82, and therefore was used to obtain a proxy

for inflation expectations in Japan.

(u.F1)
p = .015 + .00p_1 + '12M-1
(2.16) (1.05)
32 = .16 PC2,23) = 3.33 SER = .032 Dw = 1.33

where the numbers in parentheses are tsstatistics, and h is
the percentage rate of change in money~supplyx

The behavioral equations of the model were estimated
using ordinary least squares (OLS), with a Occhrane60rcutt

adjustment for those equations for which the preSence of

 

mssm. asmm. 35mm.

 

sHmH. «mmm. :mmm. Homm. floss 0 qume
mwma. mooo. omma. Head. mmao. mama. smug. Amos m quma
sass. mmoo. mmflm. wads. ommm. mess. :mflo. Azpo 2 quma
was. Has. moo. mm. mNH. mm.~ mm.H mmxm
mm.~H mm.m mm.sa ms.~ No.sa ms.oa m:.m z<m2
xcH mca qua : mg 3 a

 

moz<Mh MOh ZOHH<ADZHw A<0Hm09mHm ho mHADmmm

m.: mqm<8

90

autocorrelation were proven to be significant.

(0.62)
p = .000 + .272p_l + .2u3w - .02ur + .191pE + .215q
(1.13) (1.78) (.72) (2.00) (2.81)
+ .017c1_l + .0003t
(.23) (.30)
32 = .75 F(7,16) = 10.69 SER = .013 DW = 2.25
p=0
(H.F3)
w = .020 + .771w_l - .0190 - .0300 + .182LP + .safie
(5.7“) (1.01) (1.50) (1.03) (1.10)
32 = .31 F(5,18) =_20.91 SER = .021 Durbinrl=ln31
o=0
(0.?0)
1n L = 11.70 + .2381nY - .0361nY_1 - .0161n(¥)
(2.23) (0.17) (1.30)
+ .0331n(lL) + .0751nL + .015t
PE ‘1
(2.02) (0.00) (3.30)
R2 = .92 F(6,l7) = 02.31 SER = .010 Durbin}1=.93
0:0
(H.FS)
1n Hr = 3.86 + .3311'0 .007t
(H.35) (7.15)
2

IR = .36 FC2,20) = 27H-18 SER = .012 DW‘= 1.59
p = .623

(3.82)

91

(H.F6)
LPRl = .350 - .002t
(2.63)
32 = .90 F(l,2l) = 326.27 SER = .000 DW = 1.25
p = 310
(6.72)
(N.P7)
LPR2 = .506 + '031U—1 + .075LP - .066(§) - .000t
(3.29) (.70) (1.31) (1.H3)
32 = .97 P(0,13) = 166.15 SER = .005 DW = 1.57
p = .561
(3.25)
(4.F8)
1n E = -9.75 + 1.72 lnY + .0351n(§§) + .0271n(2§)
(3.50) (1.37) (1.70)
+ .0301n1:_l - .063t
(.66) ' (1.76)
32 = .99 F(5,17) = 2703.03 SER = .020 DW = 2.10
p = -.039
(2.3“)
(H.F9)
In K = -.392 + .2651nY + .0311n(§;) 0 .039 meal)
(2.59) (.03) (.07)
+ .7361nl<.._l « .000t
(5.31) (.27)
E? =,.99‘ F(5,18) = 13860.8 SER = 1013' Durbin11=.68

p=0'

92
(LI.F10)
._. a: i:
Lt Mt Hrt (365/7)
(0.611)
u = (1 - 31) * 100 0
LP °

where LF = LFl + LF2 is the total civilian labor force. The
estimated coefficients of the model generally have the
theoretically expected sign, except for the percentage change
in the price of capital (:3) in the price equation, and the
ratio of wages to the price of energy, ln(%'g) in the demand
for labor equation. Also, the coefficients in the equation
for labor participation of "secondary worker" LPR2, do not
have the "a priori" expected sign. The sign of the coefficient
of the ratio of price of energy to the wages, 1n(.%§), and to
price of capital, 1n.(.-P%}:-) in the demand for energy equation
are also wrong. The estimated coefficients of the model are
statistically significant in most cases.

Autocorrelation is not a significant problem in the
price equation, wage equation, and demand for labor equation,
while other equations of the model have been corrected for
serial correlation of the error terms.

Using the above system of equations, historical

simulation for Japan failed to converge.

 

CHAPTER FIVE
THE EFFECTS OF ENERGY PRICE SHOCKS

5 . l - Introduction

 

. In this chapter, we will examine the effects of
energy price shocks since 1973 on the economies of five
industrial countries included in this study. The
resulting effects of energy price shocks are obtained by
simulation of the model developed and estimated in the
previous chapters, in turn, for each country. As a first
step, we will prepare a base case'describing the possible
evolution of each country's economy since 1973 in the

absence of the abrupt increases in the price of energy.

5.2 «- The Effects of Energy Price Shocks

 

A. ‘ U.S.A.

Table 5.1 summarizes the actual performance of the
(1.8:. economy over the period 19.73 to 1979. Values of the
endogenous variables of the model obtained in the historical
Simulation are given in Table 5.2. As it can be seen
from these values, the model's performance and its ability
to I‘ePlicate the actual data for the U.S. economy has been

rather» good .

To see the effects of energy price shocks on the

93

a
owns doom 0 m
Aanpx CH I ex ca 0. max "mm Umcflwmmv ma
003m aoaymNeafipp Hmpflamo u x
AcowvaESmcoov ocmEmQ mwnmcm u m
.Aoosom gonna cmew>wo mo av mumm HQmEAOHQEmca u D
awpmcm mo moans n ma
vcmfi>oamsm Hmpos u z
.Apsoacwz 9mm psmvzov muw>wposoopm sonma n ma 096:3

 

9M

 

sm.:~ H.Ho: ms.m No.ms om.m mNOHm 5H.m :~.H :m.m sm.m msma
as.m m.~Nm mm.~ ma.ms No.w Hosms mH.m =~.H :m.m mo.s mums
ms.mfl N.Nom ss.~ mm.ms 30.0 mazms so.m NN.H Ho.m mm.m puma
om.m m.ms~ sm.H Hm.:s mo.s seams om.» mH.H ss.m mo.m mesa
sm.sfl H.m:~ Hm.H as.os m:.m swoop mm.s mH.H mm.m mo.oH mesa
oH.mm m.mo~ as.m ms.ms mm.m mmsas mm.s mH.H mo.m Nm.m shad
:~.mH n.3ma mm.: Hm.:s mm.s ssmos «5.5 mH.H mm.m ms.m mesa
Awe m Ase has Assess a “so Ass

mm a max m o 2 “3 m z . a

 

H6 5mg

msumsma szozoom .m.p mme so moz<zmommmm q<aeo<

95

TABLE 5.2
HISTORICAL SIMULATION, U.S.A.

 

w _w_ LP M U E KUR
(3) (5) p (1000) (3)' .

 

21973 5.38 7.38 .13 7.70 70976 4.69 73.12 ".77

1970 10.06 9.07 .12 7.03 72721 5.20 72.01 3.29
1975 3.50 3.15 .11 7.53 71320 3.01 70.23 2.11
1976 5.01 3.30 .10 7.79 72935 7.37 71.30 2.99
1977 3.15 7.70 .15 7.97 76032 6.61 70.30 3.63

1&378 5.92 9.01 .19 8.15 78306 5.87 77.00 3.86

H F‘ +4 H +4 +4 H

1&379 7.8” 9.72 .21 8.10 80932 5.55 79.01 3.H7

 

where LP = Labor Productivity (Output per Manhour)
M = Total Employment
U = Unemployment Rate (96 of Civilian Labor Force)
E = Energy Demand ‘
KUR = Capital Utilization Rate
(Deflned as: KUR = 1n Kt - 1n Kt-l)
% = Real Wage

 

96

U.S. economy, we deve10p a base case describing the evolution
of the economy since 1973 in the absence of sharp increases
in the price of energy. In preparing this base case, we

make the assumption that the nominal price of energy
increases at a rate of only five percent per year, which

is the average of the inflation rate over the five preceding
years. The model's base case projections are given in

Table 5.3.

The energy price shocks generate a sizable inflationw
ary pressure in the U.S. economy, as can be seen from Table
5.0, which shows the deviations of the historical simulation
of the model from the base case forecasts.

The predicted effects of higher energy priceson the
price level is substantial in the first two years of the
energy price shocks. In terms of inflation, the effect is
by far largest in 1970 with a 0.75 percentage point added
to the inflation rate.

The higher energy prices have induced more labor
intensive methods of production, as it is apparent from
increases in the level of employment, and thereby a
reduction in the rate of unemployment — given that we have
kept the level of output fixed in simulations of the model.
Also, higher energy prices have resulted in a reduction
of energy consumption. However, the reduction in theiCapital
utilization rate does not seem to have been significant.

