PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 comma.“

MODELING FUTURE DEMAND FOR ENERGY RESOURCES:
A STUDY OF RESIDENTIAL ELECTRICITY USAGE IN THAILAND

By

Prapassara Nilagupta

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Department of Resource Development

1999

ABSTRACT

Modeling Future Demand for Energy Resources:
A Study of Residential Electricity Usage in Thailand

By

Prapassara Nilagupta

Thailand has a critical need for effective long-term energy planning because of the
country’s rapidly increasing energy consumption. In this study, the demand for
electricity by the residential sector is modeled using a framework that provides detailed
estimates of the timing and spatial distribution of changes in energy demand. A
population model was developed based on the Cohort-Component method to provide
estimates of population by age, sex and urban/non-urban residency in each province. A
residential electricity end user model was developed to estimate fiiture electricity usage in
urban and non-urban households of the seventy-six provinces in Thailand during the
period 1999-2019. Key variables in this model include population, the number of
households, family household size, and characteristics of eleven types of electric
household appliance such as usage intensity, input power, and saturation rate. The
methodology employed in this study is a trending method which utilizes expert opinion to

estimate future variables based on a percentage change from the most current value.

This study shows that from 1994 to 2019 Thailand will experience an increase in
population from 55.4 to 83.6 million. Large percentage population increases will take
place in Bangkok, Nonthaburi, Samut Prakarn, Nakhon Pathom and Chonburi. At a
national level, the residential electricity consumption will increase from approximately
19,000 to 81,000 GWh annually. Consumption in non-urban households will be larger
than in urban households, with respective annual increases of 8.0% and 6.2% in 2019.
The percent increase of the average annual electricity consumption will be four times the
average annual percent population increase. Increased electricity demand is largely a
function of increased population and increased demand for high-energy appliances such
as air conditioners. In 1994, air conditioning was responsible for xx°/o of total residential
electricity demand. This study estimates that in 2019, air conditioning will account for
yy°/o of the total demand. Low, base, and high scenarios were modeled to provide
estimates of the potential range of electricity demand that could be expected. The low
scenario assumes the adoption of energy efficient appliances and reduced usage through
conservation measures. Total demand in 2019 under the low scenario is xx, yy°/o of the
demand under the base scenario. Through policy, education, and technology options, the
potential exists for Thailand to reduce the fixture demand for energy. This study
identifies key factors related to residential electricity demand that should facilitate

planners in the development of these options.

Copyright by
Prapassara Nilagupta
l 999

 

ACKNOWLEDGEMENTS

I wish to express my sincere thank you to Dr. Cynthia Fridgen, who was the chairperson
of my guidance committee and dissertation director. Her patience, support and
understanding made it possible for me to finish this dissertation. I would like to thank
Dr. Larry Leefers, Dr. Tom Edens, and Dr. Jon Bartholic who contributed their time and
expertise, and served on my dissertation committee. I also would like to thank the
Thailand Institute of Environment, National Statistics Office, Office of Civil
Registration, and Electricity Authority of Thailand. I appreciate the data and help
provided from members of these organizations.

To my friends, Soavapa Reungchinda, John Hartzell, Chaweewon Wongwarangkul, and
Linda Greasser, I would like to thank-you all for being supportive, keeping my spirit high
and sustaining me through the difficult times. My special and sincere thanks go to
Thomas Moen who initiated my interest in Geographic Information Systems and
provided professional support throughout my entire research. I value all the hours he

spent working patiently with me.

To my parents, Sirirat and Pramuan Nilagupta for their love, sacrifice, constant
encouragement and unfailing support, 1 am gratefiil always. To my grandmother, sisters,

brother, and niece, I cherish our family lives, together and apart.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... ix
LIST OF FIGURES ........................................................................................................ xi
LIST OF ABBREVIATIONS ........................................................................................ xv
CHPATERI INTRODUCTION AND OVERVIEW OF THE STUDY ...................... 1
1. 1 Introduction ..................................................................................................... 1
1.2 Overview of Thailand ...................................................................................... 3
1.3 Background of The Electricity Situations and Problems ................................... 9
1.4 Problem Statement and Justification of the Study .......................................... 16
1.5 Objective of the Study ................................................................................... 21
1.6 Overview of the Study ................................................................................... 22

CHAPTER II POPULATION MODELING: GENERAL DISCUSSION AND
LITERATURE REVIEW .................................................................... 24

I GENERAL DISCUSSION ON POPULATION MEASURES,
CHARACTERISTICS AND TRENDS OF POPULATION IN THAILAND....24

2.1 Fertility ......................................................................................................... 24
2.2 Mortality ....................................................................................................... 30
2.3 Migration ...................................................................................................... 33
2.4 Population and Growth Rates ........................................................................ 43
2.5 Population Distribution and Density .............................................................. 43
II POPULATION PROJECTIONS ...................................................................... 44
2.6 A Framework for Population Projection and Modeling .................................. 44

2.6.1 Mathematical Extrapolation and Regression ........................................ 45

2.6.2 Ratio Method ....................................................................................... 46

2.6.3 Cohort-Component Method ................................................................. 47

2.7

Population Projection in Thailand .................................................................. 50

2.7.1 National Population Projection ............................................................ 51
2.7.2 Regional Population Projection ............................................................ 54
2.7.3 Urban and Rural Projection .................................................................. 57

CHAPTER III POPULATION MODELING: METHODOLOGY AND RESULTS 58

I MODEL AND METHODOLOGY ................................................................... 58
3.1 Data Input and Sources .................................................................................. 58
3.1.1 Population Data ................................................................................... 59
3.1.2 Geographic Information System (GIS) Data ........................................ 59
3 .2 Baseline Data Preparation .............................................................................. 60
3.2.1 Population Baseline Data ..................................................................... 61
3.2.2 Migration Baseline Data ...................................................................... 67
3.3 Population Model .......................................................................................... 69
3.3.1 Overview of Population Model ............................................................ 69
3.3.2 Fertility Modeling ................................................................................ 73
3.3.3 Mortality Modeling .............................................................................. 76
3.3.4 Migration Modeling ............................................................................. 83
II PROJECTION RESULTS AND DISCUSSION ............................................. 101
3.4 Projection Results ........................................................................................ 101
3.4.1 National Level ................................................................................... 101
3.4.2 Regional Level .................................................................................. 104
3.4.3 Provincial Level ................................................................................. 109
3.5 Sensitivity Analysis ..................................................................................... 119
3.6 Discussion and Comparison of Projection Results ....................................... 123

CHAPTER IV ELECTRICITY MODEL: GENERAL DISCUSSION AND

LITERATURE REVIEW .................................................................. 129

I RESIDENTIAL ELECTRICITY DEMAND IN THAILAND ........................ 129
4.1 Population Growth and Economic Performance Outlook ............................. 129
4.2 Electricity End Use Efficiency ..................................................................... 134
4.3 Electricity Supply Outlook .......................................................................... 138
II RESIDENTIAL ELECTRICITY DEMAND MODELING ............................. 142
4.4 Studies on Residential Electricity Demand Modeling .................................. 142
4.5 Factors Effecting Residential Electricity Demand ........................................ 145
4.5.1 Household Characteristics .................................................................. 145

4.5.2 Housing Characteristics ..................................................................... 147

4.6 Residential Electricity Demand Projection in Thailand ................................ 148

vii

CHAPTER V RESIDENTIAL ELECTRICITY MODELING: METHODOLOGY

AND RESULTS ................................................................................ 155

I METHODOLOGY ......................................................................................... 155
5.1 Introduction and Overview of the Model ..................................................... 155
5.2 Data Sources and Inputs to the Model .......................................................... 157
5.2.1 Housing Characteristic Inputs ............................................................ 157

5.2.2 Electric Appliance Data Inputs .......................................................... 158

5.3 Preparation of Baseline Data ....................................................................... 159
5.3.1 Population and Housing Data ............................................................. 160

5.3.2 Electric Appliance Data ..................................................................... 160

5.4 Residential Electricity Demand Model ......................................................... 162
5.4.1 Numbers of Electrified Household for each Province ......................... 163

5.4.2 Ownership or Saturation Rate ........................................................ 166

5.4.3 Annual Electricity Consumption By Types of Appliance ................... 171

5.4.4 Projections of electricity demand per household ................................ 172

H PROJECTION RESULTS .............................................................................. 175
5.5 Projected Households .................................................................................. 17 5
5.6 Projected Total Residential Electricity Demand ........................................... 178
5.6.1 National Level ................................................................................... 178

5.6.2 Regional Level .................................................................................. 194

5.6.3 Provincial Level ................................................................................. 203

5.6.4 Electricity Consumption by Type of Appliance .................................. 210

5.7 Comparison of Residential Electricity Demand Projections ......................... 214
5.7.1 Comparison of Number of Households .............................................. 215

5.7.2 Comparison of Total Residential Electricity Demand ......................... 216
CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS ............................. 220
6.1 Summery of Results .................................................................................... 220
6.1.1 Population Estimates ......................................................................... 221

6.1.2 The Residential Electricity Demand ................................................... 223

6.2 Implications for Energy Planners ................................................................. 227
6.3 Future Research ........................................................................................... 235
6.4 Final Remarks ............................................................................................. 235
APPENDICES ............................................................................................................ 236
APEENDIX A ....................................................................................................... 236
APEENDD( B ....................................................................................................... 246
BIBLIOGRAPHY ....................................................................................................... 255

viii

LIST OF TABLES

Table 1: Migration Streams of Migrants: 1965-1970, 1975-1980, 1985-1990 ................ 37
Table 2: An Example of Baseline Population Data in Urban Area by Age Group ........... 80
Table 3: Calculation of Number of Births, Deaths by Sex and Age Group of Urban
Population ........................................................................................... 81
Table 4: Calculated Total Urban Population by Sex and Age Group, Time 1 ................. 82
Table 5: Population Change at National Level by Area, 1999-2019 ............................. 103
Table 6: Effect of Birth Variable on Total Number of Population ................................ 121
Table 7: Effect of Death Variable on Total Number of Population ............................... 122

Table 8: Projected Total Fertility, Crude Birth and Crude Death Rates, 1999-2019 ...... 125

Table 9: Thailand Selected Appliances Owned by Area (Percent of Households),
1991 .................................................................................................. 134

Table 10: Potential bor Efficiency Increases in Major Electrical Appliances ................ 135

Table 11: Birth And Death Rate of Population in Thailand by Age, Sex, and Area,
1995 .................................................................................................. 236

Table 12: Migration Distribution of Females and Males in Thailand by Age, 1990 ...... 237

Table 132 Projected Population by Region and Area, 1994 And 2019 .......................... 237
Table 14: Projected Population of Thailand by Province, 1994 And 2019 .................... 238
Table 15: List of Provinces in Thailand, 1990 .............................................................. 240

Table 17: Number of Households in Thailand by Types of Appliance Owned and
Area, 1990 ......................................................................................... 253

Table 18: Electric Household Appliance by Types And Usage bor Average Scenario .. 257

ix

Table 19: Projected Residential Electricity Consumption in Thailand, 1994-2019 ....... 258

Table 20: Projected Residential Electricity Consumption in Thailand by Region,
1994 And 2019 .................................................................................. 259

Table 21: Projected Residential Electricity Consumption in Urban Areas of Provinces in
Thailand by Types of Appliance, 2019 .............................................. 260

Table 22: Projected Residential Electricity Consumption in Non-Urban Areas of
Provinces in Thailand by Types of Appliance, 2019 .......................... 263

LIST OF FIGURES

Figure 1: Political Boundaries of Thailand and Neighboring Countries ............................ 4
Figure 2: The Seventy-six Administrative Provinces of Thailand ..................................... 7
Figure 3: Contribution of Each End User to Total Electricity Consumption in Thailand. 10
Figure 4: Electricity Consumption per Capita and GDP at Constant 1992 US$ .............. 11

Figure 5: Thailand Population and Residential Electricity Consumption: 1974-1992 ..... 14

Figure 6: Thailand Growth, Birth, and Death Rates, 1950-1995 ..................................... 15
Figure 7: Thailand Crude Birth and Death Rates: 1982 — 1994 ...................................... 26
Figure 8 Thailand Age Specific Birth Rates: 1990 - 1994 .............................................. 29
Figure 9: Thailand Death Rate by Gender and Age: 1994 .............................................. 32
Figure 10: Thailand Migration Distribution by Regions: 1990 ....................................... 34
Figure] 1: Thailand Age and Gender Specific Migration Rate, 1990 .............................. 41

Figure12: Number of Male Migrants Per 100 Female Migrants by Migration Streams... 41

Figure 13: Political Boundary of Provinces in Thailand by Urban and Non-Urban Land 64

Figure 14: Flow Chart of Baseline Population Preparation ............................................. 66
Figure 15: Flow Chart of Processes in Developing the Population Model ...................... 72
Figure16: Flow of Population Under Cohort-Component Method .................................. 74

Figure 17: Diagram of Calculation of Female Within and From Other Provice Migrants93

Figure 18: Diagram Illustrated Calculation ofOrigin Province Female Migrants ............ 94

 

Figure 19:
Figure 20:

Figure 21:

Figure 22:

Figure 23:

Figure 24:
Figure 25:
Figure 26:
Figure 27:
Figure 28:
Figure 29:
Figure 30:
Figure 31:

Figure 32:
Figure 33:
Figure 34:

Figure 35:
Figure 36:
Figure 37:

Figure 38:

Flow Chart of The Structure of The Population Model ................................ 100

Projected Total Population at National Level, 1999-2019 ............................. 96
Projected Population and Urbanization Rate By Region and Area,

1999-2010 ................................................................................................... 107
Percent Change of Total Population by Province, 1999-2019 ...................... 113

Political Boundary of Provinces by Non— Urban Areas by Province,
1999-2019 ................................................................................................... 117

Percent Change of Population in Urban Areas by Province, 1999-2019 ...... 118

Population Projection Results from Different Studies ................................. 124
Projected Total Fertility Rate, 1991-2019 ................................................... 126
Thailand Residential Electricity Consumption, 1974-1993 .......................... 130
Per Customer Residential Electricity Use .................................................... 131
Per Capita Residential Electricity Use ......................................................... 132
Distribution of Residential Electricity by End Uses (1987) ......................... 135
Electric Generation Mix by Types of Fuel Source, 1980 and 1990 .............. 140
Projections of the Residential Electricity Demand for Bangkok Metropolitan
Region by MEA (Very Low Case): FY 1997-2001 ..................................... 152
Projections of the Residential Electricity Demand for 73 Provinces by PEA
(Very Low Case): FY 1997-2011 ............................................................... 153
Change of Appliance Saturation Rate of Households in Urban and Non-urban
Areas over Time .......................................................................................... 167
Trend of Saturation Rates of Electric Appliances ........................................ 168
Projected Percent Annual Increase of Household in Thailand ..................... 177
Total Electricity Demand by Areas from Three Scenarios, 1994 and 2019.. 179

Residential electricity Consumption and Annual Growth Rate of Population
and Electricity Consumption, 1994-2019 .................................................... 182

xii

Figure 39: Share of Electricity Consumption from Appliances, 1994 and 2019 ............ 183
Figure 40: Share of Electricity Consumption from Appliances Types in Urban Areas,
1994 and 2019 ............................................................................................ 184
Figure41: Share of Electricity Consumption from Appliance in Non-Urban Areas, 1994
and 2019 ...................................................................................................... 185
Figure 42: Saturation Rate by Appliances at National Level, 1994 and 2019 ................ 188
Figure 43:Saturation Rate by Appliances in Urban Areas, 1994 and 2019 .................... 189
Figure 44: Saturation Rate by Appliances at Non-Urban Areas, 1994 and 2019 ........... 190
Figure 45: Residential Electricity Consumption by Region, 1994 and 2019 ................. 199
Figure 46: Share of Residential Electricity Consumption by Region, 1994 and 2019 200
Figure 47: Residential Electricity Consumption by Region in Urban Areas,
1994 and 2019 ............................................................................................ 201
Figure 48: Residential Electricity Consumption by Region in Non-Urban Areas,
1994 and 2019 ............................................................................................ 202
Figure 49: Projected Residential Electricity Consumption of Provinces in Thailand,
2019 ............................................................................................................ 204
Figure 50: Percent Electricity Consumption Change in Thailand,
1994-2019 .................................................................................................. 206
Figure 51: Percent Change of Residential Electricity Consumption in Urban Areas
by Province, 1994-2019 .............................................................................. 208
Figure 57: Percent Change of Residential Electricity Consumption in Non-Urban Areas
by Province, 1994-2019 .............................................................................. 209
Figure 53: Change of Electricity Consumption by Air Conditioner, 1994-2019 ........... 211
Figure 54: Change of Electricity Consumption by Air Conditioner in Non-Urban Areas,
1994-2019 ................................................................................................... 212
Figure 55: Change of Electricity Consumption by Air Conditioner in Urban Areas,
1994-2019 ................................................................................................... 213
Figure 56: Projected Total Number of Households (Customers) in Thailand,

1999 -2019 ................................................................................................. 216

xiii

Figure 57: Residential Electricity Consumption Projected under Different Scenarios... 217
Figure 58: Age Distribution of Projected Population in Thailand, 1994 and 2019 ........ 242

Figure 59: Age Distribution of Projected Urban and Non-Urban Population in Thailand,
2019 ........................................................................................................... 243

Figure 60: Age Distribution of Projected Non-Urban Population in Thailand,
1994 and 2019 ............................................................................................ 244

Figure 61: Age Distribution of Projected Urban Population in Thailand,
1994 and 2019 ............................................................................................ 245

xiv

LIST OF ABBREVIATIONS

ADB .............................................................................. Asian Development
CDR ................................................................................. Crude Death Rate
DSM ........ ' ............................................................. D emand Side management
EGAT ................................................. Electric Generation Authority of Thailand
FGD ........................................................................ Flue Gas Desulfirrisaton
GIS ............................................................... Geographic Information System
Gwh .................................................................................... Gigawatts-hour
HOMES ................................. Household Model for Economic and Social Studies
IEA .................................................................. International Energy Agency
IIEC ............................................................ Institute for Energy Conservation
Kwh .................................................................................... Kilowatts-hour
MEA ............................................................... Metropolitan Electric Authority
NESD ............................................. National Economic and Social Development
NESDB - ................................. National Economic and Social Development Board
NIE ............................................................... Newly Industrialized Economies
NRD2C .......................................... The National Rural Development Committee
NSO ..................................................................... National Statistical Office
OECD .................................... Organization of European Community Development

PEA .................................................................. Provincial Electric Authority

PTT ............................................................ Petroleum Authority of Thailand
TEI ............................................................... Thailand Environment Institute
TDR .............................................................................. Total Death Rate
TFR .............................................................................. Total Fertility Rate
TDRI ................................................... Thailand Development Research Institute
UN ................................................................................. United Nations

CHAPTER I

INTRODUCTION AND OVERVIEW OF THE STUDY

1.1 Introduction

Problems of increasing energy consumption persist in Thailand. Energy planners must
plan in the long term to meet firture energy consumption levels, taking into account the
country’s limited resource base and social, economic, and technological dynamics of
energy. Conventional long term planning in Thailand, using a macro approach, does not
provide detailed estimates of the timing or spatial distribution of changes in energy
consumption within the country. Such detailed estimates are critical to planners and
decision-makers in Thailand who are responsible for long range development in the
energy sector. Detailed estimates of the timing and spatial change in fiiture energy
consumption allow planners to proactively manage resources such as, allocating budgets
to build necessary infrastructures for provinces with high future energy consumption or

provinces with dramatic increase in future consumption.

Long term planning for the entire energy sector includes conducting detailed studies for

all users and resources of the energy sector which include users in industrial, commercial,

transport, and residential sectors and energy resources such as petroleum, electricity,
coal, and natural gas. This study performed detailed estimations of future energy
consumption for one energy resource by one sector. That is, this study estimates future
electricity usage of households in urban and non-urbanl areas of Thailand at the
provincial level for the period between 1999 to 2019. The estimates at the provincial
level, which never before were available in Thailand, are important information for utility
planners to assess timely and costly investment in expansion or new transmission and
distribution of electricity systems to households in each province. Additionally, the end
user approach, which estimates electricity usage from appliances owned and used in
households, utilized in this study, allows energy planners to assess the impact of
technological improvement of electric appliances on total electricity usage in households
of Thailand. The residential electricity model developed in this study can also be used as

a framework for energy planning in other sectors.

Modeling of residential electricity consumption in this study also included estimation of
households, which was performed based on detailed population information and
population modeling. In Thailand, as in many countries, electricity usage and population
levels are positively related. An increasing population translates into more households

and more demand for electricity. Estimation of population was performed at the

 

' The word non-urban used in this study does not necessarily mean rural. This study classifies urban and
non-urban population based on land use classification of Thailand, Department of Land Development.
Urban class refers to urban and built up land. Land that does not belong to urban class such as, agricultural
land, forest land, and miscellaneous, therefore, falls into non-urban class.

provincial level based on age, sex, and residency (urban or non-urban areas). Estimates
of future population at this detailed level are beneficial not only for energy planning, but
for other resource management and planning in the residential sector; because population
is a driving force of resource utilization. Future population estimates can be used to
project firture consumption for resources such as; schools, hospitals, and garbage disposal

services within a province.

The development and use of detailed models of population and residential electricity in
this study should provide a better understanding of the dynamic of demographic
processes and linkages between population and household electricity consumption. The
estimates of fiiture electricity consumption should lead us to better our ability to manage,

in the long range, our limited natural resource base.

1.2 Overview of Thailand

General information about the geography, politics, social and economic variables of
Thailand are useful for understanding this study. Geographically, the Kingdom of
Thailand occupies a land area of 5 13,1 15 square kilometers (approximately the size of
France). Thailand is in the Indochinese Peninsula, located in Southeast Asia on the Gulf
of Thailand and the Andaman Sea. Thailand shares boundaries with the Union of
Myanmar (formerly Burma) and Laos People’s Democratic Republic on the North to

form the “Golden Triangle”, Cambodia (Kumpuchea) and Laos People’s Democratic

Republic on the East, Malaysia on the South, and the Union of Myanmar and the India

Ocean on the West (see Figure 1).

 

Indian

Ocean

 

 

 

 

 

Figure 1: Political Boundaries of Thailand and Neighboring Countries
(Source: The Perry-Castaneda Library Map Collection, The University of Texas at
Austin: South East Asia Map Produced by US. Central Intelligent Agency.
[Online Image] Available
http://www.lib.utexas.edu/Libs/PLC/Map_collection/middle_east_and_asia/Southeast
Asia _poll97.jpg, December 23, 1998

Politically, Thailand has been a unified monarchy since 1350. Presently, Thailand’s
government is a constitutional monarchy with an elected parliament. There are three
levels of government created for administrative purposes: the central (based in Bangkok
Metropolis) the provincial, and the local levels. For statistical purposes, Thailand is
usually divided into five geographic regions: the mountainous and ridges in the North;
the rolling and semi-arid plateau (the Khorat plateau) in the Northeast, the valleys of
Chao Phraya and several smaller rivers in the Central area; the plains and hilly area in

the East; and the peninsular in the South.

Bangkok is generally treated separately from the five regions because of its unique
characteristics and importance as the center of politics and economics. Adjacent to
Bangkok is an area known as Bangkok Metropolitan Region (BMR). The BMR includes
provinces ofNonthaburi, Prathum Thani, Samut Prakarn, and Samut Sakom. Being close
to Bangkok, the BMR has access to the largest market of the country (Bangkok), access
to the central government services, and superior infrastructures compared to remaining
provinces in the country. The BMR is the area where the majority of Thailand’s
industries are located and it is also close to ports. For these reasons, the BMR plays an
important role in economic development of Thailand and is one of the most populated
areas of the country with a high percentage of migrants. The other area which has shown
increasing number of new industries and rapid population change due to the Thai
government policy to promote export led and labor intensive industries in this region is
the area known as Extended Bangkok Metropolitan Region (EBMR). The EBMR

includes provinces of Chonburi, Rayoung, Chachoengsao, Saraburi, and Pranakom Sri

Ayudhaya. Each region includes one or more provinces that are the center of a region as
far as social, economic, and political aspects are concerned. These provinces have high
population and superior infrastructures compared to other provinces within the region.
The centers of the regions include Chiangmai in the North region, Nonthaburi and
Bangkok in the Central region, Khon Kean, Ubon Ratchathani, and Nakhon Ratchasima
in the Northeast region, and Songkla and Nakhon Sritammarat in the South (See Figure

2).

For administrative purposes, Thailand is further subdivided into 76 provinces. Each
province is governed by a centrally appointed governor. Each province is divided into
districts (amphoe) governed by centrally appointed district chiefs. Currently, there are
699 districts in the nation. Each district is further subdivided into subdistricts (tambons),
and villages governed by locally elected subdistrict chiefs and village headmen. There
were 6,905 subdistricts, and 62,994 villages and 132 municipal areas in Thailand in 1997.
In terms of urban and rural areas in Thailand, the latest Population and Housing Census
(1990) indicates that 27.1% of Thai population live in urban areas and 72.9% live in rural

31' 8382 .

 

2 Definitions of urban and rural areas in this case are the administrative definition. Urban areas include
municipal areas and W sanitary districts. Rural areas refer to rural sanitary districts and non-municipal
areas and outside sanitary districts. For detailed definition of municipal areas and sanitary districts see
Summary Report of Population and Urbanization in Thailand from the Population and Housing Census,
1990. Office of National Statistics

 

 

 

 

 

Figure 2: The Seventy-Six Administrative Provinces of Thailand
(Source: Map With Only County Boundary. [Online Image] Available
http://www.ee.surrey.ac.uk/Personal/R.Apiwatwaja/Map/tmap1 .html,
December 23, 1993)

Thailand is located in a tropical zone. The temperature and humidity of the country is
high throughout the year. The average temperature of the country ranges fi'om 77 to 82
degrees fahrenheit. Typically, the temperature increases during the day, peaking during
afternoon hours. Thailand experiences high solar radiation throughout the year. The
solar radiation ranges from 370 to 390 cal/cmz- day. The relative humidity of the country
is higher than 70 percent year round. Climate has a large impact on the types and usage
of appliances that are unique in Thai households and different from western countries.
Approximately 37 percent of total electricity consumption in Thai households in 1997 is
from space cooling (using air conditioners and fans) while more than half of total
residential electricity consumption in western countries is from space cooling and

heating.

The economy of Thailand grew at a rate ranging from 5.5 to 11.3 percent during 1961 to
1990. During this period the nation was regarded internationally as a success. Thailand
has changed from an exporter of agricultural products to an economically progressive
state with exports dominated by manufactured goods and services. Nevertheless, the
agricultural sector still employs the majority of the Thai people, approximately 60-65
percent of the labor force (Siamwalla, A., Setboonsarng, S., and Patamasiriwat, D.,

1994). Recently, the Southeast Asian countries have experienced an economic down turn
that started in early 1997. The economic growth rate of Thailand declined from
approximately 9 percent and reached zero in 1997. Economists believe that the economic
situation of Thailand will not improve in the next few years (Thailand Development

Research Institute, TDRI in Susankam, 1998).

1.3 Background of The Electricity Situations and Problems

During the period 1975 to 1990, total electric consumption of Thailand increased
approximately 5.2 fold, from 7,782 to 40,130 gigawatt-hrs. (Gwh)2 (IEA, 1995).
According to the Electricity Generation Authority Thailand (EGAT, 1997), the
consumption of electricity is currently growing at more than 14 percent per year. Last
year, EGAT invested more than 1 billion USS in additional power generation and
transmission systems to meet the growing consumption for electricity in Thailand. It is

likely that the consumption will continue to increase in the firture.

The growing consumption for electricity in most developing countries, including
Thailand, is underpinned by the grth in economic activity, per capita income, and
population (IEA, 1995, CBC, 1993, Warr, 1993, and Ross and Thadaniti, 1993). The
growth in economic outputs underlined by rapid expansion in industrial activities is
reflected in the share of the industrial sector in electricity consumption. IEA (1995)
reported that the electricity consumption of the industrial sector in Thailand increased
from 4,765 gwh in 1975 to 17,687 gwh in 1990. In this period, the residential sector
accounted for a lesser, although still substantial, share of Thailand electricity

consumption. The consumption in this sector increased fi'om 1,490 gwh to 8,088 gwh,

 

3 According to IEA (1995), calculations of electricity production and final consumption use the energy
content of the electricity. Production of hydo-electricity uses the energy content at a rate of l gwh = 0.086
thousand ton Calculation of production of electricity from oil, gas, and coal (lignite) uses the net calorific
values). Net calorific value of oil is 7.31 barrels/ton. Net calorific value of lignite is 0.092 ton of oil

equivalent per ton of lignite.

however, its contribution to the total consumption in the nation was fairly constant and

varied from 19 to 25 percent throughout this period (Figure 3).

§

§

3

3

Electricity Consumption (9%)
1
r
l
r
l
1

S

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1976

1972 1974

 

 

 

Year

IEnernyedor llndustry
DResidertial Eanspea'tied

 

 

 

 

Figure 3: Contribution of Each End User to Total Electricity Consumption in Thailand

Note:

(Source: International Energy Agency, 1995)
Energy Sector: Electricity used by the energy producing industries, e.g. for heating, lighting and
operation of all equipment used in the extraction process.
Industry: Include non-ferrous metals, iron and steel industry, and non-specified industries.
Residential: All consumption by households.

Non-Specified: All other consumption

10

 

1200 0.7

.0
or

Electricity Consumption per GDP (kwh per SUS)

 

1000

-~ 0.5
800

 

-r 0.4

 

—~ 0.3

ikwh per person)
8
0

400

 

—» 0.2

Electricity Consumption per Capita

200 -

 

 

l

r
p
—l

 

 

 

1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
Year

 

-I-— Electricity consumption per capita + Eleclectn'city consumption per GDP

 

 

 

Figure 4: Electricity Consumption per Capita and GDP at Constant 1992 US$
(Source: Asian Development Bank (1992), p 501)

A specific explanation of the increase is not simple because the level and characteristic of
electricity consumption in each household varies within the country, especially between
urban and non-urban areas. Higher electricity consumption by households and utilization
of electricity beyond subsistence and the use of electric appliances are related to the
growth in per capita income and increasing urbanization. These two factors are highly
correlated (OECD, 1994). According to the Asian Development Bank, ADB (1994), the

per capita GDP of Thailand increased from $US 360 in 1975 to $US 1,527 in 1990. The

11

electricity consumption increased in accordance with the rise of per capita income. IEA
(1995) reported that Thailand electric consumption per GDP increased steadily from 0.31
kilowatt-hour (kwh) per $US in 1975 to 0.52 kwh per $US in 1990 (see Figure 4). The
electricity consumption per capita also showed a similar increasing trend in those years.

The consumption increased from 208 to 796 kwh per capita in both years respectively.

Changes in residential electricity consumption are also reflected in electricity sales.
According to OECD/IEA (1997), the level of total residential sector sales per capita
depends upon both the level of urbanization and the level of electrification. For countries
whose electrification rates begin to approach saturation, growth in residential electricity
use will mirror the growth in the saturation rates for appliances. The saturation rate for
an appliance is calculated as the percentage of total households that own one or more of
the appliance. These rates depend upon the rate of economic grth and per capita
income. An increase in electricity consumption has also resulted from a rural
electrification program in Thailand implemented between 1986-90. As a result of this
program, 93 percent of villages in Thailand had access to electricity compared to 66
percent in 1985 and 34 percent in 1980 by the end of 1993 (ADB, 1993). The
electrification program is accompanied by increased use of electric lighting and
appliances. It is likely that electric appliance ownership rate for households in Thailand
will increase in the future since these appliances could become available at lower cost

and urbanization is expected to increase.

12

The growth in economic development in Thailand has been accompanied by an increased
percentage of the population living in urban areas. According to the United Nations
(1993), the percent urbanization in Thailand grew from 15.1 to 22.2 percent from 1970 to
1990 (Figure 4). The projection of the United Nations for percent urbanization in
Thailand is as high as 48.6 percent by the year 2025, assuming that there will be no
inverse trend. A study conducted by OECD (1994) on energy in developing countries
concluded that electricity consumption in urban areas was higher than for rural areas in

most developing countries. This also applies in the case of Thailand.

The grth in electricity consumption in the Thailand residential sector has also been
encouraged by energy pricing policies that have kept the electricity price below its
economically efficient level. The average residential price of electricity based on
constant 1992 prices was approximately 0.072 US dollars per watt-hour in 1975 and
increased to 0.08 US dollar per watt-hour in 1992 (Asian Development Bank, ADB,

1994)

The last and very important factor that contributes to the growth of electricity
consumption in the residential sector in Thailand is the rise of population. Like other
resources, the consumption of electricity increases as the population increases. Although
the population in Thailand increased by 16.4 million people or 39.5 percent from year
1975 to 1992 (ADB, 1994), the residential electricity consumption increased by
approximately 588 percent during the same period (IEA, 1995) (see Figure 5). Currently,

the population in Thailand is 60 million and is projected to be 72 million by the year

13

2025 (National Economic and Social Development Board, NESDB, 1997). The average
population growth rate of Thailand decreased fi'om 3.08 percent during 1965-1970 to
1.32 percent during 1985-1990. The recent rate is relatively low compared to
neighboring countries. The average population grth rate for these countries during
1985-1990 was 2.58 percent for Laos People’s Democratic Republic, 2.63 percent for
Malaysia, and 3.12 percent for Cambodia. While Thailand successfully decreased the
population grth rate, the Thai government still viewed the total fertility rate and crude
death rates to be too high and unacceptable (The United Nations, 1995). The decrease of
population grth rate (see Figure 6) is due to the success of the National Family
Planning Program that reduced the population growth rate from 3.0 percent in 1970 to 1.3

percent in 1990.

70 12,000

 

 

6° ‘ - 10,000

 

 

 

 

 

 

 

 

 

E 50 *,75 .432
.3 / - 8.000% g
E 40 « ur g
g / ~ 6,000 .3; a
g 30 5 5
I)
a [v I 4'000 % g
10 X -2,000
0 . . T 0
V In (D N O O) O P N 0') V ID (0 N G 0') O 1" N
N N N N N N ‘0 Q Q Q Q to Q Q Q on a: O) O)
eeeeeeeeeeeeeeeeeee

Year

 

I + Population + Residential electricity consumption

 

Figure 5: Thailand Population and Residential Electricity Consumption: 1974-1992

14

(Source: Asian Development Bank, 1994 and The United Nations, 1993)

 

 

 

 

 

 

i

2? at e i 8
Population (thousand)

lation

 

 

 

 

 

 

 

8"”3’“

 

i

 

 

R...
a
m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 d .. ,,.J.._ A...“ ,_ o
1950-19551955—196019604965196549701970-19751975-19801980-1985198549901990-1995
Year
=r Crude birth rate - Cnrde death rate New Population growth rate

+Totai Population --0— Male Population —-i-Femaie Population

 

 

 

Figure 6: Thailand Growth, Birth, and Death Rates, 1950-1995
(Source: World Population Prospects: The 1992 Revision, The United Nations, p.624)

According to the 1985 population estimates by Mason, Phananiramai, and
Poapongsakom (1993) the structure of population in Thailand was composed of a large
number of young people (age from birth to 14) and a small number of older people. The
age composition of Thai population in that year had a triangle shape. Since the
population growth rate has decreased, it is very likely that the future population pyramids
of Thailand will have a narrower base because of a decreasing number of young

population and an increasing number of adults and nriddle-aged population.

15

The last population component that is more important at the local or provincial level is
the composition of population in urban and non-urban areas. In Thailand, this
composition has changed over time. According to the most current Population and
Housing Census (1990), 81 percent of the population in Thailand live in rural areas.
Historical migration trends indicated an increasing number of people moving from rural
areas to urban areas and a decreasing number of people moving from urban areas to rural.
However, migration from rural to rural areas still dominates other streams. The stream
that effects electricity consumption the most is the urban designation migration stream
because urban migrants have the potential to earn more income and to purchase and use
electric appliances. The census (1990) on ownership of household appliances also

indicated the growing ownership of electric appliances in urban households.

1.4 Problem Statement and Justification of the Study

As electricity consumption in Thailand continues to increase, energy planners need an
analysis tool to facilitate long-range planning to prepare for the increasing consumption.
A tool such as modeling is especially beneficial for a country with limited resources such
as Thailand because a model allows us to better understand variables and their linkages in
the model. It also provides estimates of key variables based on our best information and
assumptions formulated in the model. For planners to benefit the most fi'om the model

developed in this study, there are four important characteristics of the model that need to

16

be taken into consideration: 1) level of detail, 2) flexibility, 3) incorporation of opinions

of planners or experts, and 4) range of outputs

Firstly, determination of the level of detail of a model is one of the most critical
decisions. Long-term energy planning requires a thorough, detailed understanding of
energy consumption dynamics, at a detailed level of sectors, sub-sectors or end uses.

This study pertains to only one sub-component of the entire energy planning process.
This study estimated the consumption for one resource (electricity) by one sector
(residential) of one country (Thailand). Thailand is chosen as the study area because of
the affiliation of the author with the country and because of the country’s critical need for

prudent energy planning in the future.

The conventional approach to energy planning in Thailand is a macro-level approach.

The projections performed by this approach are usually performed by extrapolation of the
historical trend. At the macro level, energy consumption can correspond to the amount of
energy required in a country or to the amount of energy supplied to the consumers.

While macro-level studies allow us to understand the general long-term trend in terms of
energy consumption, this approach does not provide detailed estimates of the timing or

spatial distribution of changes in energy consumption within the country.
Rather than using the aforementioned macro-level approach, this study focuses on the

detailed modeling of the residential electricity usage in Thailand at the provincial level.

The model developed in this study looks at the role of technology, that is, unit

17

consumption of electrical household appliances. The model also tries to identify the
technical improvements incorporated in new equipment from the effects of energy
savings measures in the existing equipment. The economic role incorporated into the
model is the change of consumer consumption behaviors in terms of acquiring electric
household appliances (appliance saturation rate) and the appliance usage. The model also
considers the social factor by including the impact of population change in different

geographical areas of the country into the model.

