This is the html version of the file http://www.sabhlokcity.com/almaak/Dissertation/proposal.doc.
G o o g l e automatically generates html versions of documents as we crawl the web.
To link to or bookmark this page, use the following url: http://www.google.com/search?q=cache:sTt9ZtvgFUcC:www.sabhlokcity.com/almaak/Dissertation/proposal.doc+%22pre-agricultural+societies%22+%22sex+ratio%22&hl=en&ie=UTF-8

Google is not affiliated with the authors of this page nor responsible for its content.
These search terms have been highlighted: pre agricultural societies sex ratio 
 
 
 
 

A comparative study of the fertility transition

in different rural areas of

Thailand and Costa Rica

with special emphasis on the old age security motive 
 
 
 

Proposal for Dissertation  

Sanjeev Sabhlok 
 
 

Under the guidance of  

 

Preface 

      This proposal is being structured in the form in which I expect  to bring out the dissertation. Most chapters are preliminary explorations in the concerned area, and work on others (empirical chapters) has not yet begun. At this stage, the focus of the topic has been defined, a broad literature survey carried out, and a basic theoretical model worked out. As a pause in this process of research, marking the end of the ECON 790 (Directed Research) course that I have taken this semester with Professor Nugent, I thought it appropriate to acknowledge the people behind this research. 

      After Professor Nugent introduced me to this area during a discussion in January, 1997, I felt that there was a lot of interesting work to be done, and so I chose to work in this area of demographic economics. I would like to thank him for giving me the opportunity to work under him in this area where he is considered perhaps the leading authority in the world. I would also like to thank him for constantly keeping me on my toes throughout the semester, guiding me to various readings, going through my ‘preliminary drafts’ listening to my ‘ramblings’ while I was not quite clear which approach might actually be at work, and most significantly, pointing out the complexities of ‘reality’ and alternative motivations that might be at play, each time I thought I had achieved some measure of understanding.  

      I would like to thank Professor Easterlin for introducing me to the discipline of demographic economics in Spring, 1996, wherein I learnt about his synthesis theory which has made such an major impact in world literature on the economics of fertility. Professor Maurice Don Van Arsdol, Jr. of Sociology Department (Population Research Laboratory) has been a powerful force in my thinking, by showing me the alternative currents in the discipline, which is now extremely inter-disciplinary, during a course I took with him in Summer, 1997. I would like to thank Professor Kuran for introducing me to the disciplines of evolutionary psychology and evolutionary biology and of the necessity of considering their findings in modeling the human mind during the course I took with him in Fall, 1995. I would like to thank Professor Robinson for the interest he has taken in this area and the discussion I had with him on this topic. I would like to thank Professor Caroline Betts for giving me an opportunity to present this proposal on the 25th of April in the Graduate Students Workshop, and for throwing up many valuable suggestions during that presentation. 

      Among the students at this department, I would like to thank C.V.S.K. Sarma, Shailender Swaminathan, Sunanda Ray, Sripad Motiram, Atul Gupta, Lata Gangadharan and Pushkar Maitra, among many others, for helpful discussions at various stages.  

      Finally, I would like to formally thank members of my Guidance Committee for kindly agreeing to be on my Committee and critiquing this bulky document. 

28th of April, 1997        Sanjeev Sabhlok 

 

Contents  
 

Chapter Contents        Page

   

PART I THEORETICAL UNDERPINNINGS 

    1  Introduction          4 

    2  A Superstructure of Theory       16 

    3  Theoretical and Empirical Models to be used in the Dissertation  50 

PART II EMPIRICAL ANALYSIS 

    6  Conclusion         89 

            Appendices        109 

 

 
 
 
 
 
 
 

PART I  
 
 

THEORETICAL ANALYSIS

 

Chapter 1

Introduction  
 

1.1 Introduction 

      The old age security hypothesis (OASH) states that an important reason for people having children is to provide for them when old. This motive for fertility has been accorded differing significance in the literature. Some believe that it is a powerful motive and therefore, by providing for the old through some mechanism, say through old-age pension programs, it will be possible to reduce the demand for children. Others believe that its role is relatively small in determining fertility. Following the work of Nugent and Anker (1990), I will attempt to formalize the theoretical basis for this argument, and to examine whether the changes in fertility over the past thirty years or so in rural areas of Costa Rica and Thailand, could at least partially be attributed to this motive.  

