u... m iNPOPULm 0N NTERACTIONS E MD ECONOMI LOPMENI CDEVE IN TAKWAN m .fiw .Mry.~..dr... . x fir.’ 7 . aw... ., . . 2?. P mu _. .tnww.” LN V 1m...“ 6.7 0.... Q . m m u S c j . e In \ h TIM ,m This is to certify that the thesis entitled LIED Ah:L MiChiganStace niversiry ‘1 / I INTERACTIONS BETWEEN POPULATION GROWTH ANN ECONOMIC DEVELOPMENT IN TAIWAN presented by Paul Ke-chih Liu has been accepted towards fulfillment of the requirements for Ph. D. _ Economics degree 1n Date April 3, 1973 0-7 639 \ 5' amounts av 1“ HUAB & SUNS' I BUUK BINDERY WC. LIBRARY BINDERS . mmmm. moms“ lllllllll(”Ill"Ill!Hill"llllllllIlllllllfllllfllll 3 1293 10249 8031 WW % l 1 .fl“)? » mag ABSTRACT INTERACTIONS BETWEEN POPULATION GROWTH AND — ECONCMIC DEVELOPMENT IN TAIWAN BY Paul Ke-chih Liu This study attempts a theoretical and empirical analysis of the interactions between population growth and economic development. It sets forth an economic-demographic two-sector growth model with special emphasis on the characteristics pertinent to developing countries. The main conclusions that emerge from an application of the model to Taiwan data for the period from 1951 to 1970, together with the findings of a historical analysis of the population and economy back to 1905, are as follows: 1. The relationships between pOpulation growth and economic development in Taiwan have been significantly affected by the institu- tional factors of international exchange relations, income distribution policies, agrarian systems, and traditional Chinese values. 2. Given the sociocultural and institutional milieu, inter- actions between population growth and economic development are found to be extensive. The salient results of the multivariate analysis are: (a) Fertility for women over thirty years of age bears a strong positive association with the proportion of the labor force engaged in agriculture and the ratio of the agricultural to the nonagricultural wage, but fertility for women below age thirty did not respond to these factors as much. . ln‘ u. I u (b) There is a negative correlation between mortality and spending on health care for both sexes at all ages. Per capita income, however, is negatively correlated with mortality only of younger (0-15 years old) and older persons (50 and over) but not with the adult population. (c) Movement of the labor force from the rural sector is strongly and positively associated with shifts in the pattern of demand from agricultural to nonagricultural products. Such movement relates nega- tively to the urban-rural wage differential because industrial employers hired fewer workers in response, but the strength of the relation is comparatively weak. (d) An increase in the size of the population during the period from 1951 to 1970 seemed to increase demand for food, education, health care, and shelter, but to lower demand for services and clothing. (e) Ceteris paribus, population growth creates much disguised un- employment which depresses the agricultural wage rate more than that of nonagriculture. (f) An interlocking relationship may be traced from the above findings: as the level and pattern of aggregate demand change, the extent and structure of economic activities adjust accordingly. These induced changes influence the rate of population growth, mainly through fertility, which in turn affects the level and pattern of aggregate demand. (3) During the period from 1951 to 1970 some of the economic and sociocultural institutions like land reform.regulations, nature of the family planning program, services for the elderly, and certain traditional values that might have been conducive to economic develop- ment and population.modernization probably would have been more 1.“. nan effective if their character and timing had been different. In conse- quence, interactions between population growth and economic development were confined to certain limited ranges. The conclusions of this study are useful for the understanding of the conditions and mechanisms of changes in the population and the economy. It is necessary to emphasize, however, that the principal findings - the numerical parameters of the multivariate analysis -- are strictly applicable to Taiwan in the postwar period only because they are socioculturally and institutionally bound and hence do not necessarily measure demographic and economic relations alone. The general validity of these parameters should be determined by an ex- tensive comparative study of other populations in different institutional environments. INTERACTIONS BETWEEN POPULATION GROWTH AND ECONOMIC DEVELOPMENT IN TAIWAN By Paul Ke-chih Liu A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1973 To Professor Ronald Freedman and Professor Deborah 8. Freedman l I 'leu |:‘. D'n“ ....5 ’ 0 ho. ‘_ . at: F. 8d. ‘\I;’l ACKNOWLEDGMENTS This work is an outgrowth of my long and difficult graduate ”career. I want to express my deepest gratitude to all the persons who inspired, encouraged, guided, and supported me through the process and made possible its completion. 1 I am indebted to Professor warren Samuels, the co-chairman of my dissertation committee, for his help and guidance. I am also grate- ful to Professor Ronald Freedman of The University of Michigan who served as a co-chairman, for his generous assistance and sympathetic understanding of both professional and personal problems. Professor W. Paul Strassmann has made invaluable comments and suggestions at various stages of the preparation of this study. I am sincerely ap- preciative and very thankful for the care and patience in reading and discussing drafts by Professor Ronald Lee of The University of Michigan. Professor Simon Kuznets of Harvard University has kindly read Chapter III and provided invaluable suggestions. Professor Mo—huan Being, my director at the Institute of Eco- nomics, Academia Sinica in Taiwan, who instilled my interest in economics and population studies, has been a source of encouragement throughout my undergraduate and graduate training. The Population Council and the Population Studies Center of The University of Michigan have financially supported my graduate education through fellowships, for which I am grateful. The Population Studies Center also provided me with office space and related facilities for 11 ' D I' A. I 0 he... no h... J \.,.3 preparing this study. Many useful comments and various kinds of help were given by numerous other individuals, including Dr. Deborah Freedman, Dr. Eva Mueller, Dr. Albert Hermalin, Mr. Maurice MacDonald and Dr. & Mrs. T. Y. Na at The University of Michigan; Dr. Anthony Y. C. Koo of Michi- gan State University; and Dr. T. 8. Yo, Mr. T. L. Tsay, and Mr. S. M. Chen of the Institute of Economics,Academia Sinica. I am grateful to Mrs. Jane.Gleason for her careful editing and to Mrs. Carol Crawford for typing the thesis. Lastly, I would like to express my lasting appreciation to my family; Madeleine Siyu, my wife, has worked to help support the family and shoulder the responsibilities in caring for our parents, parents- in-law, and children during my long period of study abroad. I am, of course, solely responsible for any omissions and errors in the dissertation. iii v."' o n '..L'. L- et. a. l o .u.. 0.01 . n..- 'v «Q... . ‘v . . s‘. TABLE OF CONTENTS ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . Chapter I O ImODUC T I ON 0 O O O O O O O O O O I O O O O O O O 1. Objectives of the Study . . . . . . . . . . . 2. The Problem and the Conceptual Approach . . . . A Brief Review of the Literature . . . . . . . . Sources of Data. . . . . . . . . . . . . . . . . Plan of Presentation . . . . . . . . . . . . . U'I#U II. THEORETICAL FRAMEWORK OF AN ECONOMIC-DEMOGRAPHIC HwEL . O O O C I O O O O O O O I O I I O O O O C 1. Introduction . . . . . . . . . . . . . . . . . 2. Resource Allocation between Sectors . . . . . 3. The Relation between Population, Employment, and Economic Development . . . . . . . . . . . a. Population, Labor Supply, and Employment . b. Saving and Capital Formation . . . . . . . c. The Workings of the Short-Run Market Equilibrium. . . . . . . . . . . . . . . . 4. The Interactions between Population Growth and Economic Development . . . . . . . . . . . 5. Concluding Remarks . . . . . . . . . . . . . . III. ECONMIC DEVELOPMENT AND POPULATION IN TAIWAN SINCE 1895: AN WERV IEW O O O O O O O O O O O O O 1. The Colonial Period: 1895 - 1945. . . . . . . 2. The Developmental Period: 1945 - 1970 . . . . a. General Background . . . . . . . . . . . . b. Aggregate Demand . . . . . . . . . . . . . c. Economic Structure and Population Dynamics 3 O smry. O O O O O O O I O O O O O I O O C O O iv Page ii vi viii \OONUH H PJP' 23 24 31 31 46 57 59 65 70 72 93 93 95 103 113 1 _. IV. THE ECONOMIC-DEMMRAPHIC MODEL APPLIED T0 TAIWAN . 1. Intrwmtion O O O O I O O O O C O O O O O O O 2. Numerical Estimates of the Model . . . . . . . 3. Discussion of the Equations . . . . . . . . . a. Production Functions . . . . . . . . . . . b. Resource Allocation Equation . . . . '. . . c. Population Growth and Labor Supply . . . . d. Comodity and Money Markets . . . . . . . e. Foreign Sector and Identities . . . . . . V. SUMMARYANDCONCLUSIONS............. 1. Summary of Major Findings and Policy Implications O O O I O O O O O O O I I O O I O 2. Evaluation and Suggestions for Future Research BIBL ImRAPHY O O O I O O I O O O O O O O O O O O O O I O O 122 122 123 144 145 155 160 173 182 187 188 198 201 ‘0': o... in | In ' i '. Ill ' o .J .|.' q al“ "1 "L4 Table 1.1 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 LIST OF TABLES Comparison of Population and Labor Force Figures from Labor Force Surveys and the Registers............. Relative Changes in Taiwan's Exports, 1899 to 1970..... Relative Changes in Taiwan's Imports, 1899 to 1970..... Number of Household Farms, Their Mean Size and Tenancy RelationShips.00.0.0000...I0.00.0.0.0...0...... Agricultural Outputs and Inputs........................ Percentage Distribution of Net Farm Income of Taiwan... Agricultural Production per Farmer, per Capita Farm Income, Consumption, and Savings in 1935-1937, NT$IOOOOOOOOOOOOOIOOOIOOOOIOIOIOOOOOOOOIIOOOOOOOOOOOOI. Composition of Farm Family Expenditures................ Occupied Population and Major Industrial Composition... Birth Rates, Death Rates and Rates of Natural Increase’ Taimn1906.1970000000000000000OIOOOOOOOOOOOO Comparison of Total Population and Economically Active POPUlationoooooooooooooooooooooooooo00000000000000.0000 Components of the Average Annual Change Per 1,000 of Taiwan's Economically Active Male Population........... Annual Rate of Price Change and Relative Price Change.. The Composition of Taiwan's Aggregate Demand........... The Aggregate Demand per Person and Demand Eh8t1c1tie80000000000.0.0....0.000.000.0000.00.0.00... Population, Labor Force, and Income in the Agri- cultural and Nonagricultural Sectors.................. Vital Rates for Agricultural and Nonagricultural POPUIationso0.09.00.00.00oooooosooooooooosososassoc-coo vi Page 18 73 74 76 78 80 81 83 84 86 88 89 99 100 107 108 3.17 3.18 3.19 3.20 3.21 3.22 3.23 4.1 4.2 4.3 4.4 4.5 Composition of Employed Persons by Status.............. Age-Specific Death Rates for Each Sex and Percent cmnge for Taiwan, 1950-197000000000000OOOOCCOOOOCOOOOO Age-Specific Birth Rates and Other Fertility Rates for All WOmen and Percent Change for Taiwan, 1950-1970. Proportion of WOmen Currently Married, Marital Fertility Rate, and Percent Change for Taiwan, 1960-1970000000‘000000....0..00....OIOOOOIOOIOOOOOOI... Proportion of Wives wanting No More Children and Using No Contraception in Taiwan, 1965-1970............ Wife's Desired Number of Children by Wife's Age for Taiwan, 1965, 1967’ 1970......CCOI.....00....ICC... Wife's Desired Number of Children, by Wife's Age and Education and by Type of Place of Residence, 1970.. Adjustment Indexes for Level of Utilization and Quality of Labor in Agricultural and Nonagricultural satorBOOIOOOOOOOOO.I0.0.0...I0.00....00.000.000.000... Comparisons of Marginal Products of Capital and Labor and the Actual Returns to Capital and Labor between Agricultural and Nonagricultural Sectors....... Percentage Currently Using Contraception, Currently Pregnant, and Mean Length of the Open Birth Interval, by Number of Living Children and Number of Living Sons, for Wives Aged Twenty-Two to Twenty-Nine With One to Four Children, Taiwan: 1970............... Comparison of Economic Variables....................... Income, Population, and Price Elasticities of Demand for Various Consumption Components.............. vii 109 114 115 116 117 117 118 147 151 163 171 177 Figure 1. LIST OF FIGURES Labor Supply and Labor Demand at Various Stages Of Development...OOOOOOOOOOOOOOOOOOOOOO0.000000....0... Graphical Presentation of Economic-Demographic muflibrjIWOOOO00.00.00.000...0....OOOOOOOOOOOOOOOOOOOO Natural Growth Rate of Population and Labor Force and warranted Rate of Capital Accumulation............. viii Page 44 56 63 CHAPTER I. INTRODUCTION 1. Objectives of the Study_ The interactions between population growth and economic development of the developing countries have attracted the attention of some economists and demographers for the past decades. The main reason for this is the occurrence of two consecutive demographic events: first, a rapid growth in population owing to a decline in mortality with fertility remaining constant and high; and second, a high rate of urban migration which is in part an effect of the first event. An increase in p0pu1ation growth, with other economic con— ditions remaining constant, imposes an immediate burden on the economy. The rapid influx of large numbers of people into urban centers creates a series of Social and economic problems, such as shortages of public facilities, accommodations, and jobs. In Taiwan the first event oc- curred approximately in the period from 1945 to 1960, while the second has been felt in recent years. Substantial knowledge of the consequences and determinants of population trends has been accumulated since the occurrence of the first event. Once believed to be the most problematic factor govern- ing population trends, fertility, which has occupied considerable efforts by academicians and administrators, has gradually been brought under the control of human will in some of the developing countries. 1 in! ’, to. C In 3“. " “val "at; ' I%._ .u hosq, ' P ‘..4:' 1...... . sw“ ‘t (I’D y 6 IL. ' rv- "1 "I 2 Increasing research on the relation of urban migration and economic development has also been carried out. Relatively little attention has been given,however, to the integration of population growth and urban migration into one general economic analysis. Moreover, as urban unemployment, mainly originating from rapid p0pu1ation growth and migration, has succeeded population pressure on resources as the major policy problem, interest in labor absorption in the developing countries has revived, while few economists and demographers have gone beyond inquiries into the effects of the rate of p0pu1ation growth on economic progress. The primary objective of this thesis is to study in depth, both theoretically and empirically, the interactions between popula- tion growth and economic development in a broader context. First, a mmcroeconomic-demographic two-sector model is constructed with special emphasis on simultaneous consideration of the components of population growth and their dynamic effects on demographic and economic char- acteristics of the population. Next, the model is tested with eme pirical evidence. With the aid of existing knowledge in the fields of economics and demography the job of constructing a logically con- sistent theoretical model is easier than to find an adequate locale for testing the model. Fortunately, Taiwan is particularly suited to our purpose, because on the one hand it has experienced significant economic and demographic changes since World war II, and on the other this experience has been verbally and statistically documented in a rather detailed form. This, of course, is not to imply that our study is free of constraints of data. In fact, both descriptive and quan- titative analyses are important to our study of the interactions w. h: 3 between population growth and economic development. 2. The Problem and the Conceptual Approach Studies on the interactions between population growth and economic development must involve both economics and demography. The relation between the two disciplines, in the opinion of most economists and demographers, has been rather unstable and even disharmonious. The general trend of development, in W. C. Mitchell's words, is des— cribed as: Position of population problem in economic theory has undergone radical change. It had been treated by the classical economists as a dynamic factor. As economics became more rigorous and academic it became more static. Population changes had no place in a strictly static analysis. . . . They do not figure prominently even in Marshall. If we are to develop dynamic economics, population must be restored as one of the basic factors in economic theory. Theorists must consider the implications of the prospect of an aging population in respect to characteristics of demand, and also to available supply of labor.1 Recent developments in growth theory have verified Mitchell's prediction. There seem, however, to be many unreconcilable obstacles between the two disciplines. In 1957 F. Lorimer stated outspokenly: The marriage of demography and economics while both were inmature —- "Parson" Malthus officiating -- resulted in a stormy and unfruitful union. Both dynamics of interactions among economic factors and the dynamics of vital trends in relation to population structure were long neglected in a hasty synthesis that placed undue emphasis on the relation of population to resources, and the corollary theory of a theory of a hypothetical fixed "Optimum". . . . Both the demographer and the economist have important contributions to make to an understanding of interrelations between population trends and economic changes. But neither en- compasses the whole subject within his own field o. M E 4 of special competence. . . . Yet an analysis of interrelations among such complex economic and demographic process, with attention to their social and cultural context, is essential in a sound approach to this subject.2 More recently, massive efforts have been under way to explore the relationships between population change and economic development. Much of this research, however, has still been devoted to the direct relationships between population and resources. D. Kirk pessimistically wrote in 1968 in his encyclopedia article on the subject: Dissatisfaction with the Malthusian approach led to the divorce of demography from economics and to a continuing suspicion among some economists that demography overemphasizes the force of popula- tion growth and that population control in under- developed areas is in some way a diversion from, and even a threat to, the central purpose of economic development.3 This brief review of opinions on the relation between economics and demography clearly indicates that disharmony arises not because of the existence of any innate disagreeable characteristics but mainly because of the changes in emphasis on the subject matter of the two disciplines. In order to reconcile economics with demography then, one must first seek well-defined scope of the two disciplines and then link the subject matter defined therein into an integrating economic- demographic model. Among students in both fields there are considerable differ- ences of opinion on the proper areas covered by each. This is es- pecially true of demography. Some believe that demography disperses among various disciplines -- economics, history, geography, statistics, sociology, anthropology, and human biology -- and others look upon it as an independent discipline. For our purpose a comprehensive view Mr Wu ”'1 \ 5 free of any preconceived intention of affiliating economics to demo- graphy will be most suitable. According to the above criterion the definition of economics used by W. J. Samuels in "The Scope of Economics Historically Con- sidered" is selected. He circumscribes economics as follows: Economics has been historically a discipline of two traditions, Economic Theory and Theory of Economic Policy. Whereas Economic Theory is cone cerned with market resolution of the three basic economic problems of resource allocation, income distribution, and aggregate-income determination. Theory of Economic Policy is concerned with a fourth problem, the organization of the economic system, ultimately the distribution of power .4 For demography Hauser and Duncan's definition best meets our criterion. They define it as follows: Demography is the study of the size, territorial distribution, and composition of population, changes therein, and the components of such changes which may be identified as natality, mortality, territorial movement (migration), and social mobility (change of status).5 One of the special features of this definition is the omission of any reference to population quality. This omission, however, has no effect on the construction of our model. For changes in human quality result from investment in human resources which,by definition, belongs to economics. The basic elements of these definitions can be organized on the basis of modern economic and demographic theory and the inter- relationships among them explored. Let us start with the economic problem of aggregate-income determination. The rate of income (output) of the economy at time t is determined by the quantities of inputs -- land, labor, capital, and technology actually used in the production ‘1! >11 6 process -- and by the forces of the specific economic system (the structure of power). Assuming that the production relation between inputs and outputs is homogeneous in the first degree, the function of demand for inputs which relates the marginal physical products of each input to the quantities of that input can be derived. The demand for input functions together with the given functions of supply of input simultaneously determine the economic problem of income distribution in terms of factor prices and the quantities of inputs employed in the production process. If the economy contains several main productive processes, then under the assumption of perfect competition the total available land, labor, capital, and capacity of technology will be allocated among the processes to the point where the values of the marginal products of each input in process are equal to each other. Population enters the economic complex via its direct influence on capital formation and labor supply. An increase in the size of population through the increase of births over deaths exerts an im- mediate effect on level and pattern of demand for consumption goods, which in turn affect the capacity for capital accumulation. After certain time lags the increase in population contributes to the supply of the labor force on the one hand and on the other creates the problem of employment. Conversely, allocation of labor among various produc- tion processes and changes in level of income have an influence on the movement and mobility of population and then an effect on the trends of fertility and mortality, which in turn affect the size and age-sex composition of the population. The above brief sketch completes the general picture of the interrelationships between economics and demography. It will be used ‘1 -4 "1 .u s. ‘3 '0. '2 \. 