Electricity consumption is directly related to number of households and appliances
owned and used in households. Estimation of electricity consumption at the provincial
level based on fixture trends of changes in family size and appliance ownership in urban
and non-urban areas specific to each province should give us better estimates of total
household electricity consumption in each province. Estimates at the provincial level
from this study will allow electric planners to foresee the change of the consumption
pattern at a level that was not possible in the past. The information at this level is very
important for long-terrn planning for electricity distribution and transmission systems.
This is because expansion or establishment of new substation or transmission line into an
area with high future consumption or an area with high potential increase of the
consumption is a time consuming and costly process. This is especially true when new
systems are required in a remote area where other infrastnrctures such as roads are not
adequate. The electricity consumption estimates can also be used as an indication of

expansion or urbanization of a province.

18

Secondly, flexibility of a model refers to the ability for the model to handle special
circumstances or exceptions that apply to some provinces. This means that the model has
to be developed in a manner to handle these cases without difficulty. There were
circumstances that criteria formulated in the population and residential electricity models
did not apply to all provinces. An example is the calculation of net migrants from within
province migration stream. Because of the lack of data on net migration at the provincial
level, the national net migration rate was supposed to be used in the calculation of the net
migrants for all provinces. The national rate could not be applied to all provinces
because this may result in negative population of non-urban areas in provinces with low
non-urban population. Therefore, three criteria were established to handle this situation

(see the detail of the criteria in Chapter 111 under Migration).

Thirdly, involvement of planner or expert opinion in the model is a critical step for any
model to be accepted by decision-makers and planners. They are the ones who have
authority in making final decisions. Expert opinions referred to judgements of
demographers and energy planners who know the past and can assess the firture trends of
population and electricity variables in the model developed in this study. Expert opinions
utilized in this study were judgements of the author based on discussions or reading
materials related to population and energy trends in Thailand. In this study, experts
assessed firture trends of each variable in the models. Trending method was used to
assess firture estimates of the variables based on percent change of these variables from
the current estimate. A trending method is a reasonable method to communicate with

experts or planners about future changes because planners are more likely to foresee

l9

future changes in terms of percent rather than complex mathematical fiinction. With the
ability to incorporate expert opinions and be mathematically flexible, this model can be
used to make informed decisions that are responsive to the rapid changes taking place in

Thailand

Fourthly, a single estimate of residential electricity consumption is not as meaningful as a
range of the consumption estimates (low and high consumption which were calculated as
fi'action of base consumption). The electricity consumption model was designed to run
under different "scenarios", with a scenario defined as a particular set of assumptions or
input parameters of the model. A low (conservative) scenario was defined to estimate the
lower level of electricity consumption. This scenario includes, for example, low usage
and power input of electric household appliances. Alternatively, a high-end scenario was
defined and run using the upper ranges of the input variables and assumptions. The low
and high estimates provide the potential range of electricity consumption that could be
expected. Ultimately, policy makers can then assess this range of potential consumption
and determine acceptable response levels and the criteria necessary to achieve such
levels. By defining different scenarios, the model can be used to assess the impact of
various policy options (such as family planning, and consumption side management) on

the projected consumption for electricity.
In summary, the development and use of detailed models of sub-components, such as
developed in this study, should provide a better understanding of each component, and

consequently, a better understanding and management of the entire energy planning

20

process. This model can be used as a framework for energy planning in other sectors.
Moreover, the underlying detailed population model can be used as a basis for modeling
other resources used in the residential sector. Better understanding and modeling
capabilities should lead to pro-active management, not reactive management. Ultimately,
this is the basis and the justification for this research; better understanding leading to

more effective management of resources.

1.5 Objective of the Study

There are two objectives in this study. The first objective, which is a global objective or
a broader goal of this study, is to build a framework for modeling and projecting
consumption of energy resources in a manner that is both dynamically and spatially
detailed arfi mathematically flexible. The residential electricity model in this study was
developed such that it can be used as a framework for estimating consumption of other

energy resources in other sectors.

The second objective, which is a specific objective of this study, is to estimate electricity
usage by the residential sector in Thailand from 1999 to 2019. Based on the estimates of
potential range of the future household electricity consumption, policy recommendations
were made for energy planners on supply and consumption of energy resources.
Recommendations on options for energy fuel resources were made for the supply side.
For the consumption side, also known as the demand side management (DSM), policy

options to manage the increasing future electricity consumption in the residential sector

21

by seeking reduction of electricity consumption through the use of more efficient

appliances were made.

1.6 Overview of the Study

Residential electricity consumption is largely a firnction of the number of households and
the saturation rate of electrical appliances within the household. It follows, therefore, that
the modeler must be able to estimate the firture number of households in Thailand. Since
household numbers are a function of total population, a detailed population model is
necessary. Note that because average household size has decreased over time, there is
not a 1:1 correlation between household numbers and population. Population at the
Luntgy level is largely a function of the age-specific and gender-specific birth and death
rates. Because this study is done at the provincial level, however, migration between

provinces must also be considered in the population model.

Because of the differences in electrical appliance saturation rates by urban and rural
households, it is also necessary to categorize and model the population and households as
either “urban” or “non-urban”. Migration must also explicitly consider migration

patterns in four streams: rural-rural, rural-urban, urban-rural, and urban-urban.

The modeling technique employed in this study is based on the hypothesis that historical

trends combined with expert opinion of firture changes in these trends can be used to

22

accurately assess the future. The general situation is to start with the most current
estimate of some variable, (for example, age-specific birth rate) and then “trend” this
variable into the future based on expert opinion, results of other studies, local knowledge,
or educated assumptions. Note that the trend need not be linear or even conform to any
mathematical formula. This methodology allows for the use of expert opinion and
judgement in a simulation modeling setting; an important point as we consider the

potential for use of this model by planners.

23

CHAPTER II

POPULATION MODELING: GENERAL DISCUSSION
AND LITERATURE REVIEW

Chapter 2 consists of two sections: a review and discussion of population characteristics
and trends of the Thai population and a discussion of population projections and
modeling. The first section provides historical and current information on five
characteristics of the Thai population: fertility, mortality, migration, population growth,
and population distribution and density. The second section is divided into 2 subsections.
Population projection and modeling in general is discussed in the first subsection and a

discussion specific to Thailand is in the second subsection.

L GENERAL DISCUSSION ON POPULATION MEASURES,
CHARACTERISTICS AND TRENDS OF POPULATION IN THAmAND

2.1 Fertility

Fertility can be measured in many ways as summarized by Davis (1995) and Saunders

(1988). This subsection describes three fertility measures that were used in this study.

The first measure is the most widely used and most easily understood measure: the Crude

24

Birth Rate (CBR). CBR is the number of live births in a year divided by the total
population in the same year multiplied by 1,000 to eliminate decimals. The second
measure, called age-specific birth rate, is a rate of birth calculated for women in a
particular five-year age group. Age-specific birth rate is calculated as the number of live
births by women in an age group divided by the number of women in the same age group
multiplied by 1,000. An aggregation of age-specific birth rates is called Total Fertility
Rate (TFR). TFR is the number of children an average woman can be expected to have

during her reproductive years. The formula for the TFR calculation is:

i=45—49
TFR = 51 z(b./ Pry/1000
i=15—19

Where:
5 is number of years in age interval
b,- is number of births born during the yearto women in age group i

Pi:- is number of women in age group i

The fertility rate in Thailand declined rapidly in the 19805 and declined more slowly in
the 19905 (Figure 7). More recently in 1997, the CFR was 15.6 per 1,000 population
(Institute for Population and Social Research, Mahidol University, 1997). There are three

explanations for the declining fertility trend in this country.

The first explanation, which probably has made the greatest contribution to the declining
CFR is related to population policy in Thailand. The Thai government developed the

National Family Policy in 1968 and implemented family planning programs in 1970. The

25

goal of the policy was to reduce the rate of population growth from 3.0 to 2.5 percent

fiom 1972-1976, 2.5 to 2.1 percent from 1977-1981, and to 1.5 percent from 1882-1986

(United Nation, 1980). As a consequence of the success of the programs, the CFR

declined rapidly.

25

 

\\

 

 

 

 

A A
- -

W

 

 

 

1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
Year

 

—e—Crude birth rate ~—a—-—Crude death rate

Figure 7: Thailand Crude Birth and Death Rates: 1982 — 1994
(Source Public Health Statistics, Ministry of Public Health)

The second explanation involves land availability and the change in the demand for

children. Thailand has undergone profound socioeconomic change during the last 30

years. Changes taking place tend to interact with and reinforce one another. Rural

26

agrarian lifestyle in the country has gradually been replaced by a modern urban lifestyle.
The proportion of household members engaged in agriculture has decreased.
Consequently, the size of land holdings has not changed, averaging about 3.5 hectares per
person in both 1963 and 1983 (Chamratrithirong and Debavalya, 1987). The rate of
expansion of agricultural holdings has also slowed because most of the usable new
agricultural land in Thailand was occupied by the end of 1980, except in parts of the
south (Whitaker, 1981). Since activities in the agricultural sector have been reduced, the
demand for farm labor and correspondingly children has been reduced as well. The
desired family size decreased from 3.8 in 1970 to 3 in 1984 at the national level, 3.8 to 3
for rural families and 3.6 to 2.7 for urban families in the same years (Chamratrithirong
and Debavalya, 1987). The results of group discussions conducted by Chamratrithirong
and Debavalya (1987) also indicated that large families are now increasingly perceived as

an economic burden.

The third explanation concerns education and the role of women in the social context.
Generally speaking, the role of Thai women during the past thirty years has changed from
housewife to working woman. Fertility studies have demonstrated that in most
developing countries reproductive behavior is related to female education levels and the
employment of women in the formal sector (United Nations, 1992). This seems to apply
to the case of Thailand, however, in lesser extent. In the past, Thai women did not have a
chance to acquire education because females were prohibited by culture to go to schools
that were generally run by monks. Universal compulsory education in Thailand was

enacted into law in 1921. However, it was not until 1980 that compulsory education was

27

implemented and all females had to attend schools. Education opens doors for work
opportunities in the formal sector for women. Work and education tends to delay
maniage. Age at first marriage of Thai women has traditionally been moderately late
compared to other developing countries (Chamratrithirong and Debavalya, 1987).
Chamratrithirong and Debavalya (1987) stated that the contribution of age at marriage to
the reduction of fertility in Thailand was modest although the age at marriage of Thai
people has increased somewhat over the last few decades. The later study conducted by
Guest and Chamratrithirong (1992) concluded that a woman’s level of education has been
shown to be strongly and inversely related to fertility in Thailand. They cited the study
conducted by Retherford et al (1979) and reported the TFR during the period 1965-69 for
women with some primary school education to be 6.5 births per woman and for women
with some secondary school to be 2.7 births per woman. The 1980 and 1990 census also

showed that differentials continue to exist but the differentials were much smaller.

The analysis of fertility or reproductive behavior can be done in greater detail when the
age-specific fertility rate is available. According to Guest and Chamratrithirong (1992),
changes in fertility from the period 1965-69 to 1975-79 occurred for women of all ages.
They discovered that in the period 1965-1969 the highest fertility rates occurred in
women aged 25-29 followed by women aged 30-34. In the period 1975-1979 this
phenomenon shifted to women aged 20-24. Public health statistics (1994) showed that
women aged 20-24, followed by women aged 25-29 in the period 1990-1994 had the
highest fertility rate (Figure 8). The absolute fertility rate declined from 1965-69 to

1975-79 in women aged 30-34. Guest and Chamratrithirong (1992) discovered that

28

women aged 25-34 were responsible for close to 90 percent of the total change in fertility

during that period. This age group should be the focus group for family planners.

120

‘—
n W
100 X n

 

 

 

 

a
i

 

 

 

 

 

 

 

 

 

 

 

 

 

.5.
E
3
D.
8
.e
I
E
3- so
‘- 3'
82
O O
Eé at at *f i“ ii:
83 6°
5?
§.
0 c
- ________...._--I
5 40 I +— A l
>4
2 k + + + .4
5 .
20
i t.
2 - -+
.1
o e . e . c . c M e
1990 1991 1992 1993 1994

Year

 

+<15 +1549 +2044 +2529 +30-34 +3530
—-+—4o-44 —-—4549 ---->50

 

 

 

Figure 8: Thailand Age Specific Birth Rates: 1990 - 1994
(Source Public Health Statistics, Ministry of Public Health)

In terms of the fertility rate for urban and rural women, not many studies have been
conducted. According to Knodel and Prachuabmoh (1973 ), the fertility rate difference

between rural women and those from the provincial urban areas were noticeable for

29

women in all childbearing age groups between 1965-1979. The greatest difference was
found in women aged 30-34 in this period. The more current study conducted by Guest
and Chamratrithirong (1992) concluded that the fertility differentials across urban/rural
place of residence as well as socio-economic groups have been reduced or have almost

disappeared.

2.2 Mortality

The number of deaths can be described in various ways. A first measure is the Crude
Death Rate (CDR). It is the number of deaths occurring in a population during a year
divided by the number of persons in the population and multiplied by 1,000. A second
measure is the age-specific death rate. This is a measure of the number of deaths during
the year of persons in a five-year age interval divided by the number of people in the

same age interval.

In Thailand the CDR declined from 5.1 to 4.9 deaths per 1,000 population in 1982 and
1993 respectively. This change was relatively small compared to the decline from 19.2
per 1,000 population in 1950-55 to 6.3 per 1,000 population in 1985-90 (United Nations,
1992, see Figure 6 in Chapter 1). The high CDR in the past could result fiom inadequate
health care services and lack of health education. Theoretically, the improved economy of
a country should increase the quality of life of people in the country. In other words,

increased individual income should result in better health status and prolonged life.

30

According to NESDB (1985), The Thai government expenditure on health services per
capita increased by 151 percent during the period 1976-1984. NESDB also found that
individual consumption expenditure per capita on personal health during the same period
had increased. The more current statistics on environmental sanitation fiom the Public
Health Department (1994) reveals a better quality of life as there were increased numbers
of safe drinking water sanitary wells, households with pest control, and households with
improved food sanitation. With the improved quality of life and more government
spending on public health programs, it is likely that the death rate in Thailand will

continue to decline gradually.

Differences in death rate were found between Thai males and females in all age groups in
the period 1990-94 (see Figure 9) with males having higher death rates in all age groups.
The difference was more than 50 percent in age groups from 15-19 to 45-49. The
difference was highest in age groups 20-24 and 25-29 (80 percent). The death rate
difference of males and females was down to 18 percent in the population age group 70
years and older. A possible cause of death rate differences between males and females in
this period can be attributed partly to life styles. In Thailand, males have a higher
tendency to be involved in high-risk activities that could lead to death, such as smoking,

drinking and violence.

31

 

8
8

 

8 8 5‘ 8
c
WWӢTWWW\<S\
S 3

 

e

 

 

 

8 8
98 Death Rate Difference

.1.

 

Death Rate (per 1,000 population)

8

 

 

 

Id
0 8
Wm .
-d 2‘
. ‘_ PTT-7:75:55,m"""‘iz'.t"":"":i. .183 .. / .t- ' e. 324'“: ' r”
l I l I

 

 

 

 

   

I . 1O
o'c- .,, we_%-/-§I lgflgfl I _0
0-4 5-910-1415-1920-2425-2930343539404445495054555960-6465-69 >70
498 (Your)

‘i‘iri-i'vi'I-‘i‘ Total death rate ::::.....s:i:~ Male death rate
:1 Female death rate —x— % Male and female different

 

 

 

Figure 9: Thailand Death Rate by Sex and Age: 1994
(Source Public Health Statistics, Ministry of Public Health)

Regarding death rate by residence, studies in this area in Thailand are very limited and
not up to date. However, the relationship between death rate and residency established in
the past can be utilized to project firture trend. Porapakkham (1986) studied differences
in death rates between urban and rural Thai fi'om the pattern of utilization of health
resources and accessibility to health care service in terms of travelling time and cost to
visit health services. She conducted a survey in 1979 and concluded that the population
in urban areas, where modern health facilities were available and accessible, was three

times as likely to use private clinics, hospitals and government hospitals as those in rural

32

areas. She also discovered that rural people are often more likely to use drug stores for
self-treatment. The smaller proportion of the rural population utilizing hospitals reflects
the unavailability and inaccessibility of these facilities in the past. In the firture, it is
likely that the death rate of the Thai population both in urban and non-urban areas will
continue to decrease gradually because of the government efforts to improve education
and living standards, increase spending in health care programs, and support

technological advancement of the country.

2.3 Migration

This section describes migration streams by regions, sex and age group, and urban and
non-urban status of migrants. In Thailand, migration streams by region refer to a change
of residency of persons to the current province of residence from other provinces within
or outside the region since the time of the last census. Figure 10 illustrates the overall
picture of migration at the regional level between 1980 to 1990 in Thailand. According
to the most current census (1985-90), regions ranked by net migrant rate from highest to
lowest were the Bangkok, Central, Northeast, North, and South. Only Bangkok and the
Central region experienced a net gain in this period. The net migration rates were 6.7 and
2.6 per 100 population for Bangkok and the Central region respectively. The net
migration rates for regions that experienced net loss were 3.2, 9.2, and 0.2 per population

for the Northeast, North, and South region respectively.

33

 

.5
ca

 

.a
O)

 

.5
®

 

 

A
N

 

 

10

 

 

 

 

 

 

 

....................

.......
,,,,,,,,,

Percent distribution of migrants

 

 

 
     

.........

.,.. .
...\
..~-
. ‘x , .7 ..

Kingdom Bangkok Central

 

 

 

 

Regions

 

[ len province Ib/w province, wln region Dblw regions I % migrants ]

 

Figure 10: Thailand Migration Distribution by Regions: 1990
(Source: National Statistical Office)

The Northeast region experienced a net loss in population during the last 40 years. The
net loss was due to the high number of people migrating out of this region. The
Northeast has the largest population in the country. A lack of income and employment
opportunities in much of this region contributes to the high migration in this region. The
North region experienced a net gain between 1955 to 1970. This period was the peak
time for logging in the country, particularly in the North region. Job opportunities in this
region in that period also included farming. After logging was banned 1975, the North

region lost its importance and experienced a net migration loss for the first time to

34

Bangkok and the central region between 1985-1990. In this period, the number of total
migrants from the North and the Northeast region decreased. The explanation for this
was the decline in the economy due to the oil crisis in 1979. Because of the slow
economy during the oil crisis, job opportunities in the industry sector, located in the

central region and Bangkok, were also diminished.

Bangkok has always experienced a net gain of in-migrants. Bangkok experienced a net
gain of 67,045 and 372,200 net migrants in 1955-60 and 1985-90 respectively. The
increased migration to Bangkok illustrates the economic and political dominance of
Bangkok. As argued by Pejaranoda, Samipapom and Guest (1995), the Thai government
which have favored the urban dwellers, particularly those living in Bangkok, is believed

to contribute to the increase of Bangkok attraction to migrants.

Afier experiencing a large net loss fi'om 1955 to 1975, the central region has benefited
from the economic growth of Bangkok and has experienced a net gain after 1975. This
region now absorbs some of the migration stream that would otherwise be directed to
Bangkok. As a result, the central region experienced the six-fold increase of the net gain

compared to Bangkok in 1985-1990.

The South region has experienced the lowest net migration rate compared to other
regions. Being geographically isolated from other regions, this region has the least
migrants since the level of migration tends to decrease as travelling distance increases.

However, this region experiences a high proportion of movement occurring within the

35

region. Few people migrate out of the region because the job opportunities in this region
are rather high (fi'om fisheries and aquatic farming, mining, and rubber plantation and
industry). The income level of the population in this region is also relatively higher than

income levels in other regions.

The migration level in Thailand seems to be the higher within region than between
regions. In some regions, half or more of the migrated population moved within or
between provinces located in the same region. This is normal, because migration for

shorter distances involves less resource in terms of time and cost.

Economic reasons and employment opportunities are not the only reasons for the Thai
population to migrate. There also are social reasons contributing to migration such as
education and marriage. Marriage-associated migration is influenced by the Thai cultural
norm that men move into their wife’s parent’s household upon marriage. However, no
study has been carried out to determine the contribution of the marriage-associated
migration to the overall migration rate in Thailand. Married women (and children in
some families) also have to accompany husbands to new locations in the case of a job

relocation.

The other way to observe migration is to look at the urban and non-urban residency status
of migrants. This way, migration streams are classified into urban-urban, urban-rural,
rural-rural, and rural-urban. The dominant stream in Thailand has always been the rural-

rural one. Nevertheless, the percent of migrants in this stream declined - from 62.6 to

36

40.9 percent in the period 1965-1990. Rural-urban is ranked the second highest in terms
of percent migrants. The trend of percent migrants in this stream increased fi'om 10.5 to
18.4 percent in the same period. Percent urban-urban stream increased from 8.9 percent
between 1965-70 to the peak of 17.2 percent between 1975-80 and declined to 13.5

percent in the period 1985-90. The migration in the urban to rural stream also increased

from 9 to 10.5 percent between 1965-70 and 1985-90 accordingly (Table 1).

Table 1: Migration Streams of Migrants: 1965-1970, 1975-1980, 1985-1990

 

 

 

 

Migration stream Percent Migration
1965-1970 1975-1980 1985-1990
Population age 5 and over 100.0 100.0 100.0
Non-migrants 88.4 92.4 92.0
Migrants 11.6 7.6 8.0

 

 

Migration Streams

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Urban-Urban 8.9 17.2 13.5
Rural-Urban 10.5 14.3 18.4
Unknown-Urban 3.6 3.3 4.1

Total migrants with urban destination 23.0 34.8 36.0
Rural-Rural 62.6 52.0 40.9
Urban-Rural 5.4 9.4 12.6
Unknown-Rural 9.0 3.8 10.5
Total migrants with rural destination 77.0 65.2 64.0

 

(Source: National Statistical Office, 1993 in Pejaranonda, Santipaporn, and Guest, 1995)
Explanations for changes in each migration stream over time in Thailand vary depending
on the type of stream. Pejaranonda, Santipapom, and Guest (1995) provided four reasons

for migration in the rural-urban stream. The first reason is related to new employment

37

opportunities that are centered in urban-based service and manufacturing industries.
Since the fifth National Economic and Social Development plan, the Thai government
utilized the urban-based development strategy to promote their economic development.
This created job opportunities and increased migration to urban areas. The second reason
cited was the reduction of suitable land for agriculture. The traditional means of
increasing agricultural land by extending the areas under cultivation was no longer an
option. Therefore, the opportunities for employment in the agricultural sector declined.
This problem took place about the same period that mechanization in agriculture became
more common. The third reason is that the increased level of education of the population
has made aspirations for employment in non-agricultural sectors greater. Moreover, the
income in sectors other than the agricultural sector (urban areas) has risen faster than
agriculture (non-urban areas). Lastly, expansion of urban boundaries increases the
number of places categorized as urban. The expansion of urban4 boundaries increased
the number of municipalities, which were generally designated by government agencies

as urban areas, fi'om 707 municipalities in 1980, to 843 municipalities in 1990.

In the urban-rural migration stream, Pejaranonda, Santipapom, and Guest (1995) also
provided three explanations for the change of residency from urban to rural areas. The

first explanation is that some of the migrants in this stream are the return migrants from

 

4 There is no official definition of urban or rural in Thailand. Thai government usually designates
municipality as urban and non-municipal as rural. There are three types of municipality: muang, nakorn,
and tambon municipalities. A muang municipality refers to a town that has population at least 10,000
persons and an average density of 3,000 or more per square kilometer and revenue base sufficient to carry
out municipal functions. A nakorn municipality requires a population of 50,000 persons and an average
density of 3,000 or more per kilometer and sufi'rcient revenue base. There are no specific criteria for
desingation of a tambon municipality. They are established by decree (Pejaranonda, C., Santipapom, S.,
and Guest, P., 1995)

38

rural to urban areas. These migrants are those who leave the rural areas to work in urban
areas and then return to the original place of residence after the work is finished. The
second explanation contributes to the increased cost of urban land. This forces people to
relocate to the rural periphery of urban places and commute to work. The third
explanation involves rural development policies of the Thai government. These policies

promote the transfer of government employees to work in rural areas.

In terms of sex and age group of migrants, the probability of migration for both males
and females in Thailand has not changed during the period 1955-1990 (Figure 11). The
highest migration in Thailand in this period was for young adults. The percent migration
of the population aged 20-24 and 25-30 were more than any other age group. This
applied to both males and females. The concentration of migrants was greatest in both
sexes aged 20-24 and gradually declined except for males in the period 1985-90. About
one third of total male migrants in this period were in the 20-24 age group. The percent
male migrants in the next age group reduced drastically to 15 percent of the total male
migrants. Afier this age group percent male migrants declined gradually as the age group

increased.

The migration differential based on sex could also be observed from the sex ratio of
migrants in each migration stream. According to the 1965-70 census, the migrant
population of males dominated the migrant population of females in all streams. This
pattern changed in 1980 and 1990 (Figure 12) when the number of female migrants was

more than the number of male migrants in the urban designation areas (urban-urban and

39

rural-urban). This pattern also continued in 1990. However, the ratio of female to male
migrants in urban-urban stream declined by 2.45 percent. The sex ratio of female to male
migrants increased in the rural-urban stream by 3 percent. The increase in female urban
directed migration streams was evident for migration to Bangkok. Of every 100 female
migrants to Bangkok there were 87 male migrants between 1985-90. The high proportion
of young adult female migrants in Bangkok suggested a high demand for young females
and high job opportunities in a sector that was selective of females. This sector included
the service sector. Types of jobs in the service sector are housekeeping jobs, restaurant
work, and jobs in the entertainment business. The demand for female labor was also high
in some industrial sectors such as garment and textile industries, which were one of the
fastest growing industries in the country. Some other industries where the nature of work
does not require special sex related skills, such as electronics and computer industries

also prefer women to men for various reasons (dependability and responsibility).

40

 

8

 

8
II

 

 

8

 

 

 

8
I

.a
UI

 

 

 

.a
O

 

0|

 

Age-specific migration rate per 100 parosna

 

 

 

 

 

 

 

 

o. l. ‘
0-4 5-9 10—14 15—19 20-24 25-30 3064 35—39 40-44 4549 50-54 55-59 60-64 sear 70up
An-

I I All Male IAII Fermle Cl R-U Male Ci R-U Fermle

 

Figure] 1: Thailand Age and Sex Specific Migration Rate, 1990
(Source: National Statistical Office, 1990)

 

1980 1990
Year

 

[ lumamuman lRumHfibml DRumLRmd numaana j

Figure12: Number of Male Migrants Per 100 Female Migrants by Migration Streams

41

(Source: National Statistical Office, 1990)
The proportion of male to female migrants was very high in the urban directed stream to

the Central and Northeast regions. This is obvious between 1985-90 when the sex ratio
in the urban-rural stream increased by 26.7 percent from the period between 1975-80.
The increase of the sex ratio in the rural-rural stream was 5.7 percent during the same
period. Migration to the Central and Northeast regions was dominated by males

interested in work in agriculture in that period.

The age and sex structure of migrants in the past suggests that the young adult age group
will continue to be the major migrant group. Previous trends based on the age and sex
structure of migrants also suggest that the difference between percent male and female
migrants will become smaller. Currently the sex ratio in all migration streams is female
dominated (117.1 percent). It seems that this trend will continue, especially in the urban
directed streams because the demand for female labor will continue to grow due to a
policy to promote the tourism industry in the next decade of the Thai government. It is
expected that tourism will bring 2-4 billion baht to the country (Siam Post, December 15,
1996). The rural directed streams will continue to be dominated by males because the
nature of this work requires labor skills specific to men. Moreover, it is very likely that
job opportunities in the urban areas will decrease due to the current economic crisis.
There is high potential that a large number of migrants (rural-urban) will return to rural
areas and migration to urban areas will decrease in the next few years during the recovery

of the economy of the country.

42

2.4 Population and Growth Rates

The p0pulation growth rate refers to the change in the volume of population from natural
increase (birth and death) and net migration when compared to population in a given a

year. This number is usually expressed as the rate per 1,000 population per year or a

percentage.

The population grth rates in Thailand declined steadily from 32.2 to 1.98 per 1,000
population in 1960 and 1990 accordingly. In the future, it is expected to continue to

decrease more slowly because of lower fertility and mortality rates in the country.

2.5 Population Distribution and Density

According to the Institute for Population and Social Research, Mahidol University
(1998), the population of Thailand in January 1997 was approximately 60 million persons
with 18,981,000 urban population and 41,122,000 rural population. By region, there
were 20 million people in the Northeast, 12 million people in the North, 13 million
people in the Central region, 7.5 million people in the South, and 7.9 million people in

Bangkok.

The overall population density in Thailand was about 106 persons per square kilometer as

of 1997. The region that had the highest population density was the Central Plains with

43

118 persons per square kilometer. This was followed by the Northeast, the South, and the
North with a population density of 113, 99, and 62 persons per square kilometer
respectively. The population density of Bangkok Metropolitan was 3,758 persons per

square kilometer.

II POPULATION PROJECTIONS

This section provides an overview of population projections and the modeling framework
in general as well as the framework specific to Thailand. Projections for the population

components are also discussed.

2.6 A Framework for Population Projection and Modeling

In this subsection, population, fertility, mortality, and migration projection methods
utilized in the population modeling are described. This subsection focuses on the

methods that can potentially be used and serve the purpose of this study.

Projection of population can be performed at different levels of detail using various
methods. In general the base period population and projections of fertility, mortality, and
migration are required to estimate future population. There are many techniques that

have been used in population projection studies (Nowell, 1988; United Nations 1992; and

44

Dave 1995). The most common ones are (1) mathematical extrapolation, (2) ratio, and

(3) Cohort component.

2.6.1 Mathematical Extrapolation and Regression

This method is a rough approximation of future population by extension of a historical
population trend into the filture. The major assumption of this method is that the current
rate of growth will continue in the future. This assumption can lead to unreasonable
projections, especially if the projections will be carried far into the future and/or done for
non-homogeneous areas. The extrapolation method involves extrapolation along
mathematical curves such as straight lines and geometric and logistic curves. A least
squares technique is often used to fit a curve through the known data points of the past.
When regression is applied as a mean of extrapolating a trend, the future of unknown

behavior variables related to population change can be integrated into a single expression.

Mathematical extrapolation in projection models is often used as submodel of the model,
such as survival rates (equals to 1 — death rates) and fertility rates that may need to be
projected into the future. The projection can be simply performed by the extension of the
historical rate of change of the survival rate or fertility rate. More sophisticated

regression methods can also be used for projecting these submodels.

The mathematical extrapolation method is useful when historical data is available and

rough approximation is required, without detail, by age and sex. The model in this study

45

performs projections of population by age and sex in urban versus non-urban areas for the
76 provinces of Thailand. Using this method for the fertility, mortality, and migration
submodel at the provincial level is not possible because historical data of these
parameters by age, sex, and urban versus non-urban areas in each province is not
available. The non-linearity of components such as migration trend over time could not

be dealt with insufficient accuracy using extrapolation or regression techniques.

2.6.2 Ratio Method

In this method the ratios or shares of a larger population for which a projection already
exists are calculated to express the data. The past trends of the ratios is determined,

projected, and multiplied by the projection for the parent population.

For the ratio method to be utilized in this study, it would require acquiring projection of
population by age, sex, and urban and non-urban area at the regional or national level.
The projected population would then be allocated to provinces based on the projection of
the provincial, regionaVnational population ratios into the future. Although this approach
seems straightforward and flexible, it is not the best approach to perform population
projections for this study for two reasons. First, population projections at the regional or
national level by age, sex, and urban versus non-urban do not exist in Thailand. Second,
projection of national or regional population as well as the provincial ratio would have to

be performed.

46

There is a possibility to use the ratio concept in this study for of the fertility and mortality
submodel. Age and sex fertility and mortality rates specific to urban and non-urban
populations are not projected for each province. The most recent historical data could be
used to calculate the ratio of the age specific fertility/mortality rate for the province to the
nation. The provincial age and sex specific fertility rate could then be calculated by
multiplying this ratio by the projected national age and sex specific fertility rate.
However, the statistical trending method is more appropriate to project births and deaths
in this study because of the flexibility of the method and the ability to incorporate expert

opinions into the model (This is discussed more in Chapter IH).

2.6.3 Cohort-Component Method

The cohort-component method, also known as cohort-survival, deals explicitly with the
fertility and mortality components of population change. The Cohort-component method
takes into account the process of aging. 1t desegregates population into cohorts (i.e.,
population segments) of uniform age and sex. The projection is carried out based on the
surviving members in each age group over time into the succeeding age group. The

cohort-component model yields projections of firture population size and its composition.

Changing population over time in relation to fertility, mortality, and migration is

generally represented as the equation below:

P1-Po=F-M+NM

47

Where:

PrPo = the populations at times 1 and 0, respectively;
F = births during the period 0 - 1;

M = deaths during the period 0 - 1; and

NM = net migration during the period 0 -1

Davis (1995) summarized the cohort component model mathematically as will be
described below and made the following assumptions. The population is divided into n
male and n female cohorts. All cohorts other than the last of each sex consist of the
specific population of an age group of m years. The 2n population cohorts projected m
years into the future can be described algebraically. The cohorts projected m years into

the future are calculated as:

F2 = F1 ' F31
Mn-f = MID—2 1' MSn-z
Mn—t = Mn—2 . MSn—Z

Fn-t = Fn—Z . an—2

M.= i

48

where:

PM = Proportion of male births.

l

M,-F,~ = Male cohort ,~ and Female cohort ., respectively;

M8,, FS, = Survival rates for male cohort ,- and female cohort ,, and
BR,- = Birth rate for female cohort,

The Cohort component method relies heavily on assumptions made in the projection of
various components. Projections of age specific fertility, mortality and migration rates
for females and males in the area of study are required. This means that this method
demands detail and a great amount of data. However, this method should provide more
accurate projections since each component will be projected and evaluated separately.

Additional variables related to each component can be included in the projection process.

The Cohort component method is ofien used as the main approach to performing
population projections when age and sex are the major concern (Newell, 1988, and
Davis, 1995). This technique is also often used with other techniques such as
mathematical extrapolation and ratio techniques to calculate future births, deaths, and
migration. This approach has been used by the US. Department of Commerce, Bureau
of the Census and many local governments for population projections by sex, age, and
race in local areas (Bureau of the Census, 1977). This study also considers the Cohort
component method as the most appropriate method for projection since it handles

population data by age and sex.

49

2.7 Population Projection in Thailand

The National Economic and Social Development Board (NESDB) is the central planning
agency within the Office of the Prime Minister. The main function of the NESDB is to
prepare the government’s official statement on the direction of development for five-year
periods. As part of the preparation of the national development plan the NESDB is
responsible for projection of population of the country. The results from the projection
are utilized in other development planning such as infrastructure, health services, and

education.

During the course of preparing the population projections, the NESDB obtains population
data from various agencies, but mainly from the National Statistical Office (N S0). The
N80 is also a central agency within the Oflice of the Prime Minister and is responsible

for conducting the population and household census survey every 10 years.

The NESDB utilizes the cohort component method to prepare a population projection at
two levels: national and regional. Projection of urban and rural populations is also done,
but only at the national level. The section below describes three population projections
summarized fi'om the study on integrating development and population planning in
Thailand, conducted by the United Nations (1992). The projection results are not
included in this section. Some of the results are presented and discussed in Chapter III in

section 3.6 Discussion and Comparison of Projection Results.

50

2.7.1 National Population Projection

The NESDB prepared national population projections for the period 1980-2015 as part of
the Seventh National Economic and Social Development Plan. The age and sex structure
of the population was derived from the 1980 population and housing census data,
adjusted for under-enumeration and age-misreporting. The NESDB prepared national
population projections under three scenarios based on different assumptions regarding
fertility and mortality change. These three scenarios were formed based on three fertility
assumptions: high, medium, and low, and one mortality assumption that applied to all

three fertility assumptions.

The NESDB used the time trend functions to fit the historical national age-specific
fertility rate data obtained from the survey of population change, the longitudinal study,
and the contraceptive prevalence survey. The resulting functions were utilized to project
future age-specific fertility rates. Afier the age-specific fertility rates were obtained, total
fertility rates were calculated. These rates were utilized in the population projection

based on the medium fertility scenario.

The low and high fertility assumptions were formulated around the medium fertility
assumption. The low and high fertility assumptions stated that it would take less time for
the low fertility assumption and longer time for the high fertility assumption to achieve

replacement fertility than the medium fertility assumption.

51

In terms of mortality assumptions, The NESDB formulated a single mortality assumption
fi'om the life expectation at birth by sex. It was assumed that the life expectation of males
increased faster than that of females. This brought about the progressive reduction in the
excess of female over male expectation of life. The life expectancy for both sexes was
postulated to increase at a falling rate. The NESDB utilized life tables for the total
population of Thailand to derive the survivor ratios. These age and sex specific survival

ratios were used during the course of the projection.

The basic procedure to project population using the cohort component method is to
project births, deaths, and migration independently. Thailand utilized this method and
projected births using total fertility rate and deaths using survival rates calculated fi'om a
life table, however, it was not clear what migration projection method was used to
prepare firture migration rates for population projections. The common way to project
migration, when the cohort component is used as a mean to perform population
projections, is to calculate residual net migration by the forward survival method,
assuming that the migration rate observed will continue (United Nationsl992; US
Department of Commerce, Bureau of The Census, 1972). In the NESDB case, the
residual net migration was the difference between the 1990 census population and the
1990 expected population that was calculated from 1980 census population multiplied by

survival rates.

It is possible that a migration component was not taken into account when Thailand’s

population projections were carried out at the national level. Projection of migration

52

within the country is not necessary because changing places of residence from one place
to the other within the country does not effect the change of the number of people in the
country. Immigration has a negligible impact on Thailand’s national population, so it
was unnecessary to take into account this component during the course of population

projections.

The cohort-component method seems to be the appropriate method for Thailand to
project population at the national level for several reasons. First, the method is
straightforward, simple to use, and is logical if base data (population, births, deaths, and
migration) is available and correct assumptions are made. Thailand has a good set of
historical population and fertility data at the national level. Thailand Population and
Housing Censuses provide information with great detail on population (broken down by
age and sex) and fertility (broken down by age) that could be traced back to the year
1911. Because of the relatively reliable historical fertility data, using the mathematical
curve fitting function was appropriate to project firture births. Projections of population
based on the assumptions of three levels of fertility provided good information in terms

of the range of population and the uncertain future behavior of fertility.