Finally, higher energy prices have resulted in

reductions~in labor productivity and real wages.

97

 

 

a
mmmz ammm u m
Aaupx :H 1 us ca n. max ”mm emcfimmn
mpmm cowvmuwawvo HMPHQMU n max
panama zmpmcm u m
Amosom poem; cowafi>wo mo we ovum ucoesoaasmc: u D
hwpmcm mo mofinm n ma
pcmssoaaam Hopos u z
Ansogcmz 9mm wrapsov Apfi>wso=ponm nonmq n ma 090:3
o.m m.mma mm.m Ho.sm mm.oH assms sm.m mm.a os.m ma.m msma
o.m m.mmH Hm.m Hm.mm os.m momms om.m sm.H m:.m mm.m msma
o.m :.smH mmvm HH.om sm.oH memos mm.m mm.a oa.m H:.m ssma
o.m m.::a mo.m mm.ms mo.HH Hosmm Hm.m om.H s:.m :m.m msma
o.m m.smH sH.m ma.:s .oo.aa oaasm oo.m sH.H Hm.m ms.m msmH
o.m m.omH :m.m s:.:s sm.s mmmmw mm.s mH.H Hm.s Hm.m :sma
o.m m.:ma ms.: mm.ms ma.m mmmos ms.s mH.H mo.s mm.: msma
CS
. va soooav a
mm mm max m D 2 mg m. AWV Awe

 

.<.m.D I wwmmzm mo MUHmm zH mmm<mmozH BmDmm< mo mozmmm<

m.m mAm<B

 

 

98

mmm3 Hmmm

Aanpx ca 1 0x 2H n max

”mm conflmmav

3Hl

 

 

opmm cowvmnwawpb Hapwmmo u m:
fiscapmssmcoov pcmsmn smsmcm u m
opmm pcmemoamEmca u D
mwpocm mo moahm u mm
, “coshoamem HMpos u z
Apsonsmz pom vsmpsov.spwsmvonp09m gonna n ma mamas
mm.mn ma.mn om.:| floss mm.ms mm.ms No.0 Hs.H msma
mm.as ms.s1 mm.ms momm mm.ss om.mu m:.os :o.o wsma
mm.as om.sn mm.m| :m:m HH.sn om.mn wm.on :s.o ssma
sm.H| o~.mu ma.mn :mm: m~.mu oo.mn mm.o so.a msma
ms.~n mH.mn mm.~n :HN: mm.ms mH.mu :N.H ms.m msma
om.Hn om.mn m:.~u momm Hs.:u Hm.mn mm.H ms.: :sma
Hm.u :m.1 m:.1 Hmm ss.s o.o sm.ot om.o msma
“we c: 888 m. :3 2:
x m a x mg 3 3 a
sHo w .mHQ w .mHQ .mm< .mHQ .mm< mHQ w mHQ w .mHQ .mm< .mHQ .mm<

 

.<.m.D .mHm<ommOh mm<o mm<m 20mm ZOHB<H>MQ

:.m mqm<B

 

— —_. _,7

99

B. - Canada

Table 5.5 summarizes the actual performance of
Canada's economy over the period 1973 to 1979. Table 5.6
shows the values of the endogenous variables of the model
obtained in the historical simulation of the model for
Canada.

To see the effects of higher energy prices on the
Canadian economy, we develop a base case describing the
evolution of the economy since 1973 in the absence of
sharp increases in the price of energy. In preparing this
base case, we make the assumption that the nominal price
of energy increases at a rate of four percentage points
per year, which is the average of the inflation rate over
five preceding years. The model‘s base case projections
are given in Table 5.7, and1fluaTab1e 5.8 shows the deviations
of the historical simulation of the model from the base -
case forecasts.

As can be seen from Table 5.8, the energy price
shocks have generated a sizable inflationary pressure in the
Canadian economy.

As was the case in the.U.S.A., the predicted effects
of the higher energy prices on the price level is substantial
in the first two years. In terms of the rate of inflation,
the effect is by far largest in 1970 and 1975 with.7.62 and
5.00 percentage point added to the inflation rate,
respectively.

The higher energy prices have also induced more

 

100

 

m
mwm3 Hmmm u m
Aavpx ca 3 vx ca 0 .mox “mm omcwmmne m:
ovum cospmnwafius Haywamo u x
Acowuaaomcoov panama swsocm u. m
Amopom poem; unfiafl>wo mo we mpmm ucmEsoamEmc: u D
smpmcm mo mowhm 0 mm
pcoazoamEm Hmvos u z
Assoccmz pom assures spw>wposp09m sons; 0 mg mamas
m.Hmm no.3 Hmmm m:.s mmms mm.m «so. Hm.s ma.ua msma
:.msw :H.: mmmm sm.m mams mm.m 3:0. mm.m mo.m msma
m.::m m:.: mamm Ha.m wads ms.m 12o. m~.m H~.m .ssma
m.oam ~m.: mmsm ma.s swam Hm.m mzo. mw.ma o~.s msma
s.mmH :m.m mmmm mm.m smmm om.m oso. H:.mH Hm.HH msma
:.mma m:.m mm:m :m.m mssm om.m mmo. so.:H sm.:a :smH
~.sHH om.m mzmo mm.m sm:m sH.m mmo. mm.oa mm.m msma
m . ms 2: 825 m 2; so
a x m a z .3 z s a

 

mslmsmH MZozoom z<HQ<z<o mo moz<ZMOhmmm A<DBU<

m.m mqm<B

 

101

TABLE 5.6

HISTORICAL SIMULATION - CANADA

 

LP M U E K

 

 

P w ‘1 UR
(%) (%) p (1000) (%) (%)
1973 7.85 8.96 .037 6.12 6530 0.72 6381 5.70
1970 10 38 12.52 .037 6.19 6876 0.69 6053 5.60
1975 11.23 12.65 .037 6.28 6892 7.16 6365 5.01
1976 7.73 11.55 .038 6.50 7009 6.98 6600 0.82
1977 8.35 8.05 .038 6.59 7253 7.88 6626 0.18
1978 7.39 7.97 .039 6.71 7009 7.37 6719 3.85
1979 8.87 9.39 .039 6.80 7650 7.57 6789 3.75
where LP = Labor Productivity (Output Per Manhour)
M = Total Employment
U = Unemployment Rate (% of Civilian Labor
Force)
E = Energy Demand (Consumption)
KUR = Capital Utilization Rate

E
P

(Defined as: K =

UR )

1n Kt — ln Kt-l

Real Wage

 

102

 

a
mmm3 Hmmm u m
AH-px as - px ca n max “mm umcwmmno x:
09mm cowsmaaaflpb Hopflmmo u M
AGOfivQEchooo pcmEmQ smhmcm u m
mpmm vcoemoaoEmcD u D
mmnmcm mo moans 0 mm
pcossoanEm Hmpos u x
Ansoccmz pom psmpsov spfi>wuosoopm gonna n ma 090:3
ma.mma mm.m soss om.aa :mos m:.s aso. mH.m :o.: msma
mo.oNH :o.: mmos mo.oa mooo om.s «so. oo.~ Na.: osma
mo.:~H o:.: os:s oN.HH zaso mo.s «so. mo.m mm.: ssma
:H.omH ma.m mmms oo.o :Hoo mm.o «so. ms.s os.m osoa
mm.mHH N:.m msmo mo.m ammo No.o aso. ms.o oH.o msoa
oo.HHH so.o mmmo ms.o saoo m:«o oso. oo.oa os.o :soa
Ho.ooa oo.m oomo o~.m sono ma.o omo. oo.o om.m msoa
m “we Awe Access 0 Ass “so
a x m o x mg 3 z a

 

<Q<z<o u wommzm mo MUHmm zH mmm<mmUZH memm< ho mozmmm<
s.m m4m<9

owns Hmmm
Hupx ca 1 ex :H n max “mm omcwmmao

A
mpmm cowvmnwafipa Hopwmmu max

 

103

 

Aco«pmeawcoov ocmsma smpocm m m
Amosom pocma mafiaw>ao mo av mpmm Homemoamsmca u a
smsmcm mo mowpm u mm
vcmamoamam Hapos u z
Asaozcmz pom HoovSOV mpw>fivosoopm sonma n ma 090:3
no.2: mm.mau Ho.:1 oao am.su oo.:u o~.o mo.m msma
os.:u mm.HHu mo.mu omm om.sn :H.su Hm.m sm.m osma
os.os om.HHs om.mn mom mo.sn mm.mu oo.: oo.: ssoH
H:.on No.oHu mo.mn mm: om.oa mm.on mo.m mo.m osma
om.su ms.or s:.ur oom :H.ms os.os om.m :o.m msma
:s.su mo.ot mo.~n omm ms.mn om.su ms.H mo.s :smH
om.mn mo.an am.| mo so.) mo.~l NH.1 om.m msma
“my 883 Ame to :3
x m . a z .3 3 3 a
.mHa w .mHa a .mHa .mm< .mHa .mm< .mHa m .mHa a .mHa .mm< .mHa .mm<

I!"