      The key differences between Costa Rica and Thailand have been summarized in Nugent and Anker (1990). I am placing some of these differences in Table 1. In terms of key indicators of modern economic growth, Costa Rica was clearly at a higher level of human development than Thailand around 1970. Accordingly, it was not surprising that it had a much lower fertility rate than Thailand. Additionally, Costa Rica has had in place, for many decades now, a relatively strong1 governmental structure for provision of security to the elderly. Finally, it appears that in terms of development of financial markets, Costa Rica has been more advanced than Thailand, allowing for greater alternative modes of savings than in Thailand.  

      Accordingly, if the OASH were true, one would expect that Costa Rica would lead Thailand in further declines in fertility. Even other, more commonly held theories of fertility decline would tend to predict that Costa Rica would do better. However, it turns out that fertility declines in Costa Rica have been rather minimal, while Thailand has made enormous progress in reducing its fertility.  

 

Source:  Data is from World Bank (1993:20-21) unless stated otherwise.

Notes: The GNP per capita of 1965 and 1970 is in constant 1987 US dollars.

       Total Fertility Rate is defined in the Population Handbook.

       Population of Costa Rica (1965) and Thailand (1970) relates to 1960, as given in

          Estimates for TFR for 1990 are from HDR, 1991.

       Estimate for TFR for Thailand for 1970 is for 1964-65, from CICRED (1974),

       Estimate for TFR for Costa Rica for 1970 is from Gomez and Nugent (1995), who cite the

              a World Bank publication of 1995.

       Estimate for GDP per capita for 1990, is in constant 1988 USD, from HDR, 1991.

       Estimate for IMR for 1990 relates to 1989, and is from HDR, 1991.

       Data on mean years of schooling of persons above the age of 25 years is from HDR, 1991. 

      Thailand is not the only such nation. In many other traditional societies, rapid reduction in fertility has taken place despite there being no alternate system in place for the elderly, and despite the fact that people can now look forward to a much longer duration of old age, which might include infirmity. It is hypothesized in this paper that the institutional structure for satisfying the old age security motive (OASM) has a lot to do with explaining the observed facts. “Numerous studies show that a person’s earnings are higher the higher their human capital, as measured by their education and health status.”2 Therefore, a revised OASH is being suggested here: The asset demand for children is being met in Thailand by parents investing more on fewer children, in order to capture the rents from modern economic growth during their old age through these fewer, educated children. If old age care is relatively guaranteed even from very few children, then the family has a greater incentive to respond to the economic incentives provided by modern economic growth.3    

      At this stage, one should note that comparisons between two distant nations can at best be extremely tenuous. Other competing hypotheses might emerge as candidates for explaining some of the facts observed above. It could be argued, for example, that due to the increasing incomes of parents (particularly increasing female labor force participation), the rate of growth of the opportunity cost of time in Thailand was high enough to make children extremely costly. It could also be argued that some other, better alternatives to children, were suddenly made available in Thailand for transferring wealth across time. The World Bank (1993:244) tested for Granger causality and found that GDP growth rates have ‘caused’ higher savings in Thailand. But this does not mean that household savings have necessarily increased. Campbell et al (1993: xxiii) state, in the context of Thailand,  that “the household saving ratio, at the aggregate level, has not responded to favorable demographic trends.”  

      The study will therefore examine a broad set of explanatory variables which are recognized in the literature to impact on fertility, while attempting to isolate the effects of OASM and related institutions for old age care. The empirical study will also attempt to go into the relative differences in decision making within the household. Use of will be made of the old age security survey (OASS) carried out in Costa Rica and Thailand in 1991-2 under the supervision of Nugent and Anker, supported by the International Labor Organization. Since the data is extremely disaggregated, it will be possible to utilize inter-regional variation within these two nations, to test the hypotheses.  

      The dissertation will be organized in two parts: Part I will comprise the theoretical underpinnings, and Part II, the empirical analysis. In Chapter 2, based on a literature survey,  I shall establish a superstructure of theory for looking at the observed declines in fertility as well as an apparent lower limit to fertility. I will also investigate the link between the old age security motive and fertility. Chapter 3 will outline the models - both theoretical and empirical - to be used in this dissertation. 

      Moving to the empirical part (Part II), in Chapter 4, I will outline the demographics and economics of Thailand and Costa Rica, with particular reference to work done on understanding the impact of old age security on fertility in Thailand. The OASS will be discussed, and its basic results considered, in this chapter. Chapter 5, which has not even been outlined at this stage, will deal with the testing of the proposed models and the results obtained therefrom. Chapter 6 will conclude the dissertation. 