As far as the approach used for mortality projection is concerned, the calculation of age
and sex specific survival rates fi'om the appropriate life table is a common approach. The
population and Housing Census before 1990 did not provide information on the
expectation of life by sex and age groups. Age-sex-specific survival ratios were

computed using the West family of the Coale-Demeny model life tables (Ministry of

53

Public Health, 1994). Because of less reliable information on life expectancy at birth, it
is arguable whether the postulation of life expectancy is appropriate to obtain future death
rates. In Thailand’s case, this method seems to be acceptable because the life expectancy
data was the best reliable information available in the county as far as mortality is
concerned. Although stated in the assumption regarding mortality that the life
expectation of males increased faster than of females, this assumption did not agree with
historical statistics. The expectation of life by sex and age groups from 1969-1991
illustrated the faster rate of increased expectation of life of females in young age groups

and equal rates in the young adult age groups (Public Health Statistics, 1994).

2.7.2 Regional Population Projection

The NESDB prepared two sets of regional population projections for the seven regions of
Thailand based on two internal migration assumptions. The internal migration
assumptions were formulated from five-year interregional migration information in
censuses of 1960, 1970 and 1980. From the size and patterns of the migration in those
periods, the first assumption postulated a constant age-sex structure of migrants and the
constant regional distribution of migrants that were consistent with a historical rate on
internal mobility. The second assumption postulated a fixed age-sex structure of
migrants with a changing regional distribution of migrants. This change was compatible

with a declining overall rate of internal mobility over time.

54

Regarding preparation of the regional fertility rate, the age-specific fertility rates were
obtained fiom adjusted registration data on births and the number of females in each
region in the period 1972-82 for the projection of filture fertility rates. The time trend
functions were fitted to these age-specific regional time series. These functions were
used subsequently for the projection of future age-specific fertility rates of women in
each region. Adjustment was made so that these age-specific fertility rates after being
weighted by regional numbers of women in the child-bearing age would add up to the

national age-specific fertility rates under the medium fertility assumption.

In the case of mortality, assumptions were made by postulating initial assumptions on
regional-level mortality by sex using expectations of life at birth. Adjustment was also
made for the regional expectations of life at birth after being weighted by regional
populations by sex so that they would add up to the national-level expectations of life by

sex.

The preferred migration data for the migration projection for a large area such as a region
is gross migration rather than net migration because gross migration reflects the closer
relationship to the real migration process. During the course of migration projection,
Thailand obtained gross migration data from the census from five-year migration data.
The first assumptions that postulated the constant age and sex structure of migrants along
with the constant regional distribution of migrants were unrealistic. Historical migration
data based on the age and sex structure of migrants showed insignificant change in age

and sex structure of migrants. It is also arguable whether changes in regional distribution

55

of migrants were constant. Previous migration patterns showed variations in the number

of migrants in each region in different census years. However, whether the changes were
constant is still in question. The second migration assumptions regarding unchanged age
and sex structure of migrants and the changing distribution of migrants were more

compatible with historical migration trends.

The projection of fertility was prepared in the same manner as for the national projection
except that the base data was from difference sources. Adjusted Household Registration
data at the regional level was used instead of the Population and Housing Census Data.
By using the registration data, there could be problems associated with the proper
allocation of births to the place of residence of the mother. This becomes more
problematic in areas with high migration. Recent migrants in Thailand do not usually
change residency. If a migrant had a child, the migrant has to notify the local registrar
office to obtain a birth certificate. In this case, the birth cannot be correctly allocated to
the place of the residence of the mother. When age-specific fertility rates were obtained,
the denominator, which is the number of women in the childbearing age group, did not
represent the true number of childbearing age population of the region. Although there is
a disadvantage associated with utilizing registration data to project fertility, one
advantage regarding utilizing registration data is that the data on births is accurate and

much more updated than the census data.

Regional mortality assumptions were formulated using the same procedure as the

national assumption except that they had to be applied specific to the regions. It was not

56

clear how Thailand coped with the problems of limited regional mortality information.
Problems arise when a life table that relates strictly to the population of each region is not
available. Ifage specific death rates are similar through out the country, national rates
could be used instead of the survival rates specific to age and region. In the case of
Thailand, national age-specific survival rates could be applied to the regional area
because the mortality difference between regions is not too large and has decreased over

time.

Although a migration component was built into the regional projection based on the
analysis of migration patterns fi'om census 1980, there was still no attempt to prepare a

projection of population in urban and rural areas for each region.

2.7.3 Urban and Rural Population Projection

Projections of urban and rural populations were carried out at the national level using the
United Nations modified ratio method. The United Nations (1984) population projection
method was carried out based on the assumption that urbanization would follow a logistic
curve. The only requirement for this method was estimates of the rates of growth of the
urban and rural populations. Once an assumption about the urban-rural growth rate
difference was made, the level of urbanization can be determined independently of the
rate of total population growth. The 1980 population and housing census and the
medium fertility variant of the national population projections were used to project the

urban and rural population for the period 1980-2015.

57

CHAPTER III

POPULATION MODELING: METHODOLOGY AND RESULTS

There are two main sections in this chapter: methodology and results. The methodology
section is divided into three subsections. The first subsection discusses the input data and
sources. The second subsection describes the preparation of baseline data. The third
subsection discusses the development of the population model. The result sections is
divided into three subsections. The first subsection presents projection results at three
levels: national, regional, and provincial. The second subsection presents and discusses
the sensitivity analysis of the population model. The last section compares projection

results of this study with other studies.

I MODEL AND METHODOLOGY

3.1 Data Input and Sources

Two types of data for population modeling were used in this study: population data and
the spatial data in Geographic Information System (GIS) Format, which was used to

reclassify population data and to generate map results.

58

3.1.1 Population Data

Population data used in this study were data of fertility, mortality, and migration. These

data were obtained from three sources:

0 Central Office for Civil Registration, Bureau of Registration
Administration, Department of Local Administration, Ministry of Interior:
Base Population Data for 1994

0 Ministry of Public Health: National age and sex specific fertility and
mortality rates for 1994

o The National Census of Population and Housing: F ive-year regional

migration data (Recent migration data) from the 1990 survey.

3.1.2 Geographic Information System (GIS) Data

The GIS data used in this study was obtained fiom the Thailand National digital database
called Thailand on a Disc, 1996, assembled by the Thailand Institute of Environment
(TEI). Thailand on a disc is a GIS database developed for use with PC ARC/INFO and
ArcView. The scale of this digital database is 1:250,000. The datasets are organized into
five regions: Central, East, Northeast, North and South. Each region contains nine layers
that include administrative, agricultural, rural socioeconomic, environmental, forestry,
geology, infrastructure, topography and water. The first three (administrative,

agricultural, rural socioeconomic) were the main layers used in the model. Three layers

59

were the base layers used to classify the base population into urban and non-urban

populations at the district level.

0 Socioeconomic Layer: This layer has a coverage called Human and an
associated lookup table’ that contains rural population data at subdistrict
level. There are 7113 subdistrict polygons in this coverage. Each polygon
has associated 1992 population data and a unique id based on the National

Rural Development Committee (NRDZC) registration coding scheme.

0 Administrative Layer: A coverage of political boundary in this layer has
polygons of boundaries of subdistricts with associated ids of subdistricts,

districts, and provinces based on the NRD2C registration coding scheme.

0 Agricultural Layer: This layer has a land use coverage that contains
polygons classified based on seven types of land use. This coverage was

used as a source to generate urban and non-urban coverage.

3.2 Baseline Data Preparation

Baseline data in the population model in this study were population, fertility, mortality,
and migration data. Only base data of population and migration were processed and are
discussed in detail below. Base data of fertility and mortality are described in the

modeling section and under fertility and mortality, accordingly.

 

5 A Look up table is a separate data table associated with a coverage that is linked to a feature attribute
table (polygon attribute table in this case) of a coverage by unique ids.

6O

3.2.1 Population Baseline Data

Baseline population data in this study was obtained from the Central Office for Civil
Registration. The data was the 1994 population data at the district level, classified by sex
with 104 age groups fi'om new born to age group 103 years and older. This data had to
be reclassified into 5 year intervals of 16 age groups by sex and urban versus non-urban

areas.

Each district was given a new id based on the NRC2D coding scheme because this
coding scheme would be used as a common key to join the population data to the GIS
data. Building this relational database required processing of the aforementioned GIS
data and the population datasets. The process of preparation is summarized in Figure 13

and explained below.

Processing of GIS Data

The purpose of processing the GIS data was to build coverages of urban and non-urban

population at provincial level. The human, political boundary, and land use coverages

were used in this process. Steps involved in the processing can be summarized as

follows:

61

Verrfling Original Data

The human and political boundary coverages were verified for missing data or duplicate
id". Each polygon must have a unique id for tables to be joined or a relational database to
be created. Polygon attribute tables of both coverages were exported and stored in MS

Access. Data verification was performed using query statements in MS Access.

Generation of New Coverages

The urban and non-urban human coverages were created with the following steps:

1) Generate national coverages of human, political boundary, and land use.

Original GIS data was organized by regions. The national coverages were generated in
ARC/INFO using the union command to join the regional data into the national coverage.
All coverages were converted into shapefiles so that the files could be easily used in
ArcView as themes.

2) Generate district population theme

The district population theme was generated by joining the subdistrict population data in
the lookup table of the human coverage with the polygon attribute table of the national
political boundary generated in 1). The joined table was processed in ArcView using the
mergetheme operation to dissolve boundaries of polygons of subdistricts, which share the

same district id, to form new district polygons.

 

6 It is necessary that each polygon (in this case represent each tambon or subdistrict) has a unique
identification. A unique id is used to differentiate each polygon and is important when operations such as
merging and intersecting polygons are performed.

62

3) Generate urban and non-urban land use themes

These new coverages were created through Query operations in Arcview. The national
land use file generated in 1) was reclassified from 7 to 2 classes of land use type: urban
and non-urban. A query statement was built to separate urban land use type from the rest
and this became urban land use theme. A query statement was built to combine all
except urban land use types together to form a non-urban land use theme.

4) Generate urban and non-urban district political boundary coverages

These coverages were created using the Arc/Info intersect operation. Intersecting
coverages created in 2) and 3) yielded a coverage that shows polygons of district based
on urban and non-urban land use type (See Figure 13). The polygon attribute table of this
coverage was later joined with the processed population data from the Central Office of
Civil Registration.

5) Generate provincial political boundary theme.

This theme was generated in ArcView using a merge operation to dissolve boundaries of
district polygons that share the same provincial id and form provincial polygons. This

theme was used to display the results of population and electricity projections.

63

 

  
 

Bangkok

5 Province

Urban and Non-Urban Area
NorrUrhan

Urban

 

 

 

 

Figure 13: Political Boundary of Provinces in Thailand by Urban and Non-Urban Land

64

The population data from the lookup table of the human layer was replaced by the
population dataset fiom the Central Office of Civil Registration. The main reason for this
was that the human layer did not include population data for urban areas. The Central
Office of Civil Registration, on the other hand, has more current, detailed, and complete
population data of the entire country (both urban and non-urban population data). A
utility was written in MS Visual Basic to process the population dataset from the Central
Office of Civil Registration so that it could be used to replace population data in the
lookup table of the human layer. The utility did the following:

1) Replaced the subdistrict and provincial codes in the Central Office of Registration
with the NCR2D subdistrict and provincial codes in the Human lookup table.

2) Classified Data by Age Group

A five-year interval of ages was use to classify population data from the Central Office of
Civil Registration were reclassified using a five-year age interval. The resulting data was
the baseline population data grouped by sex and divided into 16 five-year span age
groups.

3) Assigned urban and non-urban codes

Urban and non-urban codes were assigned to processed data in 2) by joining this data
with the urban and non-urban district political boundary coverages created previously.
The resulting data was used as the baseline population data for the population modeling

in this study

65

 

Registered

Text file format

 

 

 

 

Population Data (1994) r—fi Management System (DBMS)

 

 

i

 

 

 

 

 

 

 

 

 

Import into 011an Geographic Data by Region
(Shapefiies)
MS Access
Human
Political Boundary
Land Use
Import to GIS ARCIINFO
Clean and build
geographic coverage

Create urban and
Non-urban coverage by
reclassify land use
cover

 

 

 

 

 

Data Manipulation
(MS Visual Basic)

match codes to GIS data
classify by sex and 5 year age

 

 

 

l

—*r

 

 

Link DBMS to GIS (ArcView)

Replace (518 population data
with registered population

 

 

 

 

Import to Gis ArcView

Convert national
coverages to shapfiies

Identify missing data

Create population coverage
bitty joiningHUMAN.GEN_POP

To POLBNDPAT

Create subdistrict urban and
non-urban population coverage

 

 

 

 

1994 baseline population data
I classified by sex. age group.
and urban and non-urban residency

 

 

Figure 14: Flow Chart of Baseline Population Preparation

66

 

3.2.2 Migration Baseline Data

Regional five-year migration data from the Regional Population and Housing Census
1990 was utilized as the base migration data. Some characteristics of the 1990 census in
terms of administrative political boundary can be described as follows. In 1990 Thailand
had 73 provinces7 (76 provinces in 1994) and the regions, according to the census, were
classified into four regions: the Central, North, Northeast, and South. The East region did
not exist at that time and provinces, which now belong to the East region, was included in
the Central region. As a result, there was no interregional migration data or data of
migrants fi'om other provinces moving into the East region. Furthermore, the census

treated Bangkok as an individual region, excluding it from the Central region.

The regional five-year migration baseline data used were available in a tabular format,
which can be described as follows. There was a total of five tables -- one for each region.
The row listed provinces (within the region) of residence at census, 1990. The column
listed previous regions of residence before 1985. The columns were consisted of the
Central, North, Northeast, and South regions, and Bangkok. Each regional tabular data
provided the number of male and female migrants fiom each region moving into each
province listed within a region. For each province, data on migrants moving within

province was also available.

 

7 As a result of administrative adjustment in 1992, there are three new provinces.

67

Notice that migration data were given at the regional and provincial level and classified
by sex. No data concerning age and urban and non-urban status of migrants were
available. In order to determine age of migrants in each province, national age
distribution of migrants for male and female was applied to all migrants in each province.
The number of male or female migrants in each age group was determine by multiplying
a probability of male or female in each age group to migrate to number of male or female
in each age group. Determination of urban and non-urban status of migrants was based

on percent urban and non-urban population in each province.

From the given data, three types of migration rates for male and female were calculated.
At the provincial level, percent migrant of a province and percent migrant moving within
province were calculated. Migration rate of a province was calculated fiom the number
of migrants in a province divided by the total population in a province. Within province
migration rate was calculated from the number of migrants moving within a province
divided by the total population of a province. At the regional level, percent migrant from
other regions was calculated from the number of migrants from each region divided by
the total population of a province. For each province, there were into province migration
rates from Bangkok, the Central, North, East, Northeast, and South region. The
calculated rates were used as based migration data in the population model. Calculation

of the rates was performed using MS Excel and the rates were stored in MS Access.

The processing of the migration data included assigning the NRC2D provincial codes to

all provinces for each table. This step was essential because population projections were

68

performed at the provincial level. All provincial codes of the migration data had to
conform to the provincial codes of other data tables or GIS data for the model to
recognize the migration data. The three new provinces with no migration data used the
migration data from the parent provinces before the change of the political boundary.

The baseline migration database was built and stored in MS Access.

3.3 Population Model

This subsection explains the population model developed in this study. Assumptions and
calculations involved in the model are discussed. This subsection is divided into four
parts. The first part describes the overall population model in terms of methodology
used. The remaining parts discuss the fertility, mortality, and migration component in

detail. Examples of calculations are also included.

3.3.1 Overview of the Population Model

Population projections were performed at provincial level by age, sex, and residency. It
is necessary to perform the projections at this level of detail. Age structure of population
plays important roles in terms of births and deaths. Age structure of female population in
childbearing age group determines number of newborns. Population in different age
group also has different probability or risk to die. Since the projection was performed at
provincial level, migration plays crucial role in determination of provincial population. It

is necessary to be able to determine residency of population and migrants.

69

The projections of population in this study were performed using the most current
estimate of age and sex specific birth and death rates, migration rates combined with
trends of these variables based on expert opinion, results of other studies, local

knowledge, or intuition of firture changes in these trends.

The Cohort-component technique was used to perform population projections (Figure
15). The base population was categorized into 5-year age groups by sex and area of
residency (urban and non-urban) in order to perform 5-year period projections from 1999
to 2019. Based on the birth trends, Births were estimated from the summation of
newborns to women in the childbearing age group. Deaths by age and sex were also
projected, based on the trend of death rates. At the end of every five years, the total
population was calculated by adding the number of newborns and subtracting deaths,
within five years, fi'om the base population. The projected number of migrants was
performed after this adjustment since migration data excludes all persons, both migrants
and non-migrants that die during the projection interval. Since the cohort-component
model has run in five-year time interval, migration estimates must also be done in five-

year interval.

Notice that the model involved the projection of each population component separately.
Multiple regression, a common approach to perform projections, is not suitable for this
study because it requires the modeler to identify causal factors of each population

component and use these to build models. A good example is the case of the migration

model. Construction of a regression model of migration would require estimation for

70

causal factors of the 76 provinces in Thailand. Regression models such as this are
typically linear or follow some continuous (increasing or decreasing) function, which is
not always the case when considering the trend of certain variables. Construction of a
regression model of migration would also require time-series data of provincial migration
in order to define the model (not available for this study). Because of the complexity of
modeling provincial migration, the lack of adequate time-series data on inter-provincial
migration, and the likely inability of the resulting model to explain a high proportion of
the variance in migration, it was decided that regression modeling would not be used to

project firture migration amounts and patterns.

Rather than regression modeling, the trending method was used for modeling provincial-
level migration (see detailed discussion under migration). This method was also used for

modeling provincial level fertility and mortality for similar reasons as migration.

Discussion of population modeling is divided into three subsections: fertility, mortality,
and migration. Each subsection discusses base data, assumptions, trending, and
calculations in the model. Migration is discussed in great detail because of the
complexity of the model, assumptions formulated, and the direct impact on urban and

non-urban populations.
The overall population model was developed as an integrated application. MS Access,
and MS Excel were used to create necessary tables and store data and projection results.

The Visual Basic programming language was used to define and run the model. The

71

projection results were also mapped using ArcView. The summary of the process

involved in developing the population model is illustrated as a flow chart (Figure 15).

 

Population and Housing Census, 1990

National Statistical Office

 

 

 

 

Data Entry

Five-year regional
Migration data by sex

 

 

 

 

+

Data Manipulation
(MS Excel)

 

Calculate percent migration
by region and sex

 

 

 

 

 

Data from other sources

Data Entry of

Public Health Statistics, 1994
National age and sex
death and birth

National Statistics Office, 1993
Age distribution of
Migrants by sex

 

 

Inport of Database
Management System

 

 

(MS Access)

 

 

 

 

 

Develop lation model

(MS Visual Basic)

 

 

 

Figure 15: Flow Chart of Processes in Developing the Population Model

72

 

3.3.2 Fertility Modeling

Base Data

The national age-specific birth rates of child bearing women age group 15-49 yeas old in
1994 were used because of the unavailability of this data at the subdistrict, provincial, or
regional level. The data was obtained from the Public Health Statistics, Ministry of

Public Health.

Assumptions

1) There is no significant dtflerence in the fertility rate specific to age group nationally,
regionally, or locally (provincial level).

2) There is no significant diflerence in the spatial pattern of age-specific fertility rates
for urban and non-urban women in the childbearing age groups.

In the absence of historical trends of age specific fertility rates of women in urban and

non-urban areas, historical total fertility rates (TFR) were used to observe the fertility

patterns between women in both areas. The TFR of non-urban women had always been

slightly higher than the rates of urban women. For example, in the period 1974-1976, the

TFRs of non-urban and urban women were 4.98 and 4.49 births per 1,000 women

respectively. Because of the small difference in the past and the tendency for lesser

difference in the firture, this study assumed that urban and non-urban women in the same

age group have the same fertility rate.

73

3) The probability of a newborn female is 0.5 and the probability of a newborn male is
0.5.

This assumption was necessary because the calculated number of newborns (births) had

to be broken down by sex in order to calculate the number of males and females born to a

particular female age group in the projection year.

4) Yhe age of childbearing women is between 15 to 49.

Selection of the childbearing age group in this study was based on population literature

that usually indicates the age group to fall within this range. Statistically very few births

occur to women aged below 15 and older than 49.

Trending of Birth Rates

The over all age specific birth rate in Thailand has not changed significantly the last five
years (1990-1994). There was a slight decline of the rate in most age group, however,
especially the high fertility age group: women aged 20-24, and 25-29 years old (see
Figure 8 in Chapter H). Based on current (1994) age specific birth rate, the rate was

trended, using expert opinion, to decline by 1 percent per year in the period 1994 to 2019.

Calculations
The national age specific fertility rates were applied to childbearing women in urban and
non-urban areas in every province. Calculations of the number of births were performed

at the end of each projection year (every five years). For every new projection year, the

74

trended rate - a five percent decrease from previous projection year — was used in the

calculation. The equations for the calculations are listed below.

10
Births = Z(brWr"5)/1000
i=4

Where:
Births,- = Number of newboms to all women in child bearing age group
b, = birth rate of women age group i (births per 1,000 women age group i),
w,- = Number of women age group i,
i = Age group (group 4 is age 15-19 years old, and age group 10
is age 45—49 years old, and
5 is the number of year during one projection period

Births were calculated only for women in childbearing age groups (group 4 to 10). After
births were calculated, the next step was to determine the number of newborn males and
females. This number later becomes the base population for age group 0-4 years old for
the next projection year. The equation used to calculate the number of males and females

in age group 0-4 years old in the following projection year includes:

PM = Births * (1-Pctm)

Mm = Births * (Pct...)

Where:
FM, = Number of newborn females,

M"... = number of newbom males, and

Pet... = Probability of newborns being males = 0. 5

75

3.3.3 Mortality Modeling

Base Data

The national age and sex specific death rates for the year 1994 was used as the base death
rates because of the unavailability of this data at the subdistrict, provincial, or regional
level. The data were obtained from the Public Health Statistics, Ministry of Public

Health.

Assumptions

1) The dtflerence between the death rate specific to age group and sex is not significant
at the national, region, and local levels.

2) Mortality rates were not significantly diflerent between urban and non-urban
population within the same sex and age group.

This assumption was formulated based on indirect social measures associated with health

and death. Improved living standards and government spending in heath care programs

should lessen the difference in death rate between the urban and non-urban population.

3) There will not be a catastrophe or major events, such as a war or new disease, that

will significantly reduce the population in the next 20 years.

76

Trending of Death Rates

In the absence of the historical age specific mortality data, the recent trend of CRD was
observed and applied to the age specific death rate. CRD in the last five years (1990-
1994) showed a much more gradual decrease compared to the rapid decrease between
1950 and 1990 (see Figure 6 in Chapter 1). Based on expert opinion, firture death rate
was determined using current (1994) national age and sex specific death rate, which was
trended to decrease 0.05% per year or 2.5% in the period 1994-2019. Notice that the
death rate was trended to decrease at a much slower rate than the birth rate (1% per year)
because it generally takes a much long period of time to change life expectancy (used to
determine age specific death rate). Therefore, there will not be a large change in the life

expectancy within 25 years.

Calculations

Calculations of deaths were performed every five years and conducted separately based
on area of residence (urban and non-urban) for males and females in each province. The
death rate specific to each age group and sex was applied to the base population of the
projection year. Trended age and sex specific to death rate, a decreased by 0.05 percent
of the previous projection year, were used for the next projection period. The equation

used to determine the number of deaths in each age group is as follows:

77

Age group Age

10

11

12

13

14

15

16

0-4
5-9
1 0-14

15-19

20-24
25-29
30-34
35-39
40—44
45-49
50-54
55-59
60-64
65—69
70—74

75 Up

 

Year
2004

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Year Year
1994 1999
Births \
\
[31 \
P2 P,
P3 P2
P4 ............ P3
- , .. - deaths “"t———~

P5 ............ P,
P6 P5
P Childbearing P

7 Age group 5
Pa P7
P9 P3
P10 + P9
P11 P10
I”12 P11
P13 P12
P14 P13
P15 “““““““““ P14

~ “*
P18 “““““““ P15
L ————— " ‘~“\“i— ______

 

 

 

78

 

Figure16: Flow of Population under Cohort-Component Method

16
Deaths = 2(dr pr "’ 5)/1000
i=1

Where

Deaths = Total deaths of population in all age group,

(1,- = Death rate of population age group i, (deaths per 1,000 persons in age group i ),
P, = Population age group i, and

i = Age group = i = 1 to 16 (group1 is 0-4 years old and age group 16 is 75 years old

and up), and 5 is the number of year during one projection period

When births and deaths by sex, age group and urban versus non-urban areas were
calculated, the total population before migration was calculated by adding projected
births and subtracting projected deaths from the current population. Figure 16 shows the
flow of population between projection period 1 and 2. An example of the calculations of
births, deaths and population of urban males and females in each age group, using the

Cohort Component approach, is given below.

79

Table 2: An Example of Baseline Population Data in Urban Area by Age Group

Baseline population: Time 0

 

 

Age Male F'emale
group Age Number of urban Number of urban death rate death Birth rate
male female rate

1 0-4 98 100 2.3 1.7 0

2 5-9 103 81 0.6 0.4 0

3 10-14 118 98 0.7 0.4 0

4 15-19 146 130 2.3 0.7 42.8
5 20-24 186 152 3.4 0.9 106.4
6 25-29 179 171 4.3 1 103.9
7 30-34 175 178 4.5 1.2 66.8
8 35-39 155 184 4.8 1.6 30.2
9 40—44 117 166 5.4 2.3 10.1
10 45-49 112 126 6.7 3.3 2.7
11 50—54 89 99 9 4.9 0
12 55-59 87 96 11.9 7.1 0
13 60-64 73 91 18.1 11.2 0
14 56-69 60 61 27.7 18.6 0
15 70-74 37 45 63. 3 54. 5 0
16 75 up 44 78 63.3 54.5 0

Total 1779 1856

Proportion of male and proportion of female birth = 0.5

 

80

Table 3: Calculation of Number of Births, Deaths by Sex and Age Group of Urban

Population Using Cohort Component Method (An Example)

Projected birth and deaths: Time 1

 

 

Age
group Age Number of death urban Number of death urban Number of urban births
male female
1 0-4 0.23 (98 x 2.3)l1000 0.17 (100 x1.7)/1000 0.00
2 5-9 0.06 (103 x 0.6)l1000 0.03 (81 x 0.4)I1000 0.00
3 10-14 0.08 (118 x 0.7)I1000 0.04 (98 x 0.4)l1000 0.00
4 15-19 0.34 (146x2.3)l1000 0.09 (130x0.7)l1000 55.64 (130x42.8/1000)
5 20-24 0.63 (186 x 3.4)!1000 0.14 (152 x 0.9)l1000 161.73 (152 x106.4/1000)
6 25-29 0.77 (179 x4.3)/1000 0.17 (171 x 1.0) I 1000 177.67 (171x103.9l1000)
7 30-34 0.79 (175x4.5)/1000 0.21 (178x 1.2)!1000 118.90 (178x66.8/1000)
8 35-39 0.74 (155 x 4.8)l1000 0.29 (184 x1.6)/1000 55.57 (184 x 30.2/1000)
9 40-44 0.63 (117 x 5.4)l1000 0.38 (166 x 2.3)l1000 16.77 (166 x10.1/1000)
10 45-49 0.75 (112 x 6.7)l1000 0.42 (126 x 3.3)!1000 3.40 (126 x 2.7/1000)
11 50-54 0.80 (89 x 9.0)!1000 0.49 (99 x 4.9)!1000 0.00
12 55—59 1.04 (87x 11.9)l1000 0.68 (96x7.1)l1000 0.00
13 60-64 1.32 (73x 18.1)l1000 1.02 (91 x11.2)l1000 0.00
14 56-69 1.66 (60 x 27.7) I 1000 1.13 (61 x18.6)/1000 0.00
15 70—74 2.34 (37 x 63.3) I 1000 2.45 (45 x 54.5) I 1000 0.00
16 75 up 2.79 (44 x 63.3)/ 1000 4.25 (78 x 54.5)/ 1000 0.00
Total 15 12 59

 

81

Table 4: Calculated Total Urban Population by Sex and Age Group, Time 1

Projected population: Time 1

 

 

Age

group Age Number urban male Number of urban female
1 0-4 29 a 29 a
2 5-9 97.77 (98 — 0.23) 99.83 (100 —0.17)
3 10-14 102.94 (103 - 0.06) 80.97 (81— 0.03)
4 15.19 117.92 (118 -0.08) 97.96 (98—0.04)
5 20-24 148.45 (146 + 2.78 -0.34) 132.32 (130 + 2.78 — 0.09)
6 25-29 193.45 (186 + 8.09 - 0.63) 159.95 (152 + 8.09 - 0.14)
7 30-34 187.11 (179+ 8.88-0.77) 179.71 (171+ 888-017)
8 35-39 180.16 (175+ 5.95-0.79) 183.73 (178+ 595-021)
9 40—44 157031 (155 + 2.78 - 0.74) 186.48 (184 + 2.78 - 0.29)

10 45-49 117.21 (117 + 0.84 — 0.63) 166.46 (166+ 0.84 — 0.38)

11 50-54 111.42 (112+0.17-0.75) 125.75 (126+0.17-0.42)
12 55-59 88.42 (89 - .80) 98.51 (99 - 0.49)

13 60-64 85.96 (87 - .1.04) 95.32 (96 — 0.68)
14 56-69 71.68 (73 -1.32) 89.98 (91 -1.02)
15 70-74 58.34 (60 - 1.66) 59.87 (61 - 1.13)
16 75 up 75.87 (37 - 2.34) + (44 -2.79)] 116.3 (45 — 2.45) + (78 - 4.25)
Total 1823 1903

 

a = Number of new born male = number of new born female = [(130 x 42.8/1000) + (152 x
106.4/1000) + (171 x 103.9/1000) + (178 x 66.8/1000) + (184 x 30.2/1000) + (166 x 10.1/1000) +
(126 x 2.7/1000)] x 0.5 = 29

In this example, the total number of urban males increases fi'om 1779 to 1823 persons
and the total number of females increases from 1856 to 1903 persons from projection

Time 0 to 1 respectively.
The population model performed the same calculations as the above example of births,

deaths, and total population for each sex in 16 age groups in urban and non-urban areas

for each of the 76 provinces of Thailand.

82

3.3.4 Migration Modeling

Base Data

For each province, the following migration rates for male and female were calculated
from the regional five-year migration data and used in this study: total migration rate,
within province migration rate, rates of migration from other provinces. These migration
rates were assumed to remain constant for the next 25 years. The migration rates were
not broken down by age and residency (urban or non-urban). Some assumptions or
exceptions were applied to estimate migrants of each province by age, sex, and area of

residence and are listed below.

Assumptions

1) Net migration rates fiom within province migration were equal to 0. 03, 0. 06, and
0. 06 for a province with rural population greater than or equal to 10 percent, 15
percent, and 20 percent respectively.
The within province net migration rate referred to the difference between migration rate
of the urban to rural stream and rural to urban stream. The net migration rate of 0.06 was
the average number of the net gain of rural to urban migration within a province at the
national level during the period 1965-1990 (see Table 1 in Chapter 2). This study
assumed that population of provinces with higher proportion of non-urban population
was more likely to migrate within a province from non-urban to urban area. Therefore,

the national net migration rate (6%) was used for provinces with 15% non-urban

83

population. The rate of 5% and 3% was used for provinces with non-urban population

equal or greater than 20% and 10% respectively.

2) The probability of migrants to settle down in urban and non-urban areas of each
province was equal to a proportion of urban and non-urban population in each
province. Provinces in the Em and EBMR were treated as special. The probability
for migrants to settle down in non-urban areas of these provinces was 0. 35 while the
probability for migrants to settle down in urban areas was equal to the proportion of
the population of in urban areas in each province.

These probability values were applied to the total number of migrants in a province to

determine the number of migrants that settled down in urban and non-urban areas.

Because of the nonexistence of urban and non-urban migration data at the provincial

level, this study assumed that the probability for migrants from other provinces to settle

down in urban and non-urban areas in a province is the same of the proportion of urban
and non-urban population in the province respectively. Provinces in BMR and EBMR
are fast growing provinces with a high potential for jobs and housing. As a result, it was
assumed that the probability for migrants to settle down in non-urban areas in these
provinces is lower than the proportion of non-urban population. However, the probability
for migrants to settle down in urban areas of provinces in the BMR and EMR is the same
as proportion of urban population in these provinces.

3) Number of migrants taken out of a province origin was a proportion of percent
province population to total migrants of each region.

Base migration data provided only migrant data from each region moved into each

province. In this case, the designation province was known, but the origin province of

84

migrants could not be determined. In the normal case when both data were available,
afier adding migrants into designation province, the same migrants could be subtracted
out of origin province. In this study, the model could not perform this task
simultaneously. The model had to add migrants into province one at a time until
completed the calculation of 76 provinces. When performing this task, the model
accumulated migrants which had to be subtracted out from origin provinces by regions.
At this point the model had total number of migrants from each region which had to be
broken by province. Determination of migrants to be subtracted out of origin province
was then determined based on fraction of province population to total population of each
region.
4) There was no significant diflerence between the age distribution of male and female
migrants at the national and provincial level. The age distribution was also the same
for urban and non-urban migrants across the nation.
The data at the national level on age distribution of migrants by sex was obtained from a
migration survey conducted by The National Statistical Office (United Nations, 1993).
The table of this distribution is shown in Appendix A. The probability for males and
females in each age group to migrate was applied to urban and non-urban migrants to
determine the urban and non-urban population in each age group. The assumption
regarding no difference in age distribution of migrants between province was in part
supported by the regional summary on census migration data, 1990 (National Statistic
Office, 1993). The results demonstrated that the percent migration of females and males

in the same age group in each region were very similar.

85

Calculations

The calculations for migration modeling involved determining four main types of
migrants: total migrants in a province, within province migrants, in-migrants from other
provinces, and out-migrants from a province. A diagram illustrating the calculation of
the first three migrants is shown in Figure 17 and a diagram illustrating the calculation of
out of province migration is shown in Figure 18. Calculation of migration in the
population model in this study was done in similar manner for both males and females in
urban and non-urban areas. When the word “population” is used in the calculation, it
refers to males or females and the word “each area” refers to urban or non-urban areas in

each province. The calculation in each step is described below.

1) Total migrants in province i

Calculation of migration in the model started by determining total migrants. Total

migrants were calculated from multiply percent migrants in a province with population in

the province.

Mi = wi * P m
Where:
M, = Total migants in province i,
W,- = Population in province i, and
P,,, = Percent migation of population in province i,

86

Migrants in province i were consisted of migrants for other province and migrants that
moved within province i. Therefore, calculated total migrants in province i was later

used to estimate within and between province migrants.

2) Within Province Migration

Based on the calculated total migrants in province i and the assumption made regarding

percent migration within a province (Pm), the migrants moving within province i was

calculated using the equation below:

X.- = M.- * Pw/n
Where
X,- - Migants from within province i,
M,- - Total mig’ants in province i, and
PM, = Percent within a province migration

The calculated within province migrant of province i was the migrants from the four
migration streams (rural-rural, urban-rural, rural-urban, and urban-urban). The migration
stream that determines population in each area of a province was the migrant in the
streams of the rural-urban and urban-rural stream. In the case of Thailand, the net
migrant resulting from within province migration was always a net gain in rural-urban

stream. The other two streams did not effect the net population in each area. From the

87

calculated within province migrants, migrants resulting from net migration between the

rural-urban and urban-rural stream were determined using the formula below.

Y,‘ = Xi * P not
Where:
Y, = Net migrants within province i,
X,- = Migrants from within province i, and
PM, = Percent within a province net migration

The next step was to extract calculated net migrant based on migration designation (urban
or non-urban areas). The calculation was performed based on the proportion of urban
population within province i for non-urban to urban migration and the proportion of non-
urban population within each province for urban to non-urban migration. Since the net
migrant was the result of the net gain to urban areas, therefore, only migrant in urban

areas was calculated.

Urban Pop = (Yr) (Pu)

i
3
é”
ll

Within province net migrants settled down in urban areas,

3’
ll

Proportion of urban population within province i, and

The next step was to determine the net migrants in urban areas by age group. It was
estimated by applying the age distribution (P,-) of migrants to the net urban migrant. The

calculation used the formulas below:

88

2,,” = (Urban Pap) ( P,)

Where
2,,“ = Number of urban migrants from within province i in age goup j,
P) = Probability of migrants to be in age goup j, j = 1 to 16

At this point the model had the number of within province female and male migrants by
age group, and urban and non-urban areas. The next step was to calculate the number of

migrants from other provinces.

3). In-Migration from Other Provinces

For province i, migrants from other provinces was calculated by subtracting the migrants

from within province i from the total migrants in province i. The calculation is shown

below:

IN.- = M, - )(i
Where:
IN, = ln-migrants in province i from other provinces,
M, = Total mig‘ants in province i, and
X, = Migrants from within a province i

The next step was to extract the calculated in-province migrants (IN, ) in order to

determine the migrants that settled down in urban and non-urban of province i. The

89

calculation was performed based on percent urban and non-urban of population in

province i. This can be written as:

1 u = M * Pu
Inu = N, * Pnu
Where
[:4 = Migrants from other provinces settled in urban areas of province i,
[nu = Migants from other provinces settled in non-urban areas of province i,
N, = Migants in province i from other provinces,
Pu = Percent urban population in province i, and
Pnu = Percent non-urban population in province i

After the in- province migrants in urban and non-urban areas in province i were
calculated, the model had to detemiine these numbers based on age group of the
migrants. The calculation was done by applying age distribution of migrations to the
urban and non-urban migrants from other provinces. Equations used to perform the

calculations were as follows:

Ow = lu * P,

Onw — Inu * P,
Where
0",, = Migrants from other provinces age group j settled in urban areas of province i,
OM, = Migrants from other provinces age group j settled in non-urban areas of province i,
P, = Probability of migrants to be in age group j, j = 1 to 16

90

At this point, the model calculated within province and from other province migrants of
all provinces broken down by age, sex, and residency. The next step was to calculate
population in urban and non-urban of each province adjusted by migrants. The

calculations were performed using equations below.

Non-urban Population

Non-urban population of province i adjusted by migration was determined from adding
migrants from other provinces that settled down in non-urban areas in the province and
subtracted by the population of the province that moved out of non-urban areas of the

province. The equation used to perform the calculation is listed as follows.