<a<z<o .mBm<ommom mm<u mm<m Zomh ZOHB<H>MQ

o.m mAm<B

100

labor intensive methods of production in Canada; as it
can be seen from the increase in the level of employment,
given that we have kept the level of output fixed in the
simulation of the model. Also, higher energy prices have
resulted in a reduction in the level of energy consumption,
as well as a substantial reduction in the capital utilization
rate.

Finally, higher energy prices have led to lower

labor productivity and lower real wages in Canada.

C. - United Kingdom

Table 5.9 summarizes the actual performance of the
U.K. economy over the period 1973 to 1979. Table 5.10
shows the values of endogenous variables of the model for
the U.K., obtained in historical simulation of the model.
Table 5.11 shows the model's projections for a base case
describing the possible evolution of the U.K.‘s economy
in the absence of abrupt increases in the price of energy.
Finally, Table 5.12 shows the deviations of the historical
simulation of the model from the base case forecasts.

The energy price shocks have generated a rather
large inflationary pressure in the U.K. economy, and the
results are very similar to those obtained for the U.S.
and for Canada.

The substantial increases in theprice of energy:
in the U.K. do not start until 1970. The predicted effects
of higher energy prices on the price level is substantial

in 1975 and 1976. In terms of the rate of inflation, the

 

105

.pr

u

3..

owns doom

 

 

x 5 .. x 5 .9... 32:38 xa
mvmm cofiPMNwvaa Hmvwmmu u x
AcowumEdmcooo pamEmQ xmsocm n m
“mopom momma :mwaw>fio mo we mpmm womEmoamEmca u a
smsmcm mo mowsm 0 mm
pamEsoamEm Hwyos n z
nanomcmz 9mm waQPSOQ >HH>HPosooxm momma n ma muons
:.NH o.moa o:.~ mooo oo.m ooswm oo.~ HH.H am.mH mo.~H osoa
:.o H.Hma oo.m sumo mo.m momma oo.m :o.H mo.:a oo.HH osoa
m.sH H.omH ms.m ammo ma.o Hoomm mo.m Ho.H o~.oH so.HH ssoa
m.sH m.sHH mo.~ soao mo.m mmmmm Ho.m mo.H oo.oH os.mH osma
o.mm o.ooa mo.m Noam :N.: osmum om.H oo.H o~.mm mm.sm msmH
m.mm :.ms o~.m oomo om.~ Hosmm mo.a Ho.H oH.sH oo.oa :soH
sm.a H.oo mm.m sooo mm.~ momma mm.H Ho.a so.~H mm.s msma
m... m “we so 883 m 2: to
a a x x a 2 n3 3 3 a

 

m.m mAm<B

mslmsma axozoom .x.a mo moz<zmommmm A<DBQ<

 

106

TABLE 5.

HISTORICAL SIMULATION. U.

 

 

 

p w 3 LP M u E KUR
('90) (00) p (1000) 96) (90)
1973 10.23 20.81 1.02 1.93 22025 .02 8763 3.52
1970 15.16 12.18 1.00 1.92 22901 .88 8767 3.30
1975 22.69 20.51 .98 1.90 22615 .39 8500 3.02
1976 17.85 21.03 1.01 2.01 22370 .03 8090 2.95
1977 10.13 16.17 1.02 2.06 22300 .92 8009 2.75
1978 8.35 10.25 1.00 2.08 22673 .30 8638 2.70
1979 11.68 15.55 1.08 2.11 22720 .63 8778 2.50
where LP = Labor Productivity (Output Per Manhour)
M = Total Employment
U = Unemployment Rate (% of Civilian Labor
Force)
E = Energy Demand (Consumption)
KUR = Capital Utilization Rate

(Defined as:

Real Wage

KUR

ln Kt - 1n Kt_

m...

puma/\rh

 

m
mwmz Hmom u m.

H ma

0x as I 0x 3x n x 3mm omcxcmoo
mpmm cofiumnwafipa Hopflomo
ACOflPaEDmcooo panama swsmcm

Amosom xocma comaw>wu mo so wwmm ucofionoEmca

A
max
m
a

smxmcm mo moflsm mm
usmEmonEm Hmpos u z
Assoccmz pom psapaoo spw>flposoosm momma ma mamas

 

107

 

o.m m.ow am.a :mom om.m NJHHN mm.m 5N.H mm.mx mm.m msmx
o.m s.ms HH.N moms Hm.m smxxm mm.m ma.x mm.m N:.m msmx
o.m H.ms ma.m cams mm.m msmom om.m ox.x :m.xx mm.m ssmx
o.m w.mw m:.m mmwm ws.m momma ma.m NH.H mx.mx mm.HH msmx
e.m m.©w mm.m some mm.s ammxm mo.m mo.H Hm.mx mx.:x msma
C.m x.ms mo.m mama m:.: momma 03.x so.H m:.mx m:.:H :sma
um.x x.om mm.m moss No.3 mmzmm mm.x No.x Hm.o~ mm.ox msma
Mme a WWV Axe xoooxa m Axe Axe
o a x m a 2 ma 3 a a

 

.x.a I wwmmzm mo monm ZH mmm<meZH Bmamm< mo mozmmm<

HH.m mqm<9

108

 

 

 

ommz ammm u ,m
Aarpx ca - Px ca n max .mm amcaxmqa ma
mumm COwmewawva Hopaamo u
ACOameamcooo ocmsma swamcm u m
Amoxom Locma coaaaaau no so mpmm pcmExoamEoca n a
mwsmcm mo wofism n ma
“amaonoEm aspos n z
Axaoacmz pom pampaoo spw>avoaooxm soama n ma mamas
ma.mm mm m- sm.:- mama :m.m- om.ma- mm.N ma.m msma
ow.om om.mI ao.aI osaa ms.oI ao.maI om.o mm.a osma
as sm mm.mI os.mI moma om.oI so.maI mo.: ms.m ssoa
o:.am ao.aI mm.mI :oma mo.ma mo.mI oo.: mm.o osma
mo.sa mm.aI oo.mI mmoa m:.:I om.oI om.: om.o msoa
om.m om.I mm.aI osm :m.mI :m.oI mm.sI mo. :sma
o o o o o o o o msoa
“we Axe xoooav m Axe Axe
-.x a a 2 ma 3 3 a
.amm a .mHa m .mHa .mm< .mHa .mm< .mHa w .mHQ w .mHa .mm< .mHa .mm<
.x.a I msm<ommom mm<o mm<m 20mm ons<H>ma

NH.m mam<H

109

effect is by far largest in 1975 and 1975 with a 8.56 and
6.22 percentage point added to the rate of inflation,
respectively.

The higher energy prices have also resulted in more
labor intensive methods of production in the U.K., as the
level of employment increases, and therefore reducing the
unemployment rate - given that we have kept the level of
output fixed in simulations of the model.

Again, there has been some shift away from energy
as an input into production, and in sharp contrast to the
result obtained for other countries, higher energy prices
have resulted in a higher capital utilization rate in the
U.K..

Finally, higher energy prices have led to a sizable

reduction in labor productivity and real wages in the U.K..

D. - Germany

Table 5.13 summarizes the actual performance of
Germany's economy over the period 1973 to 1979. Values of
the endogenous variables of the model obtained in historical
simulation are given in Table 5.10. As can be seen from
these values, the model's performance and its ability to
IPeplication the actual data for Germany's economy has been
rather good .