7 as our blueprint for the construction of our economic-demographic model. 3. A Brief Review of the Literature Ancient studies on relationships between population change and economic growth can be found.6 Plato and Aristotle both asserted that there was an optimum size of population that would allow a city-state to be capable of self-defense and economically self-sufficient. A large and growing population was favored by the'Mercantilists for its economic, political, and military advantages to the nation. More re- cently.Adam.Smith also argued that a large population would benefit the economy by supporting a large market and extensive division of labor. The view on population expressed by Malthus in his First Essay on Population 1798 has had great influence on economics and related subjects. It states that population tends to grow faster than the means of subsistence and thus to erode economic progress if not checked by vice, misery, or moral restraint. This conclusion, however, was arrived at by Malthus from a demographic point of view, not from his view of economics or population studies. Malthus made this point clearly in the §§§§y_and in his Principles of Political Econppy. He said that the purpose of the §§§§z_was to "trace the cause which prac- tically keeps down the population of a country to the level of its actual supplies,"7 while the object of the Principles "is to show what are the causes which chiefly influence these supplies."8 In the Principles Malthus claims that under a system of common property the production of food and the increase of population might go on until the soil absolutely refused to grow a single additional plant. In 8 this situation, the total society would be exclusively engaged in pro- curing the necessities of life. But under a system of private property with political liberty and popular education, the profit motive of the landlord would bring land under cultivation only to the extent that the returns would be sufficient to pay for the maintenance of labor and a profit for him. The profit would allow this class to support the progress of the arts and sciences. The benefit from this diversion is so significant that "the effect of taking poorer land into culti- vation is likely to be counterbalanced under such circumstances, that in the actual state of many countries, or in their probable state of 9 many centuries to come." At the same time this system "tends to elevate the character of the lower classes of society, which makes them act as beings who 'look before and after, and consequently cannot acquiesce patiently in the thought of depriving themselves and their children of the means of being respectable, virtuous, and happy."10 It is seen that in the §§§21;Malthus compared the natural power of mankind to increase with the actual power of increasing produce from the soil, while in the Principles be investigated the causal relation between the actual supplies of product determined by the effectual demand and the actual growth of the labor force (population) in the particular circumstances of the society. Obviously the conclusion based on the former is a tautology and that based on the latter is Malthus' economic theory of population. In his Principles of Economics Alfred Marshall had developed a dynamic system of population supply and demand in line with his classical predecessors.11 After Marshall population economics flourished with the impulse of population effects on the economy. Optimum 'lvs .n. ons- .. so I n... 'I‘s Nu m l 'KL. ‘ . -,~ a n. ‘ n. . I '6 q 't v I 9 population theory prevailed around the turn of the century. During the 19308 the Keynes-Hansen stagnation thesis dominated the academic cycle of economics, and thereafter economic-population development theory surpassed them all and thrives to the present time. The Optimum population theory is mainly concerned with finding the size of popula- tion which would maximize the average or marginal products of labor, given that all other factors remain constant. The stagnation theory states that population growth is one of the important elements in economic progress. Hansen estimated that 40 to 60 percent of the total capital formation in the western countries during the nineteenth century was due to population growth.12 Decline in population growth thus causes a deficiency in effective demand for investment goods and in turn, a slowdown in the pace of economic development. Since WOrld war II economists have been largely concerned with the problems of population and economic development in the developing countries. Kuznets, and then Coale and Hoover, chose several courses of fertility trend arbitrarily and worked through a simple and closed Harrod-Domar model to demonstrate the differential effects of alterna- tive changes of fertility levels on economic growth.13 Nelson and Hagen compared the rate of population growth with the rate of increase of per capita income and studied the interrelation between these two variables.14 Leibenstein went further in attempting to make the p0pu- 1ation factor an endogenous variable in a growth model and developed his minimum critical effort theory.15 Lewis, Fei and Rania, and Harris and Todaro constructed two-sector models to demonstrate the workings of urban-rural migration and economic growth.16 Clark, Hirschman, and Boserup may be the only exceptions in considering the growth of “I" .0"; u ‘ V “n - ..'-r< p‘o.‘ g.-.. u-' h..- I 3" a -v.“ u,,.. "L. .::'-- Nu L . n, 'I. :~ .I o. l ‘A l ‘».'| “Al . ‘A ‘. ‘x:_-‘ U ‘1 10 population a stimlant rather than an obstacle to development of eco- nomically backward countries. Enke and Zind, Lloyd, and Barlow and Davies evaluated the effects of government projects on p0pu1ation and economic growth with the use of economic models. Sauvy and Votey modified the economic optimum population theory by considering power and preference for children in the welfare function. Easterline Followed Malthus, Keyneg,and Hansen in attempting to elucidate the role of effective demand in the relation between population change and ac onomic ac tivities . 20 This brief review indicates that many specific economic- demographic relationships have been investigated. Some of these in- vestigations have been conducted in a general equilibrium framework, but any single work can be viewed as a partial equilibrium analysis in light of important relationships treated in other models. In sum, then, it can be stated that there is a need to integrate features of individual models into a single complex of variables that consists of a general equilibrium framework in the sense that it does not omit important economic-demographic interactions. 4. Sourcesof Data The data for this study are drawn primarily from three sources Which are described more thoroughly in the pages that follow: (1) na- tional income statistics compiled by the Directorate-General of Budgets, Accounts, and Statistics, Executive Yuan (DGBAS), (2) population and vital statistics derived from the population register, and (3) labor force statistics from surveys conducted by the Labor Force Survey and Research Institute. o..- M- , .u—I- . .‘J... Iat‘ '- II. 3""; yw h bah. . 9;.ar. 'Hv. . ..'. . a ' a... 1215 u, M b "s c..‘. "mi 12".,” Q'l 11 NATIONAL INCOME STATISTICS Compilation of national income sta- tistics is relatively new in Taiwan. Annual estimates have been made available by DGBAS since 1951. Basic concepts and methods of compiling Taiwan's national income closely follow the National Account System suggested by the United Nations.21 Unfortunately, the reliability of the statistics for 1951 to 1964 is questionable, mainly because many of the items for that period, such as imputed rent for owner-occupied dwellings, values of agricultural by-products, and depreciation of government buildings, were estimated in 1967 to 1969 and added to the Estimates made in retrospect in- reports for the respective years. evitably involve guesswork and arbitrary assumptions. These adjust- ments are necessary however, to make the time series of national income Statistics comparable. POPULATION AND VITAL STATISTICS from the p0pu1ation censuses of 1956 and 1966 and statistics compiled The latter gives yearly Demographic data can be obtained from the continuous population register. p0pu1ation and vital statistics in great detail and their accuracy and completeness are not inferior to those from the general population censuses. Therefore, we rely mainly on the second source. The continuous population register has operated since 1905. The hierarchy of registration is parallel to the civil divisions. A registration office is affiliated with each government of a township, a precinct, or a small city. The office, staffed approximately by one clerk for every 4,000 population, is responsible for keeping the registers. The basic unit of registration is the household. Each household must file a "Household Record" in the registration office located near its residence, and keeps a roster of all members and an .4 .- hf P A 12 "Identification Card" for each person above 14 years of age for the purpose of cross-checking. These papers also serve as legal proof of identity, family relationships, citizenship, and civil or public privileges and obligations. The head of each household is responsible for reporting changes in demographic events of and within the house- hold. Births, marriages, and migration must be reported within 15 days after occurrence and death within 5 days. The registration entries include the relationship to the head of the household, the domicile, the birth date, the birth order, the marital status, the education, and the occupation. For each person blanks are reserved for entering mi- gration, military service, and criminal records. Births, deaths, marriages, changes in education or occupation, divorces, and population movements are recorded on special forms before appropriate entries or Changes are made on the Record. Population and vital statistics are compiled by local offices from the information on Household Records and on the special forms. The tabulations are population by ageiand sex, by education, by occupation for each sex; births by age of mother and father; and deaths by age and sex. Capies of tabulations are sent to Taiwan provincial and Taipei city general registersfor final checking and publication. The registers are checked at the end of each year by a canvass. The census also functions as a thorough check of the registers. There have been relatively few omissions or duplications of events required to be registered. The system, however, is not necessarily perfect. There are three main sources of error: (1) errors in the completeness of registration, (2) errors in the registered characteristics, and (3) errors in data processing. Each type of error is briefly discussed ... u at tr > n V" ‘i 1. :s' 4‘ If 13 as follows. Errors in the Completeness of Rggistration One serious flaw, from the user's point of view, is the deliberate omission of military personnel from the published figures. It has been estimated that a military force of about 600,000 moved into Taiwan around 1950. This accounted for about one-tenth of the population at that time. All except those personnel who had families and resided outside the bar- rack areas were omitted from the registration system for administrative reasons. Thereafter the size of the armed forces has been maintained through recruiting civilian youth for two or three years of service. The retired servicemen are reported in the register. Although the recruits remained in the record, they did not appear in the published p0pu1ation figures until 1961. In 1969 the military forces moved into Taiwan around 1950 from the Mainland and not retired were also included in the register. For these reasons the published figures for the total p0pu1ation of Taiwan for the years before 1969 are incorrect. To pro- V1de consistency in this study estimates of total population for each Year from 1951 to 1968 were made in the following manner. The popula- tion by age and sex in 1970 was used as a base population and projected backward by reverse survival ratios, thus including the civilian popu- lation plus the survivors of the 1950 military forces. A second problem of the completeness of registration is delay in reporting population movement and vital events. There have been no major movements of population in and out of Taiwan since 1951. The small stream of in and out movement, largely students going abroad for advanced studies, was quite accurately reported. Delays in reporting movements within the island create problems for p0pu1ation counts in 14 local areas. In our study, however, only total population by age and sex was used, therefore errors in this respect have no effect on the numerical results of our study. Deaths of persons over five years of age were reported fairly accurately and the reports were made on time. Studies on the accuracy of the registration of deaths found that mortality rates for those :five and over were in high agreement with model life table rates.23 livaluation studies all showed that of the total number of deaths regis- tered in 1965, 96 percent were registered within a month after the date of occurence. Deaths of infants soon after birth are found to be underregistered throughout the island. This kind of error affects the £i<2curacy of the infant mortality rate more than it does the complete- tlcass of total counts, so no adjustment was made in total counts. In- ifaant mortality rates were estimated by extrapolating the mortality Treate of persons 5 to 9 years old in accordance with the relation estab- liished by the appropriate model life table.24 Registration of births, except for the case just mentioned, fines fairly accurate, although late reporting of births was not uncommon. In 1968, 83 percent of the births were reported within a month. No In fact, the percentage of improvement has been reported since then. 25 birth reporting within a month decreased to 81 percent in 1971. If the pattern of late reporting does not change significantly, annual counts of population will not be seriously affected. Errors in the Registered Characteristics The accuracy of age registration has never been a serious problem in Taiwan, due to the fact that age is calculated from the registered date of birth rather than obtained from direct questioning of persons. The procedure used 15 avoids the problems of inaccurate declarations of age and of preference or avoidance of certain digits. However, a minor defect found in an evaluation survey made in 1965 is that some older persons tend to give birth dates according to the lunar calendar.26 This yields a one- to two-month difference from the standard solar calendar. The accuracy of the education and occupations registered must be seriously questioned. The 1965 check survey indicated that the per- centage of persons with accurate education and occupation entries in the registers was 91 and 82 percent respectively. The errors in edu- cation and occupation registration were not the result of misreporting: they were mainly due to failure to keep information up to date. Since producing education and occupation tabulations is not the primary function of the registration system, registrars and the heads 0f households were not urged to make such corrections. In response to demands from academic circles, a general check and corrections for these Entries were made in 1968.28 Measures to keep the registration entries Up to date were also implemented. Our study needs a time series of education and occupation classifications for the labor force, and ad- justment for these data would be difficult, therefore we have to rely on other sources described in this Chapter for this kind of data. Errors in Data Processing The township or precinct registrars are responsible for preparation of tabulations from their registration records for three time intervals in a year. At the beginning of each month the total number of households, the population by sex at the end of the preceding month, and the registered births, deaths, mar- Every three riages, in and out migrants during the month are counted. months the registered births are classified by age of mother and 16 father, and by fathers' occupations; and the registered deaths by age and sex. At the end of the year information on the household record is twed to make tabulations of p0pulation by age and sex, by education, and by occupation for each sex. All the work is done by hand tally by the clerks in township offices under the supervision of registrars at county, city, and provincial governments. Mistakes made during the pmocess of tabulation are considered to be fairly few. A re-count of tflm:registered births and deaths in 1965 shows that the total dif- iferences between the original count and the re-count in each county c>r city have not exceeded the range of 2 percent.29 A serious defect, flowever, has been found in the tabulation of deaths of infants and czluldren. Most deaths of infants below age one registered during the Srear are misclassified into the group for ages one to four. This is partly due to delays in reporting deaths but mainly due to the Chinese (:IJstom of adding one year to each personh age on New Year's day, e.g., an infant born on New Year '3 eve will be two years of age the next uuarning. This defect again rules out the usefulness of the published (181a on infant and childhood death for purposes of this study. LABOR FORCE STATISTICS Labor force statistics that are more detailed than those derived from the register are found in the reports on labor force surveys prepared quarterly since October 1963 by the Taiwan Provincial Labor Force Survey & Research Institute. The primary objective of the survey is to collect current information on character— istics and employment situation of the civilian population over fifteen years of age. The labor force, in the reports, is defined as persons over fifteen who are able and willing to work in the production of economic goods and services. Employed persons are those who worked . n...- - a (ID ‘ 1.-.»... ' 'I-. Q a "5‘3! '- a... ‘ '1’! hivg.“ . . .. .. A " tin-"u .V'qfi, has -‘ ‘- ...., " out n.“ - :O'y '. "Iu u' . .. «Qt. ’- "‘o L. ."A " '1‘. g I“ ‘5 .‘i. ’1 III P 17 at least one hour for pay or profit, or who worked fifteen hours or umre as unpaid family workers during the week including the fifteenth day of the survey month. The p0pu1ation is selected by a two-stage sample procedure. For example, in the survey of October 1970 in the first stage six precincts from Taipei city and six small cities were randomly chosen. The remaining 349 administrative districts (town- sfldps, small cities, and city precincts) were divided into 43 strata aiccording to (1) the population density, (2) the industrial composition c>f labor, and (3) the employment rate. Two of the administrative dis- t:ricts in each stradum were selected with a probability proportional tzo size. A total of ninety-eight districts was selected. In the second astage households were randomly selected from the selected districts txxudl the sample of the population above fifteen included in the survey 'hmas two per thousand of the total population. The representativeness of the survey data can be directly evalu- eated by comparing the estimated total population over age fifteen, tJne total number of employed persons, and the persons employed in agri- cailture and nonagriculture, with those from the population register. Statistics from the register represent conditions enumerated at the end of the year, while the estimates of the survey reflect con- ditions at the end of the week including the fifteenth day in October; therefore the magnitudes of the former tend to be greater than the latter. It can be seen from Table 1.1 that the survey estimates of population over age fifteen and of employed persons were no more than 5 percent lower than those for the whole population enumerated from the register. The range of the differences did not exceed the value of one standard error. Judging from the results of registration .fd VIS- nlwn — «N. V- .