Adjusted Popnuj = Poanj + Onuj - Znt
Where
Popnu, = Current non-urban population in province i, age group j,
Onuj = Migants from other provinces age group j settled in non-urban areas of
province i,

Z,,.,,- = Z“, = Migrants within province iin age group j moved out from

non-urban areas of province i.

Urban Population

The current urban population (before migration) for province i had to be subtracted by the

number of migrants moving out from this area (urban to rural within province i ), added

91

by the number of migrants from other provinces, that settled in the urban areas. The

calculation can be represented in the equation below:

Adjusted Pop“, = Pop“, + Ouj - 20]
Where:
Popui = Current urban population in province i, age group j,
Ou, = Migants from other provinces age goup j settled in urban areas of
province i, and
Z", = Z“, = Migrants within province iin age group j moved out from

urban areas of province i.

Notice that at this point the model estimated within province migrants and in-migrants for
province 1' and adjusted them to the current population. Since the province origin of
migrants added into province i (desingation province) were not available, migrants added
into province i could not be subtracted out from the origin province. This meant that the
model had to perform the calculation of within province and in-migrants for all 76
provinces, at the same time accumulate migrants, by regions, that had to be subtracted out
fi'om the origin provinces. Therefore, the next step was to determine migrants to be taken

out of origin provinces.

92

 

Current Population of
Province i (Y)

A

Male Female

 

 

 

 

 

 

 

 

 

 

 

Projected Non-urban Female = Y - A + C

Projected Urban Female = Y + A + B

A/|\x

 

 

 

l Non-Migrants Total Migrants

 

 

 

 

 

Net Within
migrants province

 

 

 

 

 

 

 

- migrants from origin province

 

 

A

 

 

 

Other
provmce

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A A4 % Non-Urban Pop

‘——_ % Urban POPulation /

\

Non- Urban Migrants

 

 

 

 

 

 

Urban Migrants Urban Migrants
Age distribution
1,
16 age groups (A) 16 age groups (B)

 

 

 

 

 

 

 

16 age groups (C)

 

 

 

Figure 17: Diagram of Calculation of Female Within and From Other Province Migrants

93

 

Total Female
Migrant of a Region

 

% Female Population

of Province to region
Population

 

 

 

Projected Urban Female = Y +A + B - U

 

 

 

 

Female Migrants
to be taken out
of Province

 

 

Urban
migrants

 

 

 

 

16 age group (U)

 

 

 

 

 

 

Non-Urban
migrants

 

 

 

16 age groups (N)

 

 

l

 

 

Projected Non-urban Female = Y - A +C - N

 

Figure 18: Diagram Illustrated Calculation of Origin Province Female Migrants

94

 

 

4) Out of Province Migrants

Determination of migrants to be taken out from origin province was based on the
fraction of provincial population to total population in each region. The fraction of
provincial population was assumed to be proportional to total migrants of the region.
These migrants had to be extracted based on urban and non-urban residency and age
group like other types of migrants. The calculations were performed in 4 steps as

explained below.

4.1 Percent Population of Province i in Region r
The first step was to determine percent population of origin province or province I to
each region population. The percents of population in province i to population of the

Central, North, Northeast, South region, and Bangkok were calculated.

% PM Nir / Nr

where:

96 P,,, = 96 population of province i in region r,

M, = Population of province i in region r, and,

N, = Total population in region r when r refers to Bangkok, Central, North,

Northeast, East, and South region)

95

4.2 Migrants To Be Taken Out From Origin Province i

It was assumed that migrants to be taken out of the origin province was equal to the

percent of population of the origin province multiplied by total migrants of the region.

Mou, = M,* % P,.,
Where:
Mo“, = Migrants taken out from province i,
M, = Total migrants from region r to province i, and
96 P,,, = 96 population of province i in region r

4.3 Migrants To Be Taken Out of Origin Province i by Urban And Non-Urban Areas

Migrants taken out of origin province had to be extracted by area of residency. The
calculation was performed based on proportion of urban and non-urban population in the

origin province multiplied by total migrants to be taken out of origin province.

Mout,nu = Moat * Pu
Moutu = Maul * Pnu
where
Moutnu = Migrants to be taken out from non-urban areas of province i,
Moutu = Migrants to be taken out from urban areas of province i,
Pu = Percent urban population in province i, and
Pnu = Percent non-urban population in province i

96

4.4 Migrants To Be Taken Out of province Urban and Non-urban Areas of Origin

Province i by Age Group.
MOUlJlUJ = Mout,nu,i * P}
Mout,u,_i = Moutw' ' Pi
where:
Mama, = Number of non-urban population moved out of province i age group i,
Moul.u,j = Number of urban population moved out of province i age group i, and
P, = Probability migrants in age goup j to mig’ate, j = 1 to 16

The assumption of age distribution of migrants was applied to out of province urban and
non-urban migrants to extract the migrants based on their age group.

The final step was to subtract migrants taken out of province i from the current
population of province i adjusted by within province and from other province migrants.
In the case of the non-urban population, the adjusted current non-urban population for
province i had to be subtracted by calculated migrants to be taken out of non-urban areas

of province i, based on age group. The calculation can be represented in the equation

below:

Projected non-urban population of province i = Adjusted POpnu,. . .- MMWJ
Where:
Adjusted POani '-' Adjusted current non-urban population in province i, age group j, and
MouthuJ = Non-urban population to be taken liom non-urban area of province i,

age goup i

97

In the case of urban areas, the adjusted current urban population for province i had to be
subtracted by calculated migrants to be taken out of urban areas of province i, based on

age group. The calculation can be represented in the equation below:

Projected urban population of province I = Adjusted Pop”, - Mom,
Where:
Adjusted Pop“,- = Adjusted current urban population in province i, age youp j, and
Mout,uJ = Urban population to be taken out from province i age goup i

In summary, Births and Deaths of each province were calculated and adjusted to the
current province population before the calculation of migration was performed at the end
of year five. Using necessary assumptions, all population (births, deaths, migrants) in the
population model were extracted in to units of population by sex, age, and urban and non-
urban residency at the provincial level. The population model in this study had a

hierarchical structure.

The calculations of births, deaths, and migrants were carried out according to the
hierarchical order of each component in the mode and were performed in loops. The
orders from the first to the last were year, region, province, and area respectively (See
Figure 19). The model started the projection for year 1, region 1, province i in region 1,
and area 1. The model calculated births and deaths for non-urban areas in province i,
then calculated births and deaths for urban areas of the same province. Migrants within

and from other provinces were calculated and added to the base population that was

98

adjusted by births and deaths. This completed a loop for the projection of population for
province i. The model performed the same calculations for the rest of the provinces in
region 1 then moved to region 2 and repeated the same calculations until the calculations
for the rest of the regions were completed. At this point the model adjusted current
population with migrants within and from other provinces. Next, the model calculated
migrants to be taken out from each province and subtracted this number from the adjusted
current population. The model completed the circle of a projection of year 1 at this point
and repeated the same calculations for year 2 until year 5. The results of the projections
were the total number of population in 76 provinces of Thailand broken down by sex, age

group, and urban and non-urban area from 1999 to 2019.

99

 

_’FFor yeart (1 10:]

 

 

For regionr

“T, ("1105)

 

 

 

 

.__’ For pravincel

For area u
(u .1. urban
u I 0 I non-urban)

Natl area u

 

 

 

 

 

 

 

 

 

 

 

 

Next provincel

 

 

 

 

 

 

 

I Nextrogionr

 

Next year i

 

 

 

 

 

Figure 19: Flow Chart of The Structure of The Population Model

100

II. PROJECTION RESULTS AND DISCUSSION

This section is divided into three subsections. The first one presents the base year 1994
population and the five-year projection results for the period 1999 to 2019. The projected
results are presented at three levels by area of residence: 1) national, 2) regional, and 3)
provincial. The discussions of projection results are carried out in more detail under the
provincial level subsection because the projections were performed at this level and the
results were also used as input to the residential electricity model. The results at the
national and regional levels were simply the aggregation of the provincial results.
Bangkok was treated as a region and excluded from the central region because of its
uniqueness in social, economic and political aspects. The second subsection compares
and discusses the projection results between this study and the projections of the NESDB.

The last subsection presents and discusses the result of the sensitivity analysis.

3.4 Projection Results

3.4.1 National Level

The projection results at the national level show an increase of the total population from
55.4 to 83.6 million between 1994 to 2019 (See Figure 20). In 2019, Thailand is
projected to have 38.7 million people live in urban areas and 44.9 million in non-urban
areas. The percent of urban and non-urban population to the total population will be 46.3

percent and 53.7 percent respectively in this year. In the same year, the projected female

101

population is 42.2 million, and the projected male population is 41.3 million. The female
population will account for 50.5 percent of the total population. The average annual
population grth rate is expected to decline from 2.09 to 1.42 percent in this period.
The increase of the total population in Thailand will be largely influenced by an
increasing population in urban areas. The share of urban population (urbanization rate) is
expected to increase from 40.6 to 46.3 percent in the same period (see Table 5). The
percent change of urban population in each projection year will always be higher than the
percent increase of non-urban population. However, the change of population in urban
areas is expected to increase at a decreasing rate. The percent increase of urban
population is projected to be 12.8 between 1994 tol999 and decrease to 10.4 percent
between 2014 to 2019. Although the majority of the population of Thailand will still live
in non-urban areas, the share of the non-urban population will decline fiom 59.4 percent
in 1994 to 53.7 percent in 2019. The projection curves of urban and non-urban residents
in 2019, shown in Figure 20, become closer toward the end of the projection year. This
indicates that the difference between population in non-urban and urban areas will

become less in the future.

102

90.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

70.0
37‘
E ‘/
'5' 50.0
3
'2 ”’4
a 40.0
o
a. I//A———’/o/
" 30.0
g //
p.
20.0
10.0
0.0 ,
1994 1999 2004 2009 2014 2019
Yea
+ Urban + Non-urban + Total J
Figure 20: Projected Total Population at National Level, 1999-2019
Table 5: Population Change At National Level By Area, 1999-2019
Year % Total % Urban % Non-Urban % Urban % Non-Urban
Population Population Population Population Population
Change Charge Change
1994-1999 10.4 12.8 8.9 41.5 58.5
1999-2004 9.4 12.0 7.5 42.5 57.5
2004-2009 8.4 1 1.3 6.2 43.6 56.4
2009-2014 7.7 10.8 5.2 44.9 55.1
2014-2019 7.1 10.4 4.4 46.3 53.7

 

 

 

 

 

 

103

 

 

3 .42 Regional Level

Bangkok

At the base (1994) year, regions ranked by total population fi'om highest to lowest will be
the Northeast, North, Central, South, Bangkok, and East region respectively. By the end
of the projection year (2019), the region with the largest population will still be the
Northeast region, following by the Central, North, Bangkok, South, and East region (see
Figure 22). IfBangkok is included with the Central region, this region is ranked the
second largest population in the base year and the largest in 2019. Bangkok’s population
is projected to increase from 5.1 to 10.5 million in 1994 and 2019 respectively. The
percent population increase of Bangkok is the greatest in the Country. The change is
expected to be 108.1 percent between 1994 to 2019 or a grth of 4.3 percent per year.
Bangkok will account for approximately 50% of the region’s (Central) population both at
the beginning and the end of the projection years. In terms of population share, Bangkok
is projected account for 12.6% of the Thai population in 2019. The share of Bangkok’s
population to the total population in this year will be greater than the share from the
South and East region’s population. Bangkok is the only province in the country that has

an urbanization rate of 100% in this period.

Central Region

The Central region is ranked second to Bangkok as far as population change is

concerned. This region is expected to experience an approximately 83 percent increase

104

of population between 1994 to 2019 or a growth of 3.3 percent per year. The population
of the Central region is projected to increase from 9.1 to 16.7 million while the share is
projected to increase from 10.9 to 20 percent in 1994 and 2019 respectively. The
increase of the total population of the Central region will result from an increasing
population in urban areas, which is projected to be the greatest in the country. The
percent increase of urban population will be 111.4 percent in this period. Approximately
70 and 76 percent of the population in this region will live in urban areas in 1994 and
2019 respectively. Percent urbanization of this region is also expected to increase from

5.9 percent in the base year to 12.6 percent in 2019.

Northeast Region

The population in the Northeast region is projected to increase fiom 19.7 million in 1994
to 26.5 million in 2019. Although this region will have the largest population in this
period, the percent population change of this region will be the second. The estimate of
percent change of the population in this period is 34.0 percent or a growth of 1.4 percent
per year. The Northeast region is expected to experience a decline of population share
from 35.6 percent in 1994 to 31.7 percent in 2019. The decline will take place because of
the increased shares of population from Bangkok, the Central and the East region.
Approximately 82.7 percent of population in this region lived in non-urban areas in 1994.
The share of the non-urban population will not change in 2019. When compared to other
regions, this region will have the largest non-urban population in the country. Population

in the non-urban area of this region is projected to increase from 16.3 to 21.8 million in

105

1994 and 2019 respectively. The non-urban population of the region in 2019 will
account for approximately 26.1 percent of Thai population. The Northeast region is
expected to experience an increase of urban population of 35.8 percent between 1994 to
2019, 2.1 percent greater than the percent increase of non-urban population. However,
the Northeast region will experience a decline of urbanization rate from 6.2 to 5.6 percent

in this period.

South Region

The population in the South region is expected to increase from 7.2 million in 1994 to
10.4 million in 2019. The percent population increase in this region will be
approximately 44 percent or a grth of 1.76 percent per year in this period. The share
of population from this region will decrease slightly from 13 percent in the base year to
12.4 percent in 2019. In terms of urban and non-urban population mix, the South region
will not experience a significant change in this period. Similar to the Central region, the
South region is projected to have more than half of the population living in urban areas.
That is 59.5 percent (4.3 million people) and 59.7 percent (6.2 million people) of the
region population will live in urban areas in 1994 and 2019 respectively. The share of
urban population from the South region will be the second largest from the Central
region. However, this region will experience a decline of urbanization rate from 7 .7 to

7.4 percent in this period.

106

cascaded :35 «canton.

 

 

 

 

         

  

 

  

 

     

 

 

  

m. m. m m m m m
”1'

'lal
 _Il._I

II

v w"

\ A‘I
MI

I

It II

o. o. o o. o m. d
w z m w w 5 o

.25....5. 8.8.38 .38

£33m EN
530% _,
fluoctozm 5N

3853248?

xxm
23 .8Eoom «ON

xxm
23 iacoqvmmw

8.9.8898
5.9.88 48*
3888.8

ibcooivmmr

 

 

-mePopaaion

-l\bn-UrbanPopuaion

 

+pereert Urban Papddionwithin Riggion *Peroert Ubon Popddionto Totd

 

Figure 21: Projected Population and Urbanization Rate By Region and Area, 1999-2010

107

North Region

The North region will experience a population increase from 10.9 million in 1994 to 14.2
million in 2019. Compared to other regions, this region is expected to have the lowest
percent population increase (30 percent) in this period or a growth of 1.2 percent per
year. Approximately 67.2 percent of the region’s population (7.3 million) in 1994 and
66.9 percent (9.5 million) of the region’s population in 2019 will live in non-urban areas.
The population share from this region will decrease from 19.7 percent in 1994 to 17
percent in 2019. Similar to the Northeast and the South region, the North region will
experience a decline of percent urbanization from 6.5 to 5.6 percent in 1994 and 2019

accordingly.

East Region

The projected population for the East region will be the smallest of the five regions
(including Bangkok). The estimate of the population is 3.3 million in the base year and
5.3 million in 2019. The population share from this region will increase from 6 to 6.3
percent in this period. The East region is expected to have a 58.9 percent increase in
population between 1994-2019 or a growth of 2.4 percent per year. The population
growth in this region will be the result of a high change of population in urban areas
(78.8%). The non-urban population is projected to increase by 48% while the share of

non-urban population within the region will decrease from 64.6 to 60.2 percent in this

108

period. The East region will face an increase of urbanization from 2.1 to 2.5 percent in

1994 and 2019 accordingly.

Regional Population Summary

The trend of population change in all regions is that urban and non-urban population will
increase with the higher percent change taking place in Bangkok, the East and the Central
region. The remaining regions will experience an increasing urban population with
decreasing shares of regions’ population in urban area from the base year to 2019. The
Northeast region will experience the highest different between urban and non-urban
population. The difference will be 17.2 million people in 2019. The difference will also
be found lowest (1.1 million people) in the East region in the same year. In terms of
urban and non-urban population mix within the region in the period 1994 to 2019, there
will be almost no change (percent change less than 1 percent) in the population mix in the
North, Northeast, and South region. The percent changes are estimated to be 8.5 percent

in the Central region and 4.4 percent in the East in the same period.

3.4.3 Province Level

The projection results at the provincial level are discussed in terms of changing share and

percent increase of total, urban, and non-urban population within each province between

1994 to 2019. Maps are used to illustrate the change in each province. Percent change of

population by area was broken down into five ranges using a Natural Breaks

109

classification technique. Natural Breaks uses a statistical formula to identify breakpoints
between classes by minimizing the sum of the variance within each of the classes.
Natural Breaks finds grouping and patterns inherent in the results (ArcView User

Manual, 1995).

For the purpose of the results discussion percent population change was categorized into
four levels: very high, high, medium, and low. The level of change was determined
based on the mean value and the standard deviation of the percent population change of
provinces excluding provinces with percent change greater than 100%. These provinces
were excluded because this change was considered as an extreme case, which only a few
provinces will experience this change. For the first level, a very high change, the percent
change will be greater than 100 percent. After excluding provinces in this group, the
mean value of percent population change was calculated to be 36.8 percent and the
standard deviation was calculated to be 12.8 percent. The percent change more than one
standard deviation (change > 49.6%) was considered the high change level. The percent
change between i one standard deviation (between 49.6 % and 24.0 %) was considered
the medium change level. The percent change less than less than minus one standard

deviation (< 24%) was considered the low change level.

1) Percent Change and Share of Total Provincial Population

In the period 1994-2019, the 76 provinces in Thailand will experience a wide range of

change in total population, varying from as small as 12.4 percent to as high as 283.8

110

percent. Calculated percent population change of all provinces was distributed relatively
evenly, with some extreme values at the high change side (outliner). The very high
population change will take place in Bangkok, four provinces in the Bangkok
Metropolitan Region (BMR), as well as two provinces in the South region (Figure 22).
The largest change will take place in Nonthaburi, a province adjacent to Bangkok rather
than in Bangkok. However, Bangkok will still have the largest population and holds the
largest share of population in the country. Bangkok, Nonthaburi, and Samut Prakarn
deserve special attention because these provinces will experience very high change and
have large population (> 2 million). These provinces are relatively well established in
terms of economy and housing. The projected very high population change within these
provinces will result from large in—migrant population and births from the greater

population in the base year of these provinces.

The high population change group will consist of 10 provinces, mostly located in the
Extended Bangkok Metropolitan Region (EBMR). These are the areas where
development is growing because of the expansion of infrastructure into these provinces.
Moreover, these areas become more attractive because Bangkok and the BMR are getting
closer to saturation in terms of expansion of housing and industry. For these reasons,
provinces in this group will also experience large in-migrant population, which in turn
will increase total population in the province. In terms of the population share, there will
be only three provinces that will account for more than 1% (approximately one million or
more) of total population in 2019. These provinces are Kanchanaburi, Nakhon Pathom,

and Chonburi.

lll

The majority (approximately 68.4%) of the provinces in Thailand will experience
medium percent population change between 1994 to 2019. These provinces are mostly
located in the North, Northeast, some in the Central and the South region. While these
provinces are expected to experience medium population change, it does not necessary
mean that these provinces will have small populations. In fact, twenty-two out of fifty-
two provinces in this group will have a population of more than one million in 2019. The
projected second largest population in the country in this year (3.2 million), taking place
in Nakhon Ratchasima, is also in this group. While the first two groups (very high and
high change group) are expected to experience an increase or the same percent of
population in the province to the total population, all provinces in the medium group are
expected to experience a small decrease. This decrease results from the increase of the

population share of the provinces in the first two groups.

In the case of the low change level, seven provinces will experience a percent change of
population below 24%. These include provinces that have relatively small population,
small migrant population, and low economic activities (agricultural and industrial
activities). Chiang Rai is the only province in this group that is expected to have
population more than one million in 2019. The remaining provinces are projected to
have population ranging from 223,133 to 524,121. In terms of percent population of a

province to the total population, all of the provinces in this group will experience a

112

 

 

  
    

Peroe nt Total Population Change

|:]10% — 30%

- 140% - 286%

Nonthaburi - 284% ' ' 4 , . _
Pratum Thani - 168%

 

 

Bangkok - 177%
Samut Sakom - 176%

 
   
  

 

 

 

|
Phuket _ 116% Samut Prakam - 189%

 

Figure 22: Percent Change of Total Population by Province, 1994-2019

113

 

decrease of population share. Chiang Rai, in particular, is projected to have a higher

percent change of population share compared to other provinces in the country.

2) Urban and non-urban Population Change

The projection results at the provincial level show increasing urban and non-population in
all provinces. The increase of total population in most provinces will be more influenced
by the greater increase of urban population in the province. This is shown in an increase
of urban population share in most provinces. In 2019, 49 provinces (or 64.5 %) of the
provinces will have more than half of the province population in urban areas. There will
be five provinces that will have urban population greater than 1 million. Ranking for
highest to lowest urban population, these five provinces include Bangkok, Pathum Thani,
Nonthaburi, Songkla, and Nakhon Srithammarat. Approximately 63 provinces (or 83%)
of the provinces will experience an increased urban population share in the province from
1994. However, the extent of this or the change of urban and non-urban population mix
in each province will not be so large except for a few provinces. While most provinces
will experience the change of urban population share in the province approximately
below 2-3 percent, there will be six provinces that will experience relatively high change
ranging from 10.7 to 33.2 percent. These provinces are Phra Nakhon Si Ayudhaya (22.7
%), Chachoengsao (18 %), Saraburi (21.8 %), Chanthaburi (33.2%), Rayong (10.7 %),
and Uttaradit (11.7%). All of these provinces except Chanthaburi and Uttaradit are
located in EBMR. The percent change of non-urban population share is the opposite of

the change experiencing by urban population.

114

Unlike the small change of urban population share in the provinces, the percent change of
urban population in 2019 will be large, ranging fi'om no change to a change of 285
percent. Calculated percent urban and non-urban population change of all provinces was
distributed relatively even, with some extreme values at the high change side. These
extreme values range from 117.2 % to 284.5 %. When excluding the extreme value and
the zero percent change, the percent change of population in urban area of the remaining
provinces will be in the range of 13 to 98.9 percent. Provinces that will have a very high
percent urban population are Chantaburi, Saraburi, and the same five provinces will have
the very high change of total population (see Figure 23). A very high percent non-urban
population change will also take place in these same provinces. Within this group, the
first two provinces deserve special attention. These provinces are expected to experience
more than 100% change of the urban population at the same time relatively large change
of urban population share compared to other provinces. However, the urban population
in these provinces will still be small, less than 350,000 people. The zero percent urban
population change will take place in Amnat Charoen where urban areas do not exist. For
the case of the high, medium, and low change level, the level of percent urban population
change that a province will experience in the period 1994-2019 will follow the pattern of
the case of the total population. This pattern will also apply to the case of percent non-

urban population change (see Figure 24).

The different level of percent urban and non-urban population change as well as the

population share in each province is based on the different levels of births, deaths, and

migrants. The births and deaths were determined by the current population and the birth

115

and death rate. The assumptions formulated for fertility and mortality rate have little to
do with the total population in urban and non-urban area in each province because the
number of births and deaths of urban and non-urban area in each province were
calculated from the national rate. The sensitivity analysis also confirms that the change
of population is not sensitive to fertility and mortality input (see Sensitivity Analysis).
The births and deaths in each province were influenced more from the current population
in the province. The current population also determines migrant populations because
migrant populations in each province were calculated based on the fraction of the current
population in urban and non-urban areas. In the case of the very high and high percent
population change levels, the model assigned the higher probability of settling down in
urban area in the provinces the provinces in BMR and EBMR. As a result these

provinces will experience the very high and high level of change.

At the provincial level, the pattern of the percent change in total, urban, and non-urban
population, and the share of urban and non-urban population to the total population in
each province will not be different. The most distinct changes will take place in some
provinces in BMR and EBMR. These regions are the areas of the country that have high
economic potential. These areas have greater access to ports, markets, government
services and superior infrastructure than the rest of the country. For these reasons, these
provinces seem to attract population from other provinces. Note that the projection
results are in part based on assumptions incorporated into the model. The next subsection
presents and discusses the effects of some of the variables and assumptions used in this

model.

116

 

Percent Non—Urban Population Change
[:| 0% - 20%

[:1 20% - 30%

[:1 30% - 40%

40% - 70%

- 70% - 285%

 

_ . . Prathum Thani-164%"'-,_
_;Panong _ 134% Nonthabun - 284%

Samut Prakam — 130%
Samut Sakom - 173%

 

 

 

Phuket - 116%

 

 

Figure 23: Percent Change of Population in Non-Urban Areas by Province, 1999-2019

117

 

 

Peroort Urban Population Change
- 10% - 30%

 

Northaburi - 285%

     

Ranong-137%

    

Pretum Them-169%"

 
 
    

Bangkok-108%

  

 

F.

derom-18696

1
Sum: Prukam - 173%

 

 

 

Figure 24: Percent Change of Population in Urban Areas by Province, 1999-2019

118

3.5 Sensitivity Analysis

The sensitivity analysis was performed in order to better understand how sensitive the
results of the population model are to input assumptions. While there are many
assumptions formulated for the population model, only major assumptions were
examined and these include assumptions of the age and sex specific birth and death rates.
The sensitivity of the number of total population to migration assumptions was not
examined because migration assumptions will not effect the total population at the

national level.

The range of the values used in the sensitivity analysis was a 30 percent increase and
decrease of the birth and death rates, with the original rate being 100 percent. This range
was chosen in the analysis to match the situation of the birth and death rate change in
Thailand. In Thailand, it is unlikely that birth and death rates will change more than 30

percent within the next 25 years.

The sensitivity test was performed through a utility written in MS Visual Basic. The
utility was written in order to determine total population of Thailand by changing the
birth or death rate while holding other variables constant. For the birth variable, the
effect of birth rate on the total population was determined by holding death and migration
rates constant and applying different percent change of the birth rate to the original rate.
The total population in the year 2019 was computed using the values of birth rate that

increase and decrease by 10, 20 and 30 percent of the original rate within 25 years. The

119

percent changes were applied to both trended and non-trended birth rates. Trended birth
rate referred to birth rate that was assumed to decrease 1 percent per year or 20 percent
during the period 1999 to 2019. The non-trended birth rate assumed that current birth
rate used in the population model would remain unchanged in this period. The sensitivity

of the projected population to age specific birth rate was calculated using the equation as

follows:
3 = [030 - PD/Po] / [(Ro - Ri) / R01
Where:
8 = Sensitivity
P0 = Projected total number of population in year 2019 when there is
no change in the birth (death) rate or when R equals 1 or R0
P, = Total number of population when percent change in rate is i
R0 = Percent birth (death) rate of change equal to 1
R. = Percent change in birth (death) rate i where i = 0.7 to 1.3

The effect of death rates on the total population was determined in a similar manner to
the birth rate. Birth and migration rates were held constant while death rate varied by an
increase and decrease of 30 percent of the current values. The total population was
computed using the percent change applied to both trended and non-trended death rates.
The trended death rate referred to the rate that was assumed to decrease 2.5 percent in the
period 1994 to 2019. The non-trended death rate assumed a constant death rate in this
period. The sensitivity of the population model to the assumption of the death rate was

calculated using the above equation.

120

The results of the sensitivity analysis are reported in Table 6 for the effects of birth rate
and Table 8 for the effects of death rate on the number of total population. The first
column of the sensitivity analysis table shows the percentage change on the birth rate
relative to the original rates. The population fields show the projected population in 2019
for the trended and non-trended rates. The sensitivity fields show the calculated
sensitivity of the total population to the change of both trended and non-trended birth and
death rates. Level of sensitivity or responsiveness of the total population to change of

birth or death rate is greater when the sensitivity value is 1 or greater.

Table 6: Effect of Birth Variable on Total Number of Population

 

 

 

 

 

 

 

 

 

 

 

Age and Sex Specific Birth Rate
Percent Change Trended Non-trended
Population Sensitivity Population Sensitivity
-30% 75,399,592 0.326 78,489,008 0.362
-20% 78,078,159 0.329 81 £82,493 0.362
-10% 80,803,129 0.332 84,940,992 0.354
0% 83,574,623 - 88,054,270 -
10% 86,392,593 0.337 91,440,955 0.385
20% 89,257,462 0.340 94,893,032 0.388
30% 92,168,959 0.343 98,410,278 0.392

 

 

 

 

 

 

The results of the sensitivity analysis on the birth rate indicate that the projected total
population is not sensitive to changes of the birth rate used in this model. In the case of
trended birth rate, the sensitivity values change very little from 0.33 for the 30 percent
decrease and to 0.34 for 30 percent increase of the current birth rate. This means that if
30 percent less than the current age specific birth rate was used in the model, every 1

percent change in this rate will increase the projected population by 0.33 percent. For the

121

case that the rate was 30 percent greater than the current rate, the sensitivity value
indicates that for every 1 percent change in this rate, there will be approximately 0.34
percent increase in projected population. Similar results were found when sensitivity test
was performed for the non-trended birth rate. Population will increase by 0.36 and 0.39
percent for every 1 percent change of the used of the 30 percent decrease and 30 percent
increase of the non-trended birth rate. The different between the sensitivity value are

very small for both tests.

Table 7: Effect of Death Variable on Total Number of Population

 

 

 

 

 

 

 

 

 

 

 

Age and Sex Specific Death Rate
F8010? Of Rate Trended Non-trended
Change Population Sensitivity Population Sensitivity

0.7 87,412,449 -0.153 87,244,587 -0.155
0.8 86,066,843 -0.149 85,884,378 -0.151
0.9 84,789,085 -0.145 84,593,509 -0.147
1.0 83,574,623 83,367,473

1.1 82,418,485 -0.138 82,201,183 -0.140
1.2 81,316,862 -0.135 81,090,340 -O.137
1.3 80,265,210 -0.132 80,030,712 -O.133

 

 

 

 

 

 

In the case of death rate, the sensitivity test result indicates that the projected total
population is sensitive to neither trended nor non-trended death rates in this model. For
the 30 percent decrease trended death rate case, the sensitivity test shows that 1 percent
change in death rate will change (decrease) total population by 0.15 percent. The
sensitivity value of —0. 13 for the 30 percent increase case indicates that for every 1
percent change in the death rate, there will be approximately a 0.13 percent decrease in
projected population. The result for non-trended cases is very similar to the trended case

(see table 7).

122

For this model, the output, total population, is more responsive to the input birth rate
variable than the death rate variable. Take the case of 30 percent decrease as an example:
a 1 percent change in birth rate will cause total population to change 0.33 percent while
the 1 percent change in death rate will cause a total population change of about 0.15

percent.

3.6 Discussion and Comparison of Projection Results

This subsection compares the population projection results at the national level between
the model in this study and the model of the NESDB. The discussions of projected
population and total fertility rates are based on methods used to project birth and death

rate used by each model.

The results of projected population from both studies are presented in Figure 25 and
Table 8. Notice that the base year population used by both models was different. This is
because both models used data from a different source and different year. The NESDB
model used the data obtained from the Population and Housing Census, (1990) while this
study obtained the data from the Registered Population and Housing, 1994. In 1994,
projected population from the NESDB model was 58.7 million, approximately 19.1
percent (13.8 million) larger than the base line population used in this study. By 2019 the
projected population from this model will be 83.6 million, larger than the NESDB model

that has the estimate of 70.2 million. The curve in Figure 7 shows a steeper slope of the

123

projected population from the model in this study. In addition to the difference in the
base year population, the combination of the changes in births, deaths, and migration may
also contribute to the different projections. Following is an explanation of each

parameter that effects the difference in population.

 

(D
O

 

00
U!

 

\

 

 

 

 

 

 

 

 

 

E
3 75
E /
5 70 // ”WW” .4
5 ”WWW“
3. 65 FM
:2 /

60 ,2,"

I/
55
50 , . , , T
1994 1999 2004 2009 2014 2019
Year
L + Model 4» NESDB I

 

Figure 25: Population Projection Results From Different Studies

124

Table 8: Projected Total Fertility, Crude Birth and Crude Death Rates, 1999-2019

 

 

 

 

 

 

 

 

Year Total Population Total Fertility Rate Crude Birth Rate Crude Death Rate
(million) (per woman) (per 1,000 women) (per 1,000 population)
Study NESDB Study NESDB Study NESDB Study NESDB
1999 61.2 61.8 1.72 1.96 15.3 17.0 6.2 7.9
2004 66.9 64.5 1.63 1.87 13.7 15.7 6.6 8.2
2009 72.5 66.7 1.54 1.81 12.1 14.6 7.0 8.5
2014 78.0 69.0 1.45 1.78 10.9 13.7 7.5 8.8
2019 83.6 70.2 1.36 1.76 10.0 12.9 8.1 9.3

 

 

 

 

 

 

 

 

 

 

A large population in a certain area generally results from a high birth rate and low death
rate. However, low birth rate and very low death rate could also cause the population to
increase. The former case better fits the situation in Thailand as observed from historical
statistics. The basic measures of birth and death rates, TFR and CDR, are used here to
compare the variables used in both models. Projections on TFR and CDR from both
models are different. For the birth component, this study used a linear relationship to
trend the age and sex specific birth rate. It is expected that TFR will decrease from 1.72
to 1.36 births per woman in 1999 and 2019 respectively, a decrease of 23.3 percent. The
NESBD model used a nonlinear relationship (second-degree exponential function, Figure
26) to perform the projection of age and sex specific birth rates. The calculated TFR is
expected to decrease from 1.96 to 1.76 births per woman or 10.2 percent in the same
period. It is likely that the lower estimate from the NESDB study is the result of the birth

rate that was trended to decrease at a faster rate in the NESDB model than in this study.

125

Total Fertility Rate (births per woman)
2 z; a r. a; z;

.a
(a)

 

 

 

 

 

 

 

 

 

 

 

1999 2004 2009 2014 2019
Year
i +TFR-Model +TFR-NESDB J

 

Figure 26: Projected Total Fertility Rate, 1991-2019

In terms of death rate, calculated CDR in this study was derived by dividing the

summation of number of deaths in each age group and sex by total population in each

year. CDR is expected to increase from 6.2 deaths per 1,000 population in 1999 to 8.1

deaths per 1,000 population in 2019. The explanation for the increase of the CDR is

based on assumptions formulated for fertility and mortality. In this study, birth rate was

trended to decrease much faster than death rate, a 1% and 0.1 % decrease per year for

birth and death rate respectively. When this is the case, the number of deaths will

increase faster compared to number of births, and population will increase at a decreasing

rate. As a result, the CDR will increase.

126

The life expectancy at birth for males and females were not calculated because the life
table was not available for this study. The NESDB model calculated CDR from the age
and sex specific survival rate (survival rate = 1 - death rate). The survival rate was
derived from the life table that was used to calculate the expectation of life at birth. The
assumption to calculated age and sex specific survival rate was that the life expectancy at
birth would decrease at a falling rate. Based on this assumption, CDR is expected to
increase fiom 7.9 to 9.3 deaths per 1,000 population or 17.7 percent increase between
1999 to 2019. The projected CDRs by the NESDB model were higher than the projected
CRD in this study in both years. It is not surprising that the NESDB population estimate
is lower than the estimate from this study because the NESDB trended TFR to decrease at

a faster rate than this study, at the same time used higher CDR

Migration is the other variable that causes population change and deserves special
attention when population change in a small area is of major concern. In this case,
migration is not a major concern because the comparison is at the national level.
Migration within and between provinces will not effect total population. Although
migration in and out of the country will effect total population, the impact from these
types of migration on total population of Thailand is small because Thailand has a very

small rate of immigration and emigration.

In summary, the national projections of population from 1999 to 2019 revealed that the

population of Thailand will increase fi'om 61.2 to 83.6 million people and the TFR will

decrease from 1.72 to 1.36 births per woman in this period. The change in total

127

population will increase at a decreasing rate and the rate of urban population change
(increase) will be higher than the rate of rural population change. By 2019 Thailand is
expected to have 46.3 percent of the total population living in urban areas and 53.7
percent of the total population living in rural areas. Regionally, the greatest urban
population will be concentrated in the central region while the largest percent grth in
urban population would take place in the BMR and the EBMR. At the provincial level,
approximately 79 percent of provinces in Thailand will experience an increase in the
proportion of urban population while the rest of the province will experience less than 2
percent increase in non-urban population. Major interprovincial migration streams in the
period 1994-2019 will be concentrated in the BMR and the EBMR. These are the areas

where dramatic urban and non-urban population change would take place.

128

CHAPTER IV

ELECTRICITY MODEL: GENERAL DISCUSSION
AND LITERATURE REVIEW

Chapter 4 is organized into 2 sections. The first section discusses residential electricity
consumption in Thailand, in terms of its relation to three issues: population growth and
economic performance, electricity end use efficiency, and electricity supply. The second
section reviews literature related to residential electricity modeling. The review focuses
on three topics. These include previous studies on residential electricity demand
modeling, factors effecting the demand, and projection of electricity demand in the

residential sector in Thailand.

1. RESIDENTIAL ELECTRICITY DEMAND IN THAILAND

4.1 Population Growth and Economic Performance Outlook

One way to obtain a general idea of residential electricity consumption in Thailand is to

observe the previous trend of per capita electricity consumption in the residential sector

and to look at the residential electricity consumption per gross domestic product (GDP).