To see the effects of energy price shocks on the
(Sermmn economy, we deve10p a base case describing the

Phassible evolution of the economy since 1973 in the absence

<31? sharp increases in the price of energy. In preparing this

110

mmmz ammm

H ma

U.x ca I U.x ca u x ”mm amcammao
mvmm coapmmwaapa amvwmmu
Acowwmeamcooo osmEma smxmcm

Aooxom socma cmaaa>wo mo so mpmm pcwEsoamEoca

A

swxmcm mo moasm
pcmesoamEm ampos
xxaoacmz pom pampaoo spa>apoaoosm soama

m
3

max

a
a
ma
2
ma mamas

 

 

cc.ma m.mmm mw.m H.mNNHH mm.m saomm :H.wH sm.H Nb.m :m.m msma
oo.m m.mHm mm.m m.somoH mm.m ooszm Ho.mH :N.H sm.m mm.m mama
cc.m| s.mmH Hm.m m.mo:oH mo.a Hamzm mm.mH mN.H mo.s ss.m ssmH
mm.mH s.mmH om.m N.mmw0H :H.: mmmrw Hm.:H mH.H m:.m mN.m msmH
sm.:H s.ssH mo.: s.oomm mH.: mmsam mm.:H mH.H :b.s ms.m momH
cm mm H.mmH mm.: m.owmoH mm.m mmmmm mw.mH :H.H 3m.OH mm.m :bmH
ca.m m.mmH sm.m H.mmmOH :o.H Homwm mo.MH HH.H mm.OH Ho.w msma
mm as mama a 4% 20:0: .5 m Ame Awe

mslmsma MZOZOUM m.%z<zmmw mo moz<zmommmm a<aBU<

mH.m mam<H

111

TABLL 5.10

HISTORICAL SIMULATION, GERMANY

 

 

 

p 8 LP M U 1: KUR
(3) (%) p (1000) (3) (%)
1973 5.39 10.66 .11 13.01 26372 .30 10529. .90
1970 7.32 11.63 .16 13.53 26120 .16 10283. .07
1975 5.00 3.82 .10 10.15 20890 .00 9802. .07
1976 3.70 6.05 .17 10.50 20756 .05 10365. .88
1977 3.95 0.90 .19 15.10 20838 .83 10602. .88
1978 3.95 7.88 .22 15.70 20821 .92 10796. .70
1979 0.07 6.09 .25 16.13 25217 .29 11177. .50
where LP = Labor Productivity (Output Per Manhour)
M = Total Employment
U = Unemployment Rate (% of Civilian Labor
Force)
E = Energy Demand (Consumption)
KUR = Capital Utilization Rate
(Defined as: KUR = 1n Kt - ln Kt-l)

= Real Wage

112

base case, e make the assumption that the nominal price of
energy increases at a rate of 5.0 percentage points per
year, which is the average of the inflation rate over

five preceding years. The model's base case projections

are given in Table 5.15, and Table 5.16 shows the deviations
of the historical simulation of the model from the base case
forecasts.

As can be seen from Table 5.16, the energy price
shocks have generated a sizable inflationary pressure in
the German economy. The results are similar to those
obtained for the U.S., Canada and the U.K.

The predicted effect of higher energy prices on the
price level is greater in the first three years. In terms
of inflation, the effect is larger in 1975 with 2.25
percentage points added to the inflation rate.

The higher energy prices have induced more labor
intensive methods of production, as can be seen from
increases in the level of employment, and therefore
reductions in the rate of unemployment, given that we have
kept the level of output fixed in simulations of the model.
Also, higher energy prices have resulted in a small
reduction in energy consumption, however, the reduction
in capital utilization rate due to higher energy prices
are relatively large.

Finally, the higher energy prices have resulted in
a reduction in labor productivity, but they have not led

to lower real wages.

113

 

 

 

mH.m mam<H

a
mmma Hmmm n m
xaImx ca I U.x cm H xax ”mm swammmao xa
mmmm commmmmampa Hmpaamo u x
Acomvmeamcooo pcmEmQ mmmmcm u m
Amomom momma cmmaa>mo mo so mmmm mamasoamEmca n a
mwmmcm mo mommm 0 mm
mcmEsOHQEm ammos n 2
Amaomcmz mmm mammaov >pw>wvospomm momma n ma mmmmz
o.m o.oma os.m m.smmaa o:.m mmoam mo.oa sa.a oo.m mo.m osma
o.m m.mma oo.m :.o:ooa so.o oasmm a:.oa ma.a am.m mo.m osma
o.m o.m:a oo.: o.omsoa sm.o mosmm No.ma ma.a om.m mo.m ssoa
o.m a.oma mm.: s.osmoa :s.s mammm mm.ma :a.a Ho.m am.a osma
o.m m.ama m: : o.oaooa am.o wooam :o.:a ma.a mo. ma.m msma
c.m m.mma ms.: a.om:oa :o.m mmomm os.ma ma.a ms.o mm.m :soa
o.m m.oaa no.3 m.:mmoa sm. momom :o.ma ma.a am.oa aa.m msoa
mi a name 88: m s: so
a a x m. a 2 n3 3 3 a
wz<zmmw I wwmmzm mo monm 2H mmm<mmozH smamm< mo mozmmm<

110

[1; A

 

 

m
mmma ammm u m
AHIPx :a I mx ca 0 max "mm omcmmmao ma
mmmm coammmaamma ammammo u x
Acommmfiamcoov Ucmea ammmcm n m
Amomom momma cmaam>ao mo we mmmm acmemoamEmCa n a
xmmmcm mo mowmm 0 mm
momExoamEm ammos u z
>mw>mpospomm momma n ma mmmmz
.uI om.aI sa.mI amaa :o.aI :o.o om.m m:.a msoa
mI mm.aI mo.mI msoa mm.:I sa.m :m.m so.a msma
.aI oa.aI :o.mI osoa om.:I am.m ::.m om. ssma
.mI am.aI mm.mI soaa s:.aI mo.m :a.m om.m osoa
mI as.aI om.mI mmm mm.mI mm. mo.m mm.m msma
.oI o:.aI m:.aI mm: mo.aI sm. mm.m mo.m :sma
.aI :m.I mm.I ms mm.I mm.I ma. om. msma
o o a
Aav Aoooav l va va
x a a 2 ma 3 3 a
a mHa w .mHa .mm< .mHa .mm< mHa w mHa w .mHa .mm< .mHa .mm<

mz<zxaw .mm<omxom mm<o mm<m

ma.m mam<s

was 20mm ZOHH<H>MQ

L. — France

Table 5.17 summarizes the actual performance of the
French economy over the period 1973 to 1979. The values
of endogenous variables of the model for France, obtained
in the historical simulation of the model are given in
Table 5.18. Table 5.19 shows the model's projections for
a base case describing the evolution of the French economy
in the absence of abrupt increases in the price of energy.
Finally, table 5.20 shows the deviation of the historical
simulation of the model from the base case forecasts.

Energy price shocks do not appear to have had a
significant effect on the French economy. There is a
rather small increase in the rate of inflation in 1970 and
1975, that could be attributed to the effect of energy
price shocks. However, the increases in the price of
energy have not had significant effect on the level of
employment, or on labor productivity in France.

Also, higher energy prices have not reduced the
energy consumption in this country, however, they have

led to a reduction in the capital utilization rate.

5.3 - The Relative Impact of Energy Price Shocks

 

It may well be the case that the extent by which
various countries have been adversely affected by higher
energy prices differs across them. To examine this
possibility, we can determine the relative impact of a 1%

increase in the price of energy above the level used in

 

116

 

m
mmmz Hmmm u m
AaImx cm I mx am n max "mm mmcmmmao xa
mmmm commmwmaaua ammmmmo n x
ACOmmmESmcooo ocmEma mwmwcm u m
Amomom momma cmaaa>mo mo so mpmm vcmamoamEmca u a
xwmmcm mo mowmm 0 mm
ucmE>0amEm ammos u z
Amaomcmz mmm mammaoo spa>mmoaoomm momma n ma mmmms
a.omm mo.m o.oooo :o.o aoaam am.am om.a om.ma m:.oa osoa
m.amm mo.m a.omoo Hm.m ooaam :o.mm mm.a om.ma mm.m msma
o.mmm :a.o s.asso mo.a amoam ms.mm o:.a mm.ma ao.o ssma
m.oam om.o o.amso om.: aosom om.mm m:.a :a.ma so.m osoa
o.aoa :s.s :.oooo ma.: aooom oa.am om.a mo.sa a:.ma msma
s.osa mo.m m.sooo so.m omoom om.om am.a oo.oa ma.aa :soa
.m o.saa ma.o m.maoo oo.m maoom m:.ma mmqa :a.am os.s msma
we a Mme Axe moooao m Axe ass
a x a a a ma 3 3 a

 

msImsmH wZozoum m.muz<mm mo moz<zmommmm a<aho<

sH.m mam<H

117

 

 

 

TABLL 5.16
HISTORICAL SIMULATION, FRANCE
p W’ 3 LP M U E K R
(‘3) (93) P‘ (1000) (9.) (g)
1973 7.29 13.88 1.30 19.52 20863 2.09 6828.6 7.93
1970 9.86 10.33 1.00 20.01 20898 2.78 6691.6 7.62
1975 11.19 13.97 1.03 21.18 20510 0.98 6061.8 7.07 m
i
1976 9.35 9.93 1.00 22.30 20001 6.03 6839.6 6.76 5
I
1977 8.00 12.98 1.50 22.89 20779 5.52 6685.9 6.00 It
1978 8.23 11.83 1.55 23.71 20939 5.66 6807.5 5.55
1979 8.32 11.39 1.59 20.60 21003 5.99 6880.9 5.38
where LP = Labor Productivity (Output Per Manhour)
M = Total Employment
U = Unemployment Rate (% of Civilian Labor
Force)
E = Energy Demand (Consumption)
KUR = Capital Utilization Rate
(Defined as: KUR = 1n Kt - ln Kt-l)