—- -. Anz< “J, .I .o >w-auud. unuvflne a... V.» v'—<.~ F‘thp‘ - .- TLL r-.-«- .N- .. .d-. '1. v.;.. ..7L< ~4.v- -.<.-...-.~.- .-.u 2...... - v-<.-F:I. .- n u u u um.-~’\,~ l8 .vououmswom ou mo>uom mo osumm « .Hmma mumsomo .0m .02 .om3woe cHlNo>p=m mason uoomq onu so usomom hasouummo .ouousumoH cosmomom mom mo>usm munch uoooq Hmaoos>oum mom .Aamma "nouamzv xoom uumm uscmmamoamo smegma Omma .musmmw< HH>fio mo uooauumaoo Howoos>oum "mmomoom sso.s sss.s ass.s mos. ass.s sss.s sss. son.s oso.s ass. som.s sss.s sass sso.s sss.s sss.s sss. mes.s sss.s sss. oos.s sss.s sss. sss.s sos.s ssss sss.s sss.~ sss.s sss. sos.s sss.s sss. sss.s ass.s ass. sss.s sss.s ssss sss.s sss.s sso.s ems. sss.s sso.s mas. sso.s oss.s sss. sss.s sss.s asss sss.s sso.s oss.s ass. sss.s oso.s sss. sss.s oss.s sss. sss.s sss.s sass ass.s sss.s sss.s sss. sss.s sso.s ems. ssm.s sss.s sss. sss.s sss.s moss oss. sso.s oss.s sss. sss.s oso.s sss. sss.s sss.s ass. sss.s sss.s sass «osumm ho>u=m monoumwwom «osumm ho>usm monoumwwwm «afiumm HonouHSUHuwmaoz Housuasofiuwd m U m o m m o m <.H HouOH Mm>usm wououmfimom «oaumm mo>usm monoumwwom so>o one ma sossssssos assss>su mmmHmHumm mmH Qz< mwm>MDm women mom¢q XCMh mags: asses 83 as 22322 as zssssaeas .s.s Ema at. -. s "“-§ 1. “on c ‘r 'Oes "cu... v "'-I.. all ‘5':- is . . a" I. P on 4 :m. I"~ ‘ho, . ¢.~.= l9 evaluation we may conclude that both sets of data are of high quality. The large differences seen in the figures for persons employed in agri— culture and nonagriculture reaffirm the fact that the registers had failed to keep their records up to date in a rapidly industralizing economy. 5. Plan of Presentation In this Chapter, we have discussed the objectives of the study, the problem and the conceptual approach, the related studies, and the ruature of the data to be used in this study. The following chapter is (levoted to the construction of an economic-demographic two-sector growth nuxdel. In Chapter III we shall describe the history of the population .arui economy of Taiwan in order to show the significance of sociocultural aarui institutional factors in determining the relationship between p0pu- Jaation growth and economic development. Chapter IV presents and inter— prets the empirical regression results from the application of our ‘nncxiel to Taiwan in the period from 1951 to 1970. In conclusion, Chapter ‘7 summarizes the basic findings of the study and discusses their impli- <=£1tions for population and economic policies and for further research Ii‘egarding the generalization of the empirical results from this study. ' '..u.4 .l “-. “n‘r.~' a . '4 efi:v'{.u . Ir .. ~._.-' a..“. ‘ S . ~ ‘ s -u-.:, l 0.“ - - , 'H N" .a. 5.. ‘ 1; “‘u.1 o '\ “‘h 1 s“ 9 \ ‘ 20 Footnotes 1 W. C. Mitchell, Types of Economic Theogyl from Mercantilism _J§:c> Institutionalism, Vol. II, (New York: A. M. Kelley, 1969), p. 764. 2F. Lorimer, "General Survey," in D. V. Glass, ed., The Uni- versity Teaching of Social Sciences: Demography (Paris: UNESCO, 1957), 131.. 21. 3 D. Kirk, "The Field of Demography," in International Encyclo- edia of the Social Sciences, Vol. XII (New York: Macmillan, 1968), p - 334. 4 W. J. Samuels, "The Scope of Economics Historically Considered," mziJneographed (East Lansing, Michigan: Michigan State University, 1968), ‘;>.. 1. _ 5 P. M. Hauser and O. D. Duncan, The Study of Pppulation, an JExnnventory and Appgaisal (Chicago, Illinois: The University of Chicago Press, 1959), p. 31. United Nations, The Determinants and Consequences of Pppula- t::113n Trends, Document No. ST/SOA/Ser. A/17 (New York: 1953). 7T. R. Malthus, First Essay on Population 1798 (New York: Kelly, 1965), p. 3. 8T. R. Malthus, Principles of Political Economy (New York: Kelly, 1964), p. 307. 9Malthus, Principles, p. 288. 1QMalthus, Principles, p. 226. 11Cf. P. K. C. Liu, "Marshall's Economic Theory of POpulation," tnimeographed (Taipei, Taiwan: The Institute of Economics, Academia Sinica, 1970) . ' 12 A. H. Hansen, "Economic Progress and Declining Population Growth," American Economic Review, Vol. 29, No. 1 (March 1939), pp. 1‘15. 13 Cf. S. Kuznets, "Population and Economic Growth," Proceed— ings of the American Philosophical Society, Vol. III, No. 3 (June 1967); amlAn J. Coale, and E. Hoover, Population Growth and Economic Growth Eglow Income Countries (Princeton, New Jersey: Princeton University Press, 1958). 21 14Cf. R. R. Nelson, "A Theory of the Low-level Equilibrium €[firap in Underdeveloped Economies," American Economic Review, Vol. SHEELVI, No. 5 (December 1956), pp. 894-908; and E. E. Hagen, The jEZxconomics of Development (Homewood, Illinois: Richard D. Irwin, Inc., ’1: 968). 15 Cf. H. Leibenstein, Economic Backwardness and Economic Growth ( New York: John Wiley & Sons, 1957). 16Cf. W. A. Lewis, "Economic Development with Unlimited Supplies (3’15 Labor," The Manchester School, Vol. 22 (May 1954); J. C. H. Fei zaqxnd G. Ranis, Development of the Lgbgz Surplus Economy (Homewood, ililllinois: Richard D. Irwin, Inc., 1964); and J. R. Harris and.M. P. flfcadaro, "Migration, Unemployment and Development: A Two-Sector Analysis,” American Economic Review, Vol. LX, No. 1 (March 1970), pp. 126-142. l7Cf. C. Clark, Pppulation Growth and Land Use (New York: lktzacmillan, 1968); A. D. Hirschman, The Strategy of Economic Develop: Ixieent (New Haven, Connecticut: Yale University Press, 1958); and E. 13<>serup, The Conditions of Agricultural Growth (Chicago, Illinois: Aldine, 1955) . 18Cf. S. Enke and R. G. Zind, "Effect of Fewer Births on Aererage Income," Journal of Bio-Social Science, Vol. I, No. 1 (January JLSB69); P. J. Lloyd, "A Growth Model with Population as an Endogenous ‘7Eiriable," Pppulation Studies, Vol. XXIII (November 1969), pp. 463-478; Eitld R. Barlow and G. W. Davies, "Project Evaluation with a Detailed Macroeconomic Model," mimeographed, (Ann Arbor: University of Michigan). 19Cf. A. Sauvey, General Theory of Population, trans. C. Campos (Ilondon: Weidenfeld and Nicolson, 1969); and H. L. Votey, Jr., "The Chptdmmm Population and Growth: A New Look," Journal of Economic Theopy, 20Cf. R. A. Easterline, Pppulation, Labor Force, and Long Efigings in Economic Growth (New York: Columbia University Press, 1968). 21United Nations, A System of National Accounts, Studies in ‘MEthod (New York: United Nations, 1968). Provincial P0pulation Census Bureau, 1956 Population Census Reports (Nantou, Taiwan: 1958); Provincial P0pulation Census Bureau, 1966 Population and HousingyCensus Report, (Nantou, Taiwan: 1969); and Provincial Department of Civil Affairs, 1970 Demographic Fact Book (Nantou, Taiwan: Provincial Government Department of Civil Affairs, 1971). 22 23J. M. Sullivan, A Review of Taiwanese Infant and Child Mor- tality Statistics (Taipei, Taiwan: The Institute of Economics, Academia 2['Sullivan, Infant and Child Mortality, pp. 2-29. 25Provincial Department of Civil Affairs, The Monthly Bulletin of Population Registration Statistics of Taiwan, Vol. 1, No. 7 (October i966), pp. 20—21; and Ministry of Interior, Taiwan DemographLMonthly, Vol. 7, No. 4 (April 1972), p. 44. 26P. K. C. Liu, The Use of Household Registration Records in Measuring the Fertility Level of Taiwan (Taipei, Taiwan: The Institute of Economics, Academia Sinica, 1967), pp. 17-23. 27Liu, Use of Household Records, pp. 25-32. 28Provincial Department of Civil Affairs, Taiwan Area Report of Household Registration Statistics and Result of Register Check (Nantou, Taiwan: 1969). 29Provincial Department of Civil Affairs, The Monthly Bulletin of Taiwan Population Registration Statistics Vol. I, No. 7 (October 1968), pp. 15-44. 30Provincial Labor Force Survey & Research Institute, Qparterly Report on the Labor Force Survey in Taiwan, No. 30 (January 1971), p- 870 I .u l n...‘ v ...._ sun; ! 1.. :- v .v- CHAPTER II. THEORETICAL FRAMEWORK OF AN ECONOMIC-DEMOGRAPHIC MODEL 1. Introduction The basic aim of this chapter is to build an explanatory model ifcor the analysis of the interactions between population growth and eaczonomic development, with application to developing countries like 'ITEIiwan. especially in mind. Economic development has been seen as "the xreealized capacity of a country's economy to generate significant and satlstained rises in per capita product."1 To enlarge the capacity of 1:11e developing countries involves a transition from a traditional rural t:<> a modern industrial economy. Changes in the level and pattern of zigggregate demand and the ability of the economy to supply these new Ileeeds provide the force for the transition. The size and composition ()1? the population affect the demand and productive capacities in many ‘Iwespects. The latter again interact with the dynamics of the popula- t ion. To trace these interlocked chain reactions we first attempt to formulate an aggregate production function to show the structural tran- SSition. This production function is then used in conjunction with the Iharket forces governing the supply of and demand for productive factors and commodities to detect the economic and demographic effects on Capital accumulation and employment. Finally, we will be in a position 23 24 to analyze the interactions of population growth and economic develop- ment . 2. Resource Allocation between Sectors The economic system consists of two sectors: agricultural and nonagricultural. The former is characterized by traditional and the latter by modern economic features. It is assumed that in both sectors the production functions are linear homogeneous, i.e., subject to constant return to scale. Owing to the resource limitations of the island economy and the rigidity of the social and economic systems, inputs of the two sectors are different in nature and quality. In the agricultural sector, land and its improvement are considered to consti- tute a portion of the capital stock.2 Investment in farm equipment is mainly made by owners of the farm; improvement in quality has largely been capitalized in the costs. The quality of labor inputs is likely to remain more or less constant because of the outflow of well-educated Young men from farm to industrial work. The introduction of new var- ieties of seeds, new methods of land utilization, and efficient methods 0f organization, however, brings in neutral technical progress. The gross domestic agricultural product (GDPA) and product per employed Worker (gdpa) may be expressed as (1) GDPA = Ae TatLAa KAI-a (la) gdpa = Ae Tat kal-a where KA and LA stand for the total amount of capital and labor employed in agriculture respectively; ka stands for amount of agricultural capital Stock per worker. A is an arbitrary constant and t is the time trend. The parameter a is the elasticity of agricultural output with 25 ‘xrespect to labor input. Under perfect competition it is also the erelative share of labor input. 1 -t11ation, it is assumed that capital stock in efficiency units in the xxrxnagricultural sector (KNE) grows at an exponential rate (Tkn) so that (2) KNE a KNe Tknt “91mere KN represents capital stock measured in constant dollars in the nonagricultural sector. Labor input measured in natural units (man years) in the non- agricultural sector improves over time at an exponential rate (Tln). jrhis improvement may come either from the training of employed workers <1! from new entrants with better education, skills, and ability. The Ilabor input in terms of efficiency units (LNE) is then given by (3) LNE = LNe Tlnt The introduction of new methods in management and organization benefits general productivity. It is thus a kind of neutral technical progress. Assuming that it grows at an exponential rate of In , the 26 production of the gross domestic product in the nonagricultural sector (GDPN) may be written as (4) GDPN = Be Tnt LNEB KNE 1'8 where B is an arbitrary constant. B and 1-8 are the output elasti- c ities with respect to labor and capital respectively. Substituting equations (2) and (3) into (4) we obtain 'TBT n+ (1-B)T B 1'3 (4a) GDPN 8 Be (T KN 1n t 1 kn) LN In terms of output per employed worker (gdpn), (43) may be expressed as ' t T - .. (4b) gdpn = Be ( “8'16“ B)Tkn) knl 8 Assuming perfect competition, then the owners of labor and capital earn their respective values of marginal product in each sector. (5) wa - 3GDPA/3LA = o (GDPA/LA)Pa = T Ae Ta tkal-O’Pa (a) ra = scum/am = (1—s) (cum/mm = (1-a)Ae.'atka’°‘pa (7) wn = BGDPN/BLN a B(GDPN/LN)Pn - 8 Be Totknl-BPn (8) rn = acorn/arm = (1-B)(GDPN/KN) = (l-B)BeT0tkn-8Pn Where wa, ra, wn, and rn, respectively, stand for wage rate'and rate of capital return in either sector; Pa and Pn stand for the relative prices of agricultural and nonagricultural output; and To stands for (Tn + B Tln + (1-B)Tkn). Under perfect competition conditions, equations (5) and (7) are also the labor demand curves. Manipulation of these two equations tIlakes it possible to represent the proportion of employed nonagricultural Workers in total employment (pm) as (9) pm - LN/LD = — 1 1 + 3(WU7PIDDa 8(Wa/Tamn Where LD stands for total number of employed man years; pa and on stand for the proportions of agricultural and nonagricultural output, '5- ~l 27 respectively, in total output; and p + p - l. a n If labor can be transferred freely from one sector to another and no sector has reached the state of zero marginal product of labor, equation (9) gives some insight into how employers .in the economy de- c :lde to deploy the relative amounts of labor in the two sectors within the limits of the existing distribution of the capital stock. Given the product elasticities of labor in each sector (3 and B) as data, the prOportion of nonagricultural workers hired by the employers is inversely related to the ratio of nonagricultural wage to agricultural wage and positively related to the output proportion ratio, viz., p nlpa. The first portion of the result seems to be contradictory to the rationale and empirical findings of internal migration because there it is proved that the positive difference between urban and rural wages is the main pulling force causing the redistribution of workers from farm to nonfarm. Both statements are valid, for equation (9) tells how employers will act, while the migration studies illustrate how people respond to a difference in wage rates.3 When the wage dif- ference exists the economy is in disequilibrium and both employers and workers adjust their respective demand and supply. Only with the attainment of equilibrium (wa -= wn) will both of them complete their profit maximizing adjustments and migration cease. Now assume that both labor and capital can be moved freely be- twaen sectors according to the profit maximization principle. Then, in equilibrium the wage rate and rate of capital return in each sector Will be equal, i.e., wa =- wn - w and rs = rn = r. (Where w and r Stand for aggregate equilibrium wage rate and rate of capital returns.) 28 lkassuming again that there exists an aggregate production function, then w and r represent the equilibrium values, respectively, of ‘tihe marginal product of labor and capital in the production function. Aéaggregate output is a function of total capital stock per worker (k); <::onsequently, w and r are also functions of k. Denoting the aalggregate wage—rental ratio as h, then h is a strictly increasing function of k, that is (10) h = h(k), h' > 0 Given any arbitrary aggregate k0, the optimum capital-employ- ment ratio in each sector, and thus the division of labor and capital, .ealnd the proportion of output, are uniquely determined, i.e., equations (5) through (8) give (11) ka = (l -a) h (ko)/a (12) kn = (1 -B) h (ko)/B An aggregate production function for a two—sector model that c omprises explicitly the relation of changes in the allocation of re- sources between sectors is not only helpful in determining the wage- rental ratio but is also a useful analytical tool for better investi- ga tion of the process of economic development. The literature has be en silent about the formulation of such a production function. At- t:‘Et‘uipts made here are far from satisfactory: nevertheless, they can be used, with some inconvenience, for our purpose of constructing an ac onomic-demographic model. The procedures adopted to derive the ag- ggjt‘eegate output function are as follows. By definition total domestic gross product is the sum of agri- (:‘JSltural and nonagricultural gross outputs. The growth rate of total 1) roduct is then - c 29 (13) GDP = paPaGDPA + pnPnGDPN = apaBA + BanN + (1.s)pail + (1-B)oniN + T The circumflex attached to a variable represents the growth rate of that variable, i.e., CPD = dGDP/ dt / GDP, etc. T. represents the technical progress effects on the growth rate of total product, namely (14) T = Ta + Tn + BTln + (l-B)Tkn Assume that the aggregate production function at any given moment in time is also of the Cobb-Douglas type and the output elasticities of labor and capital are functions of time t, as (15) GDP = He TtLDa(t)Kb(t) The growth rate of total output derived from this production function is (16) GDP -= a(t)lI) + b(t)R + T + a(t)lnLD + b.(t)an D :Lsaggregating the growth rate of the total employed labor and capital into sectors, we have (16a) GDP = a(t) pla LA + a(t) plnLN + b(t) pkaKA + b(t) (3an + a(t)lnLD + b(t)1nK From (13) and (16a) we obtain (17) (spa-apla)LA + (Son-aoln)LN + (( 1 —a)oa -boka)KA (( 1 - B)Dn - bpkn)KN + T - ('t+ a(t)lnLD + b(t)1nK) = 0 Assuming that the growth rates of labor force and capital accumulation in neither sector are equal to zero, the parameters of the aggregate production are thus determined (18) a(t)1 = Goa/Ola or a(t)2 Boa/o1n (19) b(t)1 (1-a)pa/pka or b(t)2 = (l-B)pn/pkn 30 (20) 1' - T - ( a(t)lnLD + b(t)1nK) Substituting (9) and (10) in (18) and (l9),respective1y.we get (21) a(t)1 = 01.08 + Bpnwa/wn or a(t)2 = Bpn +0L%wn/wa (22) b(t)1 ( l —0L)pa + ( 1 - B)Onra/rn b(t)2 ( 1 —B)Dn + ( 1 - a)parn/ra Now we may rewrite the aggregate production function of our t1 two-sec tor model as apa+Bpnwa lwn (23) cup = He I‘m K< 1 -a)oa + ( 1 -B)ra/rn From this equation it is seen that the aggregate output of a dualistic economy depends not only on the amount of productive inputs and tech- nical progress but also on the proportion of economic activities and the remunerate ratios of labor and capital between the two sectors. When resource allocation between sectors has approached opti— mum equilibrium wa - wn, rs -= rn, a(t) 8 0, b.(t) - 0 and for a given wage-rental ratio, for example, ho, there are corresponding equilibrium values of the proportions of economic activities, i.e., Duo and pao' Since the wage-rental ratio is a function of the overall capital- etnployment ratio, the aggregate production function (23) under sectoral equilibrium conditions reduces to (23a) GDP - He T tLnapa'iepnx l-ocpa-Bon I t is a linear homogeneous function, and therefore may be expressed in par-employed-worker terms, that is (23b) gdp a HeTtk(1 - apa-Bp“) 1\‘Clte that in sectoral equilibrium the aggregate output elasticities of labor and capital reSpectively are (l) the activity proportion weighted averages of the output elasticities of the two sectors and (2) the 31 return to scale changed from decreasing returns to constant returns. The remnerate rates for labor and capital are also equal to their Weighted averages. 