129

The International Energy Agency (IEA, 1995) reported that per capita residential
electricity usage in Thailand increased during the period 1974-1993 with a slight drop in
1985 (Figure 27). Residential electricity consumption per GDP also increased in this
period except for 1987 and 1988. These measures indicate the increasing trend of

residential electricity consumption as the economy of the nation is expected to continue

 

 

 

 

 

 

 

 

 

 

to grow in the fiiture.
1200 0.7
3 1000 +06%
5 E
é’ 800 #052
e, 8
c: O
£0 —0.42
15.2 .9
511 600 ‘5.
2% f / ~03;
O D
0‘- g
5; 400 - _ o
.2 02:
E 8
ILI
200 r _0_1§
o I I T I I T I I j I I I O
N N N D Q Q 0 Q Q Q 00 0 O) O) O) O)
22292222292222222’222
Year

 

L+ Per Capita Electricity Consumption + Electricity Consumption per GDP I

Figure 27: Thailand Residential Electricity Consumption, 1974-1993
(Source: International Energy Agency, 1995)

130

Another way to examine the relationship between residential electricity consumption,
economics, and population is to examine the trends of grth in electricity sales per
capita and the growth in electricity consumption per residential customer. Figure 29
shows the relationship between the electricity sales and per capita income of countries in
the East and South East Asia. At a given income, there is a considerable spread of the
sales per capita over time among countries, with the Philippines and Korea at the
extremes and Thailand as well as the other countries at an intermediate level. On the
contrary, the residential electricity sale per customer or household at a given income of

these countries except Korea lies close to each other, with little change over time (Figure

 

 

 

 

 

 

28).
10% P
E o SeuitKeue
: O m
g 1W '- 0 Xx X 'W
"5 oat? a maps...
:3 09
l m
10200 1“!) 10mm
RedPerCmituGDPWBOPPPUSDellen

Figure 28: Per Customer Residential Electricity Use
(Source: International Energy Agency, 1997, p.126)

131

10000 F

1000r'

Soles (kwh/Capita)
§

10" J‘

 

I 4_l

100 1000 10000 100000
RaolPerCepitoGDPll980PPPUSSi

 

a 5 Korea 1:: Thailand I Philippines
- lndonesio x Japan

 

 

 

Figure 29: Per Capita Residential Electricity Use
(Source: International Energy Agency, 1997, p.126)

The explanation for these contradictory phenomena is that the growth in sales per capita
is caused almost entirely by the electrification of new households. The grth in total
sales per capita reflects the increasing electrification of new households in these
countries. Korea is the exception because the process of electrification in this country
was almost completed by the early 1970s. The electricity sale per customer reflects the

ownership of electric appliances and usage. Slow grth in consumption per customer

132

takes place after a household obtains electricity. The rising number of new connected
households are usually the low-income users (due to rural electrification programs) with
low usage. Consequently, the average consumption per household is pulled down by the
low usage of this group. When the electrification of households nears the saturation point
and incomes continue to grow, the total electricity use will change based on the number

of appliances per household.

The level of urbanization and electrification is related to per capita electricity use at given
a GDP. Philippines has larger per capita electric consumption than Thailand as shown in
Figure 30 because Philippine households are in urban areas, even though more

households in Thailand are electrified.

The historical and current electricity consumption in Thailand can be summarized as
follows. The slow growth of electricity consumption per capita during 1974 — 1988
resulted from an increased number of new electrified households, mostly low-income
households in non-urban areas (low usage). In 1992, 98.2% of households were
electrified. The growth the consumption has increased more rapidly in recent years,
driven by the growing numbers of appliances per household accompanied by the growing
GDP. However, appliance ownership in Thailand is still far from approaching the
saturation point as seen in most western countries such as the US, Sweden, and
Denmark. The ownership of selected appliances in electrified urban and non-urban
households in Thailand is show in Table 9. Appliance ownership is different between

urban and non-urban Thai households. Households in urban area generally have an

133

average household income higher than the household income in non-urban areas.

Appliance ownership in urban areas is also higher than ownership in non-urban areas for

most appliance types.

Table 9: Thailand Selected Appliances Owned by Area (Percent of Households), 1991

 

 

 

 

 

 

 

 

 

 

 

 

Selected housing characteristics and Total Urban Area Non-Urban

appliance Area
Electric lighting 89.7 99.0 87.5
Radio 81.3 87.4 79.9
Color television 46.2 77.9 38.6
Electric rice cooker 61.2 90.0 54.3
Electric fan 72.1 93.7 67.0
RefriLerator 36.0 70.8 27.7
lron 54.5 87.3 46.7

 

(Source: Housing and Population Census, National Statistical Office, 1990)

4.2 Electricity End Use Efficiency

The residential sector comprised 20.6 percent of total electricity consumption in Thailand

in 1994. The distribution of electricity end uses in this sector in 1987 is shown in Figure

30. The highest percentage of electricity uses in that year was for heating and cooking.

Lighting and refrigerators shared the same percentage for contribution to total electricity

consumption. Share of a use of electricity for space cooling was lowest among other

uses.

134

 

  
  
    

  

8 Lighting
I Air Conditioning
1:! RW‘

5 MM
l Others

Air Conditioning
0%

 

Refrigerant:
21 96

 

Figure 30: Distribution of Residential Electricity by End Uses (1987)

(Source: Thailand Promotion of Electricity Energy Efficiency Project, World bank and

GEF Project Document, April 1993 in International Energy Agency, 1997)

Table 10:Potential for Efficiency Increases in Major Electrical Appliances

 

(Source: International Institute for Energy Conservation, Bangkok, 1995
in International Energy Agency, 1997)

135

Electric appliances in most households in Thailand tend to be old and those purchased
before 1990 are not energy efficient. The result of a 1995 study on the investment
opportunities in energy efficiency industry in Thailand conducted by the International
Institute for Energy Conservation (IIEC) indicates significant potential for improving
efficiency in various electrical appliances and is summarized in Table 11. The study
reveals high potential to increase energy efficiency for home appliances, especially
refrigerators, fans and air-conditioners, as well as high efficiency fluorescent tubes and
ballast. Electric appliances in most households in Thailand tend to be old and those
purchased before 1990 are not energy efficient. It is expected that in the fixture more
efficient appliances will be available affordable prices. As more efficient appliances

replace less efficient ones, electricity consumption should fall.

A good understanding of the relationship between appliance efficiency and electricity
demand is essential for planning. Determination of the consumption and efficiency of
some appliances is not an easy task. Utilization of electric appliances is highly variable
among households. For example, the consumption of lighting depends on house size and
number of occupants, as well as habits of turning off unused light. Electric irons are
important in Thailand because the kinds of fabric Thai people prefer require ironing.
Consumption of electricity due to television is driven more by hours of operation and the

size of televisions rather than the efficiency which has been remarkably improved.

Residential customers should benefit from greater energy efficient appliances because

electricity bills will be lower. People’s comfort levels will increase because of less heat

136

generated from electricity used. Although it looks like the higher income group will gain
more benefit from the improved efficiency of appliances that are not common to low
income groups (such as air-conditioners, refrigerators, and other large appliances), it is
expected that these appliances will become more affordable to the lower income group in

the near future.

According to the residential energy savings plan in a 1991 Demand Side Management
(DSM) Master Plan for Thailand Electric Power System prepared by IIEC, high
electricity savings is expected from improving efficiency of home appliances, especially
refrigerators, air-conditioners, ballasts, and fluorescent tubes. This plan was reviewed by
EGAT, PEA, and MBA, and also was approved by the National Energy Policy
Committee. The DSM plan also expects potential long-term total electricity savings to
increase from 1,189 gwh in the period 1993-1996 to 22,177 gwh in the period 2007-2011.
The DSM plan will attempt to curtail total electricity demand in evening peak hours (6:30
pm. to 9:30 pm.). The program also encourages households to reduce home electricity
use by promoting thin-tube fluorescentss, energy efficient refiigerators, and air
conditioners. In 1996, EGAT spent $US 12 million on advertising to stimulate public
awareness of energy conservation with a focus on these home electric appliances. The
program also works towards creating energy-saving and energy efficiency concepts to be

a matter of course and culture.

 

8 A thin tube fluorescent refers to newly developed fluorescent tubes that provide an equivalent amount of
light and use 10 percent less energy.

137

4.3 Electricity Supply Outlook

Generation of electricity in Thailand depends on four sources of fuel input: coal or
lignite, fuel oil/diesel, natural gas, and hydro-power and other). According to the Energy
and Industry Department, Asian Development Bank, (1993 ), total electricity generation
in Thailand increased from 14,001 gwh in 1980 to 43,189 gwh in 1990. Corresponding
to the total generation, changes in generation mix in this period, by fuel type, were

illustrated in Figure 31.

Currently, EGAT (1996) reported that gross electric generation rose 9 percent from the
previous year. The country’s electricity supply was met by the generation by fire] type
consisting of 28 percent from natural gas, 26 percent from bunker oil, 20 percent from
lignite, 4 percent from diesel oil, 9 percent from hydro power and 13 percent purchased
from private utilities. These electricity shares from all energy sources except natural gas
were on the increase. The increases were 6, 19, 50 and 8 percent for bunker oil, lignite,
diesel oil and hydro power respectively. The plan for future fuel sources has to take into
consideration that fiiel choices must satisfy the demand and be environmentally,
economically, and socially acceptable. It seems that the firture power supply of Thailand
will rely on lignite and coal for several reasons. New hydropower projects will be
extremely difficult to build because of the opposition to hydro-power schemes. Any new

development will be limited to a few economically efficient, small-scale and

 

9 Other fuel input includes solar, biomass, geothermal, and wind energy. These inputs are used in small
scale to produce electricity at experimental stations.

138

environmentally benign projects. Although natural gas is a preferred the] fi'om an
environmental perspective, the availability of domestic natural gas seems to be limited.
The Petroleum Authority of Thailand (PTT) forecasted that gas production would reach a
peak in year 2000 and steadily decline. To diversify away from oil has always been the
goal of the country because Thailand has very limited oil reserves. However, depending
on the oil price, oil is considered an option for power generation due to higher efficiency.
From a supply standpoint, coal/lignite is the most favorable fiiel because of a large
domestic reserve. The availability of coal in the international market and the stable price
make it a competitive fuel. Since power plants produce sulferdioxide, installation of flue
gas desulfirrisaton (F GD) technology is now required by the Thai government for all new
coal fired projects. EGAT also considered nuclear power as an option. The World Bank
conducted a study on costs of nuclear power and concluded that it was not a competitive
option in Thailand (IEA, 1997). Nuclear power also received strong public opposition as
two proposed nuclear power plants by EGAT that were approved by the Cabinet had to
be postponed. Eventually, the plants were not included in the current Power
Development Plan (PDP) 1995-2001 which was prepared for the NESD plan in the

period 1996-2011.

139

PM. martin
0%

Hymn-Poul
12%

Cool
10%

Nerural gas
0* I Hydro-PM

I Coll

D Natural on

D OWDI“

I Privfle gen-fitter!

 

 

 

1980 Fuel Source

 

 

 

 

 

1990 Fuel Source

Figure 31: Electric Generation Mix by Types of Fuel Source, 1980 and 1990
(Source: Asian Development Bank, 1993)

140

The last alternative that can be viewed as a “new” fuel source is conservation (or energy
savings or rational energy use)”. Kleinpeter (1995) referred to the World Energy
Council that classified energy savings into four categories. The first category deals with
energy savings resulting from changes in user behavior without intervention on a
technical level. The second involves energy savings from improving control systems.
The third relates to energy savings from system improvement without fundamental
changes in the system or energy substitution. The last one focuses on energy savings
through fundamental changes in the system or the use of another form of energy or
technology. In the case of the residential sector, electricity conservation falls into the
first category. Energy policy-makers could explore strategies and incentives that will
make users react more cautiously regarding electricity use. An alternative such as
improving information or communication on energy savings can be utilized to promote
conservation behavior in the residential sector. EGAT pursued this option as it launched

the DMS program in 1992 (see 4.2. under Electricity End Use Efficiency).

 

'0 World Energy Council Energy Dictionary provides following definitions oonceming energy as below:
Rational energy use: utilization of energy by consumers in a manner best suited to the realization of
economic objectives, taking in to account social, political, environmental and financial constraints.
Energy conservation: term that defines a policy embodying the actions to be taken to ensure the most
eflicient use of finite energy resources.
Energy saving: measures or the results of measures taken by suppliers and users of energy to limit or
avoid wastage.

141

II. RESIDENTIAL ELECTRICITY DEMAND MODELING

4.4 Studies on Residential Electricity Demand Modeling

Studies conducted in the area of electricity demand in the residential sector, especially in
developing countries, are limited. Most studies conducted in the past utilized
econometric tools. Electric consumption is considered like other economic goods. That
is, income, production, substitution, and prices play a major role in electricity demand
related relations and equations. Nelson (1965) stated the possibility of deriving the
residential electricity demand from the demand for electricity consuming goods and also
found a strong relationship between electricity price and the quantity of electricity
consumed in the residential sector. His conclusion that income was not a significant
factor that determined electric consumption in households, but that consumption was
related to time, was contrary to other studies”. Anderson (1973) and Lyman (1994)
reported other factors contributing to the demand for electricity in the residential sector to
include the price of electricity and its substitution as well as climate. Climate was a
factor in utilizing electric heaters and air-conditioners and contributed to a large portion
of total electricity in the Northern countries. Another contributing factor was

urbanization, as stated in the studies conducted by Anderson (1973), and Westley (1989).

 

” Finding that was opposed to studies conducted Nelson includes studies conducted by Wilson (1971),
Anserson (1973), Action, Mitchell, and Mowill, 1976, Westley (1989), Bemdt and Samaniego (1984),
Economic and Social Commission for Asia and the Pacific (1991), Irnaran, and Barnes (1991), and Lyman,
1994.

142

Bemdt and Samaniego (1984) emphasized the effect of accessibility to electricity to the

demand.

Recently studies have examined that looked at electricity demand from end uses. The
idea was that household electricity is always consumed through electric appliances. In
this method, the residential demand is projected fiom the actual electric appliances
owned and used in households. This method is known as the end use method, also called
the techno-economic method because it considers both economic and technological
parameters. This method, in some respects, addresses the limitations of econometric
techniques one encounters when performing electricity projection (Chateau and
Lapillonne, 1991). The requirement of reliable historical data for this approach usually
constrains the level of disaggregation and the choice of explanatory variables.
Parameters estimated using econometric techniques (particularly elasticities) are not
stable in time because of the rapid change most countries have to confront. Estimated
parameters usually depend upon a regression period. Thus, long term projection is not
easy to perform using this technique. Finally, technology change, and electricity savings,

in particular, cannot be well reflected in econometric models.

The end use methods also have limitations. They require fairly detailed information that
goes far beyond what usually is necessary for econometric modeling. These methods are
best used when household appliance ownership and usage information is available.
Information from a residential electricity consumption survey and household appliance

manufacturer are good information sources. Utilization of end use methods, for

143

electricity projection purposes implies quite a number of assumptions on exogenous
variables that have to be reasonably consistent among themselves. The advantage of
these methods is that they are straightforward and the results are easy to understand. The
calculation simply involves the number of operation hours multiplied by the required watt
input to operate the appliance. The end use approach has been successfully applied in
many countries for energy demand projections (Division of Energy Information
Administration, US. Department of Energy (DOE), 1993), for the assessment of energy
conservation policies (Division of Energy Information Administration, US. Department
of Energy (DOE, 1993), as well as for historical energy demand analysis (Department of

Housing and Urban Development (HUD), 1972).

In the case of Thailand, there are very few studies that deal directly or attempt to project
residential electricity demand. Most studies were conducted by international
organizations and performed analysis of residential electricity from historical trend data.
The Asian Development Bank (ADB, 1993) conducted a study on the relationship
between electricity consumption, and real GDP. They found that electricity elasticity was
1.43 for Thailand compared to 1.15 for Newly Industrialized Economies (NIEs) in the
period 1980 - 1990. This means that if the real GDP increases by 1 unit, the residential
electricity consumption will increase by 1.43 unit. Other studies conducted on residential
electricity end use were performed by EGAT (1997) and The International Energy
Agency (IEA, 1997). Both studies concluded that the main grth in residential
electricity demand is due to the increasing consumer demand for electric household

appliances such as air conditioners and refiigerators. These and many other appliances,

144

once considered luxuries are now standard in many households because the appliances

are available at much lower prices.

4.5 Factors Effecting Residential Electricity Demand

Contributing factors effecting the residential demand for electricity in Thailand seem to
be similar to the findings in previous studies. Explanatory variables that are likely to
influence the demand for electricity in the residential sector of Thailand include income
level per household, climate, the share of households located in urban areas, electricity
price, the price of electricity substitution, accessibility to electricity, ownership of

appliances per household, and technology change.

The ideal for this study is to incorporate as many as possible variables into the model.
However, this study focuses more on household and housing variables or characteristics
because of inadequate data on other variables. The following section provides an
overview of the Thailand household and housing characteristics that effect home

electricity usage.

4.5.1 Household Characteristics

Household characteristics that are related to household electricity usage or demand are
the number of households in the study area and the average number of people in each

household. The number of existing and new households determines the number of

145

electricity users, which in turn determines the household electricity demand. Therefore,
housing stock is important information for residential electricity demand projections. In
Thailand, HOMES (Household Model for Economic and Social Studies; Mason and
Campbell, 1993) is used to project the housing demand and required residential
construction at the national level. The model takes into account population grth and
changes in composition as well as tenure composition. The projections show an increase
of Thai households from 10.2 million in 1985 to 21.9 million in 2015 with an average

annual grth rate of 2.5 percent and a population grth rate of 1.3 percent.

Following housing stock, locations of households or the percent of households in urban
and non-urban areas also effect the demand for electricity more at the regional and
provincial level. So far, there has not been any attempt to project the number of
households in urban and non-urban area at these levels. The share of urban and non-
urban households, however, could be derived from the population in each area divided by

the average of the household size.

Because electricity is a shared household resource, electricity consumption is more
closely related to the number of households than to total population. The average
household size is used to determine the number and types of households. According to
Mason and Campbell (1993 ), the average household size for Thailand was 4.6 members
in 1990 and was estimated to reduce to 3.5 members per household by 2010. The
difference of household size in urban and non-urban areas was not large. The Population

and Housing Census, 1990, reported the average number of members in households in

146

urban areas to be 3.5 — 4.6 persons per household and 3.7 - 5.1 persons per household in

non-urban-areas.

4.5.2 Housing Characteristics

As far as electricity demand is concerned, there are three housing characteristics that
effect the demand for residential electricity consumption: 1) housing type, 2) size of
housing, and 3) accessibility to electricity and ownership of appliances. First, the
housing types in Thailand are categorized into detached house, row house, town house,
and apartment or condominiums. The number of each type of housing is important for
determining housing stock. In Thailand, 96 percent of households are family households.
These households are detached, row houses and town houses (1990, Population and

Housing Census).

The second housing characteristic is the size of housing which refers to floor space of
houses. The floor space of a house determines the amount of energy used for cooling,
heating and lighting a house. The other aspect of a house that largely effects the cooling
and heating system of a house is the thermal shell of the house in which it is installed. In
Thailand an efficient thermal shell is not commonly installed in typical households. If
the thermal shell is installed, the purpose of it is for shielding the house born the outside
heat. The more important aspects of housing that relates to the comfort level of a
household in Thailand are building materials and roofing. However, no information

regarding building material has been collected. For roofing information, the National

147

Committee Rural Development’s survey (NCRD, 1994) reveals that corrugated tin
roofing is a relatively common roofing material used by households in non-municipal
areas of the nation. The tin roofing attracts heat better than other roofing material and

thus makes a house uncomfortable due to heat buildup.

The third important housing characteristic relates to electricity services or the
accessibility to electricity. Approximately 87 .5 percent of private households in non-
urban areas and 99 percent of households in urban areas in Thailand have had access to
electricity since 1992. These numbers indicate that Thailand almost approaches the
saturation rate in terms of electrification. Therefore, the determining factors for
residential demand for electricity depends on the penetration rates of electric household
appliances rather than the electrification rate. Currently, the ownership of most electric
appliances per household in urban areas is higher than non-urban areas (see Table 10).

However, appliance ownership in urban Thailand is still far from saturated levels.

4.6 Residential Electricity Demand Projection in Thailand

Thailand estimates the electricity demand in the residential sector at the national level by
extrapolation from historical trend. However, it has been proposed to use the end-use
approach in order to perform projections of household electricity demand (EGAT, 1997).
The model attempts to estimate future demand at the national level using the end-use

approach to represent household electricity consumption behavior based on electric

148

appliance ownership and the use of the appliances. The model can be described

mathematically as:

£50“: 51th . ESRj/‘jq * U1,”

where:
E = Electricity consumption
S = Saturation rate / share of appliances own per household
ESR = Number of watts required for each appliance (watts)
U = Utilization of each appliance (hours)
i = Type of appliance
j = Income
k = Types of households

t = Projection year

Data preparation for the model can be described as follows.

1. Project of number of dwelling

Estimated number of dwelling will be performed based on 111me population and number
of household. Projected population will be obtained from NESDB (EGAT, 1997).
Number of population per household was obtained by extrapolation from previous time

trend data.

Number of Household = Number of Population

 

Number of Population per Household

149

2. Classify and project dwelling by income

Classification of dwelling by income will be based on a survey conducted by EGAT and

the NSC report on social and economic of households in Thailand. The dwelling will be

classified based on household income per month into 6 groups which can be listed as:.

1)
2)
3)
4)
5)
6)

Income less than 5,000 bahtulmonth

Income is between 5,000 — 7,000 baht/month
Income is between 7,000 — 15,000 baht/month
Income is between 15,000 — 25,000 baht/month
Income is between 25,000 — 50,000 baht/month
Income more than 50,000 bath/month

Projection of dwelling by income will be made based on the assumption that the number

of low income dwellings will decrease while the number of medium and high income

groups will increase on the condition that income from each group multiplied by average

income of the nation equals to Gross National Product.

3. Classify and project types of dwelling based on income

Types of dwelling based on income were classified into 5 groups which is listed below:

1) Detach

2) Rowhouse

3) Townhouse

4) Apartment and Condominium
5) Other

 

'2 SUS l is approximately 36 baht as of November 1998.

150

Projection of dwelling by income will be carried out based on the housing construction
permit data and the assumption that the proportion of number of dwelling in each type
will be not changed from the period of conducting survey (short term) and the projection

period (long term)

4. Obtain data on electric appliance ownership, input watts required for each appliance

and utilization of the appliances.

Only electric appliances that had a high rate of ownership per household and utilization
rate will be used in the projection of household electricity demand. Appliances which
will be included in residential electricity model for MBA are air conditioners,
refrigerators, fans, light bulbs, electric rice cookers, washing machines, color televisions,
and irons. Washing machines and irons will be taken out of the list and replaced by black

and white televisions and freezers for the case of PEA.

5. Efficiency Factors

ll
A
I

Efficiency Factors

where:
C = substitution + new appliance
N = number of appliances

E = Rate of efficiency of electric appliances

151

Projections of the residential demand for electricity are currently interpolated from
historical trends. The results of residential electricity demand projections, using the

interpolation method, by MBA and PEA are presented in Figure 32 and 33 accordingly.

 

 

 

 

 

 

 

 

 

 

 

 

3000000 18000
-» 16000
2500000 -A-
A /
ii ’ *- 1m
3 ’4
/ a.
a 2000000 ,1 12000 ‘
a E
e
E .. 10000 .3
o
g 1500000 . ,I m
8 ’ ~- 8000 E
-’ o:
.6 "I 8
'- 1000000 « 4‘ 6000 .2
3 ‘T m
S m,
z ..
500000
W “ 2000
x—H—H—u—ae—Ie—x—at—ah—H-H—i‘
0 . r . Y r , . , . Y . . . . . 0
1996199719981999200020012002200320042005200620072008200920102011
Year
+Total no. Customer -I— No. of Customer <150 kWh/month
+ N0. of Customer >150 kWh/month +Total Energy Sale
+Energy Sales (GWh) <150 kWh/month + Energy Sales (GWh) >150 kWh/month

 

 

 

Figure 32: Projections of the Residential Electricity Demand for Bangkok Metropolitan
Region by MEA (Very Low Case): FY 1997-2001
(Source: Metropolitan Electric Authority, 1996)

152

The projections of MBA indicate increases in both the total number of customers and
sales in the period 1997-2011. However, if these two measures are examined by quantity
consumption, the projections reveal different results for the case of smaller electricity
users (<150 kWh per month users). The number of small users decreases gradually over
this period while the sale for this group increases very little. Both the number of
customers and sale for large quantity users (>150 kWh per month users) increases at a
higher rate than the rates of total customers and sales. The projections of MBA suggest
that the large quantity users in the Bangkok Metropolitan Region (BMR) play an
important role, in terms of electricity usage, because this group are the major consumers

of electricity in this country and will continue to grow; into the next decade.

 

 

 

 

 

 

 

 

 

18000000 40000

16000000 __ 35000
O
2 14000000 __ 30000 E
g 12000000 5
a -» 25000 ..
0 0

10000000 t -
8 X «- 20000 3
3 8000000 :3
S .- 15000 "'
r: 6000000 §
5 __ n
,2 4000000 -— 10000

2000000 ~~ 5000

0 0

 

 

 

'\ N “to '5 b '\ <6 0 N
eeeeseseseeeeses
Year
L +Total no. Customer +Energy Sales(GWh) j

 

Figure 33: Projections of the Residential Electricity Demand for 73 Provinces
by PEA (Very Low Case): FY 1997-2011
(Source: Provincial Electric Authority, 1996)

153

The projections from PEA do not show the breakdown of customers in terms of the
quantity of electricity consumed per household (Figure 33). Over all, both the number of
customers in 73 provinces, excluding BMR and electricity sales, increase over time. The

growth of energy sales increases at a higher rate than the increase of total customers.

Both MBA and PEA projections indicate the near saturation of electrified households in
the country because the sales increase at a higher rate than the number of customers. In
other words, the increasing sales are driving the increasing number of appliances per

households, rather than the increasing number of new customers.

In summary the demand for residential electricity is growing rapidly in Thailand,
primarily because of the increasing number of customers and the increasing number of
customers who acquire more appliances over time. Information concerning household
characteristics and the behaviors of electric appliance users is crucial for electricity
projections. Improved technologies, in the future, would make appliances available at
lower costs and, at the same time, make appliances that are more energy efficient. The
ability to assess the impact of changes in electricity consumer behavior and technology is
an important task, especially when demand side management is to be pursued. Modeling
electricity consumption or demand in the residential sector, using an end use approach,

serves this purpose.

154

CHAPTER 5

RESIDENTIAL ELECTRICITY MODELING:
METHODOLOGY AND RESULTS

Chapter five is divided into 2 sections. The first section discusses the methodology for
residential electricity demand modeling. The methodology section includes an
explanation of the types and sources of data as well as processing and data preparation in
this study. The assumptions and calculations involved in the model are also discussed in
this section. The second section presents and discusses the results, and compares

projections of residential electricity demand fi'om different studies.

I. METHODOLOGY

5.1 Introduction and Overview of the Model

Residential electricity modeling, in this study, looks at electricity consumption as a
function of the number of household rather than total population, because electricity is a
shared household resource. For example, a color television is shared by members of a

household, whether the number in the household is 3 or 6. Accordingly, the total number

155

of color televisions is more closely related to the number of households than to the total

population. In general,

Residential electric usage = f (Households, appliances, usage level)

Household = f (Population, household size)

This study incorporates the end-user approach to reflect the aforementioned relationship
in projecting electricity usage in the residential sector of Thailand. The model developed
in this study characterizes the long-term structure of electricity consumption for typical
Thai households in urban and non-urban areas in each province under differing
assumptions. This model incorporates the roles of changing social (population and
household), economic (appliance ownership and usage), and technological (unit
consumption of electric appliance) variables based on other studies, appliance
manufacturer guidelines, and expert opinion. Expert opinion, in this case, was based on
reading materials, the local knowledge and educated assumptions of the researcher to
project fixture demand for electricity. The output of residential electricity consumption is
the consumption in terms of total electricity usage and electricity usage by types of

appliances for urban and non-urban areas at national, regional and provincial levels.

Electricity usage of households in each area was calculated born the saturation rate of
each appliance, appliance usage intensity, unit input (watts) of appliance, and number of
households. The future changes of these input parameters were depicted using expert
opinion and applied to the model using trending methods. The scenario approach was

also used to look at the relationships between these variables under different levels of

156

parameters. Three scenarios (low, medium and high) were formulated to reflect these

changes. The results provide a range of future consumption rather than just one number.

5.2 Data Sources and Inputs to the Model

The residential electricity consumption model requires input data at the household level
in urban and non-urban areas in each province of Thailand. Two major types of inputs
were identified and obtained from multiple sources. These include housing

characteristics and electricity appliance inputs.

5.2.1 Housing Characteristic Inputs

Households are the basic unit of the residential electricity consumption model. The
number of households was used to derive the electric utility customers in the residential
sector. There are two variables that determine the number of utility customers. The first
variable is the household size or the average number of persons per household by areas
(urban and non-urban). These data were obtained fi'om the 1990 Provincial Population
and Housing Census. The second variable is the total population in urban and non-urban
areas for each province, which was obtained from the projection results of the population
model. Dividing the projected population number by the average household size yields

the number of utility customers.

157

5.2.2 Electric Appliance Data Inputs

There are two types of data associated with electric appliances in the model: ownership,
and unit input power and the intensity of usage. The first type, appliance ownership data,
was obtained from the 1990 Provincial Population and Housing Census. These data are
in a tabular form for urban and non-urban areas in a province. The rows contain a list of
nine types of electric appliances. For each row, there is a column with information on the
number of households that own the appliance listed in the associated row. Appliances
included in the model are radios, color televisions, black and white televisions, VCRs,
electric irons, electric rice cookers, electric fans, refrigerators, and air conditioners. Two
additional appliances were incorporated into the model. Computers were added because
of the increasing use of personal computers in households. The other, so called
appliance, is lighting. It is necessary to include lighting into the model because lighting
accounts for a large portion (24% in 1995) of the total residential electricity consumption

in Thailand.

The second type of electricity appliance data is the unit input power (or size, also known
as wattage) and the intensity of usage of appliances. Both sets of data are usually
obtained from the household survey on electricity. In the absence of current survey data
on electric appliance usage for Thailand, the data used in this study was obtained from

other sources listed as follows:

158

- Metropolitan and Provincial Electric Authority Thailand: Electricity sale
data, number of customers, and total electricity consumption at national
level.

0 Asian Development Bank Study on Energy End Use: India Household

Survey

0 Hawaiian Electric Company: Consumer Guide to Energy Costs for
Household Appliances.

0 BECi: Residential Services and kWh usage of electric household
appliances.

0 Edison (International) Electric Institute: Energy house consumption

patterns and appliance descriptions

0 Solardome: Energy Consumption of Household Appliances

Information on the power input and intensity of usage of electric appliances from the
above sources was used as a starting point for a base data in the model in this study.
Average power input of appliance and intensity of each appliance usage were chosen

based on educated assumptions of the researcher.

5.3 Preparation of Baseline Data

Baseline data preparation involved processing two types of data: population and housing

data and electric appliance data. Both types of data were entered and processed in MS

Excel. After the processing, both data were exported and stored with other data in MS

Access.

159

5.3.1 Population and Housing Data

Population and average household size data were used to determine the number of
households. Population data used in residential electricity model was obtained fiom
projected population in urban and non-urban areas from the population model at the
provincial level. Data on number of persons per household, obtained fiom the provincial
Census, was not classified based on urban and non-urban areas in each province. The
classification of this data was based on administrative boundary, which was divided into
five areas: municipal area, sanitary district, urban sanitary district, and non-municipal
area and outside sanitary district. This classification needed reclassification so that the

household size data would conform to the projected population data.

Based on the Ministry of Interior definition of municipal areas and urban subdistricts, the
definition is similar to the definition of urban areas in the population model. Therefore,
these areas within in each province were grouped together and the NCR2D provincial
codes were assigned for each province. The sanitary districts, and non-municipal area

and outside sanitary district were also grouped together and recoded.

5.3.2 Electric Appliance Data

A table of appliance ownership was created with the rows containing a list of provinces

by urban and non-urban areas and 11 columns of appliances, indicating the number of

160

households that own the appliances. The NRC2D codes were assigned to provinces. The

ownership data was used to calculate the saturation rate for each appliance in each area.

Since the census did not provide computer ownership data, this data was assigned to
households based on household location. It was assumed that households in non-urban
areas did not have personal home computers. Ownership of computers (number of
household with computer) for households in urban areas was assigned based on the
percent of total households in each area. The criteria for assigning computer ownership
value was that urban areas with major universities and areas closer to Bangkok received
higher values than urban areas located firrther away from Bangkok. The assigned
numbers were 10 percent for Bangkok, 5 percent for urban areas in the 3 provinces in the
Bangkok Metropolitan Region, 5 percent for the urban areas of provinces with major
universities (Chaingmai, Khon Kan, Nakhonratchasima, and Songkhla), and 1 percent for
the rest of the urban areas of the country, and none for the rural areas. The criteria and

percent were formulated based on the educated assumption of the researcher.

In the case of lighting (light bulbs), data on the average lighting used per household by
urban and non-urban areas in each province was not available. Therefore, this number
was calculated from the total electricity sales divided by the total number of households
and multiplied by the proportion of lighting to the total electricity consumption (24%).
The derived value was assumed to be the same for every household with access to
electricity. This assumption was necessary because baseline data of variables for

projection of average lighting consumed per household such as, floor space, types of light

161

bulb and tubes were not available. Moreover, modeling detailed household lighting is

tedious and complicated and is beyond the scope of this study.

Because of limited available data, some input variables that effect ownership and the
intensity of usage of appliances are not included in this study. These variables include
lifetime of appliances, purchase price, substitutability, efficiency of appliance, and
electricity price. The first four variables influence ownership and the last variable effects

the intensity of appliance usage.

5.4 Residential Electricity Consumption Model

The residential electricity consumption model in this study was developed by integrating
the results from the population model and using this result as the base for projecting the
electricity consumption in a household. Calculation of the residential demand for
electricity was performed based on average use of electric appliances owned by an
average single family in urban and non-urban areas in each province of Thailand. The
relational database MS Access was used to create necessary tables, to store data, and to
report the results. The Visual Basic programming language was used to define and run

the model.

Projection of residential electricity demand involves modeling kilowatthours consumed

from appliance usage in Thai households. The model includes three key variables for

162

appliances throughout the course of the projection: ownership, and intensity of usage and
power input/size. When combined with the number of electrified households, these
variables determine the residential sector electricity usage of Thailand. The relationship

of these four variables are described in the equation below:

E, = 2' (Usage * Sizea * Ownership, * Number of electrified households) (5.1)

Where:
Ea = Total electricity consumption of an appliance per year

a = A household electric appliance type a, a = 1 to 1 1 types

Projections of E, in this study were performed at the provincial level for the urban and
non-urban households of 76 provinces of Thailand. The process involves formulation of
assumptions and the projection can be described in the steps and orders as the
calculations were performed in the model. The assumptions and calculations are

summarized below.

5.4.1 Numbers of Electrified Household for each Province

Calculation

The calculated number of electrified households in urban and non-urban areas was used
as a representation of number of electrified households. The calculations were performed

by dividing the projected number of urban and rural populations in each province by the

163

average number of people per household in urban for the urban household case and rural

for the non-urban household case. This calculation is presented as equation follows:

Number of electrified householdA = _&ojeded.populatiam_ (5. 2)
Average People per house),
Where:
A = Areas in a province, A = urban and non-urban area.
Assumptions

Projections of the number of electrified urban and non-urban households were based on
three assumptions regarding the household types and sizes.

1) Electrification rate in urban areas is equal to the rate in non-urban areas and has
reached the saturation rate.

National statistics on electrification showed relatively high electrification rates of
households in both urban and non-urban areas of the country. The rates are 99 % for
urban areas and more than 90% for the non-urban areas. This assumption was formulated
because there is no information on electric accessibility at the level of disaggregation
required in this model.

2) The major of type of households in Thailand is single family housing where the
average household size varies from 3.5 — 4. 6 and 3. 7 — 5.1 persons per household in

urban and non-urban areas respectively.

164

The above household sizes are the ranges of the number of family members per
household from the Provincial Census. Since there is limited data on types of households
at the provincial and urban versus non-urban level, the above household size ranges were
then assumed to belong to single family households that are the majority of the housing
type in Thailand.

3) The household size will decrease by 25 percent within 25 years and the rate of
decrease is constant.

The 25 percent decrease was used and was based on the NESDB average household size
projection. The NESDB reported the percent decrease of household size from 1990 to

2015 to be 28 percent

It should be noted that the projection of electrified households or housing stock14 should
consider other variables such as housing decay rates” and vintage blocks“. These
variables were not included in the model partly because of inadequate data. It is also not

the purpose of this study to perform detailed projections of housing stock.

 

‘4 Housing stock refers to number of houses occupied in the base year (1994).
‘5 Decay rate is the rate at which existing houses are removed from the housing stock.
‘6 Vintage blocks are houses existing in 1994 versus houses built after 1994.

165

5.4.2 Ownership or Saturation Rate

Calculation
An electric appliance saturation rate represents the proportion of households owning the
appliance to the total number of households. Saturation rates of all appliances in urban

and non-urban households were calculated using the equation below:

Saturation Rate“ = Number of household with appliance“ (5. 3)
Number of household with appliancea,A . Number of household without appliance“

 

Assumptions

1) The maximum number of each type of appliance owned in a household is one. In other
words, the saturation is less than or equal to one.

According to fhe census, saturation rates for most types of appliance are still low for Thai
households. Therefore, it is reasonable to assume that no household owns more than 1 of
each type of appliance.

2) Consumer behavior in terms of appliance usage intensity is the same in urban and
non-urban households across the country.

This assumption was formulated because of inadequate information on electricity
consumption behavior of households in Thailand.

3) The adoption rate of electric appliances in urban areas increases at a higher rate than

the non-urban areas.

166

This assumption deals with fiiture saturation rates of each electric appliance. In general,
the change of appliance saturation rate over time of households in urban areas increases
more rapidly than the change of the rate of households in non-urban areas (Figure 34).
Future rates were determined and trended by applying percent increases of saturation
rates to the current rate. Two curves that determine percent increases of saturation rates
were created as shown in Figure 35. Because of insufficient historical data on appliance
ownership, no mathematical equation was derived to represent the relationship between
saturation rate and time. Instead, the current saturation rate of appliances was classified
into 3 ranges: low (0-0.33), average (0.33-0.66), and high (0.66 to 1.0). Increase of
saturation rate was assumed to increase at a higher rate and decreases as saturation
reaches 1. That is, for non-urban households, the percent increase of the saturation rate
(the slope of the curve) over the five-year period was assumed to belO percent for
appliances with a current rate within the low range, 5 percent for appliances with a
current rate within the average range, and 0.5 percent for appliances with a current rate

within the high range.