      In the remaining sections of this chapter, I discuss some basic policy questions about population policy, and examine the origin of the old age security motive. 

1.2 Is high population bad for economic growth / human development? 

      It is useful to take a brief look at the possible policy implications of this study. A basic question that arises is: is high population growth bad for economic growth and human development? This is a fertile area of debate. Baland and Robinson (1996) point out, “The aggregate evidence about the effects of population growth on economic development is ambiguous. The empirical evidence, extensively surveyed by Kelley (1988), Birdsall (1988) and Robinson and Srinivasan (1996), is that there is no robust relationship between population growth and development one way or the other.”  

      “Since economists have not been able to construct a convincing rationale for the existence of a population problem on the basis of a specification of an intertemporal social welfare function, the search for the existence of a population problem has concentrated on the possibility that population growth might generate significant externalities or be connected with other market failures. Thus far neither the theoretical nor the empirical literature has reached a consensus on the issue” (Baland and Robinson, 1996). 

      “One common argument is that if all parents increase their families then the effect on labor supply will cause the wage rate to fall. As Willis (1987) showed, however, this effect is a pecuniary externality (in the terminology of Scitovsky (1954)) which does not imply inefficiency when markets are complete. While the assumption that markets are complete is clearly unrealistic, thus far a convincing synthesis of market incompleteness with population growth has not been made. In particular, it is far from clear that the second best policy intervention in such a world would be to attempt to restrict fertility. An alternative argument, studied by Nerlove et al. (1987) is that population growth may have adverse effects by crowding public goods and infrastructure, a counter argument being that a larger population reduces the average cost of provision. ... Recently there has been much interest in the issue of whether or not population growth has important externalities by causing rapid environmental degradation as in the model of Nerlove (1991). Indeed, the empirical evidence of Allen and Barnes (1985) and Cropper and Griffiths (1994) shows that population growth is positively related to deforestation. ... [But these issues are best handled by environment policy rather than by population policy]” (Baland and Robinson, 1996). 

      However, most policy makers still continue to believe that there is a significant population problem. It appears to me that there is a definite problem here, but that it is environmental rather than economic. There is no way that a particular species of animal can suddenly increase its population at the rate at which humans have done in the past hundred and fifty years without disturbing the delicate natural equilibrium in nature. Many species of natural life have come under great stress from this population growth. A Darwinian approach might treat this dominance by humans as a natural extension of  intelligence of human beings; but there are tremendous risks to this untrammeled dominance, given the tendency in mankind toward opportunistic behavior and negative social and global externalities. It is this behavior that has resulted in the savage cutting down of forests, elimination of species, and a major burst of pollution. It is not know how nature, with its complex interactions, will interact to this tremendous change, coupled with economic growth. That is the primary concern, as I see, and steps must be consciously taken to bring back population to its equilibrium of replacement levels of population.  

      In other words, I believe that though an economic proof of the existence of a population problem is in question, we cannot neglect population policy till we officially ‘discover’ such a problem. The present study has implications for such population policy. 

1.3 Evolution of the old age security motive 

      It is useful, in this introductory chapter, to lay down a possible story about the origin of the OASM. This would bring in the institutional framework which we wish to ultimately analyze.  

      When we look at the growth of animal populations on whom natural constraints have been lifted, we find that there is usually a sudden spurt in their population growth rates. Ultimately, they reach another constraint, and thereafter a new equilibrium sets in, which is almost a ‘steady state.’ Three shifts in equilibria of human populations have been identified by Easterlin (1996), who, following the methodology of Kuznets and others, has classified human economic-demographic experience into three epochs. The principal characteristics of these three epochs relevant to us, are given in Table 2. 

      Table 2: Distinctive Characteristics of Economic-Demographic Epochs 

                              Epoch I Epoch II Epoch III             _

1. Principle occupation   Hunting Farming Diverse

                               gathering

2. Principle type of

    settlement    Nomadic Village  Urban 

3. a. Initial date of epoch           -  8,000 B.C. 1750 A.D.

    b. Terminal date   8,000 B.C. 1750  A.D.     ?

    c. Duration, years   2 million ? 9,750    247 

4. a. Population at start              -     7.5    770

    b. Population at end        5-10    770  5,300

    c. Years to double population   90,000  1,459      90 

Source: Easterlin, 1996.

Notes: Population is in millions. 

 We shall see how the age structure of the population has shifted inexorably toward the elderly, over these three successive equilibria. We shall therefore motivate the existence of an OASM which will then be shown to have possible effects on fertility. 