3. The Relation between PopulationLEleoyment, and Economic Development a. Population, Labor Supply, and Employment The available labor force is governed by the size of the popu— lation, its age—sex composition, and the pattern of participation. The components of population growth--fertility, mortality, and external migration—determine the size and characteristics of the population. Since large-scale international migration is not likely to occur in Taiwan given current political circumstances, fertility and mortality are the main factors that determine population trends. Empirical evi- dence suggests that two components, population growth and changes in patterns of participation, are functions of economic development. a As result, economic development is the main factor determining the size Of the labor supply. We shall examine fertility, mortality, the na- CIJIal growth rate of population, and labor supply and employment. FERTILITY The existence of the relationship between fertility and economic development has long been recognized in the economic and population literature. Conclusions on the strength and direction of associations, however, are rather confused. The main source of am- biguity seems to be the insufficient specificity of the definition of eQonomic development. As a rule, economists use per capita output or it1eome as an index of economic development. This aggregate measure re- presents not only the level of purchasing power, but also the concomitant 32 changes in economic activities, structures, and income distribution, which do not necessarily occur in the same direction or with the same force as growth in per capita output. Fertility trends are related to changes in the whole complex of variables. The influence of any one or any subset of the variables cannot be a proxy for the influence of the whole. The aggregate production function for the two-sector model developed in the previous section may be used to illustrate this point. b d From (23) it is seen that per capita output depends not only on tech- nological change and productive capacity but also on proportion of product, ratio of wages, and rate of return on capital in each sector. The functional relationship between fertility and economic development therefore should be a function of all these variables, namely, (24) F = F( ‘t, k, on, wa/wn, ra/rn ) where F is the vector of age—specific birth rate for all women between age x and x + n, i.e., F = {an}. This measure of fertility is free from the effects of changes in the sex and age structure of‘ the popu— la tion. The rate of change in fertility is then (25) F = Ef'l' T + efk k + Efnpn + efwwa/wn + efrra/rn where 6 stands for the elasticity of i with respect to j, e.g., 13 8f R is the elasticity of the age-specific birth rate with respect to the productive capacity of the economy (OLF/ at/ ak/ at) (k/F). It is obvious that changes in fertility level depend on the S:I—gns and relative weights of the rates of growth of fertility deter- minants and fertility elasticities with respect to these determinants. me growth rates of technological progress and per capita productive caI>acity are both positively related to per capita total output, i.e., 33 the index of economic development. The growth rate of the proportion of nonagricultural output in total production, and the respective ratios of the renunerate rate of labor and capital are also positively re- lated to per capita output. The rate of this growth rate is decreas- ing if an agricultural eonomy is nearing an industrial one. When development has resulted only from an increase in agricultural pro- duction or if development has already attained maturity (i.e., fulfill- F‘ ment of sectoral equilibrium), the growth rates of these determinants become zero . According to the nature of their influence on fertility, the elasticities of technological progress may be decomposed into three types. First, among the elements contributing to the increased pro- ductivity of agricultural and nonagricultural output, Ta and Tn’ the advance in medical science and public health is no doubt of the greatest importance. These advances affect the level of infant and child mor— tality and the introduction of efficient methods of contraCeption as well. As these mortality levels decrease, couples who want to achieve the goal of a certain desired number of adult children need fewer births to do so. The introduction of efficient inexpensive contraceptives facilitates the achievement of their goal. Therefore, this kind of technological progress (labelled as T1) has a negative effect on fer- t:Llity as long as the desired is lower than the actual number of chil— dt‘ en for the population as a whole. The second type of technological progress (T2) consists of increases in the efficiency of labor p123 s3, i.e., Tln' Such progress o"Dwiously originates from a rise in educational levels. finpirical e\':l.dence uniformly suggests an inverse relation between educational .AV" :, 7‘. ; “.thv ‘s" u 5" no "" If a.‘.~ .Qp; .\v pa n! H». rt pm... \ I It L» b .. A .‘Id .91 PL .Vk o s uh V s .vt O u . ,~ A v If». . I .. a. I On I .7. 4v \ 12.. «In .4 .. no t u u— . N ..Q 5.- M. v oi. I. I‘ I A... D 1'. 4‘ ‘ I I I ‘ Qt. ll. .Db ‘i:.i‘ 34 JLervel and fertility; therefore, the fertility elasticity of T2 is nega- tive.5 The remaining elements of aggregate technical progress may be Egrouped under the third type (13). They effect an increase in the Itotal output that is equivalent to an increase in the productive ca- [pacity of the economy. The sign of the fertility elasticity of T3 *&9111 be in the same direction as that of productive capacity. Therefore ‘wae turn to a discussion of the latter. The term, the productive capacity of the economy, as used here eexpresses a narrower than usual sense that indicates the amount of czaapital at the disposal of each employed worker. A change in the pro- citnrtive capacity thus implies a corresponding change in the standard c3f'11ving as well as in prospective future disposable income. Therefore, ‘tfhe productive capacity connotes the "permanent income" or "potential income," which is more pertinent to fertility decision-making than is income at a given point in time.6 In low income countriesan increase :i_1n k (total capital stock per worker), as observed by classical and neoclassical economists, increases the marriage rate, improves health ‘<=-<:nditions, and consequently raises the birth rate.7 In high income <=L<3untries the effects of classical determinants may not be as important 53-53 in low income countries but the factor recently discussed by Becker -- tl'ae function of children as consumer durables -- is likely to be in— ‘zilreasingly influential as income grows.8 Moreover, annual observation (DNf? the adjustment of the timing of births in response to economic con- <1thions has also strengthened the belief in the positive income- fGartility relation. 0n the whole, we may assume that the fertility eElasticity of productive capacity is positive. u! r P .-» m. r-'~ «v . ‘lv.‘-O u “a... ""“I' '~-» ‘v- .1; o .. h... l. v (I. r“. 05“,. "‘~L‘ A ‘0. ~“‘ “who: "‘ i“ ' cg N ."‘:‘h I. I.o ML"... " 35 The well-established inverse correlation of fertility with industrialization may be appropriately applied to the explanation of the negative sign of the fertility elasticity of proportion of non- agricultural output, for the latter is one of the best measures of industrialization. It should be emphasized, however, that changes in economic activities from agricultural to nonagricultural and changes in life styles from rural to urban provide a favorable environment for the spread not only of birth control but also of attitudes favor- able to smaller family size. The ratios of the remunerate rates for labor and capital re- spectively, (wa/wn), (ra/rn), and the implicit equilibrium value of the total wage-rental ratio (h) may be viewed as indicators of income dis- tribution. The values of the first two indicators are probably between zero and unity -- the larger the value the more equally distributed is the income. The effects of income equality on fertility have recently a ttracted the attention of both economists and demographers. Because of the lack of statistical data empirical evidence for the exact re- la tion is still insufficient. In an industrializing economy if the increased income were distributed to a relatively small portion of the urban population, the positive income effect on fertility would be rela- t iVely smaller than if the increased income were evenly distributed among the general population. Therefore, we may assume a negative relation between fertility and income inequality. In sum, the signs of fertility elasticities with respect to determinants contained in the fertility function remain the same through- out: all stages of development. The growth rates of these determinants, 1“(>‘Wever, vary with changes in economic structure and income distribution x n. u- '\U-' «at .. f. .. b- u 1" be "I 5. up ' at?" ‘a. .. A a. [y I. 36 as the economy evolves from an agricultural to an industrial base. (Zonsequently, the effect of economic development on fertility varies in direction at each stage of deve10pment. The directions of individual (elasticities and the combined effect at various stages are presented in the following: Agricultural Industrial Mature Fertility elasticity of development development economy Technical Progress T1 - - 0 Technical Progress T2 0 - - Technical Progress T3 + + + Productive Capacity + + + Income Inequality 0 - 0 Economic Structure 0 - 0 Combined Effect + — :_ The combined effect of economic development on fertility ob- ‘xziously depends on the relative weights of opposing forces. In the éagricultural development stage of Malthus' and Ricardo's time, the ttmflical techniques developed for limiting infant deaths and controlling Iffamily size were inadequate. The positive influence of productive 1::apacity dominated fertility trends. On the other hand, the intro- duction of efficient inexpensive methods of birth and death control jLn the developing countries may outweigh that positive effect even in E1.peasant economy. When an economy begins developing industrially t:he negative effects of industrialization, income inequality, decline iLn infant mortality, and spread of family limitation gradually gain momentum and tend to outweigh all the positive effects resulting from tlhe increase in productive capacity and technological progress 13. -,.\.1 ~ ' u. .. .n-I» u r unbvb guy-“o p Iv.-.“ v... - 7-1:;‘v 9-5. ‘t1 .- ."A" h" ‘V- A: V .‘ . o‘, .H L K . . H n. ' 1 ‘l. “s . 37 However, as the economy approaches the mature stage, these negative effects are less effective and fertility trends are increasingly de- termined by the net effect of productive capacity and educational at- tainment (technological progress 12). The sign of the combined effect, thus, depends on the relative weights of Opposing forces. MORTALITY Compared to those of fertility trends, the factors cf mortality trends are reasonably well understood. Ruling out natural disasters, the accumulated evidence suggests that mortality level is negatively associated with per capita income, educational attainment, healthful living and working conditions, and the state of public health measures. Except for living and working conditions, which may de- teriorate for a brief period in the process of rapid industrialization, all other factors move in parallel with economic develOpment. There— fore, we may express mortality as a function of per capita income (26) D = D(gdp) where D is the vector of age—specific death rate by sex between age x and x + n. The mortality elasticity of per capita income in general is highly elastic at the initial stage of development because of im- provements in nutrition and health conditions brought about by incre- ments in income. This elasticity tends to lessen as mortality declines to a certain low level. In the developing countries the inverse relation between in- come and mortality represented by equation (26) was interrupted by the application of low-cost public health techniques and new drugs imported from the advanced countries at times between 1940 and 1960. The conse- quence was a sudden drop in mortality from about thirty or more deaths per thousand population per year to less than fifteen per thousand, OI vi -C '4 (I: 5< 38 without regard to economic conditions. Therefore, the public health variable might better be treated separately in mortality functions. NATURAL GROWTH RATE OF POPULATION Given the age-sex composition of the population at any particular time, the total number of births and deaths can be calculated by multiplying the vector of population by age into the vectors of age-specific fertility and mortality rates respectively. The crude birth and death rates and thus the rate of natural increase in that year can be obtained. These rates have the advantages of being easy to understand and to handle. One serious de- fect, however, is that these rates conceal the effects of the age and sex structure. Therefore it is desirable explicitly to incorporate the indexes of age-sex structure effects in the functional relationship of these rates. As noted earlier, the vectors of fertility and mortality rates are determined by the contemporary level of economic development, i.e., the variables determining per capita output. The age composition of a closed population is the result of the trends of fertility and mortality over a long period. Therefore, age composition reflects the effects of the former variables of economic development. Accordingly, the crude birth and death rates and finally the rate of natural increase must be functions of current and earlier values of these variables. Denoting the per capita output determinants in year t as yr, the natural growth rate of population (n) may be expressed mathematically as (27) n = n(yt, yt _ 1, ..... yt _ n) The relative importance of various coefficients of the variables in (27) varies because the time structure of the interaction for re- spective fertility and mortality rates is not the same. In a closed 39 population, persons at age twenty-five in year t, for example, are the survivors of the births occurring twenty-five years earlier, i.e., the product of the survival ratio from birth to twenty-five years of age and the number of births twenty-five years ago. The survival ratio, in turn, is the cumulative product of the respective product of the respective survival ratios for persons at age zero to twenty-five years old from year t - 25 to t. The latter is calculated from the mor- tality rate of that year and is the function of the per capita income in that year, therefore, the survival ratio from birth to twenty-five years of age is related to the per capita income for the twenty-five preceding years. Applying this relation to p0pu1ation for all ages it is seen that the current mortality rate (and thus the income) af- fects the population of persons of all ages. Mortality rates of certain past years, however, affect only the population of persons of a certain age and older. Therefore, the recent variables have great weight in determining the death rate. The relations between death rates and incomes of current and earlier years are, of course, negative. The time structure of the weights for income in crude birth rate function is more important and complicated than that of death rate. The variable vector of the per capita output y for a particular year has an immediate effect on the fertility rate and thus the crude turth rate in that year. The number of female births in that year, however, can also influence the birth rate at least after a period of fifteen years -- when the girls will have reached the reproductive stage. The influence reflected fifteen years later increases gradually in strength, reaches its peak after about twenty-five to thirty-five years, and wanes, then vanishes in later years. This time structure 40 indicates that the current year and the twenty-five to thirty-five year lagged variables in the crude birth function are most relevant to the determination of today's birth rate. The direction of the effect of current income on the birth rate, as discussed above, depends on the current stage of development. The direction of a laggEd development effect is determined by the stage of development. Therefore, the effect of the age structure of reproductive women implies a built-in factor that intensifies the cyclical fluctuations of the fertility rate and shifts the cycle into the next generation. Given the size of the population in the initial year (No) an equation for rate of natural growth can be used to calculate the popu- lation for the current year, that is (28) Nt = Noe{:“(yt"'°yt-n)dt LABOR SUPPLY AND EMPLOYMENT Studies on the pattern of labor force participation rates universally show that nearly all adult males, but only part of adult women, and young and old persons, engage in or are available for income-producing activities}1 Related to this variation in participation is a complex of inter-related demographic, social, and economic factors, some of which are common to all sections of the p0pu1ation and some that are especially relevant to a particular group.12 Degree of urbanization, level of earnings, and the volume and compo— sition of employment opportunities are generally found to be associated with both sexes at all ages. The school attendance rate has a signifi- cant effect on the activity rate of young peOple. Marital status and maternal responsibilities are especially important in the determination of the number of female workers. This array of factors, as diverse as they are, may be adequately represented by the vector of variables 41 contained in our two-sector aggregate production function (y). The degree of urbanization can be measured by the proportion of nonagricul- tural employment on; levels of earning and employment opportunities by the productive capacity k; school attendance rate by index of labor efficiency Tln; and marital status and maternal responsibilities by the fertility level F. As a result, we may express the age-sex specific participation rates or the gross participation rate (1) as (29) l = LP/N = 1(y) where LP stands for the total economically active population which by definition includes: employed persons, either full or part-time work- ers; self-employed persons; unpaid workers; the unemployed; and persons looking for a job for the first time. N stands for the total population. In general, the gross participation rate is relatively high during the stage of agricultural development because of the pre- dominance of household farms and enterprises. As the economy proceeds toward industrialization an initial decline sets in due to urbanization, the extension of education, the rise of organized nonagricultural em- ployment, and the increase in womens' childcare burden. A reversal trend is likely to follow the initial decline. The release of women from household responsibilities contributes to this increasing trend. Such release occurs because of the decline in fertility, the spread of modern household appliances, and the expansion of employment cp— portunities for women in the service industries. Substituting equation (28) into (29) and rearranging the terms, we obtain the functional relationship between the size of the econ- omically active population and variables of development, that is (30) LP = 1(y)Noe!nt 42 The size of the labor force available to the economy in year t sets the maximum limit of man years that may be utilized by the economy. However, this limit does not indicate how much work the members of the labor force are willing and able to do. In a given sociocultural environment the amount of work performed is determined mainly by the choice of individuals between income-producing work and leisure. Time spent on preparation for future work ability and the prevailing wage rates also affect labor supplied. With the assumption that leisure is a noninferior good and the prospect that better pre- paration pays off better, the economy's labor supply curve is expected to slope upward in an increasing rate as supply approaches the limit of total availability. Labor supply may thus be written as an in- creasing function of the wage rate and the size of the available work force. Assuming this relation to be loglinear, we have: (31) LS = Uw¢ LPW, ¢>O, 13w>o where U is an arbitrary constant and the parameters ¢ and. W re- spectively represent the elasticities of labor supply with respect to an increase in the wage rate and the size of the available labor force. On the demand side, the marginal revenue product of employed labor function (in terms of man years) derived from the two-sector aggregate production function (in sectoral equilibrium) gives the profit maximization relation between the number of employed man years and the wage rate. This equation is rearranged and presented as follows: (32) LD = (opa + Bpn) GDP/Pw In equilibrium supply equals demand, LS = LD. The equilibrium Solutions of wage and employment for given LP and GDP are respectively 43 1 (33) we = ((apa + egmnp p/ ULP )1“? 