Real Wage

118

mmmz ammm

mm.
3

 

 

xaImx cm I Px :a n max .03 emcmmmao xa
mpmm commmmmampa ammmmmo u x
AcoamQESmcooo pcmEmm xwmwcm u m
Amomom momma cmaam>mo mo so mmmm mcwE>0amEmca u a
swmwcm mo moamm n ma
mcmemoamEm ammos u 2
Amaomcmz mmm mammaov sma>mmoapomm momma n ma mmmms
o.m o.sma as.m :.mooo mo.m :moam ao.:m os.a No.aa ms.s osma
o.m a.oma mm.m o.omoo mm.m mmoom ms.mm :o.a ma.ma mo.s osma
o.m o.m:a om.o m.osoo sm.m mssom om.mm sm.a am.ma am.s ssma
o.m a.oma oo.s m.mmmo mo.m mmaom om.mm om.a sa.oa mm.o osma
o.m s.mma mo.s o.omoo No.3 :amom ma.am s:.a ma.:a mm.o msma
o.m m.mma mo.s o.mooo oo.m oomom o:.om m:.a 3:.aa :o.o :sma
cs.m o.saa mo.s o.omoo oo.m mooom mm.ma am.a oo.ma mm.s msoa
m: m . Mime $3 883 m E :3
a a x m a a n3 3 3 a

 

moz<mm..%wmmzm mo MUHmm 2H mmm<mmozH Hmamm< mo mozmmm<

ma.m mam<H

119

 

 

. m
mwmz Hmmm u m
AaImx cm I mx ca 0 max "mm omcmmmav ma
mmmx scammumamma ammmamo u x
ACOMmmEchooo pcmEma xwmmcm u m
Amomom momma cmmam>ao mo so mmmm mcmE>0amEmca u a
mwmwcm mo mommm n ma
mcmEAOHQEm ammos n z
Amaomcmz mmm mammooo >ma>mpoapomm momma n ma mmmmz
aI.mI mm. am. o o s:.oI m:.I om. msma
5a.:I mm. mm. o o ma.mI mm.I mm. osma
mo.mI am. mm. a o o:.aI mm.I mm. ssma
mm.aI am. om. m o oo.:I :N.I om. osma
om.aI ma. om. o o ms.mI ma.I am.a msma
am.aI ma. oo. o o a:.aI aa.I mm.a :soa
o o o o o o o o msoa
”we x3: 88: m 3.; :3
Ca , a
x m a 2 ma 3 3
.mmm a .mHm w .mHa .mm< .mHa .mm< .mHa a .mHa w .mHm .mm< .mHm .mm<

 

moz<mm .mHm<ommom mm<o mm<m mmh 20mm ZOHB<H>MQ

om.m mam<B

120

obtaining the base case forecasts for each country described
above. Tables 5.21 through 5.25 summarize the effects of
this one percent increase in the price of energy. Changes
are stated in percentage terms in the tables and below.

Germany's economy has been adversely affected by
higher energy prices, relatively more than other countries.
Higher energy prices have substantially reduced labor
productivity and the capital utilization rate and created
a sizable inflationary pressure.

The effects of higher energy prices on the U.K. and
Canada are similar to those for Germany. Productivity and
capital utilization rates have been reduced less than in
Germany, while the inflationary pressure has been greater.
In Canada, the induced reduction in energy consumption has
been greater than in other countries.

The effects of energy price shocks on the U.S.
economy are also similar to those for other countries, but
the induced changes have been smaller. Inflationary pressure
created by higher energy prices in this country has been
less than in Canada, the U.K., and Germany. However,
labor productivity has decreased more than in other
countries, while the relative reduction in the capital
utilization rate has been less.

Finally, the adverse effects of higher energy prices
on the French economy have been less than in all other
countries examined here. Higher energy prices have

resulted in a very small inflationary pressure in France,

 

121

mos. mom.I was. ows.I soo. mOH. msmH

 

a:w.I sam.I was. mms.- smo.I was. msma
:am.- omN.I sew. m:m.- mao.I mom. ssma
mmm.I mON.I ems. mm:.I coo. 0mm. msma
msm.I sma.I mam. mmm.I Nmo. 03a. msma
ama.I oaa.I mam. mam.I Nmo. mam. :sma
Nso.I wmo.I mom. moa.I mmo.- was. msma
:3 33
m a 2 ma 3 m
mma x .mma .mm< mma x mma x .mHa..mm< .mma..mm<

 

.<.m.a I wommzm mo monm mmh 2H Mm<mmUZH ma < mo H0<mZH a<Zme<z mms

Hm.m mam<B

 

122

msH.HI mmm.I mms. mmm.I mas. :am. msmH

 

mmo.HI mmm.I mmo. amo.I mmm. Jam. msmH
oamf mamf mam. ommf omm. 13m. ssoa
mmsf mom; amz. s13... oam. omm. osma
amm.I sma.I mmm. s:m.I mma. mmm. msma
mo:.I HHH.I smm. mmm.I sso. mom. :smH
oom.I smo.I moo. mHH.I mHo.I omm. msma
$3 33
m a 2 ma 3 m
mam .3 .mHm .mm< HS .3 Ham 33 .mHm .mm< .mHm .mm<

 

<Q<z¢o I wwmmzm mo MUHmm mmH 2H Mm<mmqu wH < mo HU<mZH a<szm<X mmh

mm.m mam<B

123

 

 

 

msm.m maN.I sam.I mmm. :sm.I mam. mam. msma
m:a.m msa.I mmN.I ass. m::.I msz.I mas. mama
53m.a mma.I msa.I mam. msm.I mam. mms. ssma
3mm. smo.I ama.I mam. mm~.I mam. mam. ssma
mam. mmo.I mmo.I mam. maa.I saw. 00:. mmma
Nam. mmo.I m:o.I omo.I amo.I mam. 3am. :sma

o o o o o o o msma
max a a a ma Awe Ame
Lma a ama x .mma .ma< Lao x mma x .mma..mm< .mma..mm<

.x.a I maxaza Lo momma mam 2m mmmaxQZH xx < mo ma<azm a<szx<a mam

mm . m mam<H

120

 

 

 

oam.a- mmm.I :mw.I mam. som.I was. mam. msma
amm.aI mom.I Nam.I mss. mos.I mom. was. msma
mom.aI asN.I m::.I ams. sNQ.I msm. arm. ssma
a:m.- maN.I o:m.I was. os:.I 5mm. amm. mama
amm.I mma.I m:m.I mam. Nam.I ass. was. msma
som.I mNa.I sma.I mom. mom.I was. ama. 5sma
mom.- smo.I mas.I ass. :mo.- amo. :mo. msma
xax a a a ma Awe Awe

Lma x .xma a .mma .ma< mma x ama a .mma..mm< .mma..mm<

mz<zxma I maxazm mo momma mam 2m am<mx02m xx < mo mo<aam a<szx<a mam

:m.m mam<H

125

 

 

 

mas.I was. mmo. a a mmo.I sac. msma
asa.- mma. was. a a mao.I was. msma
ms:.I mac. mas. o a mmo.I was. ssma
mN3.I gas. was. a o mao.I :mo. wsma
mmm.- zoo. moo. a o moo.I mso. msma
mma.I o moo. a o moo.I ama. :sma

a a o o a a o msma
max m a 2 ma Ame Awe
Ema x mma x .mma .ma< ama x mma x .mma..ma< .mma..mm<

aoa<xx I maxazm mo momma mam 2m mm<mxozH ma < mo mo<32m a<2mox<z mas

mm.m mam<H

126

with no effect on employment and labor productivity.
However, the induced reductions in the capital utilization

rate are comparable to those for other countries.

5.0 - The Effect of Higher Energy Prices on the Shape of

 

Aggregate Supply

 

Using the historical simulation and the base case
projection of the model for each country, it is possible to
estimate the change in the elasticity of the aggregate
supply curves due to higher energy prices. The elasticity
of the aggregate supply is obtained by estimating the per-
centage change in the prices for a ten percentage point change
in income. These elasticities of the aggregate supply are
summarized in table 5.26.

Higher energy prices appear to have resulted, not
only in shifts in the aggregate supply curves to the left,
but also in the aggregate supply curves becoming more
flat in all countries except in Canada, as is apparent
from the increases in the own price elasticity of supply
resulting from higher energy prices. In Canada, the
estimated own price elasticity of supply is infinity implying
a horizontal short—run aggregate supply curve. Furthermore,
higher energy prices have not changed the slope of the
aggregate supply curve in this country. Also, in France,
the estimated increase in the elasticity of supply is found

to be very negligible.