1.3.1 The first equilibrium: Hunting and gathering epoch 

      Consider the epoch hunting and gathering epoch. During this period, human populations grew very slowly, doubling every 90,000 years. This slow growth in population can be characterized as an equilibrium very similar to that achieved by most animal societies:  given the constraints of the environment, the populations of various species fluctuate around a relatively constant value. It has been hypothesized by some4 that in such primitive societies, people had children to the maximum extent possible, biologically. The biological maximum is about 15-17 children for a woman. This is also known as fecundity. It is extremely difficult, however, to come to such a conclusion.5  One would hypothesize that human beings have evolved a biological maximum of about 15 children as a response to crises: this capacity provides a primitive society with a reserve reproductive capacity: in usual situations, much lesser number of children were sufficient to enable the survival and even moderate expansion of the species; however, after natural checks decimated large populations, this ‘reserve’ capacity was utilized by women to repopulate their area within a relatively short time.6 Another argument could be that human fecundity might itself have evolved over the millions of years of pre-agricultural societies, adding to the complexity of the issue.  

      In such societies, the weak - i.e., the infirm and the old - were non-existent,7 as there were no means to transport them around while shifting from place to place. Since there were no old, there was no problem of old age security.8  

1.3.2 Agricultural epoch 

      Then, suddenly, one constraint on human population was removed: by discovering agricultural production technology, humans were able to reduce the variance of their food supply and increase the survival rate of the population by overcoming possible death from starvation. This technology put a constraint on their mobility, however, forcing them towards a less nomadic lifestyle. Agriculture also necessitated cutting down forest areas, and increasing the ‘exposure’ of the peasant to wild animals. This exposure - which is frequently fatal even in the agricultural societies of many developing countries today - was minimized by the creation of villages communities where cooperative solutions to the problems of security and insurance were worked out. The size of the community could now be much larger than the small hunting and gathering groups. Norms of cooperation were developed initially to ensure the survival of children when their parents had gone to the fields. Coterminous with the to the creation of villages, property rights on land came into existence, and it became it became possible to ‘own’ assets, in howsoever primitive a way.  

      It is extremely rare in nature to find animals that live their natural life span. But now, for the first time in the history of human beings, the older members now began to survive in ever increasing numbers. The infirm and the old were able to receive positive externalities from the safety and cooperative spirit of the village. But getting this cooperation was not quite easy. Since most species do not live to old age, evolution can be thought of as not providing suitable ‘instincts’ to children to provide for care for the old.9 It therefore needed all the experience, intelligence, imagination, and guile10 of these old folks to create man-made rules and constraints (institutions) designed to ensure their extended survival. Thus arose the problem of old age security.  

      I would imagine that sympathy generated for the old and infirm by existing maternal instincts transferred over to this task to some extent. Also beneficent to the old are the psychological traits of humans whereby children tend to internalize much ‘obedience’ towards parents (transactions analysis literature, e.g., Harris, 1969). This would have made it easier for parents to devise institutions intended to gain power for themselves and to get support when they were very old and infirm. But it would appear that this ability to free ride on the goodwill or obedience of children could not depend exclusively on the existence of charitable and altruistic individuals, just as the poor of today cannot depend on the charity of the rich. Therefore there  came into existence an incentive to design institutions to create an assured mechanism of old age care. The following institutions were therefore designed by most village communities, motivated, surely, by the self-interest of all members, including the young - who could then utilize these institutions to further their own self interest when old.  

a) The extended family. By this mechanism, the ‘portfolio’ of the household was diversified. The old were able to become ‘useful’ economic assets by providing child care to the extended family, and in return they could expect to get some attention to their needs. It did not matter whether the society was patriarchal or matriarchal: what mattered was that the elders gained increasing importance in the extended family, primarily by virtue of their experience in managing the young.  

      Sociological theory confirms this view. The development of the extended family is closely related to the development of an intergenerational role relationship, as pointed out by Entwisle and Winegarden. This approach goes beyond the economic value of children to parents; however, one can see an implicit argument about old age insurance in this approach. “This approach focuses on the expectation of support in old age rather than its economic value to parents. The expectation of old age support denotes an intergenerational role relationship that plays an important part in defining extended family structure. It includes the willingness of children to help their elderly parents as well as the extent to which parents are justified in expecting this help. Thus it provides the context in which the value of children is determined. The substitution of pension benefits for help from children is thought to weaken the strength of the intergenerational role relationship. Sociological theory posits that this shift in family structure is a crucial aspect of major and long-term fertility decline.”  