1 (34) me = ((onpa + BngP p)¢’ 11pr )“¢ The equilibrium locus of wage rate and employed man years to- gether with the curves of labor supply and demand at various stages of deveIOpment are depicted in Figure 1. There are two significant phases for the agricultural development stage, the subsistence phase and the growth phase. In the subsistence phase the practice of primitive culti- vation techniques permits only the level of output and thus of demand for labor that can absorb a relatively small portion of the available man years. The persistent insufficient demand for labor tends to en- courage certain social arrangements under which the majority of the members may have opportunities to engage in economic activities. These arrangements do not necessarily provide an equal share of work for all members, namely some may be fully employed but others not. Assuming rationality, there is a tendency to employ the number of man years necessary for increase up to zero marginal revenue product, and to pay each member in accordance with his contribution (actual working time). A reduction in man years will surely raise the marginal product and cut down the total output. But withdrawal of workers may not have any adverse effect on production as long as their share of work and income can be taken over by the rest of the members in such a manner that the disutility derived from the increments of work does not ex- ceed the utility from the increments of work.13 Under the severe pressure of shortage in work opportunities this arrangement is most likely the case. In Figure l, Oan and anLPa represent the equilibrium 44 Industrial Mature Development t e W Agricultural Development Stage stage 8 ag 1 LD e LS m LD m Wm- subsistencephase growth phase LS l LD1 Wl~ LD L8 8 8 W _ LD LS \\\\ g s o 15 fir? E LP E LP E LP Man a0 a g g l l m m years Figure 1. Labor Supply and Labor Demand at Various Stagesof Development 45 employed and unemployed man years respectively. There is no way to tell the actual number of persons who are fully or partly employed or unemployed from this graph. From the above argument, however, we know that the magnitude of the fully occupied workers must be somewhere to the left of Eho,and from there to somewhere to the right of Eao measures the number of underemployed persons. The remaining distance up to the limit of the availability, LPao measures the openly unemployed workers. The loss from vagueness in the measurements in natural units of workers is small in comparison with the advantage to be gained from an exact gauge of the extent of the employment situation as developed in terms of man years. In the growth stage agricultural output is pushed up by the progress in techniques and can therefore support a relatively high level of demand for labor. This demand curve (as shown in Figure l, labelled LDg) intersects the supply curve (LSg) in its increasing portion. Consequently, the equilibrium wage rate (wg) becomes positive and the equilibrium employed man years (E8) is relatively high as compared with total availability. Underemployment is not likely to be completely eliminated, because growth of the economically active population in this stage is faster than that of the demand for labor. During the industrial development period, the growth of modern nonfarm establishments provides more and more full time jobs. At the same time the rate of increase in the economically active population tends to slow down because growth of population is moderated by a de- cline in the fertility rate. As a result, the underemployment situation improves. In the mature economy stage, the normal level of demand for 46 labor is sufficiently high relative to supply. Therefore, the extent of underemployment is likely to be rather low, although open unemploy- ment may still be serious because of economic maladjustment. In the most serious case, i.e., a depression, the demand curve may shift to such a low level that highly developed techniques and sophisticated capital equipment and skills are not needed. Under this circumstance, irrespective of their training and skill,the workers would be willing to perform and to share whatever work is available with others to meet the conditions prevailing in the labor market. This kind of underem- ployment is not different in nature from that of other stages of de- velopment. Therefore the term "disguised unemployment" (coined by Joan Robinson) for the industrialized countries is not incompatible with the one used by development economists for developing countries of today.14 The above discussion implies that irreSpective of the stage of development the problem of under- or unemployment of labor is only a matter of degree. Sustained full employment, like the sustained growth of output, is one of the ultimate objectives of economic development. Attainment of this objective depends on the performance of the popula- tion and the economy. Our purpose is to construct an economic- demographic model to show these interactions; therefore, we shall closely follow the line of the unemployment growth model.15 b. Saving and Capital Formation We now turn to an analysis of the effects of economic and demo- graphic factors on capital formation. Capital increases by investment. The latter results from the interplay of the forces of gxuag£g_invest- ment and saving. Both of the forces function simultaneously in all stages of development. The shortage of capital during the earlier 47 period of development does not necessarily imply the existence of an investment demand which would be sufficiently high to make the desired level of saving and/or foreign assistance realizable. In fact, economic considerations (especially absorptive capacity) can support only a relatively low demand for investment. This low investment demand not only limits the realization of real saving, but also encourages unproductive consumption. Extensive hoarding of precious metals and stones, and extravagant ceremonial consumption in most of the developing countries are inevitable consequences. Due to these considerations, our model shall deviate from the neoclassical assumption that all sav- ings are productively invested. Instead, realized investment in our model is determined simulatenously by a system of equations for supply of and demand for investment and money. The supply for investment originates from saving. Saving may be undertaken by households, business enterprises, government, or may come from abroad. Household saving, by definition, is the amount of the surplus of personal disposable income over consumption expenditure. Both of its components are subject to the influence of fiscal and other government measures, but the main effect derives from the desire for consumption. This desire varies with a number of factors: the share and distribution of income (wLD, rK), the stock of wealth (K), and the size and composition of the p0pu1ation (N) are found theoretically and empirically to be the most relevant factors.16 Business saving originates largely from the total real profit (rK). Its relative importance in the bulk of aggregate saving becomes increasingly crucial as corporate ownership spreads in the process of industrialization. Saving generated by the government is measured by an excess of revenue 48 over current expenditure. In the long run, taxation provides the bulk of the government revenue, and taxes are generally collected in ac- cordance with the principle of ability to pay, i.e., income shares. Government expenditure (and therefore saving) tends to be associated with the level of development. The amount of foreign saving available to a country depends on a number of factors outside our system and is therefore treated as an exogenous variable. Symbolically, the existing knowledge of the determinants of aggregate net saving function may be summarized as follows: (35) S - S(wLK, rK, K, N) Assuming this function is homogeneous to the first degree.then the net saving per employed worker is (35a) 8 = S/LD = S(w, rk, k, N/LD) Under perfect competition w and r are marginal products of labor and capital. Then we can obtain (35b) 8 = S(k) Differentiating (35a) and (35b) partially with respect to k, and using the ratio of the share of wage income (a) to the elasticity of substitution (6) which is a/d = -k(dr/dk)/Y, we find (36) as/Rk = Y(as/3wLD a/d + as/arK( l -a/5)) + BS/ak + 8.8/8(N/LD) (aN/LD/ak) The expression in the parentheses on the right hand side repre- sents the marginal propensities to save when wages and profits rise and capital stock and p0pu1ation remain constant. In a constant return Cobb-Douglas type of world 6 a l,00, therefore the sign of the marginal propensity to save out of income shares depends on as/arK and as/aw. It has long been argued, with some statistical 49 evidence, that owners of capital save a fixed proportion of their earn- ings and wage earners are devoted only to consumer expenditure, i.e., as/arK = constant and as/aw = 0.17 Recent studies on saving behavior, especially those based on the developing countries, cast doubt on the proposition. LFirst, if the supply of net saving is defined as wealth being diverted from immediate consumption to pay for the purchase of additional future income,one would expect that after a certain amount of saving the negative "income effect" will outweigh the positive "substitution effect" so that the propensity to save will increase with a decreasing rate at first and then decline, or even turn out to be negative as savings are accumulated. Second, the speculation that there exists a consumption pres- sure on voluntary saving caused by an international demonstration effect on low income countries has not been fully supported by the empirical findings. To the contrary, evidence shows that those who spend on modern consumer objects tend to save more and to utilize modern financial facilities to make funds readily available for investment purposes. In addition, they are also shown to have modern attitudes and behavior with respect to innovative practices and family planning.18 Spending on education, health care, and even on food is found to be attributable to the productivity of the working force and consequently to saving capacity. Judging from these findings, some of the consumption ex- penditures classified by the conventional definition as savings should be treated as funds directly channelled to investment flow. Last but not least, the prevalence in the developing countries of the desire for support from one's children in one's old age perhaps reflects the fact that outlays on raising and training children, made SO voluntarily or involuntarily, are still one of the best forms of saving accessible to the majority of people in such circumstances. Due to the increasingly high population pressure on land and the changes in attitude toward family relations, rewards for this kind of saving now tend to be low and risky. This form of saving is also likely to be a burden to the economy as a whole. Saving via children, however, is not different in nature from funds saved for financial markets.19 Ac- cordingly, expenditures on children for this purpose (mainly carried out by wage earners) should be treated as saving,too. In fact, de- ve10pment economists have long recognized the existence of this source of saving in terms of surplus labor. The fundamental problem, however, is not that workers do not save, but rather that their saving must be directed into more productive uses. All of the above findings seem to support the belief that both capital owners and wage earners are willing to save part of their income, and that low income impairs ability to save efficiently, while high income may dilute saving motivations. The second term on the right hand side of equation (36) is equivalent to the real balance effect on saving. It is known to be negative. Like the income effect, its significance would perhaps show up only after the attainment of a relatively high income level. The last term of equation (36) represents the effect of changes in the population-employment ratio on perdworker saving. The first component of the term is the marginal propensity to save with respect to the population-employment ratio. It is negative because the larger the population relative to productive workers the heavier the burden on the ability to save. The second component is the responsiveness of 51 the population-employment ratio to the increasein capital stock per worker. Its sign depends on the stage of deve10pment. As noted before, in the agricultural deve10pment stage population tends to grow faster than employment and as a result it tends to be positive. This tendency reverses to negative in later stages because the growth of employment is likely to be faster than that of population. The combined effect of the two components is consequently negative in the stage of agri- cultural development and positive in the industrialization and mature stages. The positive effect is not unexpected, because the population- employment ratio is not the same as the dependency ratio -- the ratio of p0pu1ation to persons fifteen to sixty-five years of age. The latter represents the demographic effect; its effect on saving is likely to be negative throughout all stages of deve10pment.20 The former contains both the demographic and economic effect. Therefore, the effect may be negative or positive, depending on the relative strengths of popula- tion and economic factors. The determinants governing the demand for investment may be seen to operate on two levels. The first level consists of those de- terminants related to the ways in which investors make their decisions about the employment of capital equipment. The second covers those connected with absorptive capacity, i.e., the availability of the in- vestors and complementary factors of production with which capital is to cooperate. Investment theories and their empirical verifications mainly concentrate on the explanation of the determinants at the first level, with the implicit assumption that the variables at the second level remain unchanged. On the contrary, in the economic development literature,investment demand is recognized as an effective force in 52 the determination of the actual magnitude of capital formation. But most of the literature concentrates on discussion of the factors at the second level and neglects those at the first. Each treatment may be appropriate to its own substance, but for a discussion of the course of economic development we should consider the variables at both levels. According to the recent integration of the acceleration principle and neoclassical investment-demand theory, in an output market character- ized by excess supply, the profit maximizing investor's decision depends upon the interest rate (q), the real wage rate, the existing capital stock, as well as on the changes in the level of output.21 In terms of per—person employed we may have an investment function as follows: (37) i = I( q, w, k, gdp) where gdp = dgdp/dt. Although this is a result of rigorous logical reasoning, it deals with the factors influencing the investment decision Only; those influencing the capital absorptive capacity are not explicitly expressed in the function. The important limitations to an economy's capacity to absorb capital are shortages of entrepreneurship, management, and technical skill as well as the lack of efficient public administration. As diverse as they appear to be, these complementary factors are, in fact, closely associated with the level of deve10pment. Therefore, per capita output may be used as a summary index. In addition, demo— graphic factors are also found to have significant influence on the level and the composition of investment.22 Taking all relevant argu- ments into account we may rewrite the investment function (37) as (378) i = I'(q. W. k. gap. gdp. N) 53 Since w, the marginal productivity of labor, gdp, and N are all functions of k,equation (37a) can be made a function of q and k. (37b) 1 = i(q, k) The partial derivative of (37b) with respect to q,ai/aq, is the marginal prOpensity to invest with reSpect to interest rate which is known to be negative. Differentiating (37a) and (37b) partially with respect to k we find (38) ai/ak = (ax/aw)(a»/ak> + (ax/agép)(ag&p/ak> + (BI/agdeaI/Bk) + (BI/BNMaN/ak) + 31 lak It can be seen that the marginal propensity to invest with re- spect to per capita capital is the sum of the effects of factor price substitution, acceleration, absorptive capacity, population, and capital accumulation. The first three are positive since the effect of changes in wage rate and in the rate of changes in response to changes in per capita capital stock are both positive. The higher the rate of popu- lation growth, the more investments are required merely to keep in- come per capita constant -- the so-called "demographic investment." Therefore the signs of the first components of the population effect are positive. As discussed in the previous section, except at fairly high income levels, i.e., a large magnitude of k, further increase in k might bring a decrease in population growth, so that in general the changes in population are positively related to changes in income. Therefore the population effect on investment demand with the exception of the highest extreme point is positive. The capital accumulation effect is negative, but as one would expect, it would be significant only when capital stock relative to population becomes redundant. On the whole, ai/ak is a positive and decreasing function of per capita 54 capital. Up to this point we have discussed the supply of and demand for investment in the commodity market. The final forms of the func- tional relationships are (35b) 3 s(k) (37b) 1 i(q, k) The equilibrium condition in the commodity market is that saving equals investment, i.e., s = i. This relation allows us to solve q in terms of k; that is (39) q = Q(k) Equation (39) represents the commodity market-clearing interest rate as a function of the capital-employment ratio. It is akin to the IS curve in macroeconomic theory. The slope of the curve is dq/dk = (ai/ak - as/ak)/ - ai/aq. Since ai/aq< 1, it is positive when 1>ailok>oslak>0 and negative when l>aslak>ailak>0. In order to obtain the equilibrium values of the capital- employment ratio and the interest rate, it is necessary to have a money market. The supply of money is assumed to be exogenously determined by the monetary authority, i.e., M8 = M: The demand for real cash bal- ance (MD), in accordance with the Keynesian theory, is a decreasing function of the intere"t rate and an increasing function of income. Meltzer, Bronfenbrenner, and Mayer have found, however, that the var- iable of nonhuman wealth better explains the demand for real cash balance in the United States than does the income variable.23 Accordingly, we may write the demand function as: (40) MD pL(q, k) when at equilibrium M'= pL(q, k). Eliminating p by substituting wage 55 rate equation (32) we have (41) L It is seen that the long-run equilibrium capital-employment ratio is at the point at which the "natural" growth rates of population and labor force equal the warranted rate of capital accumulation ad- justed for the supply elasticities of labor with respect to the real wage rate (<9), to the total size of the working population (U)), to the level and changes in economic structure ( a, b), and to the technical progress rate ( T). In the differential equation (46) all variables are functions 62 of k. Therefore the natural and the adjusted warranted rates may be plotted against k to obtain a diagram similar to the well-known Malthusian trap model developed by Nelson and Leibenstein (see Figure 3).24 With all other determinants specified in functions for fer- tility, mortality, and participation rate, namely equations (24), (26), and (29), remaining constant, with an increase in k from a very low level to a level corresponding to the subsistence level 5, the natural rate curve NN will increase rapidly from negative to zero mainly due to falling mortality. As k moves to the right of S, the NN curve con- tinues to rise at a rather fast rate until fertility and mortality reach the upper and lower physical limits respectively. Before these limits are reached, the participation rate is likely to rise from the bondage of low economic activity characterized by insufficiency of effective demand. However, the rise of the NN curve will gradually taper off as k reaches a very high level for the participation rate -- especially because participation of the female labor force will also approach the conventional limit. Similarly the adjusted warranted rate curve GG will also in- crease as k increases. The GG curve, however, is likely to be S—shaped. It tends to be higher relative to the NN curve when k is below and around the neighborhood of S, a point equivalent to the subsistence level. As k moves beyond 8 the growth of the CG curve is generally lower than NN at first and then after a certain point outstrips NN and finally reverses the trend at the high end. The main reason for this change is that the structural effect on the efficiency of capital utilization varies with the average amount of capital stock per worker. 63 gm.) k /, capital-employment zfi/zr ratio Figure 3. Natural Growth Rate of Population and Labor Force and warranted Rate of Capital Accumulation 64 When capital is scarce,an increase in investment is most likely to be directed to the deve10pment of agriculture because of the lack of the complementary factors in other industries. The gains in scale economics and in p0pulation-induced technological improvements in agriculture at this stage tend to surpass the rates of population and labor force growth.25 Yet these gains will soon be exhausted as the law of di- minishing returns becomes effective. This situation might change only after sufficient capital has been accumulated in the industrial sector. At very high levels of k, the negative income effect of goods for future consumption is likely to work against the rate of capital ac- cumulation. The graph in Figure 3 shows that point 8 on the k axis repre- sents the "low-level equilibrium trap," point U is an unstable equi— librium, and point H is a high—level equilibrium. The movements along the NN and GG curves are not hampered by the ratchet effect.26 In our model, k represents the productive capacity of the economy only, and the effects of other variables accompanying the economic development have been isolated. Hence, there is no ratchet effect. The changes in the latter variables will induce shifts in the positions of the NN and GG curves. As noted earlier, the interactions of changes in the variables of our system differ in strength and direction at various stages of development. A deliberate impetus, small or large, on a certain se- lected set of variables would thus produce a more influential force to bring the economy toward a higher level of equilibrium than would other selected sets. A critical minimum effort is not implied by our model to be a necessary condition for rescuing the low income countries from 65 the Malthusian trap. It should be noted that the growth of capital accumulation per employed worker is jointly determined by supply of and demand for in— vestment, labor force, and capital in the short run. There- fore, our model does not offer a direct comparison of the growth rates of population and labor force with capital accumulation. The rates of population and labor force growth have been explored to their sources: fertility, mortality, migration (allocation of labor force), and par- ticipation, and they are part of the system. Factors affecting the rate and efficiency of capital utilization are also incorporated. As a result, our model gets us a step closer to an explanation of "the variety of individual and institutional adjustments to problems of 2 economic growth." 