 

127

TABLL 5.2L

THE ESTIMATES OF OWN PRICE ELASTICITY
OF AGGREGATE SUPPLY IN FIVE COUNTRIES

 

 

A%Y A31
A61) pop
Absence of Energy Actual Energy
Price Shocks Prices
U.S.A. 7.867 10.167
Canada m w
U.K. 2.831 3.659
Germany 20.393 29.067

France 0.651 0.657

 

CHAPTER SIX

SUMMARY AND CONCLUSION

The purpose of this study has been to evaluate the
effects of large increases in the price of energy since
1973 on the economies of major industrialized countries. To
this end, a model of short-run aggregate supply with
explicit treatment of energy prices was developed for major
oil consuming nations. We have concentrated on the supply
side of the economy since the important effects of energy
on macroeconomic activity occur largely through aggregate
supply. This model of short-run aggregate supply is based
on:

(i) A price determination equation which relates

the price level to the wage rate and other
input prices via some sort of mark up.

(ii) A wage rate equation or short-run Phillips
curve.

(iii) A short-run demand for labor equation.
(iv) A labor supply equation.
(v) A demand function for energy.

(vi) A capital utilization rate equation.

An increase in energy prices represents an increase
in the cost of a major factor of production. Consequently,

an increase in energy prices relative to other prices

128

129

results in a decline in the amount of goods and services
supplied by the economy at any given level of price, that
is, a shift in the aggregate supply curve to the left. An
attempt has been made to measure the extent to which the
aggregate supply curve shifts to the left (at any level of
output) due to higher energy prices.

We began by estimating and simulating the model for
five industrial countries: the United States, Canada,
United Kingdom, Germany, and France. DeSpite the simplicity
of this model, its simulation performance and its ability
to replicate the actual data series has been rather good
for all countries included in this study.

In order to examine the effects of energy price
shocks we then developed a counter factual "base" case
describing the evolution of each country's economy since
1973 in the absence of the abrupt increases in the price of
energy. Large increases in the price of energy were found
to have had substantial disruptive effects on the
economies of most of the countries we have examined. In
particular, the energy price shocks generated a sizable
inflationary pressure in the U.S., Canada, U.K., and Germany's
economies. Furthermore, the adverse effects of higher
energy prices on labor productivity, real wages, capital
utilization rate, and energy consumption were found to be
significant and similar across countries, except for France.
In the German economy the effects of higher energy prices

appear to be relatively larger than elsewhere, while the

130

the effects in the U.S., Canada, and E.L. are similar but
smaller than those in Germany. In France, energy price
shocks do not appear to have had a significant effect on
the economy.

Also, equally important, has been the effects of
higher energy prices on the demand for labor input. Sharp
increases in the energy prices have resulted in a shift
toward more labor intensive methods of production, and with
labor substituting for energy as an input in production,
there has been an increase in demand for labor across
countries.

Finally, higher energy prices have resulted, not
only in shifts in the aggregate supply curve to the left
in the various countries, but also in aggregate supply
curve becoming more flat, as is apparent from the increases
in the own price elasticity of supply resulting from higher
energy prices.

More work needs to be conducted, especially on an
expanded macroeconomic model to include the demand side of
the economy, as well as the supply side. This would allow
us to evaluate the fiscal and monetary responses of policy
makers to adverse effects of higher energy prices. These

expansions are left to future research.

APPENDIX

APPENDIX

DATA DEFINITIONS AND SOURCES

The data utilized in this study are taken from
individual country's sources, as well as international
publications. The sources of all the data that we have used
are presented in this appendix.

The percentage change in any variable is calculated

as:

 

except for the change in the unemployment rate which is

defined as:

Labor productivity is defined as:

Y
M*Hr

 

where Y is the output (or some measure of economic activity),
M is an indicator of total employment, and Hr is the number
of hours worked per employee.

The price of capital input per unit r is defined as:

r = pK (1+6 — 6K)

131

132

where pK is producer price index for machinery and equment

(or some similar measure), 6 is depreciation rate obtained

as capital consumption allowances (in constant prices) divided

by gross capital stock (in constant prices), and i is a

measure of interest rates.

Data Definitions and Sources

 

A. — U.S.A.

Y:

(1909-1979)
GDP originating in private business sector (P.B.S.)

(1972 constant dollars). The National Income and

 

Product Account of U.S. and Survey of Current

 

 

Business.

Implicit price deflator for output of P.B.S. (1967
= 100) — current dollar output of P.B.S. divided

by constant dollar PBS's output. Economic Report

 

of the President, 1980.

 

Industrial Production Index, (1967 = 100). Economic

Report of the President, 1980.

 

Constant-dollar (1972) net stocks of fixed non-

residential business (equipment and structure)

 

capital. Survey of Current Business, (April, 1976).
Producer price index for machinery and equipment

(1967 = 100) Economic Report of the President, 1980.

 

Corporate bonds rate (Moody's) AAA. ‘Economic

' Report of the President, 1980.

 

Money stock (billions of dollars). 'Economic Report

 

'of the President, 1980.

 

Hr:

LFl:

LF2:

POPl:

POP2:

133

Total energy consumption. Quadrillicn B.C.L.

U.S. Department of Energy, Energy Information

 

Administration. "Annual Report to Congress".

 

 

Producer price index for fuels and related
products and power (1967 = 100). Economic

Report of the President, 1980.

 

Compensation per hour - P.B.S. Economic Report

 

of the President, 1980.

 

Total number of persons employed in P.B.S.,
obtained as: total civilian employment of the
non—institutional population 16 years and over

minus government's employees. Employment and

 

Earning, BLS (December, 1980).
Average weekly gross hours per worker in private

(non—agriculture) sector. MonthlygLabor Review

 

BLS (May 1980).
Total primary labor force, male 25-50 years,

Handbook of Labor Statistics, BLS.

 

Total secondary labor force, male 16-20 years and

25-over, and female 16—over. Handbook of Labor

 

Statistics, BLS.

 

Population (non-institutional) male 25-50 years,

Handbook of Labor Statistics, BLS.

 

Population male 16—20 and 65-over, and female

16 and over, Handbook of LabOr Statistics, BLS.

 

130

B. - Canada (1956-1979)

Y:

Real Domestic Product originating in commercial

industries (1971 Canadian dollars).

1950-1960: Indexes of Real Domestic Products by
Industry (1961:100). Statistic
Canada, Catalog 61-515, pp. 50-51.

1961-1970: Real Domestic Product by Industry
(1961:100) Statistic Canada, Catalog
61-516.

1971-1979: Real Domestic Product by Industry
(1971:100) Statictic Canada, Catalog
61—213, 1979.

Deflator for total GDP less GDP originating

in public administration and defense and

community sectors (1971:100)

Nominal GDP:

1950-1960: National Income and Product Account
S.C., Catalog 13-531, Table 28.

1965-1979: National Income and Expenditure Account
S.C. Catalog, 13-201, Table 28.

Real GDP at factor cost: see source for Y.

Index of industrial production (1971:100).

1950—1970: Canadian Statistical Review. Historical

Summary, 1970. C.S. Catalog 11-505.

 

1971—1979: Canadian Statistical Review. Annual
Supplement to Section I, 1979,

C.S. Catalog 11-206.

135

Capital stock (end of year) machinery and equipment
plus non—residential construction, (1971 Canadian

dollars) Canadian Statistical Review, January 1979,

 

pp. IX.

Implicit price index for total gross fixed capital

formation (1971:100).

1950-1960: Industry selling price index. S.C.
Catalog 503.

1961-1979: Canadian Statistical Review (September,

 

1980).
Government of Canada average bond yield 10 years
and over.

1950-1960: Annual Supplement to Canadian Statistical

 

Review (1973).

1961-1979: Canadian Statistical Review (September,

 

1980).
Total (commercial) energy consumption, 1000
terajoules.

1950-1972: World Energy Supplies, 1950-1970, U.N.

 

1973~1978: World Energy Supplies, 1973-1978, U.N.

 

Industrv selling price for petroleum and coal
products industries.

1956—1976: Statistic Canada: Industry selling price

 

indexes, (1971:100). Catalog 62—503.

1977-1979: Industry price indexes. S.C. Catalog

 

62-011.

Compensation per man hours. Aggregate Productivity

 

Measure, 1979. S.C. Catalog 10-201.

M:

LFl:

LF2:

POPI:

POP2:

C. — U.K.

Y:

136

Number of persons employed in commercial industries.

Aggregate Productivity Measure, 1979. S.C.

 

Catalog 10-201.

Primary labor force; Male 25-50 years. Historical

 

Labor Force Statistics. S.C. Catalog 71-201.

 

Secondary labor force; Male 15—20 and 55-over,
and female lS-over.

Population; Male 25050 years.

Population; Male 15-20 and 55—over, and female,

lS-over.