b) ‘After-life’ concepts: Since the economic value provided by the old was extremely limited, the ever-present ‘mystery of death’ was utilized by the leaders among the old to promote the concept of ‘after-life.’11  
 
 
 

i. Duty toward parents:  

      Consider a simple, infinitely repeated game. If I create a set of beliefs in my children about the existence of an after-life, then I can implant on them the concept of ‘duty toward old parents’ and other such concepts and impose the following rule: “Do good to your parents, else their ghost will haunt you not only in this life, but you will be also punished in your after-life.” The existence of such beliefs will ensure a stable sequential equilibrium, wherein the old are supported in their old age by an off-the equilibrium path threat of strange retribution by the spirits. Even today, in most agricultural societies, ancestor worship is a major religious task. To this observation, one can add that this was the time when major religions sprang up all over the world, promoting the care of the infirm and weak, and attempting to ‘tame’ the violent nature of man evolved over the previous millions of years. Since everyone has occasion to be weak and infirm sometime in his or her life, religion became an extremely successful institution. 

ii. Religious need for a living son: 

      In some societies, one of the relatively successful institutions which promoted care for the old was the promotion of a religious desire for a living son at the time of one’s death. The belief in after-life meant that parents had an incentive to have non-zero children, as well as the son had a fear of displeasing his parents. This can be modeled as a sequential decision making model (e.g., O’Hara and Brown, 1976) with the decision to have a marginal child being dependent on the outcome in the last period. These models seem to generate relatively interesting results, including a non-zero fertility. But there is no justification of why a living son is desired by a parent. That justification can now be arranged through the old age security motive. 

c) Property rights and strategic bequests:  In agricultural societies, it must have taken a lot of work to create and maintain property rights institutions. It is possible to imagine the formation of groups of powerful youth in a village, which, over time, as they grew older, created mechanisms to enforce property rights, and who ultimately became the typical ‘council of elders’ which administers justice in most agricultural societies. Once the property rights institutions were put into place, yet another mechanism came into existence for use by the old: that of bequests. If bequests were made ‘strategically,’ with only those who took better care of the old getting these bequests, then the old could ensure that the young had a further incentive to look after them, and were able to enhance the quality of this care.  

      Some societies built an even more grandiose edifice of man-made rules in order to ensure the success of old age security. The status of elder members of a society was enhanced successively over time in places like China. In other societies, norms were developed such that if a young person did not take care of his or her parents, that person would face ostracism from the village community. 

      Having designed these institutions, it now became feasible for a community to support old members until their natural death.  

Old age security and fertility   

      Once such institutions had been established, and the probability of living out a complete natural life-cycle became a reality to be taken seriously, the utility of a child was enhanced, since there was now a ‘returns’ component to the ‘investment’ made in a child. Till now, i.e., in the hunting and gathering epoch, there was a uni-directional flow from the parent to the child (contrary to what Caldwell hypothesized), since the parent was sure never to reap the rewards of ‘care’ from his or her children. But with the new institutions put into place, bi-directional flows became a reality.  

      It is very difficult to say whether human beings became ‘rational’ utility maximizers at this stage (or whether they are even now), but it is safe to say that in whatever primitive ‘calculation’ that was made about a marginal child, a positive incentive was now built up: of a return from the child in terms of old age security. Therefore, parents had a stronger interest in the survival of their existing children, and also perhaps, in producing more children. This is the essence of the OASH, and this makes sense.  

      If the OASM was sufficiently strong, and the marginal value of a child was positive, then we would expect an even larger growth in population, approaching the biological limit of 15.  But it is not clear is whether the TFR increased as a consequence of the OAS factor. That is because now there were many other constraints to be considered: On the supply side, there was lesser fear of children being eaten up by wild animals, and hence the mortality of children reduced considerably. Further, there was lesser fear of death due to starvation, which also increased the survival of the children. On the demand side too, there were some opposing tendencies. (a) It is not obvious that the marginal value of a child needs to be positive to have this relationship between old age security and fertility,12 and if so, there is a natural trade-off between the increasing costs of an extra child and the insurance benefits expected from that child. (b) Marginal costs of children can be argued to be increasing. Limitations on land, given the production technology, could reduce the productivity of the farm, thus making the extra child more costly. (c) There is an associated problem of free riding by children. If there are too many children, then each child will expect the other one to look after the parents in old age, thus reducing the marginal return from the extra child. (d) Certain social norms might also restrict the return from children, for example, the institution of primogeniture, which was designed to prevent fragmentation of holdings, but which could reduce the effectiveness of strategic bequests.  