167

Sdurdion
rde

 

Figure 34: Change of Appliance Saturation Rate of Households in Urban and Non-urban
Areas over Time

ric'tliiim

 

 

 

 

 

Figure 35: Trend of Saturation Rates of Electric Appliances

168

For urban households, the percent increase of the saturation rate over the five-year period
was 20 percent for appliances with a current rate within the low range, 10 percent for
appliances with current rate within the average range, and 5 percent for appliances with
current rate within the high range. The percent increase of saturation rates in this model
was assigned based on the researcher’s opinion. These numbers could be changed later,
based on other expert or policy maker opinions on various factors that may effect the

appliance ownership situation.

The percent increase of saturation rate was applied for every five-year period. Future rate
was estimated based on the current rate and the percent increase. An example of
determining the increase of the saturation rate for an appliance can be illustrated as
follows. The following example assumes a current saturation rate for air conditioners in
a non-urban Thai household in province A in 1994 to be 0.02 or 2 percent. From Figure
37, the rate of 0.02 is considered a low saturation rate (falls in the range of 0 — 0.33). The
increase of the rate for this range is 10 percent (0.1) based on the assumption in this
study. Therefore, the saturation rate for air conditioners in non-urban province A in 1999
is the 0.02 + 0.1 = 0.12. This rate still falls within the low saturation rate range. The 10
percent increase is still applied in this case when calculating the saturation rate for the

year 2004, which would be 0.12 + 0.1 = 0.22.

As far as calculations and assumptions for electric appliance input data is concerned,

there are three points to be considered. The first one concerns applying trend and types

169

of appliances. This study grouped electric appliances into 3 types: luxury or superior”,
normal and inferior goodsls. Luxury goods include air conditioners, computers and
VCRs. Normal goods include refrigerators, color televisions, radios, irons, and lighting.
Inferior goods include black and white televisions and fans. All the inferior goods and
some normal goods, (irons and lighting), were not trended. The reason for not trending
these goods was based on the fact that these goods are relatively low price or affordable.
Therefore, most consumers (households) will have already purchased these goods if they
have access to electricity. This means that ownership of these goods will not
significantly change in the future. Refrigerators and color televisions were trended based
on the above mentioned criteria. For superior goods, the percent increase of the adoption
rates in non-urban areas was assumed to be the same as the percent increase of the rates
in urban areas. The percent increase of luxury goods in urban areas also followed the

slopes in Figure 35.

The second point involves an issue of the substitutability of some of the household
electric appliances. The appliances in this study that can be considered as substitute for
each other are color televisions and black and white televisions, and air conditioners and
fans. However, these goods are imperfect substitutes. For example, an “average”
household that purchases a new color television will not discard an old black and white
television. The household will use the old black and white television in a different room

in the house. For these reasons, this model treats these goods as non-substitute.

 

'7 Superior goods are goods that demand increase with higher incomes.
'8 Inferior goods are goods that decrease in demand when incomes increase.

170

The third point is related to the issue of migration between urban and non-urban areas.
By assigning different percent increase to appliance saturation rates based on areas, it is
assumed that people who migrate into a new area will develop the consumption behavior
in both appliance ownership and usage as residents of the new area within five years (the
projection is performed every five years). Because household income was not factored
into the model, the study assumed that the income earning potential in urban areas is
higher than non-urban areas. Therefore, people who migrate to urban areas will have
more income and are able to purchase and use more appliances than people in non-urban

areas.

5.4.3 Annual Electricity Consumption by Types of Appliance

The calculation of annual electricity consumption was performed by types of appliance
based on the intensity of appliance usage and the power required to operate each

appliance. The equation used for the calculation was as follows:

Annual Electric Consumptiona or kWh, = (Kilowatts Power Inputa) (Hours of operation per daya)
(Days of operation per year,) (5. 4)

where:

a = A household appliance type, a = 1 to 1 1

By summing up electricity usage of each appliance for all households in each area of a

province, the model calculated total electricity consumption in the residential sector for

171

each province by urban and non-urban areas. The calculation is represented in the

equation as follows:

Tate, kWha = kWha1 + kWh32+ ....* kWha12 (5. 5)

Where:

kWha

kilowatthours of appliance type a, while a = 1 to 1 1

5.4.4 Three Scenarios of Household Electricity Consumption Projection

The projections of annual residential electricity consumption in each province were
performed based on changing household size, saturation rate of appliances, number of
hours of appliance usage (intensity of usage), and the unit power input of the appliances.

Three scenarios — low, medium (base), and high - were used to reflect these changes.

Base scenario

This scenario looks at total household electricity consumption in each area when the
average household size decreases one percent per year or 25 percent within the next 25
years (1994-2019), and the usage intensity and unit input power of each appliance remain
unchanged in this period (see Appemdix B for information on the intensity of usage and

the unit power input of each appliance)

172

Low Scenario

The low scenario estimates total electricity consumption in each area based on a 20%
decrease of current appliance usage hours and power input (energy efficiency) of all

appliances except for air conditioners.

High Scenario

Under the high scenario, total household electricity consumption in each area was
projected based on a 20% increase of current appliance usage hours and energy

efficiency of all appliances except for air conditioners.

Note that trending of usage of all appliances, except air conditioners, was done based on
fraction percent of the current usage. Most end use energy consumption studies,
especially in industrialized nations, treat air conditioners separately and in great detail
because of the high saturation rate, high contribution to the total electricity demand, and
the complex physical and economic interactions that characterize air conditioner systems
in a household. For the case of Thailand, a detailed end use model for air conditioners at
present is not necessary because the saturation rate of this appliance is still very low
(lowest among other appliances). However, the intensity of usage and power input of air
conditioners were trended separately in this study, but not in great detail. Because of its
high input power compared to other household appliances, it is highly likely that the

decision on the intensity of usage of air conditioners will be different from other

173

appliances. In the case of air conditioners, three scenarios were developed to reflect the
change in the number of hours for operating air conditioners in a household. For the low
scenario, the number of hours per day and the number of days per year operating air
conditioners decreased from 8 to 6 hours a day and 300 to 200 days per year respectively.
For the high scenario, the number of hours per day running an air conditioner did not
change, but the number of days in a year was increased fi’om 300 to 365 days.

All changes in each scenario were formulated as fractions or percent change from the
current values (of household size, electricity consumption fi'om each appliance) rather
than changing individual values of each parameter because of the lack of historical data
and insufficient current information on household appliance usage. It is not possible to

determine accurately the fiiture intensity of appliance usage.

It is also crucial that planners are able to assess the impact of technological improvement
of high power input requirement appliances such as air conditioners. The model in this
study was built in a manner that is capable of assessing such impact by changing values
of power input variable in the model. However, assessment of technological impact was
not part of this study. Therefore, this variable was assumed to remain unchanged during

1 994-2019.

In summary, the residential electricity consumption model in this study projected the total
electricity usage in urban and non-urban households that was based on the change in the
mix of population in each province and the change in household electric appliance

ownership and usage. The relationship among these variables is not linear because of

174

different rates of adoption in different areas. The model utilized the projected urban and
non-urban population number from the population model as the base for the calculations
of the average number of electrified households. The projections of the residential
electricity demand were performed based on a low, medium, and high saturation rate
scenario and percent change of appliance usage. The projection results included total
urban and non-urban electricity consumption of 11 appliances from year 1999 to 2019 at
the provincial level, which was the aggregated to regional, and national level as well.
The results were reported in terms of the share of electricity consumption by appliance

types and total residential electricity consumption.

H PROJECTION RESULTS

This section is divided into two subsections. The first subsection presents base year
electricity consumption as well as the projection results from year 1999 to 2019. The
projection results are reported in two ways. The first shows the results in terms of
projected number of households, total electricity consumption at national, regional, and
provincial levels. The second presents the projection of electricity consumption by types
of appliance. The results are reported in terms of the share of each appliance to the total
electricity demand at the national level. Unless it is mentioned, the results reported the
medium scenario. The second subsection is a comparison of residential electricity

consumption projections between this study and the utility authorities in Thailand.

175

5.5.Projected Households

Households in Thailand were projected to increase from 12.7 million households in 1994
to 25.7 million households in 2019 or 102 percent increase in this period. In 2019,
provinces in Thailand will experience a range of household increase from 2 to 16.4
percent per year (see Figure 36). Provinces that will face large percent increase within
the next 25 years will be the same provinces that are expected to experience large percent
increase in population. These provinces include Nonthaburi (16.5%), Samut Sakom
(6.9%), Samut Prakam (10.5%), Prathum Thani (10.3%), Ranong (8.6%), Phuket 7.5%),
and Bangkok (7.1%). In terms of percent increase of households by areas, Thailand will
experience an increase of households of 81.7 percent in urban areas and 127 percent in
non-urban in the same period. The increase of household in the country will be

dominated by higher increase rate of household in urban areas.

176

 

Annual Houshehold Increase,
1994-2019

       
 

Pathum The 1; 10.3%
Bangkok: 7.1%

Nonthaburi 16.5%
Samut Sakhon: 10.6%

Ranong: 8 6%
Annual Houahold Change
2.0% - 2.8%
2.8% - 3.5%
- 3.5% - 5 0%
5.0% - 8 6%
Phuket 7 5% -

- 8.6% - 16.5%

 

 

 

Figure 36: Projected Percent Annual Increase of Household in Thailand, 1994-2019

177

5.6 Projected Total Residential Electricity Consumption

5.6.1 National Level

The results of the residential electricity consumption projections at the national level are
presented in Figure 37 (see Appemdix B for result tables). The results can be
summarized by scenario and area as follows. The high scenario shares the total
residential demand to increase from 26.4 to 127 GWh, a 382 percent increase between
1994 and 2019. Under the medium scenario, it is expected that the demand will change
from 18.9 to 81 Gwh (330 percent increase) in the same period. The low scenario shows
the increase of 3 12 percent of the demand or the change from 14.4 to 59.5 Gwh in the

next 25 years.

By area, the demand for electricity in urban areas will increase from 12.8 Gwh in 1994 to
49.4 Gwh in 2019. The demand in 2019 in urban areas is higher than the demand in non-
urban areas by 11 Gwh. However, the percent change of the demand for electricity in
non-urban households in this year from the base year is expected to be higher in the non-
urban areas. The increase will be 424 percent in non-urban area and 285 percent in urban

areas.

178

 

 

 

 

 

 

 

 

 

0.
m

 

 

 

 

 

0.0.0.000
m .m. m m m m

. 3.. use. 58.5.2.8 Esteem

283.9:
18-52:
88-52...
38-5.:
687:9:

39.8....

$8.528:
38.528:
88.528:
38-528:
82-528:

v8 TEE—go:

m «8.30..

v P8130;

88.33

38.33

mam 730..

8733

m

D

 

IUrban

 

Figure 37: Total Electricity Demand by Areas fiom Three Scenarios, 1994 and 2019

179

In terms of the share of electricity in each area to the total consumption, the share of
electricity demand in urban areas is projected to increase from 32 percent in 1994 to 39
percent in 2019 or 7 percent increase in the share within the next 25 years. This indicates
that the increase of the future consumption for electricity in the residential sector in
Thailand will be driven more by the demand of households in non-urban areas of the

nation.

The total residential electricity consumption or the average consumption per household
will increase at a decreasing rate over time. It is expected to decline from 9.53 percent in
1999 to 4.98 in 2019. In terms of total electricity consumption by areas, it is projected
that the annual residential electricity consumption in urban areas will be much higher
than the consumption in non-urban areas. The percent increase of the consumption in
urban areas is expected to increase at a slower rate than the percent increase in the non-
urban areas. The difference between the annual household consumption in these two
areas will become larger from 1994 to 2009 and start to decline until 2019. The
explanation for this is related to the saturation rate. Based on the assumption of percent
increase of the adoption rate in urban areas, the model uses the 5 percent increase for
appliances with a current saturation rate greater than 0.6 (high range) instead of the 10 or
20 percent for the medium and low range respectively and for the projection in the urban
areas. In the period 1994 to 2009, most high contribution electric consumption
appliances (refrigerators, irons, fans and lighting) except air conditioners in urban areas
will have adoption rates close to saturation. Therefore, the impact of residential

electricity consumption from these appliances in urban areas will be small. Although, the

180

adoption rate of air conditioners in urban areas will be lower than other appliances, the
rate of this appliance since 1999 will be greater than 0.3 and become larger than 0.6 in
2019. This means that the model will use the percent increase of the medium saturation
rate for the period 1994 to 2014 and use the high saturation rate for the period 2014 to
2019 to calculate the electricity consumption. The saturation of the same appliances in
the same period in non-urban areas will still be in the low saturation range for air
conditioners throughout the next 5 years and will be in the medium saturation range for
other appliances in this period. Therefore, the model will use the higher percent increase

of adoption rate than the urban area to calculation the residential electricity demand.

When comparing the change of annual growth rate between residential electricity
consumption, population, and household, the annual grth rate of electricity will be
relatively higher than the rate of population and household growth throughout the next 25
years (Figure 38). The trend of the decline of the population grth rate will be similar
to the trend of household growth rate because the number of households was calculated
from the population. The decline of these two trends will be much slower than the
decline of annual electricity consumption. This is an indication that the increase of
residential electricity consumption in Thailand will be more influenced by the increase of

appliance ownership than the increase of (electrified) households.

The projected consumption in terms of total kWh usage or consumption provides the

overall picture of the amount of the electricity required in the typical household in each

area of Thailand. This information is necessary for production planning, but is less

181

meaningful to energy policy makers to direct energy plans and programs (i.e., energy
conservation) in the country. The information on electricity usage by end use appliances
can better serve this function. In this study, the model projected adoption or penetration
rates for each appliance by areas and determined the electricity consumption share of
each appliance to the total demand for electricity for a household in urban and non-urban
areas in Thailand. The model performed this task at the provincial level, however, the
results were aggregated and reported at the national level by urban and non-urban areas.
The results are illustrated in Figure 39, 40, and 41 (also see Appemdix B, for detail

results).

12

 

 

// .. 1o

 

i633
r
\

1

 

 

 

Electricity Demand per Household
(kWhlhouaahold)
s i
1\ \
* .. 3..
Annual Growth Rate (Percent)

 

 

 

 

 

 

 

 

 

 

1W // v : : "T":
M . 2

500 ‘ - ~—-
0 . T r . v 0

1994 1989 2004 2009 2014 2019
Year
+Non-Urban Consunption +Urban Consumption +Nationd consunption
+Population Grown Rae +Eledricity Growth Rate ---O----House Growth Rate

 

 

 

Figure 38: Residential electricity Consumption and Annual Growth Rate of Population
and Electricity Consumption, 1994-2019

182

 

 

 

 

 

 

 

 

 

 

 

 

 

I

 

 

1(1)“)

m.
m

m. m. m. m m. m m.
m m m m n m m
scented. eoucu=n5< E2“.

635360 3.03005 3 225

10.“) 1

m0)

.oSano

28m

3.26

5.808».

3.3.00

.52..

can

:2.

3.80030”.

I 2019

I1994

 

 

Figure 39: Share of Electricity Consumption fi'om Appliances, 1994 and 2019

183

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m.
m

m.
m

m m. m. m. m. m. m.
m m w m m m m
3520...: 805:9? E2“.

6.383300 3.03003 3 053m

mo>

35.3800

0.31

3.30

55.3002".

3.3.00

32..

can.

:2.

3.8350".

.m
w

.2019

I1994

 

 

Figure 40: Share of Electricity Consumption fiom Appliance Types in Urban Area, 1994

and 2019

184

 

 

 

 

 

 

 

 

 

 

 

 

 

r

 

 

m.
m

m
m.

mmmmmmm
mmmmmmm

3:020... 30:0..00< EOE
cOzOEaecoo b.0.beo.m .0 2:00

m0>

650500

2.5.

3.26

9.308.”.

3.3.00

23

SE

:2.

00.80023”.

.
a

I2019

.1994

 

 

Figure 41: Share of Electricity Consumption fiom Appliance in Non-Urban Areas, 1994

and 2019

185

The share of electricity consumption by appliances indicates that the major end use of
electricity in Thailand’s residential sector for both urban and non-urban areas, ranging
from highest to lowest share, will be air conditioners, refrigerators, irons and lighting.
The level of consumption from each appliance is projected to be different in both areas.
The difference is found to be larger for these four appliances and smaller for the rest of
appliances. The share of electricity consumption from all normal and inferior goods is
estimated to be higher in urban areas in the base year. The share from these appliances,

except the iron, is also expected to remain higher in non-urban until the year 2019.

Among the major end use appliances, there will be a distinct different between the share
of electricity consumption from air conditioners and lighting in urban and non-urban
areas. It is expected that the contribution of electricity consumption from air conditioners
will increase from 3.4 percent to 39.5 percent in non-urban areas and 37.5 to 55.9 in
urban areas in1994 and 2019 respectively. In the case of lighting, there will be a decline
of the share from lighting in both areas with the higher reduction from 21.7 to 7.4 percent
in non-urban areas and from 7.9 to 4.7 percent in urban areas in the same period. The
difference between the share from refrigerators and irons is also expected to decline. The
decline of the electricity consumption share in non-urban and urban areas will be 5.6 and

4.58 percent from refrigerators and 12.2 and 5.12 percent from irons accordingly.

The difference in household electric consumption between urban and non-urban areas in

this study can be looked at through the four variables in the model: number of

households, appliance ownership, penetration rates, and the intensity of appliance usage

186

in the households. At the household level, this study assumed that the intensity of
appliance operation for each appliance is constant over time and is the same throughout
the country. In addition, the input power for each type of appliance is also assumed to be
the same in every household. Therefore, the increase of consumption is dependent on the
saturation rate and the changing number of households. Both variables were trended to
increase linearly for households and non-linearly for saturation rate over time. The
impact of the increase of number of households to total electricity consumption is a linear
relationship. Therefore, it is important to look at the change in saturation rate of each
appliance over time and how this change contributes to the total electricity consumption
in the residential sector. Figure 42, 43 and 44 presents the change in saturation rate by

type of appliance in total, urban and non-urban areas respectively.

187

 

 

 

 

 

 

 

«am 8:238

>._. 3.525.005

.2an00

32>

E .28

50.000 8.x

28m

.3:

can.

:0...

08209.33.

55228 ._<

Millstone

J

.2319

I1Q4

 

 

Figure 42: Saturation Rate by Appliances at National Level, 1994 and 2019

188

 

>P casts—uni

52.58

>._. 3.00

3:80 8E

0.31

Ea:

can

:2—

8.8833.

520.2300 .2

I319

I194

 

 

Figure 43: Saturation Rate by Appliances in Urban Areas, 1994 and 2019

189

 

 

 

 

 

1.2)

ND

er

a

0
8.1 5:823

04%

>.r 82:5.an

.oSnEoo

I319

I19

 

 

Figure 44: Saturation Rate by Appliances in Non-Urban Areas, 1994 and 2019

For the purpose of the analysis, appliances are categorized into 3 groups based on the
change of the saturation rate from base year to 2019. The first group is referred to as the
low change rate group. This group includes appliances that are not trended in the model.
These appliances are irons, fans, lighting, radios, and black and white televisions. The
change in the saturation rate is estimated to range from 0.0 to 0.03. While these
appliances were not trended to change, the increasing rate is the result from the number
of households that were trended to increase. Of all the appliances belonging to this
group, irons and fans are the ones that deserve special attention because of their relatively

high contribution to total electricity consumption.

In the case of electric irons, the explanation for the moderate high contribution from this
appliance is technologically and culturally involved. The high required power input of
irons and the relatively high intensity of its usage in households contribute to the high
share to the total electricity consumption. An iron is a small household appliance that
requires power input as high as 1,000 watts. Ironing clothes is also one of the major
household chores in Thailand because of the preferred cotton fabric that usually requires
ironing. This applies more to households in urban areas as the saturation rate of this
appliance is higher and its share of the total electricity consumption will be higher in

2019.

Unlike the case of irons, the contribution of fans to the total electric consumption is not

the result of high power input and the intensity of usage, but more from a high saturation

rate. In 2019, the saturation rate of fans is expected to be high in urban areas (0.93) and

191

relatively high in non-urban areas (0.64) as well. Although a fan is not a perfect
substitute for an air conditioner, the expected high increase of air conditioners make the

share from this appliance decline fi'om 10.5 to 5.2 percent in the period 1994 to 2019.

The second group is referred to as the medium-change saturation rate group. This group
includes air conditioners, rice cookers, VCRs, and computers. All appliances in this
group except rice cookers have very low saturation rates in the base year (0.07 to 0.16)
and medium saturation rates (0.29 to 0.43) in 2019. Other than air conditioners, it is
expected that the electricity consumption share of these appliances to the total electricity
consumption will be less than 4 percent in 1994 and decrease to less than 2.5 percent in
2019. The appliance that will most impact future demand of electricity in the residential
sector is air conditioners. At present, air conditioners have one of the lowest saturation
rates among other appliances (0.01 percent in non-urban areas and 0.25 percent in urban
areas) but contribute the most to total household appliance usage, especially in the urban
areas where the contribution is as high as 37 .5 percent. In 2019 the saturation rate of air
conditioners is projected to be 0.26 with a share to the total consumption of 36.2 percent
in non-urban areas. In urban areas in the same year, the saturation rate is expected to be
0.63 with a share to the total consumption of 55.9 percent. The high contribution to total
electricity consumption from this appliance is partly caused by the high power input of
the typical air conditioner used in Thai households. Improvement of the efficiency of air
conditioners will certainly reduce the total electricity demand in the residential sector.

However, changing the intensity of air conditioner usage such as setting the air

192

conditioner at a comfortable temperature (not too low) will also lessen the total

residential electricity consumption.

The last group of appliances is referred to as the high change rate group. The appliances
belonging to this group are refi’igerators and color televisions. Color televisions are not a
major concern in this case because the saturation rate of this appliance is projected to be
relatively high while the share to total electricity consumption will be relatively low.
Refrigerators also will have high share and saturation rate in the future. The share in the
non-urban areas will be as high as 30.7 percent, with the saturation rate of 0.83 in the
base year. The share is expected to be 20.1 percent, with the saturation rate of 0.92 in
urban areas in the year 2019. A refrigerator is a relatively high priced and a non-
substitute good. It is the only electric appliance in a household that is operated 24 hours a
day and 365 days a year. The fix in the intensity of usage and the near saturation of this
appliance in Thai households explains the slight change in the share from this appliance
in the period 1994 to 2019. The share is expected to decrease from 24.9 percent in this
period. Because all households operate refrigerators all day and all year long, the only
way to reduce the amount of electricity consumption from this appliance is through

efficiency improvement or a decrease in the unit power input of the refrigerator.

At the country level, the increase of residential electricity consumption will be driven by
the increase of consumption in the non-urban areas of the country where the majority of
the population (and households) are. The projection of the increasing saturation rate for

all appliances, especially in non-urban areas, will be a key factor that drives the higher

193

electricity consumption in Thai households. The near saturation rate of lighting since
1999 confirms that the increase of residential electricity demand will result from the
increase of appliance ownership in non-urban areas in particular. The share of electricity
consumption in the residential sector from appliances also indicates the high impact of air

conditioning to the total residential electric consumption in Thailand.

5.6.2 Regional Level

Bangkok

It is crucial to look at projection results of Bangkok separately fi'om the Central region
because of differences in the social, economic, and political structure of Bangkok
compared to the rest of the country. The estimate of consumption of electricity in the
residential sector in Bangkok is approximately 5.42 Giga kWh in 1994. It is estimated
that the consumption will increase to 18.1 Giga kWh in 2019. The electricity
consumption in 2019 will be approximately 22.3 percent of the total electricity
consumption in the residential sector of the country. Although the electricity
consumption in Bangkok is projected to be second highest from the Northeast region, the
percent change of the consumption in Bangkok is projected to be the lowest in the
country. A large percent change of electricity consumption is not expected in Bangkok
because appliance ownership of most appliances in Bangkok households was relatively
high compared to households in other regions in the base period. When the saturation

rates of appliances were in the high range, the model used a low percent increase of

194

saturation rates to estimate total electricity consumption. Therefore, the change of the

consumption in Bangkok will be small.

Central Region

If Bangkok is included in the Central region, residential consumption for electricity of
this region in 2019 will be 43.1 percent of Thailand’s total electricity consumption in
residential sector. By not including Bangkok in the Central region, this region is
projected to experience a change of the residential electricity consumption from 3.4 Giga
kWh in 1994 to 16.9 Giga kWh in 2019. The region will experience a change of a
regional share of the consumption from 18 to 20.8 percent in this period. The increase of
the residential electricity demand of the Central region was calculated to be
approximately 15.9 percent per year. The residential consumption for electricity within
this region will be dominated by urban areas. The urban areas will account for 64.4 and
69 percent of total electricity consumption of the region in 1994 and 2019 respectively.
The region’s contribution from the urban area to the country’s consumption is projected
to change from 2.7 to 14.4 percent in this period. This is a significant increase compared
to the rest of the region, which is expected to experience a decline of the share from
urban area in this period. Share of the residential consumption of non-urban areas of the

Central region in this period is also expected to increase from 1.5 to 6.4 percent.

195

Northeast Region

The result from this study reveals that the Northeast region will experience a change of
electricity consumption in the residential sector from 3.6 to 18.6 Giga kWh in 1994 and
2019 respectively. In 2019, the Northeast region is expected to be the largest electricity
consumer in the country. The change of the consumption of this region will be
approximately 16.6 percent per year, the highest percent change in the country (see
Figure 47). The increase of the residential consumption in this region will be dominated
by the demand in non-urban areas. There will be a change of approximately 20.4 percent
annually in these areas of the region. The residential electricity consumption share from
the non-urban areas of the Northeast reigon in 2019 will be as high as 74.6 percent of the

region consumption and 16.8 percent of the national consumption.

East Region

The East region is ranked the lowest in terms of electricity consumption and the share of
consumption of the region, as well as the consumption in urban and non-urban areas. It is
normal for the East region to experience the low change because this region has the
lowest population and household in the country. Estimates of total electricity

consumption in this region are 0.96 Giga kWh in 1994 and 4.75 Giga kWh in 2019.

The consumption was calculated to be a share of the consumption in this region

approximately 5 percent in this period. It is interesting that projection results show that

196

this region will experience the second highest percent change of residential electricity
consumption. The estimate is 15.8 percent per year. An explanation for this is because
of the relatively high increase of population, households, as well as saturation rate of
appliances. While most regions will experience a decline in the consumption share in
urban areas and an increase in non-urban areas, the East region will be the only region
that will experience an increase of urban and non-urban shares of the consumption from
1994 to 2019. However, the region’s contribution to the total electricity consumption in
the residential sector will not change in this period. In terms of consumption for
electricity by area within the region, neither urban nor non-urban areas will dominate the

region’s consumption in the period 1994 to 2019.

South Region

The estimated residential electricity demand of the South region is 2.24 and 9.55 Giga
kWh in 1994 and 2019 respectively. This increase will be approximately 13.1 percent
per year. Percent residential demand for electricity of the South region to the national
demand will remain unchanged at approximately 11 percent at the beginning and the end
of the projection period. The residential demand for electricity in this region has always
been concentrated in households in urban areas. The urban households in the South
region are expected to be the second largest electricity consumers in the country
compared to remaining urban households in each region. Domination of urban area in
this region will decrease as shown in Figure 48. Urban households of the South region

will experience a decreasing electricity consumption share within the region, from 74.6 to

197

69.6 percent in 1994 and 2019 respectively. However, the region’s share from urban area
to the total electricity demand in Thailand will be less significant than the former case. It

is found to decline from 8.84 to 8.2 percent or a decrease of 0.64 percent in this period.

North Region

The North region is expected to experience an increase of electricity demand in the
residential sector from 3.2 Giga kWh in 1994 to 13.2 Giga kWh in 2019. The increase of
the demand will be approximately 12.3 percent per year in this period. Excluding
Bangkok, this change will be the smallest percent increase of the demand compared to
other regions. In terms of the region’s share to the national consumption, the
consumption in the North region is expected to account for 16.3 percent in 2019, a
decline from 17.2 percent in 1994. A consumption pattern within this region will change
from a higher consumption in urban areas in 1994 to a higher consumption in non-urban
areas in 2019. The difference between the share of urban and non-urban household
consumption in this region will be 5.2 percent in 1994 and 12.5 percent in 2019. The
difference of 7.3 percent is relatively small compared to other regions except the East
region. The difference between the electricity demand in urban and non-urban areas is
expected to be less than 2 percent for the East region and more than 20 percent for the

other regions.

198

 

m womabaom

v8 7539.0,

 

 

a Smétoz

3m 75.02

 

o 5N.fiao£._oz

«mm 380532

 

m «outflow w

«mm zoom m

a wow
.Xxm 23 .850

3a w
.53 23 fibcoo

 

 

 

 

a «owe—8.95m

3229.28

 

 

 

 

 

a Emu—abnoo

3m 7.2500

 

 

 

 

 

 

 

 

 

m. m m. m m. m. m. w. m
a a m a m .._.. w s o
.55.. ago. 5.55880 3.9585

lNon-me

199

lumen

 

 

Figure 45: Residential Electricity Consumption by Region, 1994 and 2019

 

 

 

 

8 8i
8

.8

 

 

 

 

B
8

 

 

Share of Region (Percent)
hi
8

.a
p
8

 

50m

 

0.00 -

 

 

i I 1994 I2019 j

 

Figure 46: Share of Residential Electricity Consumption by Region, 1994 and 2019

200

 

 

 

 

 

a
3

 

 

3

 

8.8.8
8

3

 

 

 

 

Share of Region (Percent)

   

 

 

    

 

 

   

 

10.oo~
5.0m I r l I
cool
Certal 8m CerlralWo Est Nutheast Northeast Sarah
Baumk
Recon
L .1994 I2019

 

Figure 47: Residential Electricity Consumption by Region in Urban Areas, 1994 and
2019

201

 

Share of Region (Percent)

 

18.00

 

16.00

 

 

 

14.00

 

 

12.00

 

10.00

 

 

 

 

 

 

 

 

 

Central wlo East Northeast Northeast Somh
Bangkok

Region
[ I 1994 I2019

 

Figure 48: Residential Electricity Consumption by Region in Non-Urban Areas,
1994 and 2019

202

 

The electricity consumption in the residential sector at the regional level can be ranked
from the highest to the lowest as the Central (including Bangkok), Northeast, North,
South and East region during 1994 to 2019 (Figure 45). By areas, the consumption in
urban areas is found to be the highest in Bangkok, followed by the Central, the South,
North, Northeast, and East regions in 2019. For non-urban areas, the major consumption
is expected to take place in the Northeast region, followed by the North, Central, South
and East region in the same year. In terms of the percent change of total demand within
each region, the highest and the lowest change will take place in the Northeast and
Bangkok respectively in this period. The region that will experience the highest and the
lowest percent change of the demand in urban areas will be the East region and the North
region. In non—urban areas, the projection results reveal that the lowest percent change in
the residential electricity demand will take place in the Central and the highest percent

change will take place in the Northeast region in the same period.

5.6.3 Provincial Level

Projection results at the provincial level are presented as the consumption in 2019 of each
province (see Figure 49) and percent increase of the consumption from 1994-2019 (see
Figure 50). Spatial change of the household electricity consumption has to be examined
in terms of actual consumption as well as percent increase. Planners should know
estimates of future consumption of each province in order to prioritize resources for
provinces with the high future consumption. At the same time, planners have to be

informed and aware of provinces that show a high potential for increased future

203

consumption. These provinces include provinces that are estimated to have large percent
increase of the consumption in the period of 1994-2019. Estimates of the residential
electricity consumption of each province are also presented in terms of percent increase

of the consumption in urban (see Figure 51) and non-urban areas (see Figure 52).

High residential electricity consumption is expected to take place in provinces that are the
center of each region. High consumption is expected to be largest in Bangkok, followed
by Nonthaburi and Samut Prakarn, the three center provinces of the Central region. The
fourth and fifih largest consumption will be in Nakhon Ratchasima, the center province
of the Northeast region, and in Chiangmai, the center province of the North region
respectively. The consumption in these provinces is expected to be high because these
provinces have high population, households, and urbanization rate. However, the percent
change of the consumption in some of these provinces is not large. In fact, Bangkok,
which shows the highest fixture consumption, will experience very low percent increase
of the consumption. This is because in the base year Bangkok was already highly
urbanized and saturation rate of most appliances was also the highest in the country.
Therefore, Bangkok will not show large change in the consumption, but the consumption

will still be very high.

204

 

Household Electricity Consumption. 2019

      
  
 

Nakhon Ratchasima: 2.384 Million kWh
Chiangmai:
2.032 Million kWn

Nonthaburi:
2.907 Million kWh

Electricity Consumption (million kWh)
184 - 652

[:1

- 652- 1.115
- 1.115— 1.810
- 1.810— 2.907
- 2.907- 18.067

 

 

 

Figure 49: Projected Residential Electricity Consumption of Provinces in "Thailand, 2019

205

 

Percent Residential Electricity Consumption
Increase by Province, 1994-2019

it Bedricity Dem-1d Change
1 1 - .

22 275 6
275.0 - 337.1
337.1 - 304.2
394.2 - 403.9
403.9 - 745.4

 

   

Norlhabu'l - 745.4% _

SantrtSakhorn - 640.2%—

 

Scrum-56%

 

 

 

 

Figure 50: Percent Electricity Consumption Change in Thailand, 1994-2019

Other provinces that will experience a large percent increase of electricity include the
BMR provinces. It is not a surprise that these provinces will face high future
consumption as well as a large percent change because these provinces are still under
going development process. Unlike Bangkok, these provinces have high percent increase
of households as well as adoption rate of appliances, which makes the future

consumption and the percent increase in the consumption to be high.

206

The interesting provinces that show large percent increase but relatively low future
electricity consumption are provinces in the Northeast region, which include Amnat
Charoen, Surin, and Si Saket. The large increase in the consumption in these provinces
will be the result of large increase of the consumption in non-urban households.
Households in these provinces had low adoption rate of appliances in the base year.
Therefore, percent increase of saturation rate of appliances of households in these

provinces will be high which in terns makes the change in the consumption to be high.

Electricity consumption of each province for urban and non-urban areas is shown in

Figure 51 and Figure 52 respectively.

207

 

Residential Electricity Demand Change in
Urban Areas Between 1994 and 2019

Elec Chg Urban Change (Gwh)
0 - 230

230 - 580
- 580 - 1050
- 1050 - 2230
- 2230 - 12650

   

Nonthaburi - 2.076 Gwh
l

    

Bangkok - 12.648 Gwh

Samut Prakam - 2,234 Gwh

 

 

 

Figure 51: Percent Change of Residential Electricity Consumption in Urban Areas by
Province, 1994-2019

208

 

 

Residential Electricity Coneumption Charge in
Non-Urban Areas Between 1004 ltd 2019

Bee Chg Non-Urban (Gwh)
o - 130

 

Ubon Ratehamaniu,
804 Gwh -

      
 

Khon Kaen - 994 Gwh

Nakhon Ratchasrma -
1.432 Gwh

Sun'n — 871 Gwh

 

 

Si Sakit. 821 Gwh

 

Figure 52: Percent Change of Residential Electricity Consumption in
Non-Urban Areas by Province, 1994-2019

209

 

The result from the regional level indicates that the highest change will take place in the
urban areas of the Central region. The results at the provincial level shows that provinces
in the Central region that will experience a large increase in residential electricity
consumption, ranking from high to low, are Bangkok, Nonthaburi, and Samut Prakam
accordingly. Figure 51 confirms that the high change of the total consumption in these
provinces will result mainly from the change in urban areas and from the minor changes
of the non—urban areas, as these provinces will experience a small change in demand in

non-urban areas.

The regional results indicate the highest change in residential electricity consumption in
non-urban areas to take place in the Northeast region. The provincial results show that
these changes will take place in non-urban areas of five provinces in the Northeast region
(see Figure 52). Most non-urban areas of the provinces in this region will also experience

relatively medium to high changes in residential electricity demand.

5.6.4 Electricity Consumption by Type of Appliance

The other way to look at changes of residential electricity consumption is to look at the
consumption in terms of contribution of each household appliance to the total electricity
consumption. The consumption from each appliance was projected and the difference of
the consumption in the period 1994 to 2019 was calculated and presented as total change

in each province, and change in urban and non - urban areas within in a province.

210

 

Change in Residential Electricity Demand
from Air Conditioner, 1994-2019

   
   
 

Total Change (Gwh)
[j 80 - 290
- 290 - 530

530 — 850

-850- 1300
- 1300- 8310

  
   

Nakhon Ratchasima - 995 Gwh ’ i ’ '

Nonthaburi - 1.303 Gwh

Bangkok 8.306 Gwh

 

 

Samut Prakam - 1.205 Gwh

 

 

 

Figure 53: Change of Electricity Consumption by Air Conditioner, 1994-2019

211

 

Air Conditioner - Urban Areas

 
 
  

Urban Change (Gwh)
0 - 14

1200-8310

Nonthaburi - 1.901 Gwh

Samut Prekam- 1.199 Gwh

 

 

 

 

Figure 54: Cinnge of Electricity Consumption by Air Conditioner in Non-urban Area,
1994-2019

212

 

Air Conditioner - Non-Urban Areas

Non-Urban Change (Gwh)
0 - 60
80 -140
140 - 230
230 - 360
360 - 700

   

KhonKaen-Meem
. UbonRachdhmi-401Gwh'"

    
 
 
 

Nakhon Rachasima - 695 Gm

Su‘in - 426 Gwh

   
 

Si Selcet -396 Gwh

 

 

 

 

Figure 55: Change of Electricity Consumption by Air Conditioner in Non-urban Area,
1994-2019

213

 

Contribution of air conditioners to total electricity usage of households is expected to be
the highest among other appliances. Therefore, air conditioning is used as an example to
illustrate the capability of this model to estimate detailed electricity consumption usage of
. each appliance at the provincial level by urban and non-urban areas. The results are
presented as change of the consumption as total in Figure 53, change in urban areas in
Figure 54, and change in non-urban areas in Figure 55. Provinces with a large increase in
the residential electricity usage from air conditioners will take place in the same province
that will experience a large increase in total electricity consumption, both in urban and
non-urban areas. This is because air conditioners are a major end use appliance with the
highest share to the total consumption. The contribution of this appliance is expected to

be approximately 50 percent of the total consumption of the country in 2019.