(1959-1979)

GDP at factor cost (1975 pound sterling). OECD

Main Economic Indicator: Historical Statistics

 

1960-1979 and 1955-1971.

 

GDP implicit price deflator (1975:100). OECD

Main Economic Indicator: Historical Statistics

 

1960-1979 and 1955-1971.

 

Industrial production index (19752100). Annual

Abstract of Statistics. Central Statistical Office.

 

Gross capital stock at 1975 replacement cost.

Central statistical office. National Income and

 

Expenditure, various issues.

 

Wholesale price index of the output of mechanical

engineering. Annual Abstract of Statistics,

 

various issues.

Government bond yield: Long-term. International

 

Financial Statistics, I.M.F.

 

Hr:

LFl:

LF2:

POPl:

POP2:

137

Total (commercial) energy consumption. 1000

terajoules U.N. World Energy Supplies 1950-1970

 

and 1973-1978.

 

Wholesale price of fuel. OECD, Main Economic

 

Indicators, Historical Statistics, 1960-1970.

 

Weekly earnings: manufacturing (all employees)

(1975:100). OECD Main Economic Indicators,

 

Historical Statistics 1960;1979 and 1955+1971.

 

Total civilian employment.

1959-1976: British Labor Statistics, Yearbook,

 

1976.

1976-1979: Monthly Digest of Statistics.

 

Weekly hours per worker in manufacturing. U.N.

Monthly Bulletin of Statistics and various

 

U.N. Yearbook.

 

Total number of male employees (employed and
unemployed), i.e.; civilian male labor force.

British Labor StatiStics, Yearbook 1976 and

 

various issues of Monthly Digest of Statistics.

 

Total number of female employees (employed and
unemployed), i.e.; civilian female labor force.
Total home or "defacto" population of male age

lS—over. Annual Abstract of Statistics, various

 

issues.
Total home or "defacto" population of female

age lS-over.

138

D. — Germany (1959-1979)

Y:

GDP in purchasers value, at 1970 prices. National

Accounts of OECD Countries, 1950-1978, Vol. I.

 

GDP price deflator index (1970:100) - GDP at current
prices divided by GDP at constant prices.

National Account of OECD Countries, 1950-1978,

 

 

Vol. I.

Industrial production index (1975:100). Inter—

national Financial Statistics, Yearbook 1980, i
J

Net Capital Stock (end of year). Wirtschaft Und

 

Statistik, June 1979, pp. 001.

 

Producer price index of machinery and equipment,

(1970:100). Statistisches Jahrbuch, various

 

issues.
Yield of government bonds: Long-term loans,
percent per year, end of year. OECD Main Economic

 

Indicators, Historical Statistics 1960-1979 and

 

1955-1971.

 

Total (commercial) energy consumption, 1000
terajoules.

1950-1972: World Energy Supplies, 1950-1970, U.N.

 

1973-1978: World Energy Supplies, 1973—1978, U.N.

 

Wholesale price index for fuel and energy (1970

= 100). Allegemeines Statistishes Bulletin, ECSO,

 

various issues.

Hourly earnings (1975:100) - gross wages in industry

Hr:

LFl:

LF2:

POPl:

POP2:

139

including construction. International Financial

 

Statistics, IMF, (Yearbook, 1980).

 

Total civilian employment. OECD Labor Force

 

Statistics, various issues.

 

Hours of work per week, all industries. Yearbook

of Labor Statistics, I.L.O., various issues.

 

Civilian labor force, male-number of employed

plus number of unemployed. OECD Labor Force

 

Statistic, various issues.

 

Civilian labor force, female.

Population, male, age lS-over. OECD Labor Force

 

Statistic.

 

Population, female, age lS-over.

E. — France (1959-1979)

Y:

GDP at purchasers value, at 1970 prices.

National Accounts of OECD Countries, 1950-1978,

 

Vol. 1.
GDP price deflator (1970:100) — GDP at current
prices divided by GDP at constant prices.

National Accounts of OECD Countries, 1950-1978,

 

Vol. I.
Industrial production index (1975:100). Inter-

national Financial Statistics, Yearbook 1980, IMF.

 

Net Capital Stock. J. Artus (1977) "Measures
of Potential Output in Manufacturing for Eights

Industrial Countries, 1955-1978." IMF Staff_

 

Papers, 20, pp. 1-35.

 

Hr:

LFl:

LF2:

POPl:

POP2:

100

Implicit price deflator for gross capital

formation (1970:100). National Accounts of OECL

 

Countries, 1950—1978, Vol. I.

 

Government bond yield, percentate per year.

International Financial Statistic, Yearbook, 1980,

 

IMF.
Total (commercial) energy consumption, 1000

terajoules. World Energy Supplies, 1950-1970 and

 

1973-1973, U.N.

 

Wholesale price index for fuel and energy (1970 =

. W a

100). Allegemeines Statistishes Bulletin, various

 

issues.
Hourly earnings in manufacturing (1970:100).

Statistical Yearbook, U.N., various issues.

 

Total civilian employment, OECD Labor Force

 

Statistics, various issues.

 

Hours of work per week, all industries. Yearbook

 

of Labor Statistics, I.L.O., various issues.

 

Civilian Labor Force, Male - number of employed
plus number of unemployed.

Civilian Labor Force, Female.

Population, Male, age lS-over.

Population, Female, age lS-over. OECD Labor Force

 

Statistics, various issues.

 

F.

101

- Japan (1955—1979)

Y:

Hr:

LFl:

LF2:

GDP at 1970 prices. OECD, National Accounts of

 

OECD Countries, 1980, Vol. I and II.

 

GDP price deflator (1970:100) - GDP at current
prices divided by GDP at 1970 prices.

Industrial production index, (1975:100).
International Financial Statistics, Yearbook 1980,
IMF.

Gross Capital Stock of private enterprises based on

(average) prices of 1970. Japan Economic Yearbook,

 

various issues.
Wholesale price index for capital goods (1975:100).

Japan Economic Yearbook.

 

Average interest rates on loans and discounts —

"long-term credit banks". Japan Economic Yearbook.

 

Total (commercial) energy consumption. World

Energy Supplies, 1950-1970 and 1973—1978. U.N.

 

Wholesale price of fuel and power (1970:100).

Japan Economic Yearbook.

 

Wages; monthly earnings. International Financial

 

Statistics, Yearbook, 1980, IMF.

 

Total number of people employed. Japan Statistical

 

Yearbook, various issues
Average weekly hours of work, all industries.

Yearbook of Labor Statistics, ILO, various issues.

 

Total male labor force - employed plus unemployed.

Total female labor force — employed plus unemployed.

102

POPl: Total population of male, age lS-over.
POP2: Total population of female, age lS-over. Japan

Statistical Yearbook, various issues.

 

 

BIBLIOGRAPHY

BIBLIOGRAPHY

Anderson, R.G., "Disaggregate Aggregate Supply, Consistent
Aggregation, and the Hudson-Jorgenson Model", 1979,
Econometrics Workshop Paper, Department of Economics,
Michigan State University.

Ball, J. and M. Duffy, "Price Formation in European
Countries", in The Econometrids of Price Determination,
1970. If '

Black, S. and H. Kelejian, "A Macro Model of the U.S. Labor
Market", Econometrica, (September, 1970), pp. 712-01.

 

Boelaert, R., "Unemployment-Inflation Trade-offs in EEC-
Countries", Weltwirtschaftliches Archiv.. pp. 018—
051.

Bowen, W.G. and T.A. Finegan, "Unemployment and Labor Force
Participation: Time Series Results", in The
Economics of Labor Force Participation, 1969. Ch. 16.

 

Briscoe, G. and D.A. Peel, "The Specification of the Short—
Run Employment Function: An Empirical Investigation
of the Demand for Labor in the U.K. Manufacturing
Sector, l955~l972", Oxford Bulletin of Economics and
Statistics, (May, 1975), pp. 115—102.

 

Cagan, P., "The Monetary Dynamics of Hyperinflation", in
Studies in the Quantity Theory of Money, edited by
M. Friedman, 1956, University of Chicago Press,
Chicago, IL.

. P., "The Reduction of Inflation and Magnitude of
Unemployment", in W. Fellner, ed., Contemporary
Economics Problems, 1977, American Enterprise Institute,
Washington, D.C., pp. 01—50.

 

 

Chow, G.C. and G.H. Moore, "An Econometric Model of Business
Cycles", in Econometric Model of Cyclical Behavior,
1972, Vol. II, edited by B. Hickman.

Cukierman, A., "A Test of the No Tradeuoff in Long-run
Hypothesis", Eponometrica, (November, 1970), pp.
1069—1080.

 

103

100

deMenil, G. and J. Enzler, "Prices and Wages in the FRB—MIT-
Penn Econometric Model", in The Econometrics of
Price Determination, 1970.