      Consequently, the net effects of the OASM on the TFR are ambiguous.  But that does not matter. It is not necessary to demonstrate a very high TFR in order to include the effect of the old age security motive on fertility. Even if TFR remained the same as in the hunting and gathering epoch, it can be argued that the old age security motive had an influence on not reducing the fertility rates. Be that as it may, at the commencement of the fertility transition all over the world, TFRs were in the range of 6-8 in almost all societies of the world, and life expectancy ranged between 20-30. According to Preston,13 “Most of these records suggest that life expectancy from prehistoric times until 1400 or so was in the range of 20-30 years. … Confidence in the range of 20-30 for life expectancy in the era before 1600 is enhanced by the use of demographic models. Since the world’s population was growing very slowly during this period [doubling approximately in every 1400 years], life expectancy at birth was, to a very close approximation, the reciprocal of the birth rate. Given the age pattern of fecundity and the apparent absence of significant anti-natal practices, the birth rate was quite unlikely to have fallen outside the range of 0.33-0.50 births per capita per year, implying life expectancies in the range of 20-30 years.” These estimates are in line with those in Table 2. With greater survival of the children and the old, human population growth received a big boost. It now became possible to double the population within  1459 years (the World Bank, 1993:8, has estimated a growth rate of 0.06 percent in the past two thousand years). This new equilibrium growth rate was however, not permitted to continue for long, and received another technology shock. 

1.3.3 Epoch of Modern Economic Growth and the mortality revolution  

      With the onset of the technology of public hygiene and modern medicine, the mortality revolution, dating from about the end of the 19th century, took place.14 It was now possible for an even larger number of children to survive into adulthood. A disequilibrium was created in the population of the world. Since the recent “habit” of having 6-8 children did not change instantly, the accompanying mortality revolution led to a sudden explosion in the growth rates of population, both in the European nations,15 and more so in the developing nations. The recent enhancements in the technology of medicine have given a major boost to the life expectancy of the elderly. So, now we have even more elderly than ever before in the world (see Box 1).  
 
 
 
 
 
 
 
 
 

Box 1

Unprecedented increase in numbers of the old  

      It can be expected that the human race will once again establish an equilibrium, in which the population will once again grow at a small but stable rate, but establishing this equilibrium could take till about the end of the 21st century for the entire human race to arrive at (Wilson, 1975:574). Whatever be the equilibrium, it is clear that it will include a much greater proportion of the old than in the agricultural epoch. 

      In the meanwhile, the question that interests us, is: now that people can expect much greater life expectancies, what is the effect of their desire for old age security on fertility? To analyze this, broadly, we note that in the first place, in most societies, the industrial revolution preceded the mortality revolution. The increase in incomes consequent to the onset of modern economic growth has meant that parents are more capable of supporting the increased number of surviving children, at least in most cases. At the same time, there are now other factors which increase the costs of children. Children have become costlier to ‘rear’ than in agricultural societies human capital development in modern societies requires very high investment in education of the children, which increases direct costs (on books, etc.) while at the same time reducing benefits (child labor benefits). Migration to urban areas also increases the probability of default (in old age care) by children, while at the same time diversifying the ‘portfolio.’ When we examine the three major institutions designed in the agricultural period to cater to the newly developed need of ‘old age security,’ we observe that in the modern period, the first two are breaking down considerably, viz., the extended family and the belief in after-life. The third institution, viz., the institution of strategic bequests, is also breaking down, with the promulgation of laws in many nations which prevent discrimination between offspring in the distribution of bequests. Therefore the old have difficulty in guaranteeing the returns from children.  

      In the situation of tremendous flux in incomes, costs, and institutions, observed over the last 100 years - and particularly in the past 50 years in the developing nations, it is very difficult to arrive at a clear effect of the old age security motive on fertility. All that can be seen is that the problem of old age security is being mitigated by increasing incomes while at the same time it is being increased by the collapse of the traditional institutions while others have not yet been designed, tested and established. The commonly proposed solution to this problem is to enhance government social security programs. But these programs are not only very costly to implement, but wherever implemented so far, have run into trouble in terms of management and finances. Therefore, today, some of the best economists, sociologists, and psychologists of the world, are trying to find out a solution to this ever increasing problem of old age security.16  

      It has been hypothesized in the literature from a very long time that despite the strong economic incentives to reduce the number of children consequent to the changing environment of the modern epoch, and despite the breakdown of the three institutions cited above, one major reason why fertility decline has not been rapid enough is because people see no immediate solution to their problem of old age security but to have more children than would have been otherwise necessary. If this were to hold true, then, policy makers would need to devise a new institutional mechanism that takes care of this concern, assuming that this sudden disequilibrium in population is deleterious to the society and environment.17 In any case, the problem of old age security is very real, and needs to be solved sooner or later. We shall however, not concern ourselves with the solution to the problem, but with the hypothesized effect of the problem of old age security on fertility.  