5. ConcludingfiRemarks The economic-demographic model constructed in the present chapter is characterized by several features pertinent to the problems of the developing countries. First, the distinction of the conceptual time horizons, viz., short, intermediate, and long run.provides a configura- tion for cataloging the intricate economic—demographic relationships in proper places. Next, the model takes structural shifts and changes in economic activities, income distribution, consumption, and invest— ment patterns into consideration. Third, the full employment of labor is not assumed in the model; therefore, the problems of under- and unemployment can be analyzed in our growth model. Finally, the equality of saving and investment is not taken for granted; one of the bottle- necks of development, i.e., absorptive capacity, has been incorporated 66 in the model. One main shortcoming of the model is that it is less than real- istic because of the lack of a foreign market. In fact, this defect can be remedied by adding an input in the production functions and import and export relationships on the aggregate demand side.28 The inclusion of a foreign market makes the model more complicated, but the basic results of the original model are not changed. The second main shortcoming is that the money wage rate and relative prices are not endogenously determined in the system. It is possible to introduce a set of functional relations between money wage (prices) and wage (prices) determinants to complete the system. Since the inclusion of wage and price determinants will not affect the basic nature of the model we purposely left them out. On balance we have confidence that our model should be adequate for use as a framework to evaluate statistically the interactions be- tween population changes and economic development. Moreover, our model should be helpful in pinpointing relationships pertaining to a par- ticular developmental stage. 67 Footnotes S. Kuznets, "Demographic Aspects of Modern Economic Growth," Background Paper for United Nations World Population Conference (New York: United Nations, 1965), p. l. 2H. A. Leibenstein, Economic Backwardness and Economic Growth, Studies in the Theory of Economic DevelOpment (New York: John Wiley & Sons, Inc., 1957), p. 20; and nnmaAdelman, Theories of Economic Growth and Development (Stanford: Stanford University Press, 1961), p. 109. 3Cf. J. H. Harris and M. P. Todaro, "Migration, Unemployment and DevelOpment: A Two-Sector Analysis," American Economic Review, Vol. XL, No. 1 (March 1970), pp. 126-142. Irma Adleman, "An Econometric Analysis of Population Growth," American Economic Review, Vol. LIII, No. 3 (June 1963), pp. 314-339; R. Weintraub, "The Birth Rate and Economic Development: An Empirical Study," Econometrics, Vol. XL, No. 4 (October 1962), pp. 812-817; D. M. Heer, "Economic Development and Fertility," Demography, Vol. III (1966), pp. 423-444; R. A. Easterline, "Towards a Socio-Economic Theory of Fertility: A Survey of Recent Research on Economic Factors in American Fertility," in S. J. Behrman, L. Corsa, and R. Freedman, eds., Fertility_and Family Planning (Ann Arbor: The University of Michigan, 1969), pp. 127-156; J. L. Simon, "The Effect of Income on Fertility," Pppulation Studies, Vol. XXIII, No. 3 (November 1969), pp. 327-341; and Ila I. Ekanem, "A Further Note on the Relation be- tween Economic Development and Fertility," Dempgraphy, Vol. 9, No. 3 (August 1972), pp. 383-398. 5R. Freedman, "Norms for Family Size in Underdeveloped Areas," Proceedings of the Royal Society, B Vol. 155 (1963), pp. 220-245; and D. M. Heer, "Educational Advance and Fertility Change," Inter- national P0pulation Conference, London, 1969 (Liege: International Union for the Scientific Study of Population, 1971), pp. 1903-1914. 6 Easterline, "Towards a Socio-Economic Theory," pp. 127-156. 7T. R. Malthus, First Essay on Population 1798 (New York: Kelly, 1965); and A. Marshall, Principles of Economics (New York: Macmillan, 1961). 8G. S. Becker, "An Economic Analysis of Fertility," in Na- tional Bureau of Economic Research, Dempgrgphic and Economic Change in Developed Countries (Princeton, New Jersey: Princeton University Press, 1960), pp. 209-231; and D. Freedman, "The Relation of Eco- nomic Status to Fertility," American Economic Review, Vol. LII, No. 3 (June 1963), pp. 414-426. 68 9United Nations, The Determinants and Consequences of Popula- tion Trends (New York: United Nations, 1953), pp. 71-97. 10K. Krishuamurity, "Economic Development and Population Growth in Low Income Countries, An Empirical Study for India," Economic Deve10pment and Cultural Change, Vol. 15, No. 1 (October 1966), pp. 70-75; and I. Adelman, "A Quantitative Study of Social and Political Determinants of Fertility," Economic Developmant and Cul- tural Change, Vol. XIV, No. 2 (January 1966), pp. 129-157. 11United Nations, Determinants and Consequences,_pp. 47-69. 12United Nations, Methods of Analyzing Census Data on Economic Activities of the Population (New York: United Nations, 1968). 13H. Myint, The Economics of the Developing Countries (New York: Praeger, 1964), pp. 86-90. 14 J. Robinson, "Disguised Unemployment," Economic Journal, Vol. 46 (June 1963), pp. 225-237. 15Cf. H. Rose, "Unemployment in A Theory of Growth," Inter- national Economic Review, Vol. 7, No. 3 (September 1969), pp. 260-282. 16D. B. Suits, "The Determinants of Consumer Expenditure: A Review of Present Knowledge," in Impacts of Monetary Policy (New Jersey: Prentice-Hall, 1963), reprinted in W. L. Johnson and D. R. Kamerschen, eds., Macroeconomics,_Selected Readings (Boston: Houghton Mifflin, 1970), pp. 59-92. 17N. Kaldor, and J. A. Mirrlees, "A new Model of Economic Growth," Review of Economic Studies, Vol. 29 (June 1962), pp. 174-192. 18D. Freedman, "The Role of the Consumption of Modern Durables in Economic Development," Economic Development and Cultural Change, Vol. 19, No. 1 (October 1970), pp. 25-48. 19C. Ohlin, "Population Pressure and Alternative Investment," International Popplation Conference, London, 1969 (Liege: Interna- tional Union for the Scientific Study of Population, 1971), pp. 1703-1728. 20N. H. Leff, "Dependency Rates and Savings Rates, American Egonomic Review, Vol. LIX, No. 5 (December 1969), pp. 886—896. 69 21H. I. Grossman, "A Choice-Theoretic Model of an Income- Investment Accelerator," American Economic Review, Vol. LXII, No. 4 (September 1972), pp. 630-641. 22P. Demeny, "Demographic Aspects of Saving, Investment, Em- ployment and Productivity," Background paper for United Nations World Population Conference 1965, pp. 38—55. 23A. H. Meltzer, "The Demand for Money: The Evidence from the Time Series," The Journal of Political Economy, Vol. LXXI (June 1963), pp. 219-246; and M. Bronfenbrenner and T. Mayer, "Liquidity Functions in the American Economy," Econometrics, Vol. 28 (October 1960), pp. 810-834. 24R. R. Nelson, "A Theory of the Low Level Equilibrium Trap," American Economic Review, Vol. XLVI, No. 5 (December 1956), pp. 894-908; and Leibenstein, Economic Backwardness and Economic Growth, p. 22. 25E. Boserup, The Conditions of Agricultural Growth (Chicago: Aldine, 1965). 6 . . Myint, Economics. 7Kuznets, "Demographic Aspects," p. 22. 28C. Khang, "A Neoclassical Growth Model of A Resource—Poor Open Economy," International Economic Review, Vol. 9, No. 3 (October 1968), pp. 329-338; and P. Bardhan and S. Lewis, "Models of Growth with Imported Inputs," Economica, Vol. XXXVII, No. 148 (November 1970), pp. 373-384. CHAPTER III. ECONOMIC DEVELOPMENT AND POPULATION IN TAIWAN SINCE 1895: AN OVERVIEW The history of economic development and population growth in Taiwan since the turn of the century is characterized by a transition from a traditional to a modern industrial society. There are dif- ferent interpretations as to the immediate causes of this transition. The basic underlying factors, however, may be the level and pattern of aggregate demand. As the level and pattern of aggregate demand change, the extent and structure of economic activities adjust ac- cordingly. These responses affect the social and economic character- istics of the population which in turn influence the growth and dis- 'tribution of the population. Again, the changes in size and character- istics of the population affect the level and pattern of demand. The last causal relation has been demonstrated by the recent experience of the developing countries. Most of the economist's concerns are thus focused on this relation. Studies that trace the relation back to aggregate demand are rare. The long-run historical evidence in Taiwan, however, shows that the level and pattern of aggregate demand have played a curcial role in the course of economic and population growth. This may be due largely to Taiwan being a small island and having to rely on the international market for its development. To be more specific, Taiwan is a mountainous island of only 13,885 square miles. The economic value of its mineral deposits is negligible. 7O 71 Arable land is ranked as the best endowment on the island. However, it consists of only one-fourth of the total area. Furthermore, the quality of the arable land is poor, and the rainfall is unevenly dis- tributed throughout the year. Because of the absence of natural wealth, the development of Taiwan has to depend on the mobilization of human resources and on the accumulation of material capital. The extent and methods of mobilization and accumulation are closely associated with changes in aggregate demand. Since domestic demand is limited by the size of the domestic market, international demand inevitably plays an important role. International demand is primarily subject to changes in external conditions and domestic demand is influenced by the stages of socioeconomic development and government policies. Taiwan's economy has experienced two distinct patterns of ag- gregate demand, one traditional, the other developmental. The tra- ditional type is characterized by agriculturally oriented demand. This traditional pattern occurred during the Japanese occupation period (from 1895 to 1945). The developmental pattern of demand may be further divided into two stages: labor-intensive, and skill- and/or capital-intensive. Both are guided by the forces of the free market. Taiwan has experienced a labor-intensive type of development in the past twenty years. It is expected that the skill— and/or capital-intensive type will predominate in the near future. This Chapter discusses the historical development of the two types of aggregate demand; their effects on economic activities and thus on population characteristics, growth, and distribution; and the respective feedback of each on the others. (.‘J It. '. (T- ‘n A.“ .,.\ (h 72 l. The Colonial Period: 1895-1945 When Japan took over Taiwan from the Manchu government in 1895 Japanese authorities decided to develop Taiwan as a supplier of agri- cultural products for her industrial deve10pment. Not until the Japanese empire began to prepare for war in the middle of the 19303 was Taiwan's economic role changed to that of a supplier of industrial goods as well. This policy is reflected in records of exports and imports. Statistics on external trade compiled by the Japanese reveal that dur- ing the colonial period most of Taiwan's trade (85 percent of Taiwan's exports and 75 percent of her imports) was with Japan and her overseas possessions. As shown in the upper panel of Table 3.1, of the total value of exports to Japan prOper the share of agricultural and processed agricultural products amounted to 70 to 95 percent. Except for the first decade of occupation only 55 to 90 percent of the claims for exports to Japan was paid back to Taiwan in the form of commodities (see the upper panel of Table 3.2). Only one-third to two-fifths of imported conmodities were industrial consumption goods; the remainder were mainly used as a means of producing agricultural surplus. For instance, the bulk of the imported raw materials was chemical fertilizer, which was distributed by government agents to farmers in exchange for agricultural products at an extortionate rate. The im- ported capital goods were mostly transportation and conmunication equip- ment which was used to facilitate the movement of raw materials and products within the island. Until the period of Japanese preparation for the Pacific war, the pattern of exports and imports changed only slightly. Imports of capital goods, mainly plant equipment, increased at an average growth FUNgvd av. Qwerrhu I.. Unnu|hmo cam coaumumaooo afiaocoom HmooaumcumuaH pom Hwoaaoo Scum woaadaou uofiumm kuaoaaoam>ov you puma .nqma .Hodama .muaumfiumum mo uaosuumnon amped» Hm ummm saw you sesame mo huwaESm Hmuwumaumum 73 .mofiumwumum mo usoauumama Hmwocw>oum =m3HmH aoum poaaafiou coaumm HmGOfiuapmuu “0m moon "mmUMbOm .haao smash ou monomxmw .voauoa Hmuaoamoam>mv you mumaaoo .m.D mo m28S0£u "UOHHmQ HQGOHufivau How Gwen mo wENmSOr—u “wean—ma £08 How mmmuw>m .AHHNONH m.sm m.a m.e m.om ~.wa e.~a H.m «meqaomwa okaa H.~m m.m m.am m.om m.~e m.H~ a.mH wom.ase mosaumeaa ~.moa a.m a.m e.mH m.ms H.ms «.ma 50¢.mom seaa-oeaa o.ms s.m m.ma a.a e.ma a.so o.H~ os~.~ma ammaummaa - - - I 0.0 H.ma «.ma ~a0.maa smaanwmaa uoanom Hmucoaaoam>oa o.~ m.a w.H m.o m.aH «.mm m.m~ oma.nmm msmauosma s.sm m.a w.~a o.ma a.oa a.mm s.m~ oss.~os amaa-mmaa a.m m.H m.o c.H ~.m o.me m.Hm as~.om~ smma-omma H.a H.~ H.am ~.o a.e o.~s m.mm mmo.ma~ ommaummaa e.m m.aa e.m ~.mH m.HH m.ma o.mH mea.~oa smaa-omaa m.m~ n.a~ s.mH m.- N.oH s.so «.ma NmH.mm mamanmaaa «.ms H.a~ n.m o.- s.mH m.me w.o~ ems.os «Haauoaaa m.a m.oo N.mm a.mm s.m H.0m m.sm a~m.a~ somaumoma - - - . m.Hm a.os m.n~ mea.o soaauamwa Naoaumm Hmaoauaamha mumuduumm muusmuoum muUDVOHm muhomunm mHMHDuomm muonvOHm muonvoum mDHm> UOfiHmm IDGmZ HNHDUHDU HmHDUHDo HmuOH. ISGNZ HmufiuHSU HNHDHHSU H -auma nauw< -auw< nauwa Bmmmuoum $9.50 wwwwmuoum 09.50 nuaouo mo oumm Hmscc< madam owmusoouom onma ou mama .memoaxm $355. 5 mmozss maefime fin mama .H.m magma How we deem use ammomDom .maao comma scum muHOQEHN .woauod Hmucoadoao>oo you mumaaow .m.D mo monomaosu macauod Hmcoaufiumuu How ooh mo madamaonu “weapon some now owmuo>m eased» 74 0 0.00 0.00 0.00 0.00 0.0 0.00 0.00 00.0 0004000q0 0000 0.00 0.00 0.00 0.00 0.0 0.00 0.00 00.0 000.000 0000:0000 0.00 0.0 0.0 0.0 0.00 0.00 0.00 00.0 000.000 0000:0000 0.0 0.0 0.00 0.0 0.0 0.00 0.00 00.0 000.000 0000:0000 : : : : 0.00 0.00 0.00 00.0 000.000 0000:0000 ooauom Housmaaoam>ma 0.0 0.0 0.0 0.0 0.00 0.00 0.00 00.0 000.000 0000:0000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00.0 000.000 0000:0000 0.0 0.0 0.0 0.0 0.00 0.00 0.00 00.0 000.000 0000:0000 0.00 0.0 0.0 0.0 0.00 0.00 0.00 00.0 000.000 0000:0000 0.0 0.0 0.0 0.0 0.00 0.00 0.00 00.0 000.00 0000:0000 0.0 0.00 0.00 0.00 0.00 0.00 0.00 00.0 000.00 0000:0000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00.0 000.00 0000:0000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00.0 000.00 0000:0000 : : : : 0.00 0.00 0.0 00.0 000.0 0000:0000 0000000 00000000000 mooou mamauou mvooo monomaH mooou mamaumu mooou muuomaH maam> voaumm nowadasm lea Hmuanmu Hmuoe coaudsbm Ina Hmuadmo ou H Icou 3mm Iaoo 3mm muuoaxm 00 000000 nuaouo mo comm Hmsnc< madam ommuamouom ohma 0» ommH .mHmomZH m.z¢3HHHoum cmaame aoum ooaumd Hmucoaaoao>mo you damn .wuwow Hm ummm Ono you cmawme mo mumaanm HOOfiumHumum aouw voauma HmaOHuwvmuu now mama "mmumsom 76 0.00 0.0 0.0 0.0 00.0 0000000 0000 0.00 0.0 0.0 0.0 00.0 000.000 0000 0.00 0.0 0.0 0.0 00.0 000.000 0000 0.00 0.0 0.0 0.0 00.0 000.000 0000 0.00 0.0 0.0 0.0 00.0 000.000 0000 vowumm HmucmeaOHO>OQ 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 0.00 00.0 0.0 0.0 0.0 00.0 000.000 0000 : 00.0 0.0 0.0 0.0 : 000.000 0000 000000 00000000000 mama A.m .mv Amcv mmu< mumxuo3 mumpewz N puma nonsaz ooauom mo mumsso Hmuaamo mama sump oaon OmmouOcH Hst mum 0:3 voxfim vOum>0uH=o Immaom Hmscc< muoum>0uHso "mo 800m you Owduo>< maumh oaonowsom mo mwmucmoumm mmHmmonHse m.H o.« «.0 o.mo« m.m0« oema 0.0 o.« 0.0 m.na« m.mm« mmma 0.0 0.« 0.0 n.mmH o.mo« omma m.o m.« 0.0 0.000 0.000 m«mH m.o m.« w.o H.000 0.0«H ««mH m.o o.« 0.0 m.mHH m.««H mama «.o o.« 0.0 m.moa m.moa momH 00.0 0.0 m.o 0.000 0.000 momH voause HsGOHOHvsuH wwwao Ease :I~.snv A.s;v Base xsvsH Hume xsocH ems» use ass: Base use use ssu< use useuao uaeuso 0s0000uuse ssn< eouo osusw0uuH Hondu03000m< asusuassfium< HsOHEsnu mHszH Qz< mHDmHDO A .socs0om 0s0oom mo 0sau560 :.0sausoaou s0s>uam usmvam 000ase mo suuoesm ssu aoum assm 00a0sz .om30sa :0 musause uasssse mo 060:00s>00 s59: .mssnu syncs: “mumDOm 0.0« 0.0 0.00 0.0 .0.00 0.00 000 «000I0000 m.0« 0.0 0.0 0.0 0.00 0.00 000 0000:0000 0.00 0.0 0.0 «.0 0.00 0.00 000 «m00l0m00 mso0>usm usu0s£m .ud0nu00u ooosnpw 000m :0 0suoa vo0use 0cm ousuoe usssm Asso0ue 0m00v mswsus>sm 0 so0m N 000ase U8“ Eflrfl HOW msunu006sexm msuau00asmxm mo sufuasouse mMMbHanmexm ~0020uom 000so0aosoum 000.0 000 000 000 000 000 000 000 0000000000 000000 000000000000 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0000 000 00 0000 0000000000 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0000000000 00000 s0sase 0.00 0.00 0.00 0.00 - 000 00 00000 00 00 0.00 0.00 0.00 0.00 . 00000 00 sw00 ws0xuoa mo muss» usz 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0000 000 00 0000 0000000000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0000000000 00000 “mo N as .eoe s>0uos 000so0aoaoom 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0000000000 0>0000 000000000000 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0000 000 00 0000 0000000000 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 0000000000 00000 0000 0000 0000 0000 0000 0000 0000 0000 0000 ZOHH<0=mom m>HHU< V000usosxm .mO0um0umum 0am mucsoou< .muswvsm mo 0susss01susuouusu00 aoum «000 so 0sssm quM0Om 00.0 00.0 00.0 «0.~ «0.« 00.0. «0.« 00.0 «0.0 00.0 00.0 00.0 00.0 00.0 00.0 0«.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 «0.0 ~0.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 «0.0 «0.0 00.0 «.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.« 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ~.« 0.0 0.0 0.0 0.« 0.0 0.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0 0.00 0.0 «.00 «.0 0.00 0.00 camaso sumwsuww< 0suoe suuoexm GO0usauoe 0su0eso 0sx0e smouo :o0umo50m £u0ssm GO0usuum0c0a0< cO0uessmcoo unsecus>oo co0usosum nu0msm mso0>usm usu0snm 00000000 vooe :00ueasmaou sus>0ue so0ue a0 smsszu s>0us0sm .mwcsnu so0ue mo summ 0ssac< 0010000 0010000 0010000 0010000 0010000 0010000 0 0 0 0 0 0 0010000 0010000 0010000 0010000 0010000 0010000 mozHH¢0mm Qz< moz000 0.0 0.00 00. 0.0 0.0 0.00 0.00 0.00 0000000 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 00000000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00000000000 0000000 mso0ue 0000 ucsumaou u< mso0ue ucsuuso u< swsussouse 0000 0000 0000 0000 0000 0000 0000 0000 0 0 0 0 0 0 0 0 0000 0000 0000 0000 0000 0000 0000 0000 Qz<2ma me¢wmmow< m.z¢3Hav %a vmusaaoo mm: vomoaaam uaoam>fioom oafiunaaama .muuomuM coaumuumawmm any sown mumv mouou Honda scum doauoaoaoa Hmuzuasuauw< .xoom snow Hmuouasowum< .NH.n manna now no muuzom 06mm 0:» 80pm sumo maoqu "mmUMDDw N~.N wo.~ HH.~ mw.H .Hwaaoz as agooaH auumaoum cu maoucH Honda mo owumm om.o and 36 36 .5925 8 .uw< aw oEoo:H muammo pom mo oaumm m.a ~.m ~.~ Haa.HH Hmm.m omo.o ~wm.m unsussosummaoz H.n ~.H H.H Hmm.n mmo.m Hmm.~ m~a.~ musuasusum< ".0 m.s o.~ mum.mH smo.HH Ham.m “oh.” Awsz cease meoucH muaamo “we o.MH A.OH m.a omm.oo mmH.mn moo.om waa.e~ mucuflauaummaoz a.s o.s m.m NNH.H~ mmo.afi BHN.QH oma.HH unauflsuauma o.HH w.m s.o NOH.~HH www.ma o-.om moa.wm Awaz seas coaHHHz .uonmq was huumaoumv maoucH m.o ~.s s.m mos.“ smO.H Nmm.H anH.H mundasoaummcoz m.ma c.ao H.oa q.se mumxuoa .auwm mo N as «comednafi uamHm>Havm mafiuuaasm a.s N.H H.~ onc.a mmm.H sw~.H OOH.H sosoaaau sauam>asuo maau-asam a.H m.H m.o HsH.N moa.a ~mm.a Nom.H unauaauauw< m.m s.~ m.H som.s moo.m st.m mnm.~ AoooV «upon bonus 0.« m.m ~.s esa.a Nms.o Nam.n osm.s .aoa .uwmaoz s.H ~.~ m.~ Hmm.m amn.n «ma.s w~m.e .eoe Haususaoaum< w.~ m.m o.m has.ma Hmo.~H ssm.oa «ma.» AOOOV aoauasssom o~-woos neuwoaa oo-wmma owns mwms owns mmua moufloaa eonomma mmuflmma seas Hams omaa Haas mmoeomm SBABHMUEOZ 924 SBABHMU¢ ma ZH NED—AH Qz< .momom moms .onthmom 21m H.548 108 TABLE 3.16 VITAL RATES FOR AGRICULTURAL AND NONAGRICULTURAL POPULATIONS 1956 . 