Electricity consumption from each appliance in each area is closely related to the
saturation rate of appliances and the number of households in the area. Similar
calculations were performed to calculate the change in number of households, and
appliance saturation rate for urban and non-urban areas in all provinces of Thailand. The

results can be found in Appendix B

5.7 Comparison of Estimated Residential Electricity Consumption

This subsection compares the residential electricity demand projection results obtained

from this study and the combined projection results of the Metropolitan Electric

214

Authority (MBA) and Provincial Electric Authority (PEA). This includes comparisons of
projected numbers of households or customers and the annual household electricity
demand. While the projection period of the three studies is different, the discussions are
limited to the period from 1994-2009, which is the projection period that all studies have
in common. The discussion is based on the result and assumptions of the model in this

study since the methodology and assumptions of MBA and PEA studies are not available.

5.7.1 Estimated Households

The MEA&PEA projected an increased number of households from 12 to 18 million
households or an increase of 49 percent from 1994 to 2019. This study shows an increase
in the same period to be approximately 32 percent or an increase from 15 to 20 million
houses (see Figure 56). The curves from both studies seem to be linear. The steeper
slope of the MEA&PEA curve indicates a slightly higher percent increase the agencies

used to estimate future households.

215

25

15 /

10

 

 

 

 

 

Number of Customer or Households (Millions)

 

 

 

1999 2004 2009
You
i + PEA&MEA + Study Model J

 

 

Figure 56: Projected Total Number of Households (Customers) in Thailand, 1994 - 2019

5.7.2 Estimated Total Residential Electricity Consumption

Figure 57 shows a slightly more rapid increase of future electricity consumption in the
period 1999-2009 than the low and medium scenario in this study. The estimates of
PEA&MEA will be closer to the low scenario of this study at the beginning of the
projection and will be closer to the medium scenario as time progresses. The difference
between the estimates from both studies could be a function of the different percent
increase used to estimate firture households. This study yields a higher estimate in 2019

than the other study because this study used has a higher estimate of number of

216

households. However, it seems that the PEA&MEA may use a higher percent increase to
estimate fixture number of household in the low and medium scenario in this study. Ifthe
projection curve of the PEA&MEA were extended firrther, future estimate from the

PEA&MEA study will exceed the estimate from this study.

 

140000

120000 /

 

 

 

 

 

 

 

 

 

 

E 100000
3
3
u /
5 80000 ,J
E ,o”/
8 /
.Jr"
5 60000 “xv/A /’/,._
3 /?J -.. ”‘
11.1 40000 , -j x
20000 -
0 . . . . .
1994 1999 2004 2009 2014 2019

Year
L—e— MPA & PEA + Medium Scenario + High Scenario —\'— Low Scenario

 

Figure 57: Residential Electricity Consumption Projected under Different Scenarios

Saturation rate, power input and usage intensity of appliances also play important role in
the projection of future electricity consumption in the residential sector. The data on
these variables are usually obtained from an electricity consumption household survey,

which is not currently available in Thailand. These data, used in this study, were drawn

217

from other studies combined with the research’s knowledge, and educated assumptions
that may not reflect the true situation of appliance adoption and electricity usage of
households in Thailand. Types of electric household appliance included in this study
were based on the types of appliance included in the Census. The eleven types of
appliance included in the model do not represent all appliances owned in a typical Thai
household. Some large appliances (i.e., washer and dryer) as well as small appliances
(i.e., microwave oven, dishwasher, sewing machine, coffee maker, toaster, and hair
dryer) were not incorporated in this model partly because of lack of data on these
appliances and partly because of their low saturation rate in Thailand. Of all these
appliances, some appliances such as dishwashers and toasters require high input power to
operate (1190 — 1250 watts for dishwasher and 1100-1250 watts for a toaster). Not

including these appliances into the model results in conservative projections.

In summary, the residential electricity demand model was developed at the provincial
level. Projections of the demand for urban and non-urban household electricity
consumption from each appliance were performed based on assumptions formulated for
changes of number of households, appliance saturation rates, intensity usage and power
input of appliances in each area. Three scenarios were established to reflect the
relationship between residential electricity demand and the changes of these variables.
The results from the scenarios provide ranges of potential demand for energy planners
and policy makers to choose in order to meet the demand. National and regional
projections were not performed separately. They were calculated by aggregating the

projection results at the provincial level.

218

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS

This chapter is divided into 4 sections. The first section summarizes the results of this
study at three levels: national, regional, and provincial. The second section presents the
implications of this study for energy planning. This includes policy options to meet
estimates of the potential range of electricity demand in households of Thailand. The
third section provides recommendations for further study and the last section is final

remarks.

6.1 Summary of Results

Population and residential electricity consumption in Thailand were projected at the
provincial level. Projection results were aggregated to obtain the regional and national
population and electricity usage in residential sector. Because the projections involved
estimations of future values for many variables, the results should be viewed in the

context of these estimates (and assumptions in some cases). A summary of results

219

follows, presented in 2 subsections; population estimates and residential electricity

consumption estimates.

6.1.1 Population Estimates

National Perspective

The national projection results show an increasing population from 55.4 million in 1994
to 83.6 million in 2019. During this period, more than half of the Thai population lives in
non-urban areas. However, the non-urban population is expected to increase with a
decreasing rate as reflected in percent non-urban population decreasing from 58.5 to 53.7
percent in this period. Average population growth rate in the same period will decline

from 2.08 to 1.42 percent annually.

Regional Perspective

In the period 1994-2019, Bangkok is expected to face an increase of population from 5.5
to 10.5 million, with all of the population living in urban areas. By 2019, Bangkok will
account for 12.6 percent of population in Thailand. For other regions, the percent total
population will be 31.7 percent for the Northeast region, 20 percent for the Central region
(without Bangkok), 17 percent for the North region, 12.4 percent for the South region,
and 6.3 percent for the East region in the same year. The highest percentage increase of
population will take place in Bangkok, followed by the Central, East, South, Northeast,

and North region in this period. After Bangkok, the regions that will experience the

220

largest change in population in urban and non-urban areas will be the Central and the

Northeast region respectively.

Provincial Perspective

At the provincial level, population growth between 1994 and 2019 will be large, ranging
from 12.4 to 283.3 percent or a range from 0.5 to 11.3 percent per year. The mean value
of the percent increase of population in this period will be 48.7 percent with a standard
deviation of 42.7 percent. According to criteria established to group provinces based on
percent population increase (see Chapter 3 for the criteria), the most interesting group
include provinces that are expected to have a very high percent increase in population
(172 percent to 184 percent), and a population more than 2 million people. This group
includes Bangkok, Nonthaburi and Samut Prakam in the Central region or Bangkok
Metropolitan Region (BMR). Provinces in the Extended Bangkok Metropolitan Region
(EBMR) including Nakhon Pathom and Chonburi are also expected to have population
more than 1 but less than 2 million and a percent population increase in the high range,
65 percent for Nakhon Pathom, and 67.5 percent for Chonburi. Some provinces in the
Northeast region including Nakhon Ratchasima, Khon Kaen, and Ubon Ratchathani are
expected to have a population of more than 2 million people. However, the percent
increase of population in these provinces will not be high and below the average value
(38 to 41 percent). A high percent population change of 67 to 135 percent is also found
in provinces that will have relatively low population (35,000 to 70,000 people) in 2019.

These provinces include Ranong and Phuket in the South and Rayong and Chanthaburi in

221

the East region. A medium level of change is expected to take place in Saraburi, Sa
Kaeo, Trad in the East region, Phangnga in the South region, and Kanchanaburi in the
Central region. The remaining provinces will experience a low percent increase of

population in this period.

6.1.2 The Residential Electricity Consumption

National Perspective

The result of the residential electricity consumption under the base scenario is an increase
of the consumption from 18,857 Giga Watt-hour (GWh) in 1994 to 81,011 GWh in 2019.
The low scenario shows the increase to be fi'om 14,425 to 59,494 GWh in the same
period. The increase is estimated to be from 26,367 to 127,130 GWh in this period for

the high scenario.

For the base scenario, the additional annual electricity consumption of 62,154 GWh in
2019 is calculated to be an average increasing total electricity consumption in the
residential sector of 6.8 percent per year. The increase of the total consumption in the
urban and non-urban households per year in this period will be 8.0 and 6.2 percent
respectively. The average percent increase per year of total residential electricity demand
will be four times as much as the average percent population increase per year (1.7
percent) in the same period. The percent of electricity consumption by end use in 2019 is
projected to be 49.5 percent from air conditioners, 24.3 percent from refrigerators, 5.8

percent from lighting, 17.2 percent from irons, 5.1 percent fiom fans, and 8.1 percent

222

from other appliances at the national level. This compares to 26.6 percent from air
conditioners, 24.9 percent from refiigerators, 12.4 percent from lighting, 14.7 percent

from irons, 10.5 percent from fans, and 10.9 percent from other appliances in 1994.

Regional Perspective

The regional share of the residential electricity consumption in 2019 is estimated to be
43.1 percent from the Central region (including Bangkok with 22.3 percent from
Bangkok), 22.9 percent from the Northeast region, 16.3 percent from the North region,
11.8 percent from the South region, and 5.9 percent from the East region. The ranking of
the projected share of the electricity consumption, by region, corresponds to the ranking
of the share of population by region. That is, the region with the highest expected
population in 2019 will also have the highest household electricity consumption. By
area, the results indicate that the highest electricity consumption in urban areas will take
place in the Central region and the highest consumption in non-urban areas will take
place in the Northeast region. In terms of percent increase in consumption, non-urban
areas will experience a higher percent change in the consumption than the urban areas in
all regions. The range of percent change will be from 10 to 15.6 percent per year in
urban areas and from 13.2 to 20.4 percent per year in non-urban areas in 2019. The
percent change is expected to be the highest (20.4 percent per year) in the non-urban
areas of the Northeast region and to be the lowest (10 percent per year) in the urban areas

of the South region.

223

Provincial Perspective

The percent increase of the residential electricity consumption in the period 1994 to 2019
among provinces in the same region is found to be smaller than the percent increase of
provinces located in different regions. Within each region, there is at least one province
that shows a remarkably high increase in the consumption. This high increase, especially
in urban areas, will be recognized in Bangkok and some provinces in the Bangkok
Metropolitan Area (BMR) such as Nonthaburi, Samut Prakam, Pathum Thani, and Samut
Sakhon. These are provinces that will experience high population increase as well.
Large electricity consumption will also be found in the center provinces of each region
where population is already high but the percent change of the population in this period
will not be significant. These provinces include Chaingmai in the North region,
Chonburi in the East region, Songkla and Nakhon Si Thammarat in the South, and
Nakhon Ratchasima, Khon Kaen, and Ubon Ratchathani in the Northeast region. For
these provinces, the increase of residential electricity consumption will be more
influenced by an increase of ownership of household appliances than an increase of
population in the provinces. A high percent increase in the consumption will also be
found in some small provinces with relatively low population in the base period but a
large percent increase of the population estimate by the year 2019. These provinces

include Ranong and Phuket in the South region.

224

Projected Per Capita Electricity Consumption and Number of Households

The per capita electricity consumption is expected to increase from 340 kWh per person
per year in 1994 to 969 kWh per person per by the year 2019. The per capita electricity
consumption in Thailand will increase at a decreasing rate in this period. It is estimated
that the percent increase of per capita electricity consumption will decrease from 33.7
percent in the period 1994-1999 to 16.7 percent in the period 2014-2019. Over the next
25 years from 1994, per capita electricity consumption in Bangkok will always be above
the national average and is expected to increase from 1,070 to 1,713 kWh per person. By
areas, the per capita consumption will always be larger in urban than non-urban areas.
The per capita consumption will increase from 571 to 1,279 kWh per person (including
Bangkok) in urban areas and 183 to 703 kWh per person in non-urban areas. The annual
grth rate of electricity consumption per household is found to reduce from 9.5 percent
in 1994 to 5.0 percent in 2019. This increase will be larger when compared to the annual
population increase, which is estimated to decrease from 2.9 to 1.42 percent and the
annual change of number of households (housing), which will decrease fiom 3.2 to 2.9
percent in the same period. This indicates that the higher estimated electricity
consumption in the residential sector in Thailand will be more influenced by the increase
of electric appliance ownership in households rather than the increased number of

households and population.

The findings from this study reveal that provinces in Thailand will undergo changes in

terms of population and electricity demand at different rates during the period 1994 to

225

2019. Provinces with high increases of population will also face a large increase of
household electricity consumption. However, there will be provinces with moderate
population increase that will experience dramatic increases of the consumption within
provinces on a percentage basis. It is necessary that any change in energy development
strategies acknowledge areas with large increase of total and percent increase of

population and electricity consumption estimates.

6.2 Implications for Energy Planning

The electricity model provides estimates of the potential range of energy demand in the
residential sector to the year 2019. The findings provide information for planners to
manage electricity from both supply and demand sides. From the supply side, the low
and high estimates indicate the potential range of electricity demand in the residential
sector that Thailand will have to meet within the next 25 years. In 1996, EGAT
generated 74,460 million kWh of electricity from the mix of fire] consisting of 28%
natural gas, 20% lignite, 9% hydro-power, 26% bunker oil, 4% diesel oil, (with 3%
purchased). Ifunchanged firel mix and production efficiencies are assumed, rough
estimates of the fuel supply required to meet the projected residential electricity demand

can be calculated.

For the low scenario, the estimates for fuel supply to generate 60,000 gigawatt-hour in

2019 includes an increase from the base year 1994 to 2019 of 40,000 to 164,000 million

226

cubic feet of natural gas, 2.7 to 11 million tons of lignite 7,000 to 29,000 million cubic
meters of hydro-power 900 to 3,700 million litres of bunker oil and 170 to 700 million
litres of diesel oil respectively. For the high scenario, to supply 127,000 gigawatt-hour to
Thai households in 2019 will require approximately 351,000 million cubic feet of natural
gas, 23 million tons of lignite, 61,000 cubic meters of hydro-power, 8,000 million liters
of bunker oil, and 1,500 million litres of diesel oil. The increase of estimated required
fuel supply will be approximately 4 times more for each file] type in 2019 from the base
year for the low scenario and 5 times more for the high scenario over the same period.
The fuel supply estimates for the generation of electricity in the residential sector
accounts for only a small percentage of the total energy consumption of the country. The
share of residential electricity in final energy consumption of the country was
approximately 3.3% in 1994 (The share of residential electricity demand in total
electricity consumption of the country was 20.6% and the share of total electricity

consumption in total energy demand of the country was 16.2%; OECD/IEA, 1995).

From the fire] resource perspective, fuel resources are utilized and have to be allocated
for both electricity and non-electricity uses in various sectors such as coal in the
industrial sector, petroleum products in transportation and residential sectors, and gas in
the industrial and residential sectors. The fuel mix of electricity generation in 1994
shows more than 50% of electricity fuel resources utilized to come fi'om nonrenewable
resources (fossil fuels such as natural gas and oil). The reserves of these fiiels in

Thailand are relatively limited. According to OECD/IEA (1997), Thailand has proven

227

reserves of oil estimated to be approximately 206 million barrels or 327,478 million

liters.

Based on the current (1994) oil production rate of 57,000 barrels per day, Thailand will
exhaust the oil reserve within 9 years. For this reason the country must be heavily
dependent on imported oil as the energy (including electricity) resource supply.

If oil will be chosen as a medium and long-term filCl option for electricity generation,
energy planners have to take into consideration concerning the limited domestic resource
as well as the fluctuation of the price of oil in the global market. The option suggested by
OECD/IEA (1997) is to require the new oil fired power plants to be able to use both oil

and natural gas as firels.

For natural gas, the most likely reserves in Thailand were estimated to be somewhere
between 10,962 to 21,100 billion cubic feet (bcf) as of 1993 (OECD/IEA, 1997). With
the current production rate of 1,500 million standard cubic feet per day, it is estimated
that Thailand will exhaust the most likely gas reserves (16,450 bcf - EGAT’s estimate) in
the year 2020. Natural gas seems to be a better alternative fuel than oil in terms of
availability, however, energy planners must be aware that as the production of gas
continues to grow, at the same time there will be an increase of non-electricity use of
natural gas and its products such as direct use of natural gas by industries. This means
that less natural gas will be left as a fuel supply for electricity generation and that much
of it will have to be imported if natural gas is the choice of fitture fuel for electricity

generation.

228

Lignite is a competitive fuel option for Thailand because the estimated domestic reserve
of this fuel is as large as 2,069 million tons. The availability of lignite in the global
market and its relatively stable price makes lignite a more favorable option. However,
the Thai government recently has required the installation of flue gas desulfurisation

(F GD) in all new coal fired power plants. This requirement makes lignite less

competitive compared to other fuel options because of the high cost of FGD.

Another fuel option for Thailand is the renewable energy resource of hydro-power. The
total domestic hydro-power potential in Thailand is approximately 12,734 megawatts
(MW) (OECD/IEA, 1997). Currently, Thailand has utilized only 26.5% of the total
potential. However, the use of the remaining domestic water resources for hydroelectric
generation will be difficult because of the strong protests by concerned citizens and
environmentalists opposed to dam construction. The remaining water resources in
Thailand are also suitable for only medium and small-scale electricity projects. Thailand
could consider obtaining electricity through international sources by trading electricity

and forming joint projects on hydroelectric production along Thailand’s border.

The other firel resource option is nuclear power. Similar to hydro-power, the perception
of nuclear power in Thailand is that of an environmentally damaging energy resource.
The opposition to nuclear power among Thai people caused the Thai government to
announce the indefinite postponement of the two proposed nuclear power plants that had
already been approved by the Cabinet (W ongwaikul, 1994). A study conducted by the

World Bank in 1993 (OECD/IEA, 1997) on the cost of nuclear power in Thailand also

229

concluded that nuclear power is not a competitive option for Thailand when compared to
other firel resources. Because of lack of domestic resources, Thailand has to import
uranium and plutonium, the fuels resource for nuclear power. For the Thai government
to pursue nuclear power as a long-term fuel option, the Thai government has to allocate
significant resources for human resource development and training, site selection,

disposal technologies, as well as public education.

The last option, which can be considered as the other fuel option, is conservation. When
less electricity is used or conserved, less firel will be needed. Activities for the
conservation of electricity in Thailand have not been practiced in great extent because of
lack of public awareness. More information and education has to be provided to public to

create such awareness (more discussion under Demand Side Management)

Decision on future fuel supply and mix depends on the availability of current and fixture
fuel resources, the current capacity of the production, transmission, distribution systems,
potential electricity trade, as well as related laws and regulations. Energy planners must

take into consideration of these conditions when planning for the supply of the fire] in

order to meet with future estimate of the consumption with guarantee of continuity of

supply.

While an option is to continue to meet the demand by increasing the supply through

costly investment in new facilities and firel resources, this option is certainly not the best

long-term option for Thailand, where resources are limited. Another option is to reduce

230

 

electricity consumption. The residential electricity can be reduced through management
fi'om the demand side. Demand side management (DSM) attempts to reduce future
electricity demand in the residential sector by improving the efficiency of electric
appliances by consumers. The results of electricity demand by types of appliance in this
study allows energy planners to assess the impact of electricity consumption from each
appliance and look at the potential for energy conservation by focusing on the electric
appliances that have major shares of the total consumption. A study conducted by the
World Bank (OECD/IEA, 1997) shows high potential for improving electricity end-use
efficiency through DSM in Thailand. Although the DSM program has been in place in
Thailand since 1991, there are still many limitations to the success of the program. The
limitations include the high initial cost of the DSM program, weak public knowledge and
awareness of energy conservation, a lack of energy efficiency baselines (efficiency
testing), and inadequate cooperation from electric appliance manufacturers and
government agencies. There are several policy options available to the Thai government
or energy planners to reduce these barriers. These options are known to work best

together.

The first option is information and educational programs. Programs should aim to
educate the public on energy conservation, energy efficiency and its benefits. More
information relating to energy efficient appliances should be made available to the public.
For example, a guideline for consumers to purchase efficient appliances or a booklet that
contains information on how to save electricity at home could be written. The

educational programs should also be geared toward changing consumer attitudes and

231

behavior to be more efficient in everyday usage of electric appliances such as, a routine
behavior of turning the light off before leaving a room. In Thailand, a few public
education programs that are directed toward encouraging energy conservation have been
implemented. An educational curriculum was developed to educate Thai children on
energy conservation and energy efficiency (IEA, 1997). It is necessary for the Thai
government to provide continuing support for these programs because educational
programs take a long time to become effective. In the past, some programs in this

country failed because of the lack of continuing support from the government.

The second option is regulation. This option is used by many European governments to
mandate appliance efficiency labeling (IEA, 1995). A few European countries also have
mandatory efficiency standards for home appliances. Unlike these countries, Thailand
uses the market-based voluntary approach. The labeling program in Thailand is
voluntary and there is not any efficiency standard for electric appliances in the residential
sector (efficiency standard exists in the commercial and industrial sector). Appliances
that have efficiency labels include refiigerators and air conditioners. Since the program

is voluntary, very few manufacturers have participated in this program.

The third option is for the Thai government to support transfer of energy efficient
technology available in the global market. It is important that cooperation between the
DSM agency (EGAT) and other government agencies be strengthened. An example of a

usefirl cooperation is for the ministry of Finance to provide tax reductions for imported

232

energy efficient technology, equipment, and appliances for the DSM program especially

at the initial stage.

The last option is electricity pricing. In general, higher electricity prices encourage
investments of consumers in energy efficient appliances and discourage electricity-using
activities. On the other hand, lower prices stimulate increased usage of appliances by
consumers and reduce the cost-effectiveness of efficiency investments. Electric utilities
play an important role in supporting the electricity end-use efficiency program. The
decision for utilities to adopt the program will be based on the benefits and costs for
adopting the program. By adopting the program, utilities have to pay for the cost of the
program and lose the revenue as a result of the electricity savings induced by the
program. Nevertheless, utilities will gain long-term savings by not having to generate
and supply more electricity. Based on situations in each country, each government can
use electricity pricing as a market force to improve end-use efficiency. In Thailand,
electricity utilities have historically been publicly controlled monopolies. Privatization in
this business has just begun and is being encouraged by the Thai government. Electricity
pricing in Thailand is usually kept below economically efficient levels. Since electric
utilities will become more privatized in the future, the Thai government could reexamine
their electricity pricing structure in order to see if readjustment of the price could be done
to encourage investment in energy conservation or end-use efficiency programs by future

growing utilities.

233

No matter which options are selected, energy planners will have to make policy to meet
the estimated future demand. To meet the low estimate could require implementation of
policy with restrictions. With no restrictions, fitture demand will likely be in the range of

the base and high estimates.

6.3 Future Research

Future research that will compliment this study is the study of electricity fuel resources or
supply that will meet the estimated future demand for electricity in Thailand. A study
could be conducted by building a model that has the ability to track the flow of electricity
fuel supply as well as project current and potential fuel resources. The flow of the fuel
supply should include both current and future domestic and non-domestic resources. The
model should take into consideration the impact of economic, social, and technological
changes to the demand of the fuel resources. With the addition of a supply model to the
demand model developed in this study, energy planners would have an effective tool for

assessing the multiple energy options that exist for Thailand in the fixture.

The other research that will compliment this is the study of energy consumption in other
sectors by other users. The framework of the residential electricity consumption model
developed in this study can be applied to estimate energy consumption by other resources
(petroleum, natural gas, and coal) by other users (industrial, transportation and

commercial).

234

6.4 Final Remarks

It should be recognized that this study pertains to only one sub-component of the entire
energy planning process. That is, it is an estimation of the demand for one resource
(electricity) by one sector (residential) of one country (Thailand). Energy planners must
consider the supply as well as the demand for Mill—6 resources, used by multifle
sectors, influenced by local, regional and global factors. The development and use of
detailed models of sub-components, such as developed in this study, should provide a
better understanding of each component, and consequently, a better understanding and
management of the entire energy planning process. A better understanding of the energy
and enhanced modeling capabilities should lead to better planning and control, as we

endeavor to effectively manage our natural resources for the future.

235

APPENDICES

APPENDIX A

APPENDIX A

POPULATION MODEL INPUTS AND RESULTS

I. POPULATION MODEL INPUTS

1) Fertility and Mortality Component

Table 11: Birth and Death Rate of Population by Age, Sex, and Area, 1995

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Age Group Urbanfi Birth Rate Male Death Female Death
Rate Rate
04 1 0 2.3 1.7
0-4 0 O 2.3 1.7
5-9 1 O 0.6 0.4
5-9 0 0 0.6 0.4
10-14 1 0 0.7 0.4
10-14 0 0 0.7 0.4
15-19 1 42.8 2.3 0.7
15-19 0 42.8 2.3 0.7
20-24 1 106.4 3.4 0.9
20-24 0 106.4 3.4 0.9
25-29 1 103.9 4.3 1
25-30 0 103.9 4.3 1
30-34 1 66.8 4.5 1.2
30-34 0 66.8 4.5 1.2
35—39 1 30.2 4.8 1.6
35-39 0 30.2 4.8 1.6
40-44 1 10.1 5.4 2.3
40-44 0 10.1 5.4 2.3
45-49 1 2.7 6.7 3.3
45-49 0 2.7 6.7 3.3
50-54 1 0 9 4.9
50-54 0 O 9 4.9
55-59 1 O 11.9 7.1
55-59 0 O 11.9 7.1
60-64 1 O 18.1 11.2
60-64 0 O 18.1 11.2
65—69 1 0 27.7 18.6
65-69 0 0 27.7 18.6
70-74 1 0 63.3 54.5
70-74 0 0 63.3 54.5
75up 1 0 63.3 54.5
75up 0 0 63.3 54.5

 

236

' Urban code '1 " refers to Urban Areas and '0’ refers to Non-Urban Areas"

 

2) Migration Component

Table 12: Migration Distribution of Females and Males in Thailand by Age, 1990

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Age Group Percent Male Migrant Percent Female Migrant
0—4 3.3 4.1
5-9 8.7 10.3

10-14 6.5 7.4
15-19 9.7 16.4
20—24 31.6 22.4
25—29 15.2 14.9
30-34 9.7 8.8
35—39 5.1 5
40-44 3.2 3.3
45-49 2.3 2
50-54 1.7 1.7
55-59 1 1.3
60-64 0.8 0.9
65-69 0.5 0.5
70-74 0.7 1
75up 0

 

 

 

 

(Source: United Nations, 1995 from the National Statistical Office, 1993.)

II. POPULATION MODEL RESULTS

Table 13:Projected Population by Region and Area, 1994 and 2019

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Year - Region Urban Population Non-Urban Total Population
Population
1994-Bangkok 5,066,073 0 5,066,073
2019- Bangkok 10,544,488 0 10,544,488
1994-Central 4,172,771 10,031,973 14,204,744
2019-Central 6,229,005 21,044,780 27,273,785
1994-East 2,151,133 1,176,831 3,327,964
2019-East 3,182,824 2,104,474 5,287,298
1994-North 7,337,456 3,584,606 10,922,062
2019-North 9,505,796 4,697,596 14,203,392
1994-Northeast 16,317,855 3,417,500 19,735,355
2019-Northeast 21,813,646 4,641,338 26,454,984
1994-South 2,915,783 4,276,078 7,191,861
2019-South 4,176,574 6,178,590 10,355,164

 

 

 

 

237

 

Table 14: Projected Population of Thailand by Province, 1994 and 2019

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 é c é C
'53 i 5 5 2° 5 3 5 . 5-
6 = = '- = = *- 2 a 3 E g
at o 3 v o) a on on o. . c o.
8 8 a 8 ‘6 ‘5 ‘6 ‘5 ° 1‘ 5 ° “
1- : F v- N D N N a: n- 0 < n. 0
01 215432 102106 317538 319055 155094 474149 49.3 9.9
02 428873 216232 645105 648172 326874 975046 51.1 10.2
03 766450 157183 923633 972915 204612 1177527 27.5 5.5
04 440705 254336 695041 634365 371390 1005755 44.7 8.9
05 1212983 422815 1635798 1708606 601654 2310260 41.2 8.2
06 301088 109543 410631 422943 268503 691446 68.4 13.7
07 377495 196432 573927 495713 363531 859244 49.7 9.9
08 379055 506692 885747 629454 849340 1478794 67.0 13.4
09 321005 13659 334664 359875 16179 376054 12.4 2.5
10 882686 168731 1051417 1171695 224805 1396500 32.8 6.6
11 272649 112474 385123 371677 153414 525091 36.3 7. 3
12 754970 304173 1059143 887648 359257 1246905 17.7 3.5
13 699451 690581 1390032 906061 910995 1817056 30.7 6.1
14 312618 196348 508966 431772 271080 702852 38.1 7.6
15 115942 75261 191203 176634 114391 291025 52.2 10.4
16 236398 135889 372287 340489 202445 542934 45.8 9.2
17 471368 47396 518764 591590 61128 652718 25.8 5.2
18 71912 146002 217914 104523 216368 320891 47.3 9.5
19 120188 551436 671624 191861 920852 1112713 65.7 13.1
20 512259 128800 641059 686841 173861 860702 34.3 6.9
21 1865052 421417 2286469 2575553 590638 3166191 38.5 7.7
22 561904 847367 1409271 748104 1135263 1883367 33.6 6.7
23 687217 342448 1029665 895939 454803 1350742 31.2 6.2
24 159675 550045 709720 608865 2115127 2723992 283.8 56.8
25 215048 366916 581964 291617 506667 798284 37.2 7.4
26 330989 111376 442365 394395 129726 524121 18.5 3.7
27 1115737 254805 1370542 1476838 344629 1821467 32.9 6.6
28 11 1547 392077 503624 294335 1055561 1349896 168.0 33.6
29 191540 227276 418816 281819 340193 622012 48.5 9.7
30 296597 91334 387931 431080 133434 564514 45.5 9.1
31 158679 390978 549657 228675 575718 804393 46.3 9.3
32 0 5066073 5066073 0 10544488 10544488 108.1 21.6
33 492997 152272 645269 576321 254178 830499 28.7 5.7
34 77665 137043 214708 1 19255 217354 336609 56.8 11.4
35 173691 287742 461433 232886 366410 599296 29.9 6.0
36 382237 152299 534536 481233 189337 670570 25.4 5.1
37 465546 315851 781397 600887 415879 1016766 30.1 6.0
38 161739 244793 406532 211659 319974 531633 30.8 6.2
39 734910 203938 938848 1066853 299611 1366464 45.5 9.1
40 355874 124933 480807 442412 154263 596675 24.1 4.8
41 12660 163770 176430 26157 355774 381931 116.5 23.3
42 677623 206273 883896 860249 259972 1 120221 26.7 5. 3
43 154316 26346 180662 190060 33073 223133 23.5 4.7
44 48910 318493 367403 56261 358455 414716 12.9 2.6
45 1118092 143108 1261200 1402266 179408 1581674 25.4 5.1
46 76447 75137 151584 178759 178282 357041 135.5 27.1
47 219021 183108 402129 327639 353150 680789 69.3 13.9

 

238

 

Table 14 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 c 5 C c c
t: g is" z " - .2 .2

8 s 8 '9 ‘8 g o - g g

5 = 3 "' = 3 " 3 2' 3 3%

9% 33 a»: 3 33 3 3 8.: 5 8.

O. O 1- : ‘- 1- N D N N 32 O. 0 < Q 0
48 333851 379091 712942 468013 531473 999486 40.2 8.0
49 317137 380257 697394 427145 497063 924208 32.5 6.5
50 615852 168059 783911 800589 225016 1025605 30.8 6.2
51 180492 215533 396025 223706 262474 486180 22.8 4.6
52 475130 108904 584034 638178 145062 783240 34.1 6.8
53 1227649 90747 1318396 1579176 121141 1700317 29.0 5.8
54 818778 202272 1021050 1 106704 272531 1379235 35.1 7.0
55 220786 855717 1076503 322515 1281181 1603696 49.0 9.8
56 75112 170199 245311 112291 253447 365738 49.1 9.8
57 9438 851069 860507 21689 2321455 2343144 172.3 34.5
58 88108 96997 185105 109866 126659 236525 27.8 5.6
59 143288 221933 365221 361386 634995 996381 172.8 34.6
60 344765 176596 521361 458185 356874 815059 56. 3 11.3
61 152018 52112 204130 182306 62574 244880 20.0 4.0
62 270678 303499 574177 344416 387319 731735 27.4 5.5
63 504777 264160 768937 643548 338279 981827 27.7 5. 5
64 494182 251788 745970 737550 370451 1108001 48.5 9.7
65 1235681 67572 1303253 1622972 95424 1718396 31.9 6.4
66 700001 126013 826014 1007128 180969 1188097 43.8 8.8
67 219913 49893 269806 279437 65614 345051 27.9 5.6
68 1031838 375955 1407793 1425198 517535 1942733 38.0 7.6
69 426381 34669 461050 533375 49549 582924 26.4 5. 3
70 255188 49922 305110 318664 62444 381108 24.9 5.0
71 1228304 364404 1592708 1663985 492734 2156719 35.4 7.1
72 346252 150754 497006 444704 190015 634719 27. 7 5. 5
73 188809 111314 300123 249327 146225 395552 31.8 6.4
74 461935 14461 476396 699361 22125 721486 51.4 10. 3
75 445528 19791 465319 616443 29010 645453 38.7 7.7
76 343887 0 343887 457982 0 457982 33.2 6.6

 

239

 

Table 15: List of Provinces in Thailand

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Region Code Province Code Province Name

S 01 Krabi

C 02 Kanchanaburi
NE 03 Kalasin

N 04 Kampaeng Phet
NE 05 Khon Kaen

E 06 Chanthaburi

E 07 Chachoengsao
E 08 Chonburi

C 09 Chainat
NE 10 Chaiyaphum

S 1 1 Chumphon

N 12 Chiang Rai

N 13 Chiang Mai

S 14 Trang

E 15 Trad

N 16 Tak
NE 17 Yasothon

C 18 Nakhon Nayok
C 19 Nakhon Pathom
NE 20 Nakhon Phanom
NE 21 Nakhon Ratchasima
S 22 Nakhon Si

N 23 Nakhon Sawan
C 24 Nonthaburi

S 25 Narathiwat

N 26 Nan
NE 27 Buriram

C 28 Pathum Thani

C 29 Prachuap Khilikhan
E 30 Phachinburi

S 31 Pattani

C 32 Bangkok

C 33 Phra Nakhon Si
S 34 Phangnga

S 35 Phatthalung

 

 

 

240

 

Table 13 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Region Code Province Code Province Name
N 36 Phichit
N 37 Phitsanulok
C 38 Phetchabun
N 39 Phetchabun
N 40 Phrae
S 41 Phuket
NE 42 Maha Sarakham
N 43 Mae Hong Son
S 44 Yale
NE 45 Roi Et
S 46 Ranong
E 47 Rayong
C 48 Ratchaburi
C 49 Lopburi
N 50 Lampang
N 51 Lamphun
NE 52 Loei
NE 53 Si Saket
NE 54 Sakon Nakhon
S 55 Songkhla
S 56 Satun
C 57 Samut Prakam
C 58 Samut Songkham
C 59 Samut Sakhon
C 60 Saraburi
C 61 Singburi
N 62 Sukhothai
C 63 Suphanburi
S 64 Surat Thani
NE 65 Sun'n
NE 66 Nong Khai
C 67 Ang Thong
NE 68 Udon Thani
N 69 Uttaradit
N 70 Uthai Thani
NE 71 Ubon Ratchathani
N 72 Phayao
NE 73 Mukdahan
E 74 Sa Kaeo
NE 75 Nong Bua Lamphu
NE 76 Amnat Charoen

 

24]

 

 

 

 

 

 

 

(104 5m9 10014 13019 20024 251029 30034 330$ 411044 451049 50054 580$ Glad 680$ 70074 75Up
MONO")

 

L I2019 I 1994

 

Figure 58: Age Distribution of Projected Population in Thailand, 1994 and 2019

242

 

 

 

 

Forum Population

 

 

 

GM 5m9

10:01.: 151019 20:024 251on 30am amass 40lo44 45w 50to54 55m59 coma 65(069 mum 75up
No in!!!)

INon-Urban lUrbdn

 

Figure 59: Age Distribution of Projected Urban and Non-Urban Population in Thailand,
201 9

243

Pole-m Population

 

120

 

 

 

 

 

om 5m 10am 151019 20mm 25:02:; my astoaa 401044 45649 501054 55w sauna 6&069 7mm 75Up
Noun-I)

 

I2019-Nahum“ I 1994-Non-um-n j

Figure 60: Age Distribution of Projected Non-Urban Population in Thailand,
1994 and 2019

244

 

 

 

 

Percent Population

 

 

 

(104 5(09 10014 13019 20024 251029 3(1034 351039 0044 48049 5M4 551159 m4 651069 7(1074 75Up
MOW!)

 

I 2019—Urban I 1994-Urban

Figure 61: Age Distribution of Projected Urban Population in Thailand, 1994 and 2019

245

APPENDIX B

APPENDIX B

RESIDENTIAL ELECTRICITY DEMAND MODEL INPUT AND RESULTS

1. RESIDENTIAL ELECTRICITY DEMAND MODEL INPUT

Appliance Input Power and Usage Intensity

Table 18: Electric Household Appliance by Types and Usage for Average Scenario

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Types of Appliance Average Electricity Power of Number of Number of

Usage (kwh) appliance (kW) Hours Days
Radio 13 0.035 24 365
Color television 117 0.08 4.0 365
Black and white TV 117 0.08 4.0 365
Video 7.2 0.06 2.3 52
Iron 365 1.00 1.0 365
Rice-cooker 82 82.0 0.45 365
Fan 204 0.07 8.0 365
Air-conditioner 2400 2 6 300
Refrigerator 876 O. 24 365
Computer 47 0.13 1 365

_Light 187 187 kwh'

 

a: 187 kwh was total electricity consumption per household from lighting.