 

      

Chapter 2

A Superstructure of Theory  

      Two noticeable phenomena need to be explained: (a) the decline in fertility and the differential decline in fertility across nations and (b) the tendency for fertility not to decline much below 2.0. All theories of fertility that are entirely endogenous must explain both phenomenon.  

      This chapter is based primarily on a survey of literature, but one’s own understanding of this literature is also highlighted. In section 2.0, we discuss the functioning of the human mind with the intention of framing a structure for modeling. Section 2.1 examines the explanations for the fertility transition. Section 2.2 looks at the reasons for the observed lower limit on human fertility. Section 2.3 examines the relationship between old age security and fertility.  

2.0 The complex puzzle of human fertility 

      Human fertility has interested economists since Malthus (1798, reprinted, 1976). In order to evaluate the theories which seek to explain this change, we must first consider some basic mechanisms of human decision making in relation to children. What we see is a black box and we must deduce the behavior from the outcomes.   

Figure 1: The Black Box of Human Reproduction 

  Human beings → black box → decision to reproduce → child (with certain probability) 
 

      We begin with the basics. First of all we need to determine whether the of birth is outcome completely random or do human beings take a conscious decision to have a child.  

      There is a huge debate in the literature about the ability and desire of humans to invoke deliberate control of fertility in pre-transition societies. 

1. Evidence and arguments in favor of lack of deliberate control: 

b) Pre-transition people refer to children as gifts from God: 

c) Coale’s m index: No evidence of deliberate fertility control: 

d) Even if there was control of fertility, it was not based on choices by individuals: 

2. Evidence and arguments in favor of deliberate control: 

a) Pre-transitional fertility is lower than the Hutterite: 

b) If people really wanted to control fertility, they could have done it: 

c) Economic models have considerable predictive power: 

d) Rational peasants:  

e) No biological change in the transition period: 

      Therefore, one supports the argument that if human beings can take rational decisions on fertility now, there is no reason to believe that they did not take such decisions earlier, and that the observed TFR of 6-7 children in pre-transition societies was a personal optimum - or perhaps a satisficing19 number.  The only way to test this is to see if the same model can predict the TFR in both pre- and post- transition societies.  

2.0.2 Do we model this decision as “strictly rational” or “boundedly rational?”  

      Assuming that we decide that human behavior can be characterized as being rational in both the pre- and post- transitional period, a question arises: whether the decision is rational in the sense used in modern economics, or whether people are merely “satisficing” and thus “overshooting” or “undershooting” because of bounded rationality and imperfect foresight.  

      It is more convincing to argue that the individual decision to have children is boundedly rational. On the other hand, if we assume that the mis-conceptions and over-shooting of different individuals cancel out in the aggregate, then a good approximation can be made by rational decision making. If however, there are waves of optimism or pessimism (as took place in the Great Depression or after the World War II), akin to speculative bubbles, then there could be consistent biases away from the individually optimal solution for some duration of time. But on the whole, a rational decision making model should be able to capture most of the change in fertility. Therefore we shall model rational decision making and then allow for a margin of error in the observed outcomes. If possible, at a later stage, one would like to incorporate a fuzzy logic model to try to capture the bounded rationality of the decision. 

2.0.3 Does the utility function change? Role of family planning messages 

      A simple model of changing fertility would be to allow the utility function to change (not just shift). Leibenstein (1980) believes that there is at least some component of decision making which is evolutionary, in which decision making becomes increasingly rational as modernization progresses. In the context of fertility, Inkeles (1969) examined the characteristics of the personality of a modern man, an important one of which was “attempting to control births.” This was seen as an outcome of greater independence of decision making. There is thus an entire set of properties of ‘modern man’ which are reflected in these shifting preferences. This aspect also takes into account the strength of the motivation to change, including the desire to ‘convert’ others to their new viewpoint.  