1961 1966 I I I 1969 1965 1970 Agricultural Population Birth Rate 42.1 37.5 35.5 Death Rate 8.4 6.7 5.9 Rate of Natural Increase 33.7 30.8 29.6 Net‘Migratioanate 8.3 3.8 16.0 Rate of Increase 25.4 27.0 13.6 anagricultural Population Birth Rate 39.5 33.5 27.1 Death Rate 5.9 5.0 4.6 Rate of Natural Increase 33.6 28.5 22.5 Net Migration Rate 13.4 10.5~ 17.4 Rate of Increase 47.0 39.0 39.9 SOURCES: Based on Registration Reports for the respective periods. PERSONS BY STATUS TABLE 3.17 COMPOSITION OF EMPLOYED Self-employed Unpaid Paid Total Employers Workers Family workers Employees Male Agriculture 1956 100.0 1.8 56.6 30.7 10.9 1966 100.0 1.3 57.7 31.9 9.1 1971 100.0 1.1 56.8 24.2 20.9 Nonagriculture 1956 100.0 5.6 20.3 4.1 70.0 1966 100.0 3.8 12.4 2.7 80.1 1971 100.0 4.6 19.2 3.1 73.1 Female Agriculture 1956 100.0 0.6 17.2 74.8 7.0 1966 100.0 0.5 23.2 70.5 5.5 1971 100.0 0.2 11.8 76.4 11.6 Nonagriculture 1956 100.0 2.8 14.7 10.2 71.0 1966 100.0 3.1 11.6 8.3 77.0 1971 100.0 1.8 12.4 16.0 69.8 SOURCES: Data for 1956 and 1966 taken from the Population Census Reports for the respective years. Data for 1971 taken from Taiwan Provincial Social Affairs De- partment, Labor Force Survey, 1971. 110 In addition, the rapid expansion of the nonagricultural sector (the gross domestic product grew 12 percent per year in the period from 1951 to 1970) raised the per capita income of nonfarmers about 6 percent per year on the average. Consequently, the difference in earnings between the two sectors increasingly widened (see Table 3.15). This constituted a force pulling rural workers from their usual economic activities. Both the push and pull forces acted to accelerate rural- industrial mobility. The problem of urban unemployment recently prevailing in most of the developing countries, especially those in South America and Africa, did not seriously strike Taiwan. The basic reasons why Taiwan was able to evade this plague may be explored from the agricultural and nonagricultural sides. In the 19508 Taiwan was confronted with unemployment problems of the same nature as those faced now by most of the developing countries, because the initial phase of "demographic revolution," i.e., constant high fertility matched with declining mortality, had taken place in Taiwan about twenty years earlier. The fact that this did not turn out to be a disaster was largely due to the implementation of land reform, serving as a buffer in the course of readjusting the economy. Thereafter, the rural sector benefited so that it was able to provide the redundant labor force with adequate education or training while withholding workers until nonagricultural employment was available. In this, economic cal- culation was not totally absent. There was a tendency to coin a norm with regard to the newly emerged problem. In general the eldest son on the household farm took up the family business with full-time participation anm.8upported younger brothers who continued their formal education or training beyond the compulsory requirements as far as their abilities 111 permitted. The younger brothers helped with farm chores in the busy seasons, and usually reported as farmers if they were not enrolled in a formal school. After they had been employed or had established themselves off the farm their income and prestige were higher than those of siblings remaining on farms. Economic compensation to the latter was made more or less according to their losses, although usually not explicitly in the form of cash payments. The common practice was that those who worked on the farm had the privilege of acquiring the absentee brothers' shares of land with little or no payment in return. Those who left the farm also took a large share in supporting their parents. These in- formal arrangements helped to solve the problems of fragmentation and the uneconomical size of land holdings on the one hand and facilitated a less painful approach to the underemployment problem on the other. 'Taiwan was the most densely populated country in the world, 408 persons per square kilometer in 1970, and its cities were crowded. The expansion of industries in or around urban centers therefore was pro- hibited by an exceedingly high land price. On the contrary, there were some advantages to locating industries in the countryside. First, land prices were relatively lower. Next, transportation on the island was convenient and inexpensive. Third, location in rural areas gave ac- cessibility to a labor force. Last, a government policy of preparing industrial sites and infrastructures throughout the island to aid local development provided facilities which would usually be found only in urban areas. As a result, most of the new manufacturing plants were constructed in the countryside, hence opportunities for nonagricultural employment in rural areas increased with the progress of industrialization. The primary effect of the widely dispersed industrialization 112 was that the change from traditional to modern economic activities was necessarily associated neither with change in residence from rural to urban, nor with the need to break cultural and social contacts in the rural community. According to two consecutive surveys on rural labor mobility, among those who had nonfarm jobs, the conmuters, i.e., those who worked in the nonagricultural sector but stayed with their parents or brothers on a farm, increased from 20 percent in 1963 to 48 percent in 1968.8 This form of transformation, in contrast to rural-urban migration, greatly alleviated the urgent pressure on resources for housing and public services in the cities. Instead, a gradual improvement in living conditions was carried out in accordance with the individuals' and communities' abilities enhanced by an increase in productivity. More importantly, the commuters brought new ideas and values from their places of work to rural society, causing a gradual adjustment with little conflict between the modern and traditional personalities. We have noted that under the present social and economic ar- rangements, economic considerations have been gaining importance in the determination of the size of the family. It is postulated that in the earlier period of demographic transition the change in attitudes and hence behavior did not involve a confrontation with the traditional familial values, i.e., having three or four children with at least one or two sons. As mortality had declined steadily for some time, the ;persistently high fertility produced more surviving children than were needed for the ideal family size supported by social values. The economically justified small families could therefore be achieved by tliéa control of excess births. A8 economic development proceeded eco- noun 1c aspiration intensified as a growing population (resulting from 113 traditional values) still constituted a threat to economic welfare,so that a change in attitudes toward ideal family size occurred in due time. Statistical evidence, contained in Tables 3.18 to 3.23, appears to support this argument. First, there was a continuous marked decline in mortality rates, especially for the younger age groups, during the period from 1950 to 1970. Then, after a short time-lag, a fertility decline occurred among women of almost all ages. The rate of decline was increasingly high for women over thirty years of age. This de- cline, as shown by Table 3.20, is partly attributable to the decline in the proportion married in younger age groups, but mainly attributable to the sharp decline in the rate of births to married women over thirty. The decrease in the number of wives who wanted no more children while using no contraception was obviously related to the decline in the marital fertility rate. The sluggishness in the decrease in the number of children desired by wives after a substantial fall in fertility seemed to imply that economic pressures had not yet gained their full momentum. However, the educational and rural-urban differentials in ideal family size in 1970 are a signal that traditional values are giving way to rationalized calculations. 3. Summary The preceeding sections demonstrate that the level and pattern of aggregate demand dominated the relationships between demographic trends and socioeconomic development of Taiwan for the past seventy years. There were two distinct types of aggregate demand: the tradition- al occurred during the period of the Japanese occupation from 1895 to 1945; and the developmental type, continuing at present, developed at ] 1 " " I1- IIIII‘ IL‘ :I‘IEE 114 TABLE 3.18 AGE-SPECIFIC DEATH RATE FOR EACH SEX AND PERCENT CHANGE FOR TAIWAN, 1950-1970 _Age-Specific Death Rate 1950-60 Percent Change Age 1950 1960 1970 1960-70 1950-70 Male 0 41.56 34.11 10.49 -l7.95 -69.25 -74.76 1-4 24.28 7.55 2.28 -68.90 -69.80 -90.61 5-9 3.24 1.23 0.71 -62.04 -42.28 -78.74 10-14 2.34 0.92 0.62 -60.68 -32.61 -73.50 15—19 2.21 1.36 1.26 -38.46 - 7.35 -42.99 20-24 3.47 2.18 1.82 -37.16 -16.51 -47.55 25-29 4.26 2.42 2.02 -43.19 -16.53 -52.58 30—34 5.44 3.03 2.54 -44.49 -15.89 -53.31 35—39 6.31 4.06 3.20 -35.66 -21.18 -49.29 40-44 10.44 5.82 4.13 -44.25 -29.04 -60.44 45-49 14.01 7.74 6.52 -44.75 -15.76 -53.46 50-54 20.44 13.23 10.36 -35.27 -21.69 -49.32 55-59 30.82 21.30 15.02 -30.89 -29.48 -51.27 60-64 43.62 31.57 28.12 -27.62 -10.93 -35.53 65-69 58.04 51.33 44.12 -11.56 -l4.05 -23.89 70+ 127.75 112.76 96.59 -11.73 -14.34 -24.39 Female 0 38.03 30.66 9.73 -19.38 -68.26 -74.41 1-4 26.35 8.24 2.05 -68.73 ‘-75.12 -92.22 5-9 3.53 1.02 0.53 -7l.10 -48.04 -84.99 10-14 2.04 0.66 0.46 -67.65 -30.30 -77.45 15-19 2.24 1.26 0.73 -43.75 -42.06 -67.41 20-24 3.49 1.79 0.94 -48.71 -47.49 -73.07 25-29 4.23 1.95 1.15 -53.90 -41.03 -72.81 30-34 5.16 2.44 1.61 -52.71 -34.02 -68.80 35—39 6.26 3.06 2.00 -51.12 -34.64 -68.05 40-44 8.85 4.02 2.75 -54.58 -3l.59 -68.93 45-49 10.95 5.05 4.14 —53.88 -18.02 -62.19 50-54 15.84 7.99 6.60 —49.56 -l7.40 -58.33 55-59 21.32 12.38 9.22 ~41.93 -25.53 -56.75 60-64 30.18 19.59 17.20 -35.09 -12.20 -43.01 65-69 41.46 32.30 27.98 -22.09 -13.37 -32.51 70+ 96.78 88.00 79.28 - 9.07 - 9.91 -18.08 SOURCES: Based on data from Demographic Fact Book 1970 and Taiwan Demography, Vol. 5, No. 7, July 1970. .OBGH Ou HO¢H.¢&OO¢ UUflh UfiANGHMOEUQ BORN fluwfi GO UOQO umMUMDOm 115 a.mm- a.aa- a.oa- a.aau m.m ooo.s mam.s oma.m omm.o omo.o mama auaaauama asoa a.am- a.aa- o.ma- o.m . m.m oaa ama oma ama ama mama awaaauume kuwfimu o.om- o.on- w.mm- o.msn 5.8a- m 0 ma ma on maums a.~m. a.anu a.ms- m.mau o.w - oa as ma moa Naa sauce a.aou o.aa- w.os- m.-u a.aa an ooa sea aaa ama mmumm «.ms- o.s~- m.m~- o.man a.m asa mma mma mam sea «anon m.a . a.oa- a.~ . m.~ . m.sa mom can mam asm ama manna m.m - w.» . ~.m m.a u o.aa mma aoa mma maa sea sauoa s.smu a.aa- o.m~- o.s . o.man as on we on as sauna cauomaa caumoma noncoma ooummaa mmuomma oama mesa oaaa mama omma mw< moumm a.“ afloafiu mumuowoumm awn—03 oooqw pom 3mm osmalommfl .zstflH mom muzfimo Hzmofimm a 2563 AA< mom mmausm oouom wanna any so muHOQUMJNauouumad scum one monomou noooou coauoammom omma souu ozone ouoo sown vousmsoo "anomaow A.ooou Housmaoa any Esau ceauoauousa manu woaumaooou wow camconuox .: on saw one oxoonhv om.N anc.om am wouuooouw owoaaoo om.a nom.oc oca couscouuw aoocuo swan uoaoom Na 8%: 3a .333...» aooao. a»: sauna. oo.a 202 an 63263» aoofiu Sneak No. nmm.na mm coauouovo amah0w oz unmask may: momav maouoa mMMOmuOQAQwM. owouo>< aooood condo mo .02 "ozoaaou on cue oucwaoa moo mwsaouoo use .mo>u=o muaaauuom no mouwaouuoo uaaosoom ooazlocmama so aouu moma 8a omaouu soauouovo coozuon mwoasumo amaucouowuav mo manna onu so ooumaauma mum mucwaoa soauoosvo echo .oomoaoau mama a mo uocu «o m. mo unmaoa w muoxuoa haaamu mamas: one .a.a mo unwaoa I muoxuoa vomoamsonuaom .momoamsu vase a as once «.a ou assoc unwaoa w nuumoaoao mo>aw ooauom manhu .ucouuom as No mamas: u nouoom amusuasoaummcoc ocu ca muoxuos oamsow mason one .oama uaooo no mo mono uo ucoouom om mo unwaoa o wouuoo amusuaooauwm ocu ca muoxuoa «anamu uaooo use uo>o moo mo muoxuob mama .oana mason on no mono ouaculuoo ou amavo ucwamz a uo>o mam mo muoxuoa 0a860u can and mo sumo» on wove: mhoxuoa mo>aw moauoo masks .muoxuoa oomOamso no woman: oo>uouoo amuOu ozu Ou moaumauouomuocu uauaoomo NO 888 moonwaoz 0:» mo oaumu ocu ma sauna unusumsnoo span No.a No.~ mcom. namN.a amwm. mums. oooo.a mmmm. ncmm. mcmw. occm. mcom. ohma oo.a na.~ wcmm. wouu.a mowm. Nwmn. oooo.a «cum. Nmmm. acNm. mach. amon. moma ch.a ha.N omnm. -o~.a comm. moan. oooo.a wmmm. mon. mmaw. anmn. «Now. woma oh.a a~.~ mcwm. c~o~.a mmmm. comm. oooo.a ooom. cham. Ncaw. mmcs. mmma. noma mo.a o~.~ Nomm. com~.a oamm. mmom. oooo.a mmmm. oaam. awmw. Noun. mocn. ooma an.a sn.~ maom. nmaN.a onwm. Naom. oooo.a c~om. ammo. no~w. cmoh. maNN. moma nm.a ow.~ wmmo.a coma.a mmsm. mmww. oooo.a comm. camm. mama. Noon. Omoo. omma v. n v. n n t o a .3 a. a .6. a. a .. 0ft 0 2 ft 0 ft v. o a.w 4. mm. o a 1. my» 0 u to. 2 e t B 2 t S 18110.8 1 a on 1.1 1 a on 1.1 .1 n e.1 n a c .1 t .D e.i a c 1.r. .0 enu r 3... 33 a a ma a; 3 a a pm a; a... a Wcumiwuac t e as as lu t e as as In v.. ooumoaom muoxuoa huaamrd. .Nwaamnd aououaouaumm ou you xooCa ucoEumofio< new oxooca ucosumofiw< muoxuo3 amuouaso o lauwocoz mo Oauom ousuasuauwmdoz uuouaouauw< maoaumm a<¢DHADUa¢oma mom mmxma2a Hzmthbhn< a.c manand have benefited from investing in farmland. These special privileges indeed enhance their incentives to invest in land and land improvements. Investment in land, which is a kind of transfer pay- ment, has no direct effects on land productivity. Investment in land improvements is relevant to marginal product of capital. Therefore, we shall look into its magnitude and nature to understand the trends of marginal product of capital in agriculture. By law the size of landholding of each farm cannot exceed a maximum of 2.8 hectares of paddy field or the equivalent. In fact, due to the increasingly high population pressure on land the average size of farms had already decreased to 1.4 hectares in 1951 and in 1970 was only about one hectare. The smallness of farm size hinders the develop- ment of a modernized and mechanized agriculture and thus limits the agricultural capacities to absorb investments. At the same time small- scale cultivation also limits farmers' ability to accumulate capital. 155 In the first ten years after the implementation of land reform the ability to invest in land improvement was especially weak due to installment pay- ments on land purchased from former landlords by most of the farmers. In addition the accessibility of modern financial markets to farmers was not fully develOped. Under these special institutional conditions demand for and supply of private investment in agriculture, being isolated from the modern sector, equalibrated at a low'magnitude. This small stream of investment over the period under review brought only an insignificant pressure on the diminishing returns from capital formation. Therefore, the increase in complementary inputs, especially in the labor force, and technical progress produced a counter force sufficient not only to overcome the diminishing returns from capital but also to raise pro- ductivity of capital in the agricultural sector. b. Resource Allocation Equation From the above discussion of the marginal product of capital we found that perfect competitive conditions of the capital market were partially distorted by the enforcement of the land reform program. Therefore, irrespective of the fact that the marginal product of capital was higher in the agricultural sector than in the nonagricultural before 1965, the rate of capital accumulation in the nonagricultural sector was substantially greater (see Table 4.2). Capital market distortion had some beneficial impact on the economy because it diverted scarce resources to industries with great potential for develop- ment. Once the marginal product of capital in nonagriculture surpassed that in agriculture, as happened around 1965, this distortion would enhance the competing power of the nonagricultural against the trickle of investment in the agricultural sector. lack of statistical data prevents us from a detailed quantitative analysis of the problem. Our 156 evidence on the sluggish increase in the rate of capital accumulation in agriculture in contrast to the rapid increase in nonagriculture lends support to this point. As noted in Chapter III the implementation of land reform and use of labor-intensive technical innovations in the agricultural sector alleviated the underemployment problem in Taiwan by withholding a large part of the labor force. These measures were beneficial to development of both agriculture and nonagriculture as long as there was a rural labor surplus. In fact, only a few years after the beginning of economic growth, rates of labor absorption in both sectors were greater than the rate of increase in the labor supply. As a consequence, the surplus rural labor was gradually brought into effective use and the influence of the relative labor scarcity became increasingly great. The general influence of these agricultural policies on economic and population growth has been discussed in Chapter III. Here we shall concentrate on statistical analysis of their effects on labor allocation between sectors, that is, rural-urban.migration and labor mobility problems in the field of population studies. Our calculations in Table 4.2 show that (1) in 1951 marginal products of labor input employed in agriculture and nonagriculture were about the same; (2) in the process of economic growth agricultural labor productivity remained constant while that of nonagriculture tripled by 1970. The first part of our finding seems to agree with the actual situation of the developing countries in the initial phase of develop- ment. Research on labor productivity finds that industrial employees at this stage are characterized by low morale, lack of discipline, and high rates of absenteeism and turnover. As a result, their productivity is not likely to be better than that of their counterparts on farms. In 157 the case of Taiwan, conditions were slightly different. Colonial control and extension of primary education had corrected most of those deficiencies, but the same factors had also equally affected farmers' attitudes toward work. Therefore, no difference was found in labor productivity between the two sectors. This level of marginal product of labor was the result of a long period of economic stagnation. It may be looked at as a long- run equilibrium institutional rate of compensation to labor at which people are willing to render their services to production in either sector. The explanation for the second part of our empirical findings may also be found in institutional distortion originating from land re- form. Unlike capital, the effects of land reform on labor allocation are in favor of labor-intensive agricultural development. In the process of economic expansion during the period from 1951 to 1970 Taiwan's agri- cultural growth benefited from the accessibility of the source of surplus labor; in fact, it consisted of the farmers' own family members. Before the depletion of the surplus labor pool any increase in demand for labor could be readily met with a supply of labor at a constant institutional rate of compensation. For modern industrial expansion, however, the source of additional labor was not so conveniently located: it consisted of the rural surplus labor. Transfering from rural families to industrial employment generally involves a rise in real wages because of structural distortions. Reasons for structural distortions commonly found among other developing countries again are not fully applicable to Taiwan. For instance, government wage policies and trade unions in Taiwan had virtually no effect on the labor market. Throughout the period under review the minimum annual wage rate promulgated by the 158 government remained at NT$ 6,000 (at 1966 prices). This was less than half the lowest marginal product of labor in 1951, therefore it had no effect on wage determination. Trade unions in Taiwan also had no effect on wage rates for they never functioned in wage bargaining. Lack of accurate knowledge of job opportunities probably was a deterent to rural labor movement in Taiwan, for statistical evidence shows that in 1967 85 percent of the migrants to Taichung City received job information from friends and relatives and much of it was of a general nature.14 Government and private employment services and modern mass media, how- ever, were widely used toward the end of the period under study. Insti- tutional distortions caused by land reform therefore remain the primary explanation. Under the land reform system, (1) rent from tenanted land after the deduction of taxes was not profitable and had been becoming less so rapidly because it was fixed at 37.5 percent of the 1948 average yield, while taxes paid for owing land were assessed according to in- creasingly high land productivity; (2) tenantship is inheritable and cannot be terminated without the consent of tenants or their heirs; (3) farm land can be purchased only by those who are able to cultivate the land themselves, and the size of a holding shall not be more than 2.6 hectares of paddy field; and (4) farmland cannot be changed to other uses without government approval and without agreement between tenants and landlords if the land is leased. A cursory review of these arrange- ments show that tenants have enjoyed implicit subsidies for holding tenantship, while owner-cultivators bear losses if they choose to sell or lease their land. That is, farmers entitled to tenantship or owner- ship of land who intend to change their occupation will