 

, G., "Aggregate Price Dynamics", Review of Economics
and Statistics, (May, 1970), pp. 129-100.

 

Denison, E., "Explanation of Declining Productivity Growth",
Survey of Current Business, 59, No. 8, Part 2 (August,
1979), pp. 1-20.

Dhrymes, P.J., "A Model of Short-run Labor Adjustment", in
J.S. Duesenberry, et. al., The Brookings Model: Soma
Further Results, North-Holland, Amsterdam, 1969.

 

Eckstein, O. and R. Brinner, "The Inflation Process in the
U.S.A.", A Study Prepared for the Use of the Joint

Economic Committee, 92nd Congress, Washington, D.C.,
1972.

, 0., "The Great Recession, with a Postscript on
Stagflation", Data Resources Sepies, Vol. 3, 1978
New York, North-Holland Publishing Company.

Fischer, 8., "Long-Term Contracts, Rational Expectations.
and the Optimal Money Supply Rule", Journal of
Political Economy, (February, 1977), pp. 191-205.

 

Friedman, M., "What Price Guide Posts?". in G.P. Shults
and R.Z. Aliber, eds., Guidelines: Informal Controls
and the Marketplace, 1966, pp. 17-39.

, M., "The Role of Monetary Policy", American Economic
Review, (March, 1968), pp. 1-17.

 

Goldstein, M., "The Trade-off Between Inflation and Unemploy~
ment: A Survey of the Econometric Studies of Selected
Countries", IMF Staff Papepa, (November, 1972).

Gordon, R., "Inflation in Recession and Recovery", Brookings
Papers on Economic Activity, 1971, pp. 105-158.

 

, R., "Wage—Price Controls and the Shifting Phillips

Curve", Brookings Papers on Economic Activity, 1972,

9 R-I "The Effect Of Aggregate Demand on Prices",

Brookings Papers on Economic Activity, 1975. pp. 613—
622. ——

, R., "Recent Developments in the Theory of Inflation
and Unemployment", Journal of Monetary Economics.
1976, pp. 185-219. _

105

, R.J., "The Theory of Domestic Inflation", American
Economic Review, (February, 1977), pp. 128-130.

 

Hall, R.E., "The Role of Prevailing Prices and Wages in the
Efficient Organization of Markets", (1979), Unpublished
paper, Stanford University.

Hamermesh, D.S., "Econometric Studies of Labor Demand and
Their Application to Policy Analysis", The Journal of
Human Resources, (1975).

 

 

Hoel, M., "Rational Expectations and Rigid Wages: A Model
of Inflation and Unemployment", Scandanavian Journal
of Economics, (1979), pp. 387—399.

 

 

Jorgenson, D.W. and E.A. Hudson, "U.S. Energy Policy and
Economic Growth, 1975-2000", The Bell Journal of
Economic and Management Science, (1970), pp. 061-510.

 

 

, D.W. and E.A. Hudson, "Energy Prices and the U.S.
Economy, 1972-1976", Discussion paper 637, Harvard
Institute of Economic Research, Harvard University,
(July, 1978).

Lahiri, K. and J.Immx "Rational Expectations and the Short—Run

Phillips Curve", Journal of Macroeconomics, (Spring,
1979), pp. 167-190.

 

Lipsey, R.G., "The Relation Between Unemployment and the Rate
of Change of Money Wage Rates in the U.K., 1862-1957",
Economica, (February, 1960), pp. 1—31.

 

Liu, T.C. and E.C. Hwa, "A Monthly Econometric Model of the
U.S. Economy", International Economic Review, (1970),
pp. 328-365.

 

Lucas, R.E., "Expectations and the Neutrality of Money",
Journal of Economic Theory, (1972), pp. 103-120.

McCallum, B.T., "Rational Expectation and the NRH: Some
Consistent Estimates", Econometrica (1976), pp. 03-52.

 

, B.T., "Monetarism, Rational Expectations,
Oligopolistic Pricing, and the MPS Econometric Model",
Journal of Political Economy, (1979), pp. 57—73.

 

Mork, K.A. and R.E. Hall, "Energy Prices, Inflation, and
Recession, 1370-1975", The Energy JOUrnal, Vol. I,
(1930a), No. 3, pp. 31—63.

 

, K.A. and R.E. Hall, "Energy Prices and the U.S.
Economy in 1979—1981", The Energy Journal, Vol. I
(1980b), No. 2, pp. 0l~53.

 

106

MUllineaUXI D.J., "Inflation Expectations and Monetary
Growth in the U.S.A.", American Economic Review,
(March, 1930), pp. 109-161.

 

Muth, J.F., "Rational Expectations and the Theory of Price
Movements", Econometrica, 1961, pp. 315-335.

 

Nordhaus, W.D., "The Political Business Cycle", Review of
Economic Studies, 1975, pp. 169—190.

 

 

Papademos, L., "Optimal Aggregate Employment Policy",
Ph.D. Dissertation, M.I.T., 1977.

Perry, G.L., "The Determinants of Wage Rate Changes and
the Inflation-Unemployment Trade-off for the United
States", The Review of Economic Studies, (October,
1960). pp. 237-303. "

 

3 G’L'9 "Changing Labor Markets and Inflation",
Brookings Papers on Economic Activity, 1970, pp.

 

, G.L., "The United States" in Higher Oil Prices and
the World Econom , edited by E. Fried and C. Schultz,
1975, The Brookings Institution, Washington, D.C.

 

 

, G.L., "Potential Output: Recent Issues and Past
Trends", U.S. Productive Capacity: Estimating the
Utilization Gap., 1978, Working Paper No. 23, Center
for the Study of American Business, St. Louis, MO.

Phelps, E.S. and Taylor, J.B., "Stabilization Powers of
Monetary Policy Under Rational Expectations", Journal
of Political Economy, 1977, pp. 163-190.

 

Phillips, A.W., "The Relation Between Unemployment and the
Rate of Change of Money Wage Rates in the United
Kingdom, 1862-1957", Economica, (November, 1958),
pp. 283-299. —.'

Pierce, J.L. and J.J. Enzler, "The Effects of External
Inflationary Shocks", Brookings Paper on Economics
Activity 1, 1970, pp. 13~61.

 

 

Rasche, R. and J.A. Tatom, "The Effects of the New Energy
Regime on Economic Capacity, Production and Prices",
Federal Reserve Bank of St. Louis Review, (May, 1979a),
pp. 2~12. 7“

 

, R. and J.A. Tatom, "Energy Resources and Potential
GNP", Federal ReServe Bank of St. Louis Review, (June,
1977b), pp. 10—20. —

 

107

, R. and J.A. Tatom, 'Energy Price Shocks, Aggregate
Supply and Monetary Policy: The Theory and the
International Evidence", Carnegie-Rochester Conference
Series on Public Policy, 1971, North—Holland Publishing
Company, pp. 9-90. -

 

 

Tatom, J.A., "The Productivity Problem", Federal Reserve Bank
of St. Louis Review, 1979, No. 61, pp. 2-11.

 

 

Taylor, J.B., "Monetary Policy During a Transition to
Rational Expectations", Journal of Political Economy,
1975, pp. 1009~1021.

 

, J.B., "Staggered Wage Setting in a Macro Model",
American Economic Review, (May, 1979), pp. 108-113.

 

, J.B., "Recent Development in the Theory of
Stabilization Policy", in the Proceedings of a
Conference, Federal Bank of St. Louis, (April, 1980),
pp. 1—01.

 

Tella, A., "Labor Force Sensitivity to Employment by Age,
Sex", Industrial Relations. (February, 1965). pp.
69-83.

 

Tobin, J., "Keynesian Model of Recession and Depression",
American Economic Review, (May, 1975), pp. 195-202.

 

Turnovsky, S., "The Expectation Hypothesis and the Aggregate
Wage Equation: Some Empirical Evidence for Canada",
Economica, 1972, pp. 1-17.

 

, S. and Wachter, M., "A Test of the Expectations
Hypothesis Using Directly Observed Wage and Price
Expectations", Review of Economics and Statistics,
1972, pp. 57—50. ‘

 

Samuelson, P.A. and R.M. Solow, "Analytical Aspects of
Anti—Inflationary Policy", American Economic Reviey,
(May, 1960), pp. 177-190.

 

Sargent, T. and N. Wallace, "Rational Expectations, the
Optimal Monetary Instruments, and the Optimal Money
Supply Rule", Journal of Political Economy, 1975,
pp. 201-257.

 

Suits, D.B., et. al., "Birth Control in an Econometric
Simulation", International Economic Review, (February,
1975), pp. 92-1110

 

Vanderkamp, J., "Wage Adjustment, Productivity, and Price
Change Expectations", Review of Economic Studies,
1972, pp. 61—70.