246

 

 

RESIDENTIAL ELECTRICITY DEMAND MODEL RESULTS

National Electricity Consumption in the Residential Sector

Table 19: Projected Residential Electricity Consumption in Thailand, 1994-2019

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Low Scenario— Total Urban Area Non-Urban Area
Yea' Giga kWh % Change Giga kWh % Change Giga kWh % Change
Low-1994 14.4 9.6 4.8
Low-1999 21.0 45.6 13.2 37.2 7.8 62.3
Low-2004 28.6 36.3 17.7 33.9 10.9 40.4
Low-2009 37.7 31.7 23.2 31.0 14.5 33.0
Low-2014 47.8 26.7 29.0 25.3 18.7 28.9
Low-2019 59.5 24.5 35.9 23.6 23.6 25.9
Base Scenario-Year Total Urban Area Non-Urban Area
Giga kWh % Change Giga kWh % Change Giga kWh % Change
Base-1994 18.9 12.8 6.0
Base—1999 27.8 47.7 17.8 38.9 10.0 66.3
Base-2004 38.4 37. 8 24.1 35.2 14.3 42.5
Base-2009 51.0 32.8 31.8 31.9 19.2 34.4
Base-2014 64.8 27.3 39.9 25.6 24.9 29.9
Base-2019 81.0 24.9 49.4 23.8 31.6 26.7
High Scenario- Total Urban Area Non-Urban Area
Year Giga kWh % Change Giga kWh 96 Change Giga kWh % Change
High-1994 26.4 19.0 7.4
High-1999 40.6 54.1 27.3 43.8 13.3 80.6
High-2004 57.7 42.1 37.8 38.5 19.9 49.3
High-2009 78.4 35.7 50.8 34.3 27.6 38.6
High-2014 100.9 28.8 64.3 26.6 36.6 32.8
High-2019 127.1 26.0 79.9 24.4 47.2 28.8

 

 

 

 

 

 

 

247

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mam
mo+w...o B+mvos mo+mvvé oc+mmoN mc+wmc.. 3+wmm.. mo+mo..m oo+wmod 3&on oo+mmmd affirm... o.+w.w.. 806on .xoxmcam
vmm.
mo+m®NN 8+mo..m mo+wmmN S+m3d mo+w.o.v warms...” gamma... mo+wv... sewn—t. wo+mmfi~ ao+msvd ao+wuv.m 09.2.... .xoxmcam
mo+wvom no+mvmb oo+w.o.m wo+w.m.w mo+mood gamma. no+wmmd wo+wmmd B+Mmmd oo+wovN mo+m~mé ac+wmmd mmmdgd 98-525
mo+wvon Rimmed mo+wooN memo... mo+wo..v oo+meN ~o+wom§ mo+m.m.. ~o+mmmw. oc+wmvs oo+wno.m mo+wvmd Rmfimo... vmm.-5=ow
mo+mmod no+mmms mo+wmws 22¢va 8+de ho+wmmé mo+wom.. mo+w3.m Newman Gem—.0.” mo+mood o.+wmm.. 3062.... 93.532
mo+wmm.v 8+w.o.o wo+mn..v 279%.. 9796.... 8+m§~ 3.598 2750.. so+mwafl oo+wmnd oo+m...v ao+wvmd zodmfim 33-5.62
m.o~
8&8. 8&o... 8&8... 8&8...“ 8&m... 8&2. 8&2N 8&va 3&2.» 8&..~.m 8&m..o 2&8. «8.8%. 382.32
com.
3.535 0993.05 mo+wom.m 8.7mm... mo+momd 8+mood 27w”... oo+wooé sew-8d oo+wmsd oo+wo_..n mo+wj56 madam... $005.02
8&8.” no+womN mo+mmnd mo+wvm.. 8.53.0 09mg... B+wwm.v memo»... B+wmné memo“. mo+m=N 27mg... vmm.vmo.. 2.8-58.
8+9}. 8+w~né oo+wom.. 3,65... 8.500.. mo+mom© ~o+wo~N no+msoé 8+wmmd mo+mowd ~o+mvvd mo+womd .mvdg 3m.-.maw
98-55
mo+wmm.m 8+9... mo+wmod mo+wwm.v mo+wvv.. B+mmm.. oo+mom.. 8+de ~o+w8.m ao+mo~.v mo+wvos o.+wmm._. mwminmd 2335.80
32.0.65
mo+w~md ~o+mm... mo+woo.v mo+wom.. mo+wmo.m 836$.” B+mvoh 27mm». 59¢va 27mg... mo+mmu.m mo+wovd mmvftfi 0.3.2.200
m.o~
8&8. 8&8. 8&8. 8&5. 8&8N 8&8.m 8&8... 8&8.a 8&8... 8+8... 2&8. o.&m...n «8.266 4.223
cam.
mc+wmm.w B+wmné oo+wmmd mom—3N mo+mofim oo+wmos mo+w.~... mo+mnmd ~o+wmmd mo+wmo.. ao+mmad 8.4906 m.m.~mm.m 42.50 .
.938 >. 222, a 23...
Ea... 359.50 can. 005 :0: m0) new .805 3.0.00 0.00”. «809.com 0< 9.0500 20:
Aesxvucnsao £88m Ea. .33: 858
2cm 28 v9.8. dame: .3 23:2... E :ochsmcoU gorse—m 23:02QO 880.953 03:
560m 3.528... as. E scans—8.59 36302.". Ecemom AN

248

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.00.00.. 0.000.0 00..0.... 000.0.0 .0.00.0. ..0.00 .000.0. .00.000. 0.0.0. 000..000 0000.000. .000 0.000 0.
000.000. 0.0000. 000000.. .000.00. 00..00.0 ..000.. 0000.0.. 00..000. 000000. 000.0000. 000000.00 0.000 00000. ..
0000000. 00.0.00 .000000. 0.000.0 .0..0.0. .0.000 00000.. .000.0.. 0.00.0 00000000 00000000. 0.0.0 .0000 ..
0..0000 ..000.. ...0000 0.0.00. 0.0000 00.00. .....0 0000000 00000. 0.000.0. 0.0.00.0 0.0. 0.00. 0.
0.08.0. ....000 000.0.00 .00.000. 0000..00 .00.00 80.0. .000000. 00.00.. .0888. .00000.00 ..... 00.00. 0.
000.0.. 000000. 0000.0. .000.00 000.000. 0.00.. 000.... 0.0...0 0.00.0 .0.00.00 .00.00... 0.0. 00.00 0.
00.000. 0.0000 0.0000. .00000 ...00.. 00.00 00..0. 0.00.. ...00 8.88 0000000. 0.00 .00. ..
000..00 0...00. 00..000 000.000 0.0.000 0000.. 00.000 0000000 0000.0 0000.000 000.0000. 0.... 00000 0.
.000.00 00.000 000.0.0 .0000. 0.00000 0000.. 0.0.0. 00.000. 0.000. 000..00. .00.00.0 0.0. .000. 0.
000000. 80.0. ...00.. 0000..0 0000.... 0.000. 00.000 000.000 00...0 000.0.00 0.0.0.00. 0.0.. .0000 0.
0000.... 000000. 0.0000.. 0..0000. 0.000.00 0.000 0.00.. 000..... .00000. 000.0000. 000000.00 ..000 0.000. 0.
0..0000 00.0.0. 0.80.0 .00.000 0000000. 00.000 00000. 0.000.. .00..0 0..00000 00000..0. 0.00. 00000 ..
.000000 000.0.. 0.0.0.0 ..0.0.. 0.80.0 .00.0. 00.0.. 0.00000 0000.. 000..0.. ..000000 0.00. 00.0. ..
00.0000 0.0000. 0000.00 000.000 00..0.0 .00... 000000 00..00 00..00 .0000000 .000.00.. ..00. 0...0 0.
..0.00 ..00.. 0..0.0 00000. 000000 .000. 000.0 0.0000 0000. 0.0.00. 0....00 0.0. 00.. 00
.000.00. .00.0.0 0.00000. 0000.00. 0..0.000 300.0 00..00 000.000. 000000. 00.0.0... 00.0.0.00 0.00. 00.00. 00
0.800.. 0..000. 000..0.. 0000..0 0000000. 000000 0..000. 0.00000 .000.. 000..000 .000000.. 0.000 00000 .0
000000.. 00.000. .0000.. 00000.0 .0000.0. .00.00 .0.00. 0..000. .0008 .00.0000 000.0.00. 0.00. 0.0.0 00
000.00.. .000.00 0.00000. 00000.0 .000..0. 00.0.0 00.0... 000.00.. 000000 .00.00.0 .000000.0 0.000 .0000 00
0.008.. ...000. 0000..0. 0..0..0 0....00. .00.00 ..0000. 00000.0 .0..00 000.00.0 00000000. 0.000 0..00 00
.000.00 0.00.. 0.00000 0.0.0.. 0.00.0. 00000. .080 000.000 00..0. 00.00... 0..0.000 ..00. .0.0. 00
0...000 .0000.. .000000 0.00000 .00000.. .00..0 .00.... .0000.0 ..0000 00.0.80 .00.0.00. 0.00. ..000 .0
0.00000 00000.. 0000000 00..00. 00000.0 00.... 000.0. ...0000 00.00. .00..0.. 0000.0 ..00. 0.0.. .0
09.08 >. 222, .3 88

£0... .2:an0 :00. 00.”. c2. 10> 0:0 .005 2.8.00 0.00.. .oEomCEm 0< umquu 20.. 8:3
235000800 00500.0 .20. .0000: -20

 

 

 

 

 

 

0 . om 0000.750. ..o 009.... B 000..0..... :. 08:380. .3 002.0. :35 :. 00.00.000.50 3.0.0.005 3.5.0.00..— .00.oo_.o.m gm 0.00...

5.383800 36.585 30:38..

.0

249

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00000.0. ...0000 0.8000. 0.0.000 .008... 000.00 0000... .000000. 0.0000 080.08 00000.00. ..0.0 0000. 00
0000000. 0000.00 0000000. 00000.. .000000. 000000 000000. 00080.. 0.0000 .0000000 0.00.00.. 0..00 0000. 00
8000.0. 0.0000. 0000000. .000000 .0.0000. .0800 ....00. .0.0000 0000.. 000...00 ..000000. 0.0.0 0000. .0
0.00.0. 0.00... .000.0. 000.000 .0.000.. ..000. 00..00 0000000 00.0.0 ..000000 00000.00 0.... .0000 00
880.0 0.000.. 000.0.0 0000.0. 0000000 0.000. 0.00.0 .000000 ..0.0. .8..00. 00.0...0 ..00. .00.. 00
00.000. 00.000. 000.00. ..00000 00.000.. 0..0.0 .00000 0.00000 0.0000 000...00 000000... 0.00. .0000 00
00.0.0 .0000. .00000 800:0 000000. 00000 0000. 00000. 00.8 0.00000 .000.00. ..00 ..00 00
0.00.00 .000... 0..008 .000000 88.8. 0.000. 0....0 000.000 000000 00000000 8.88.. 0.00. 0.000 .0
0000000. 0.0000. 00.0000. 00.0000 .00..... 0..000 ..00.0. .0.0000 0.0.0. .00.00.0 00000000. 0...0 .80. .0
00000.0 0.000.. 00000.0 .088. 00..0.0 00000. 000.00 0.00000 000.0. 00000.0. 0008.00 0.00. 000.. 00
.0.0000 ...000. .000000 0.00.00 0000000. .0000. 000000. .00..00. 000000 .0000000 00000..0. 0.0. 00000 00
.00.000 00000.. 0.0000. 0..0..0 0.008.. 00000. 0.0000 ..00.00 00.0.0 000.00.0 0000.0.0. 0.0. 0..00 00
...000.. 000000. ..00.0.. .00.000 0.8.0.. .....0 .00000. .000.00 0000.0 000.0000 0.000000. 0.000 .00.0 .0
0.0.000 .0.000. 0000000 0000.0. 0000000 000.0. 0808 .00.0.0 000... .00000.. 000.00.0. 0.00. 0000. 00
8.8.0 0...00. 0.00000 0000000 0..0..0. 0.0000 0...00 .00000. 000000 0.00.0.0 00.0.00.. 0.00. 00000 00
.00..0. .880. 00000.. 0.00000 0.0008. .0000. 0.0000 000.08 008.0 0..00000 0.880.. 0.00. .0.00 00
.000000 0000... .0.0000 ..0.000 0.0..00. .0..0. ...00.. 8.8.0 0.0.00 .000.000 0000000.. 0.0.. 00000 00
....0..00 0..00... 8.0....0 0000000.. 00.00.00. 000000.. .000008. 0000000.. .000..0. 0.80.0.0. 0.00000000 0.00.. 000.00. .0
0000.00. 0000000 .000000. .00000. 0.0.000. 000.00 00...0 000.00.. .0.000 00000.0. 08.08.. 0.000 000.0 .0
8.00.0 .00.00 00..0.0 .0..00. ..0.0.0 .0000. 000000 0.08.0 0.000. ..0.0.0. ..080.. 0.0.. 00... 00
0.0.000. 000.0.. 0...000 0..0.00 00.080. 0...00 .000.0 ...80. 0.0.00 .0888 0000.0.0. ..0.. .0000 0.
.0.00..0 00..000. 00000000 0.000... 00.00000 00..0.. 0.00.00 00.0.0.0 000000. 00.00000. 00...0000 0.000. 0.0.0. 0.
.00.000 00.00.. 0000000 0.00000 000.00.. 0.00.0 000000 00.0000 00..00 000.0000 0000.0.0. 0.00. 00.00 ..
00..00. .00.00 ...000. 0000.0. 0800.0 0.0... ..0.0. 000000. ..00.. .0.080. 0.0.0.00 0.00. 8.0. 0.
88.0. .0.0000 000..0.. 0.0000. .88... ...000 .00000. 8.0000. 00000. 00..0... 000.0000. 0.000 0000. 0.
.0..0.00 000..00. 000..0.0. 00.8.0 .00.000.. 000000. .000.00. 000.0..0 0.08.0 0..080.0 000000000. 0.0.0. 0.00.0 0.
.968 >.. 222. Mg .08

Em... 550800 can. 00... co: mo> 0:0 .805 >._. 5.00 200.. .28353. 0< 2.0.50 Bo: 85>
23808500 8.00.8.0 .80. .0000: .20

 

 

 

 

 

 

3.0.53 .N 030...

250

6.. 00:32.. E 020 :55 02..

       

 

iiiiiiiiiii 0.
E828 %Efifiamfl'0380 animal!
a Eagagaafiggfiaia
Egaggggaggofiz gala!
Egg??? 0...00 %Eggggg
Egaggaggaggaaa

 

E00000. Eggggsfifl %%
Eggggggggg 00.0....
BEE—Egg 00000.0. Eggggassa.
Eaggaeos Egg?

0000.... agagaagagfis. 000...00
Egaaglaggggsésm
E 5.000. Egan-ERIE- .00.0.0 gamma... .0.
Eggaggsz EEEESEE:
€300.88. 5.0000 Egg???
30.000. @2000. %figgggflssa
a??? gaggaegsa.

vow

            

      

(DO)?
Vl‘
OO
00
(D

N O
(O N

O
0

IO

888888

N N
LO L0

..00.
a0
a
..0

0

ommv
I
a. 00..
m. «a:
N

 

II?

0
Egagagggggfigss E
%figggggéga E00000... 0000080 fli
%E..00.000.§E.8000.§%§ . 00.00000..E
Eggagaaaaaefifi. a
a

gagggaflggggaség
aggggggggg
agggggggggggfim Hi

%figgggaass. mania“. mun—'00...
Eggaagggggészz a
Eggggagagg 3i

8|

,—
In

              

 

     

E
a

 

        

   

    
  
    

   

  

 

    

s>>x
Em... ESQEoo :0“. 02m 00: 10> >0 8.00 0:00.. 5.0.09.8... 0< 20.. 85>
-0000: -20.

05>... 000.50 3.00.8.0

 

.0008. _. 0.00.

251

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.00.000 000.00 00000.0 ..00.00 0.0..0. 00000. 00000. 0..0000 ...80 _080.000 _000008. 0.8. 00.8 8
00000.0. .08... 0.0.8.. ..00.... .0...000 000000 0000000 000.00.. 0.0.... _00000000. 000000... ..000 .800. 0.
80000.. 0000000 .000..0. 008.8. 000.000. 0.00.0 8.0000 0000000. 0...00. 0.000000. .0.0000... ..08 .000.. 0.
.000000. 0.0000. 00.000.. ..00000. ..0000.. 00.000 000000. 000.00.? 00...0 00...000. _0.00.80. ..000 0.0.0 ..
000.00.0 .0.0000 08000.0 .00.00.0 00.00000 000000. 0000.... ..000.00 00000.. 000....00 ...00.000 0..00. ..0000 ..
000.000. 00..0.. 0000000 0.00000. 00000.0 ..0.00 00.0... 0000000. 0.0000 0..0.000. 00000.8. 88 8.00 0.
0.0.0.0 .00000 ..00.00 000000. 00..0.0 0000.. 0.0000 00..000 000000 00.00000 0000.0.0 ..... 0000. 0.
000..00 0.00.0 00..00. 0000... 000.000 0.000 0.0.0. 0.0000. 00000. 00000... 0000080 0.00 00... 0.
0.000... .000.0. 0....00 00.0..0. 00.0..0 0000.0 0.00.0. 0000.!00. 0.000. 0.000000 ..0.0000. ..0.0 00000 ..
0000000 .0000.. 0000.8 .0880 0.0.0.. .0.00. .0..00 800.0 .00.00 00.00..0 000.0000 0.00. .0000 0.
0.00000 .00000 0000.00 2.8.. .000000 0000.. .0800 .0..000 00000. 0.000.00 _0.00.000 ..... .00.. 0.
00000.0 00.000. .0880 .00...0 .0..0.0 ..0... ..0000. ..0080 .880 0.0008. ...0.000. 0.0.. 00000 0.
.00..00. 000800 0000000. 00008.. .0.00.0. .0.000 000000. 0.000.0. 0..000. .0000000. 0...0000. ..000 .00.0. 0.
0000000. 0.00.00 0000.00. 0.000.0. 0000000. 0000.0 ..0...0 00..000. .000... 0000.80. .00..0.00. 0.000 .0..0. ..
0000.00 ..0000. 0.000.0 0008.0 0.00000 00000. 00.0... 0..0.00 .000.0 08.80. 80000.0. 0.... ..000 ..
000.00.. 000...0 .0..0... 80.0.0. 000.000. .0800 ..00000 000..00. 0000.0. 0.000000. ..0.0...0 0.000 00000. 0.
00000.. 000000. 0000000 ..0...0 000..00 00000. .00.... 0000.00 000000 .000000. 0.0000... 0.... 0..00 00
.0..000. .0..0.. 000..00. ..00.0.. 0000000. .0.000 0000000 0000000. 00.0.0. 000.000.. 8880.. 0.8 00000. 00
00.00... 0.00... 00..0.0. 00000.0 00.8.0. 00000. 0.0.00. 000..... 0.00.. .000..00 000..000. 0.00. 000.0 .0
00...... 000..0. 0.00.00 0000000 .00..00. 0.000. .00..0. ..0.0... .0.000 0000...0 00.80.. 0.00. 000.0 8
0.80:0 00.0.00 .0.000.0 0..0.00. 00.0008 000..0 0..0000 00.0...0 .88.. 00000..0. 0.000.000 0.000 000.0. 00
00.8.0. 00...0. 00800.. .0.00... 0000.2. 0000.0 00000.0 00000.0. .000.0 000000... 0.8.8.. 0...0 ...00 00
.00.000. .0808 0000.00. ..00000. 0.00.00. .00000 00.0000 .088. .00000. 00000000. 0.00.000. 0.000 000... 00
..0..0.. 000.00. 0.0..00. 000..0.. .00.000. 000..0 .0..0.. .00000.. 000000. 0000000.. 000.0000. 008 00000 .0
0...... 000.00. 00..000 8.00.0 0000000 0000.. 000000. 80.8. .00000 .00..0.0 00....00 0.... 00.00 .0
.00.000 >0 22>) $3.... 0000
.00.. 8.8.80 00”. 8.0 02. 00> 000 ..0.0.0 >228 0.00.0 3.200.000 0.0 unison 22. 85>
23.005800 0058.0 .20. .008: .20

0.0.

0000.390. .0 000.: .3 000:2: E 000030... ..0 000:. 0.0.3-002 E 00..0...—.0..00 b.0502...— _0_.=00_0om 00.00.00... ”mm 030.0

252

fin 000320 0. 020 035.000 02:.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.00.00. 0000..0 0000.... .00..00. 0.0..0.. 000.00 0...000 .00000.. 00.00.. .000..0F00....00. 0.000 00.0.. 00
00000... 000..0. 0000.00. 0000000 ..0.000. 000000 .00000. .0000... 0.0... .0000000 00000.00. 0.0.. 00000 00
.000.... 000000. 000000.. .0.0..0 .0.00.0. 000.0. 0000.0. 000000.. 0pm.. .0000000 000.0000. 0.00. .0000 00
0000..0 .0..... 000.000 000.000 0.0.00.. 0000.. .0.000. 00...00 0.0000 0.0000.0 .00..000 0.... 000.0 .0
00.0000 00..00 0.000.0 .000000 00000.0 0..00. .00.00 8000.0. .0.000 ..000000 0..0.0.0 ..0.. .00.0 00
.0.0000. 0..0.00 000.00.. 00....0. 00000... 000.0. 00000.0 0000.0.0 000.00. 000..0... ..0.00000 0.0.. 0.000. 00
.00000 .00.0. 000000 0.00.0 000000. ...00 .0.... 00.00.. 00000 0.00..0 0...000. 0.0. 0000 00
.....00 000.00 00.0.0. 0000000 000000. 00000. 0000. 0.00000 00000. 000.0.00 00.000.0 ...0. ..0..! 00
00....0. 0000.0. .0000.0. .00.0.0. 00000.0. 0..000 .00...0 00...00. 0000.0. 000.0000. .0000.0.. 0.0.0 000.0. .0
0.0..0 .00.0 ...000 ..00.0 000.00. 00.0. .0.00 000.00 00000 ..000.0 000000. 0... .000 .0
0000000. ....... ...00.0 .0.00.0 .00000.. 0.000. 0..0.0. 000..0.. 000..0 .0.000.0 00008.0. 0.0. .0000 00
00.0000. .000000 0.00000. 00...... 0000000. 000000 0000000 0000.00. 00.0.0. 00000000. 0000.0.0. ..000 0.000. 00
.000000 00000. 0000000 00.0.0. 0.00000 ..0.0. 00.000 0.000.0 00.0.0 80000.0 00000000 0.... 000.. 00
.000000. 0.0000. 00.000.. .000..0. 00..0.0. 000000 00..00. .000000. .0.000 0000.00.. 0..0..00. 0.0.0 0.000 .0
0000.0: 0.0000. ...0000. 0...00. .0..0... 0.0000 000.0.. 0.0000.. 00..00 00008.0 0.0.0.00. 0..00 .0000 00
000.000 00.0.0 .0.00.0 00.0000 ....000 000... 0000... .0.0000 .0000. ..0..0.0 0.0000.0 0.00. 00000 00
......0 00.0.0 0.000.. 0...00. 000.0.0 0000. .0.00. 0.00000 .0000. 0000.0.. 000.0..0 0.00 0.00. 00
00.0.0: ...000. 00.000.. .0000.0 0000000.. 0.0000 ..0000. 0.00000. 0.000. 00000000. 00.00..0. 0.000 00000 00
0 0 0 0 _0 0 0 0 0 0 _0 0.0 0 :.0
....000 .0..0. 0000000 0000000 _0.0...0 00.... .0.000 00.0.00 ...000 ...0.000 00000000 0.... .000. .0
0.000... 000.00. ..00000 00000.. 000.00.. 0.000. 000000. .0.0.00. .00.0. 00.00000 0..0000.. ..00. 00.00 00
0.0.0.. 0...00. 00.0..0 .000000 00.0.00 00000. .0.000 00..00. 00.000 00.0.000 00080.0 ..... .0000 0.
.0..00.. 000000. ..000... 0...000 0000000. 0.00.. 000000. 0000000 .0.000 00000000 0..000.0 ..0.. 00000 0.
00000000 000...0 00000... 0.0000.. 0000000. 00..0. .00.000 .00.00.0 00.000. 0..0.0.0. 00.000000 0.0. 00.0.. ..
000.0.0 00...0. 000.000 00.000. 0..0000 0.0.0. 000000 0.00..0. 00.0.0 0.00.00. 0000.00.. 0.0.. 00000 0.
.00.000 >0. of; «“30” 0000

20: 02000.00 :00 8.0 02. 00> 000 ..00.0 53.8 0.000 3.200300 0< 0900.00 0.00 85>
235000500 0058.0 .000. .0000: .20

 

 

00.0080 .. 0.00.

253

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

000..00 0.0000. 00.000 ..00.0. .000000 F008. ...000. $010.3 00..000 0000000.! 03.1.02. ..0.. .0000 0.
.00...0. 000000. 00.0000 .0.00.0 0..000.0. .0.000 .0000.. 000000.. .0000. 0000...0 00.0.0.0. 0.000 0000. 0.
..0.000. 0..000. 0000000. 0.0.00.. 0...0.0. 0.0000 00..000 00000.|.. 00000.. 00008.0. 0..00..0. ..000 000... 0.
00.0000 ..000. 00000.. .000.00 ...0..0 .0.0.. 000000 0..00.0 0.000. 0000.R0 .0.00000 0.0.. 00.0. 0.
00000... 0.000.. ...0000 000.0.0 .000..0. 0.000. .00000. 000.00...0....0 0...0000 .00.0000. 0.00. 00.00 ..
00000.00 0.00..0 0.0000.. ...0000. 0.0.000. 0.00.0 00..00. 0...0000 00..00. .0.0...0. .0.0.0.00 0.000 00000. ..
000000. 00000.. 0000000 00.0.00 0000000 0..00. 000.0.. .0.0..0 000.00 00000000 00.0..00 0.00. 0.0.0 0.
0..000.. 000000. 0.000.0. 000.0.0 00000... .00000 0...0.. 0..0.00. 00000. 00000000. 0.0..0.0. 0.000 .0... 00
00.00.00 .000.00 000000.. .00.00.. 000..00. 00000. ..00.0 .000000. 000.0.. 000.0.0.. ..0000.00 0.00. 00.0.. 00
0.00000 0.0.... .0..0.. 00.0.00 .0000.0. 000.0. 000000. 00000.. 000.00 0000.000 000.0.00 0.0. 00.00 .0
000000.. .0000.0 0000..0. 0.0.000. 0000000. ..0000 ..0..00 0000.0.. 000000. 000.00.0. 00000.00. 0.000 .000.. 00
00000000 ..0..00 00000.0. .0..000. .000..0. 000000 0.0.00 000.0000 000.00. 05.00000 00000.0.0 0.0.0 .00.0. 00
00.0.00. .0..00. ..0...0. 00.0.0: 0..0.0.. 000000 000.000 0000000. 00000.. 0000020. 0.000.00. 0.0.0 0..00. 00
0..0000. ...000. 00000.0. .00.0.0 .0.00.0 ...000 0..0.0. 0000.00. 000000 00.000... 000.00.0. ..0.0 00..0 00
00000.0 0000.0. 00.0.00 0.0.000 0000000 00.... .00..0. 0.00000 0.0000 .00000.0 .000.0.0. 0.00. 00000 .0
.0..0.0 .00.0. .0..000 ..000.. .00.000 00.0.. 000.00 000.000 00.00. .0.00000 00.00000 ..0.. 0000. .0
000.00.. 00.00.. .0..00.. .00.000 000.000. 0...00 0.0.00. 000.0.0. 000000 00.0.000 0.000..0. 0.00. 0.000 00
0.00000. 00.00.. 000.000. 0000000 000000.. 00.00. .0.000. .0.0000 3E0 ..00000. 0000008. .0..0. 00.0. 00
000..0. .0.000 000000. 00.00.. .000000 .0000 00.000 000.00. .0000. 00.0000. 000.0000 ..00 ..00. 00
00000. ..... 00..00 00...0 00.000. ..00. .0...0 00.0.0 00000 0.00000 00.0000 ..0. ..00 .0
000000. .00.00 000000. 000000. 0.00.0. 00.00 .0.000 000000. 00000. 0000000. .0..0... 0.0 ..00. 00
0000000 0.0000. 00000.0 00..000 00.000. 00.00. ..0000. 000..0. 0....0 0000...0 00.0..00 .0.00. 00000 00
0000.00. .000000 00.0000. 0.000... 0.0..00. .00000 .0....0 0000000. 000000. 0000.00.. 000000... 0..00 00000. 00
00..00.0 00.00.0 ..00.00. 000.000. 0000000. ..000. 00...00 .000..00 0.0.00. 00.00000. 00.000000 ...0 000.0. 00
00000.0. 00..... 0.00000 0000000. 0.0..00. 0.0000 .0.000. 000.000. .00.00 00000000. .000.000. 0..00 .000. .0
00.0000 0...00 00000.0 0000.00 000.0.0 00000. 00.0.0 .000000 0000.0 .0.00000 00.0.00. ..00. 000.0 .0
03000 >. 2005 220. 030 0000

29.. 0059000 cm... 00.1 co: m0> Uco 0.00-m >._.0o_00 050m Engages“ O< ucupmmo v.0... ooE>
235000500 0058.0 .30. .0000: .20

 

 

 

 

 

 

6.0.80 .. 0.00,.

254

BIBLIOGRAPHY

 

BIBLIOGRAPHY

Action, J ., Mitchell, B., and Mowill, R. 1976. Residentgl Demand for Electricity in Los
A_ngeles: An Economietric Study of Disaggregated Data. Prepared Under a Grant
from The National Science Foundation. The Rand Corporation, R-1899-NSF.

Anserson, SK. 1973. A MethodologLfor Projectirgthe ElectricaliEnergy Demand in
California: a User’s Guide to the Model. Prepared for the Resources Agency of
California with support from the National Science Foundation. The Rand
Corporation, R-l lO7-NSF/CSRA.

Archavakel, K. 1980. Migration and Urbanisation in Thailand, 1988: The UrbakRural
continuum Analysis. IPSR Publication Number 112.

Asian Development Bank. 1994. Energy Indicators of Developing Member Countries of
Asian Development Bank, Manila: Energy and Industry Department, Asian
Development Bank.

Asian Development Bank. 1994. Energy End Use, An Environmentally Sound
Develognent Pathway. Manila: Asian Development Bank.

Asian Development Bank. 1993. Electric Utilities D§t_a Bool_< For The Asian and Pacific
Region. Energy and Industry Department, Asian Development Bank. Manila:
January 1993.

Barnes, D. F. 1988. Electric Power for Rural flowth. Colorado: Westview Press.

Bemdt, E. R and Samaniego, R. 1984. Residential Electricity Demand in Mexico: A

Model Distinguishing Access from Consumption. L_and Economics. 60 (August):
268-276.

Commission of the European Communities. 1992. Electricity System Perform_ance:

Options and Opportunities for Developing Countries. Synthesis Report of the
Cooperative Programme on energy and Development. (EUR ‘14577 EN).

255

 

Committee on Population and Demography. 1988. Fertility and Mortality Changes in
Thailand. 195—1975 Report No. 2. Washington, DC. Office of Publication
National Academy of Science.

Davis, C. H. 1995. Demographic Projection Techniques for Regions and Smaller Areas:
A Primer. Vancouver: UBC Press.

Debavalya, N. 1980. Decline in Fertility by District in Thailand: An Analysis of The
1980 Census. Asian Population Studies No. 62-A. Bangkok: Economic and
Social Commission for Asia and The Pacific.

Department of International Economic and Social Affairs. 1992. Preparing Miggation
Data for Subnational Population Projections. New York: The United Nations
Publication.

Department of International Economic and Social Affairs. 1991. Integgating
Development amd Popul_ation Planning in Thailand. New York: The United
Nations Publication (ST/ESA/SER.R/110)..

Division of Heath Statistics Bureau of Health Policy and Plan Office of The Permanent
Secretary. 1994. Public Health Statistics 1994. ISSN 08570-3093Bangkok:
Government Publishing Office.

Division of Heath Statistics Bureau of Health Policy and Plan Office of The Permanent
Secretary. 1994. Important Public Health Statistics 1994-1995. ISSN 08570-
3093Bangkok: Government Publishing Office.

Office of the Assistant Secretary for Policy Development and Research, Department of
Housing and Urban Development. 1972. Residential Energy Consumption -
Phase 1, Report. Report No. HUD-HA1 -— 1, Contract No. H - 1654. March
1972. Hittman Associates, Inc.

Economic and Social Commission for Asia and the Pacific. 1991. Sectoral Energy

Demand Studies: Application of the End-Use Approach to Asia Countries:
Energy Resources Development Series, No. 33. (ST/ESCAP/977.

 

Efficiency and Alternative Energy Branch. 1994. Statistical Report, 1993 Survey of
Household Energy Use, National Results. Ontario: Energy Publication.

Electric Generation Authority of Thailand. 1996. Annual Report.
Energy Information Administration, Department of Energy. 1983. _R_es_identi_al Energy

Consumption Survey: Reggession Amlysis of Energy Consumption by End Use.
October 1983. Washington, DC: US. Government Printing Office.

 

256

Imran, M. and Barnes, P. 1991. Energy Demand in the Developing Countries: Prospects
for the Future. World Bank Staff Commodity Working Paper, Number 23.

Washington, DC. The International Bank for reconstruction and Development.
The World Bank.

 

Inhaber, H. 1997. Why Energv Conservation Fails. Westport, Connecticut: Quorum
Books.

International Energy Agency. 1997. Asia electricity Study. Paris: Head Publications
Service, OECD.

International Energy Agency. 1995. Energy Statisticsfiand Balances of Non-OECD
Countries. 1992-1993. Paris: Head Publications Service, OECD.

 

International Energy Agency, et a1. 1992. Proceedings from International Conference on
Energy Efficiency in Asian Countries. Tokyo, Japan, 4"‘-5th November, 1992.

International Energy Agency. 1981. Energy Conservation: The Role of Demand
Management in the 1980’s. Paris: OECD Publication Service.

Knodel, J. Chamratrithirong, A., and Debavalya, N. 1987. Thailand’s Reproductive
Revolution: Rapid Fertility Decline in a Third-World Settmg. Madison: The
University Of Wisconsin Press.

Knodel, J. and Prachuabmoh, V. 1973. The Fertility of Thai Women. Bangkok:
Institute of Population Studies of Chulalongkom University.

Koomey, G. J. et al. 1995. Residential sector End-Use forecasting with EPRI_REEPS
2.1: Summary Input Assumptions and Results. Energy Analysis Program,
Energy and Environment Division, Ernest Orlando Lawrence Berkeley National
Laboratory, University of California. December 1995.

Koopsup, P. 1993. Energy Policy in The Tlgi Economy in Transition, ed. Warr, P. G.
296-324. Cambridge: University Press.

Kleinnpeter, M. 1995. Energy Planning and Poligy, ed. United Nation Economic
(UNESCO) Energy Engineering Series. West Sussex, England, John Wiley &
Sons Ltd.

Lyman, R. H. 1994. Philippine Electric Demand and Equivalence Scales. Southern
Economic Jourm 60 (January): 596-611.

257

Mason , A. and Campbell, B. O. 1993. Demographic Change and the Thai Economy; An
Overview in The Economic Impagt of Demcrgraphic Change in Thailand, 1980-
2015: An Application of the HOMES Household Forecasting Model, ed.
Campbell, B. 0., Mason, A., and Pernia, E. M., 1-52. Honolulu. East-West
Center Published in Cooperation with the Asian Development Bank.

Mason, A., Phananiramai, and Poapongsakorn, N. Households and Their Characteristics,
An Overview in The Economic Impact of Demographic Change in Thailand,
1980-2015: An Application of the HOMES H_ousehold Forecasting Model, ed.
Campbell, B. 0., Mason, A., and Pernia, E. M., 1-52. Honolulu. East-West
Center Published in Cooperation with the Asian Development Bank.

 

Munasinghe, M. and Meier, P. 1993. Energy Policy and Analysis and Modeling
Cambridge: Cambridge University Press.

Munasinghe, M. 1987. Rural Electrification for Development Policy Analysis. Westview
Special Studies in Natural Resources and Energy management. Colorado:
Westview Press.

Nelson, BC. 1965. A Study of the Elasticity of Demand for Electricity by Residential
Consumers: Sample Markets in Nebraska. Land Economics (February): 92-96.

Newell, C. 1988. Methods and Models in Demography. London: Belhaven Press.

Organization for Economic Co-operation and Development (OECD). 1994. Energy in
DevelopingLountries: A Sectorfi Analysis. Paris: OECD Publication Service.

Porapakkham, Y. 1986. Mortality and Health I§§ues: Levels and Trends of Mmlity in
Thailand. Asian Population Studies Series No. 77. Bangkok: Economic and
Social Commission For Asia And The Pacific.

Prachuabmoh, V. et al. 1972. The Rural and Urban Populations of Thailand:
Comparative Profiles. Research Report No. 8. Bangkok: Institute of Population
Studies of Chulalongkorn University.

Provincial Electric Authority 1996. Annual Report.

Rimmer, P. J. Urbanization Problems in Thailand’s Industrialization and Its
Consequences, ed. Krongkaew, M, 267-371. New York: St. Martin’s Press.

Ross, H. and Thadaniti, S. 1995. The Environmental Costs of Industrialization in
Thailand’s Industrialization and Its Consequences, ed. Krongkaew, M, 267-371.
New York: St. Martin’s Press.

Saunders, J. 1988. Basic Demographic Measure. MD. University Press of America.

258

Taylor, L. D. 1975. The Demand for Electricity: A Survey. The Bell Journal of
Economics. 6 (Spring): 74-108.

The United Nation. 1995. Global Population Policy: Database 1995. New York: United
Nations Publications. (ST/ESA/SER.R/l45).

 

The United Nations. 1993. World Populaktion Prospect; The 1992 Revision. New York:
United Nations Publications. (ST/ESA/SER.A/135).

Westley, G. D. 1989. Nontraditional Partial Adjustment Models and Their Use in
Estimating the Residential Demand for Electricity in Costa Rica. Lari!
Economics. 65 (August): 254-271.

Wilson, J. W. 1971. Residential Demand for Electricity. Qu_arterly Review of Economic
and Business. 11 (Spring): 7-22.

Wongkul, P., ed. 1994. Environmental Situation in Thailand. Bangkok: Amarin
Printing and Publishing.

Yamamoto, T. 1996. R&D Options for Demand Side Management in Japanese Electric
Utilities in Conference Proceedings, New Electricity 2]: Designing a Sustainable
Electric System for the Twenty-First CenttLv. 22"‘I-24‘h May 1995. Paris: Head
Publications Service, OECD.

259