suffer a 159 substantial loss. The amount of this loss steadily increased during the period of study because land productivity and thus land tax increas- ed. Consequently, the real wage level, i.e., the marginal product of labor, had to rise in order to attract rural surplus labor, in spite of the low and constant marginal product in agriculture. As just noted, the labor market in Taiwan was not one of perfect competition. The distortion, however, did not seriously affect the validity of the labor demand curves derived from the agricultural and nonagricultural production functions. The functional relationship estab- 1ished by these demand for labor curves to determine labor allocation between the two sectors, i.e., equation (9) in Chapter II should still be applicable to Taiwan. In estimating this equation the proportion of nonagricultural workers had been broken down into age groups for each sex. The measure for the ratio of nonagricultural to agricultural wages was replaced by the ratio of nonagricultural wages to compound farm income per worker because agricultural wages were not given separ- ately in the national income statistics. The output prOportion ratio was also substituted with ratio of aggregate demand for nonagricultural product to agricultural product because the latter was closer to reality than the former. From the estimated equations it is noted that except for the coefficient of wage ratio for proportion of nonagricultural male workers at age twenty to twenty-four, all signs of the coefficients are in agree- ment with the requirements specified on theoretical grounds. The only exception may reflect the fact that in spite of relatively high wage payments industrial employers strongly preferred to employ young male adults. On the whole the coefficients of the aggregate demand ratio are 160 more highly significant by a E_ test than are the coefficients of wage ratio. This finding is another support for our argument that the pattern of aggregate demand is a driving force in the determination of economic development and population growth. c. Population Growth and Labor Supply Both longitudinal and areal demographic data are available in Taiwan for empirical analysis of the effects of economic development upon p0pulation growth and labor supply. To make them consistent with the nature of our long-run unemployment growth.model and the available national income statistics, our estimated equations for the components of population growth and labor supply are the results of the longitudinal data for the period from 1951 to 1970. A common problem of applying the time-series technique to establish the long-run effects of economic factors on population changes is that the range of variation in characteristics and the degree of interaction among the explanatory variables in a twenty-year period may not be great enough to give accurate determinations of regression coefficients. There is, however, a series of excellent cross-section studies on popu- lation changes in Taiwan for this period which.may be used to cross check our results. FERTILITY FUNCTION For the purpose of eliminating the effects of differences in age and sex distribution of the population, age-specific birth rates for all women in Taiwan were used in the estimations. One important demographic factor, the proportion of married women, was not included because these data had not been made available before 1960. A preliminary regression analysis of the factor governing the proportion married from the published data from 1960 and on suggests that younger women are more sensitive to variables of educational 161 attainment and older women to level of economic development. Since these factors were also presented in the fertility equations, the effects of the proportion married had been implicitly taken into account. All independent variables specified in our theoretical model had been regressed on the fertility of women in each of seven five-year age groups. However, due to the small number of observations and es- pecially the lack of variation in fertility of women in the age groups fifteen to nineteen and twenty to twenty—four years old,the estimated regressions become unstable if more than three or four independent vari- ables are used simultaneously in a single equation. Therefore, the regression results presented in our study are those equations with al- ternative combinations of three variables for age groups fifteen to nineteen and twenty to twenty-four and four principal variables for the remaining age groups. These variables are: technical progress 11, technical progress 12, productive capacity, income distribution, and economic structure. Technical progress I is measured by the cumulative 1 number of married women who had a first intrauterine device (IUD) insertion in each five-year age group. Technical progress I is measured 2 by per capita private and government spending on education. Productive capacity is measured by estimated capital per worker. Income distribution is represented by the ratio of the nonagricultural wage rate to the average compound income per farmer. Except for the variable of technical progress Tl’ the scatter diagrams of fertility and our independent variables show nonlinear relationships, therefore these variables were transformed into reciprocals. Reviewing the results of regression equations for age-specific birth rates we find that the signs of coefficients for women at ages 162 thirty to fortyIfour are all in agreement with our expectations. The few exceptions, however, for women below thirty and beyond forty-five are likely to imply certain details in the process of family building rather than contradictions to our hypothesis. The positive relation between the cumulative rates of IUD in- sertion and birth rates for women in age groups fifteen to nineteen and twenty to twenty-four years old suggests that young mothers tend to use contraception for spacing. This surmise is supported by data from individual couples. TWO recent fertility surveys reveal that in 1966, 16 percent of IUD acceptors who were young and of low parity reported that they were using IUDs for the purpose of spacing, and by 1970 this proportion increased to 20 percent. Detailed tabulation of the 1970 survey also shows that a significant number of women were practicing family planning after having one child and the proportion increased rapidly as they were nearing their goals of having three or four children and one or two sons (see Table 4.3). The positive sign of the coefficient of economic structure in the fertility regression function for women twenty five to twenty-nine, though the 5' test is insignificant, may suggest that due to the in- creasing tendency toward later marriage and the practice of spacing births (both were positively associated with changes in the economic structure from agricultural to nonagricultural),some births that would have occurred to women at younger ages had been postponed to women at ages twenty-five to twenty-nine. This surmise is based on the following facts. Between 1956 and 1970 the proportion of married women at fifteen to nineteen and twenty to twenty-four decreased respectively from 11.3 percent and 89.6 percent to 8 percent and 50.3 percent. This 163 TABLE 4.3 PERCENTAGE CURRENTLY USING CON- TRACEPTION, CURRENTLY PREGNANT, AND MEAN LENGTH OF THE OPEN BIRTH INTERVAL, BY NUMBER OF LIVING CHILDREN AND NUMBER OF LIVING SONS, FOR WIVES AGED TWENTY-TWO TO TWENTY-NINE WITH ONE TO FOUR CHILDREN, TAIWAN: 1970 NO' living children, c::::::1y Percent Open birth no. living sons using currently intervala (N) contraception Aprggpgnt yjmonths) 1 child 0 sons (36) 2.8 39.2 23.2 1 son (31) 10.7 37.8 30.4 2 children 0 sons (67) 16.5 14.1 15.4 1 son’ (145) 25.4 17.4 19.2 2 sons (76) 31.8 12.4 24.2 3 children 0 sons (38) 11.6 18.8 p 10.9 1 son (93) 26.7 14.4 18.8 2 sons (116) 44.9 5.8 22.9 3 sons (19) 30.9 5.2 16.2 4 children 0 sons (5) 43.9 0.0 8.3 1 son (32) 18.5 6.2 15.0 2 sons (39) 56.0 2.5 17.8 3+ sons (42) 49.2 8.0 20.6 aFor zero parity women, OBI I months married SOURCE: Shanta Danaraj, "Son Preference in Taiwan," unpublished paper (Ann Arbor, Michigan: University of Michi- gan, 1973). 164 decrease, of course, reflects a trend toward marriage at a later age for women. In addition, by 1970 the length of the open interval, i.e., the average number of months since the last birth, for women of low parities was significantly longer than natural fertility would imply (see Table 4.3). As already noted in Chapter III fertility reduction in Taiwan was achieved mainly by regulating excess births while the tra- ditional values of having three or four children with at least two sons remained intact. In the absence of deliberate birth spacing and with the low postwar mortality level a time span of six to eight childbear- ing years is needed to realize such traditional values. This time span implies that in 1951 most women who were married at the average age of twenty would achieve their fertility goals in their middle twenties. As the average age at marriage of girls and the practice of birth spacing among young married women increased, the time span for having the same traditional number of children gradually shifted and stretched toward women in older age groups. A A plausible explanation for the positive sign of the cumulative rate of IUD insertions in the regression equation for fertility at ages forty-five to forty-nine is that after controlling for all other vari- ables, it is likely the higher the birth rate the stronger the moti- vations to limit the family size. The signs of the variables of pro- ductive capacity and income distribution coefficients for fertility at forty-five to forty-nine are also contradictory to our expectations. The reason for these may simply be that for the majority of women at very late reproductive ages who had achieved their familial goals, an increase in permanent or relative income would lead them to spend more on gratification from.improving the quality of children already born, 165 rather than from having more babies. On the whole, our results from the regression equations for age- specific birth rates tend to confirm our hypotheses about the relation- ships between fertility and economic development. They are also con- sistent with the findings of regression analyses of aggregate data for 361 local areas in Taiwan for 1961 and 1966 to 1969.15 All of the areal analyses have observed a negative correlation between fertility and education, and the proportion of the male labor force in agriculture. It is worth noting that the finding of little correlation of fertility at ages twenty-five to twenty-nine with that at other ages in the areal analyses for 1966 to 1969 has the same implication as our finding of the positive sign for the coefficient of variable economic structure in the equation for fertility at ages twentyIfive to twenty-nine in the time-series analyses.16 Our findings are also consistent with the cross-country results from studies of various countries made by Weintraub, Adelman, and Beer.17 A fundamental difference, however, is that ours were based on observa- tions from a homogeneous environment of Taiwan for a relatively short period and theirs were based on observations from countries with a vast variety of institutional and sociocultural backgrounds. Because of this difference our results tend to provide a more precise picture of the relations between fertility and economic development. These relations, however, are subject to the influence of environmental and historical factors of the particular economy in a particular time period. The factors with special significance for Taiwan are: (1) Land reform program. In our discussion of resource allocation ‘we have already found that the existing agrarian system had controlled 166 the flow of capital into and labor from the agricultural sector by imI posing implicitly an opportunity cost on industrial employers. As land fragmentation continues and rural surplus labor diminishes this oppor- tunity cost will rise. In consequence, the changes in the whole complex of variables in economic development will slow down and so the fertility decline will decrease, because of less rural labor moving to urban centers. (2) Family planning program. An organized effort to reduce the birth rate through providing family planning services starting in 1964 was an important social intervention in the history of Taiwan's popula- tion. By 1970 about one-third of all married women aged twenty to fortyI four, i.e., a cumulative total of 658,000, had had a first IUD insertion. As shown in our fertility regression equations, this organized effort had had an independent influence on reducing birth rates of women at ages thirty to forty-five and on spacing for women below thirty, even after controlling for all economic development variables. Since the purpose of the program is to make present knowledge and means of fertility regulation available to the general population rather than to change their attitudes toward fertility behavior, an increase in the number of IUD acceptors represents only that more women are able to achieve their ideal number of children or to avoid having more unwanted children. Therefore, if the family program continues to function as it has been, its influence on the birth rate will diminish when the number of women eligible for the program decreases. (3) Desire for joint family living. In Taiwan the traditional family arrangement for parents to live with their married sons is still thought to be most desirable. An islanddwide sample survey in 1965 revealed that 78 percent of wives aged twenty to twenty-nine expected 167 to live with their children in their old age, even if they have sufficient means to support themselves.18 In 1970 a reinterview of their husbands indicated that only 10 percent of them do not expect to live with their grown-up sons and another 20 percent doubted if the circumstances would allow them to live with their married sons.19 Insufficient development of satisfactory nonfamilial institutions for the aged, as reflected by the strong desire for joint family living, on the one hand, and the increasingly high uncertainty of fulfilling their desire under the fore- seeable socioeconomic conditions on the other, is probably an important deterent to lowering the ideal number of children and therefore to further decline in the birth rate of Taiwan. (4) Importance of male heirs. The traditional value given to having male heirs to carry on the family name and to continue the rite of ancestral worship are still prevalent and highly valued in present-day Taiwan. Eighty-three percent of married women aged twenty to twentyf nine interviewed in Taichung City in 1962 said that to have a male heir was very important; 13 percent rated it moderately important, and only 4 percent said it was unimportant.20 Additional evidence on this point (see Table 4.4) indicates that in 1970 the percentage currently using contraception and the length of the Open birth interval were significantly greater and the percentage currently pregnant was lower for married women aged twenty to twenty-nine who had achieved the ideal goal of having two sons than for those who had not. These traditional values are so deeply rooted in the Chinese that even in communist China the use of most drastic nmasures, such as removal of family ancestor plaques from the lineage hall in each house- hold, forbidding gatherings of lineage members, and so forth, could not 168 prevent people from honoring their ancestors and having a preference for having four children.21 MORTALITY FUNCTION Mortality statistics tabulated from the con- tinuous population register of Taiwan have enjoyed a good reputation for completeness and accuracy for deaths of persons over five years of age. Tabulations of deaths of those under five have three main defects: underregistration and delay in reporting and misscount of ages for infant deaths.22 These defects are serious but they can be and were corrected by use of the techniques of model life tables.23 In our estimates the corrected survival ratios by sex and age for the years 1951 to 1970 were used to take the place of age-specific death rates, the main reason being that the survivor ratio can be readily applied to population projections. In the period under review Taiwan benefited from the implemen- tation of a low cost public health program. Therefore, an indicator of public health conditions -- the sum of real per capita expenditure on public health -- was included in the mortality functions. Since Taiwan's mortality rate in 1951 had already achieved a fairly low level -- only 11.5 deaths per 1000 population -- further decline was bound to be slower. A nonlinear form of reciprocal values of the explanatory variables is found to represent this tendency better than a linear form. It is seen from the estimated functions for each five-year age group of both sexes that the regression coefficients of survival ratios with respect to the reciprocal of the total expenditures on health care are almost all negative (except 2 out of 32 cases) and statistically significant. The signs of the regression coefficients of the survival 169 ratios with respect to the reciprocal of per capita real disposable income tend to be positive for working age populations of either sex, and negative for younger and older persons. In general the coefficients are statistically significant. In brief, our regression results indi- cate a negative association between death rates and expenditures on health care for the population of both sexes at all ages, and a negative partial correlation between per capita real disposable income and death rates for population at younger and older ages, but a positive correlation for the working-age population. These findings are consistent with our expectations but roughly speaking, they partially conflict with both Leibenstein's contention and Adelman's statistical findings of the relation between the rate of growth of real per capita income (ours is level of income instead) and mortality rates. Leibenstein advances a basic hypothesis in his economic population theory: The greater the rate of per capita income grOwth, the greater the rate of investment and, hence, the lower the level of consumption per unit of income. On the presumption that lower levels of consumption imply higher mortality rates, we arrive at the conclusion that the greater the rate of per capita income growth the less the rate of population growth, other things being equal.24 In a regression analysis of aggregate data for thirty-four countries during the period from 1947 to 1957, however, Adelman found a uniform negative correlation between the rate of growth of real per capita income and age—specific death rates. She explains that when the rate of growth of income increases, the relative share of income and thus consumption by the laboring population increases, therefore death rates diminish; while when the rate of economic expansion declines, 170 the share of profit increases at the expense of wage share, therefore leads to an increase in mortality.25 Table 4.4 presents the data of elements, which appeared in their arguments, on Taiwan for the period from 1951 to 1970. It is seen that Leibenstein's contention on the relation between income,investment, and consumption, to some extent confirms Taiwan's experience but his hypothesis about mortality and the rate of income growth does not. On the other hand, Adelman's hypothesis about the relation between rate of income growth and mortality agrees with what happened in Taiwan but her argument on the relation of the rate of income growth to wage share is weakly supported. According to our analysis the explanation of the phenomena shown by the elements in Leibenstein and Adelman's arguments may be that the greater the rate of investment, the greater the rate of per capita income growth and, hence, the greater the level of consumption. In general, an increase in consumption level has favorable effects on lowering mortality, but as the consumption level rises beyond the level of basic nutritional requirements as that in present-day Taiwan did, its favorable effects tend to be offset by hazards owing to the rapid economic expansion for those in the labor force. The hazards commonly observed in this stage of development were long and exhausting hours of work, congested and insanitary working conditions, and occupational accidents. LABOR FORCE PARTICIPATION EQUATIONS An important modification lmas been made in the estimation of the labor force participation equations -- that is, in order to eliminate the effect of changes in age and sex composition, age-specific participation rates rather than 171 .ma.m manna now on made one ”mMUMDOm m.o h.o c.a «.0» o.mo m.oo m.ao mausom c.o w.o ~.a n.co m.mo m.oo m.nm 0am: nuuao um mucouoomxo owaa owmuo>< c.o a.ou a.a mo.o wo.o no.0 mo.o osoooa amuou ca OBOOOH Honda mo madam I I I mc.o mc.o cc.o ~c.o sapwooo ou muamsomowm omnuo>< a.o m.m a.m sam.o omo.m mm~.e ~66.m awaz coma sac coaumaomooo ouammu pom m.- o.aa ~.o oom.~ no~.a com cmm Amaz ooma mo mooaaaaa oav uoosumo>oa mo doom m.o m.c o.~ Num.ma cmo.aa anm.m mon.m Amaz ooma sav maoooa Ouammo Hum onlooma molaoma ooaomma onma moma ooma mmma a a a a a a a monaoma oulooma mmlamma ooma aoma omma amma nuaowo no snow amoooc camauamaa .zazaes za aesma whaa UHZUZOUH mo 20mam