      Similarly, Pritchett and Summers (1994) and many others believe that one of the consequences of increased levels of education is to cause preferences to change toward lower number of children, other things remaining the same. Rutstein (1995), however, empirically finds that women’s education is not critical to this change in preferences. As a consequence of modernization, attributed to shift in the “attitudes current in society,” a large shift seems to occur in the preferences even of uneducated women. Rutstein (1995) has systematically examined this issue and believes that “family planning programs do more than just act as passive providers of access to contraception… Through increased contacts with persons using contraception, gains in knowledge about contraception and legitimization of the expression of small family values and the use of contraception, through the mass media, cohort experience and official policy, family planning programs actively affect the fertility desires of couples”(italics mine). The World Bank (1993) also seems to believe in the efficacy of family planning messages in changing preferences at least to some extent: hence the emphasis on quality of the family planning programs. 

      Nabli and Nugent (1989) admit the possibility of deliberately changing the preferences of people. In the context of institutional change and development, they mention, that “[e]nhancing some preferences or changing them may have an effect similar to that of setting norms or rules in helping to resolve free-rider problems and reducing transaction costs” [emphasis added]. Kuran (1995) allows for such change and in fact, models such changes in preferences. While not completely relevant to the field of fertility, Easterlin and Crimmins (1991) have provided evidence, based on time series analysis, that “material aspirations increase over time with the level of income,” or, in other words, that “real income is being deflated by rising material aspirations” (Easterlin, 1996). What we want to note from here is the feasibility of the utility function itself changing over generations. Easterlin (1996: chapter 10) also talks about how the socialization of people shapes the way their tastes develop. According to Landes (1990), “Values are an especially thorny problem for would-be developers … Values and attitudes do change, but slowly, and their force and influence vary with circumstances. Many religious values operate, for instance, to impede the mobility and openness conducive to efficient allocation of resources and rational economic behavior.” 

      Therefore, using these very persuasive arguments, it is quite plausible that fertility in Thailand has declined rapidly compared to Costa Rica simply because a rapid change took place in the preferences for children. But such an explanation fails to tell us why this rapid change took place in Thailand and not in Costa Rica. Further, the analysis not only becomes tautological, but also nearly intractable. Therefore, most neo-classical economists object to such an analysis, and Stigler and Becker (1977) even wrote that “De gustibus non est disputandum,” or “tastes are not to be disputed,”  wherein they proposed a strictly ‘rational’ methodology for dealing with many issues of non-economic human behavior. Using their methodology, we can think of children as representing two commodities:  “success in transmission of genes” and “care in old age.” Conflicts can arise between the consumption of these two commodities, such as when an agent is old and his son is wealthy but the son does not take care of the agent. In such a situation, the agent will have the satisfaction of seeing his genes carried forward to the future, successfully, but will have the dissatisfaction of his son not taking care of him.  

      There is clearly a substitution between these two utilities, as can be shown by an example from West (1997). Elders of the Sumo Association, who are themselves retired sumo wrestlers, are required to sell their share in the Association by the time they turn 65. Usually, this share is sold to a top ranking wrestler who can afford the high price of this share. But if the wrestler marries the elder’s daughter, then the price of the share is considerably lowered. This can be interpreted as a substitution between the above two needs. When the future success of an agent’s genes (through his daughter and her progeny) is assured, then the agent is willing to forego a little of the old age care motive which would have been fulfilled with earnings from the sale of the share at its market value.  

      It is worth noting here that neo-classical economics did not begin with such a strict interpretation of human preferences. “In his Principles, Book III, ‘On Wants and Their Satisfaction,’ Marshall begins Chapter 2 by taking an anthropological view of the development of ‘wants’ or tastes. Those of the savage are few; those of civilized man become increasingly complex, varied, and subtle with the passage of time and institutional development” (Klein, 1994). This implies a changing utility function. Unfortunatel, by the next chapter, Marshall decided to keep taste constant while analyzing demand.  

      The key problem, I perceive, is not that economists do not recognize that preferences can change, particularly across generations, but that there are no mathematical tools available to model this change. Also, there is very little predictive power from such a model. Therefore, we too shall restrict our attention to neo-classical models in which the utility function is kept constant.  

2.0.4  Is the utility function of various individuals inter-related?  

      It is also possible to assume that the utility functions of each individual is inter-dependent on that of others. Therefore, the effect of social norms and pressures is extremely high in such cases. In such a situation, it is possible to have a ‘critical mass’ of opinion which is sufficient to enable a change in the preferences of all individuals at the same time. These changing attitudes can be modeled as per Granovetter (1978) or Kuran (1995). This aspect was initially touched upon by Duesnberry (1949), whose argument was taken further by Leibenstein (1950) who then showed the existence of bandwagon effects, such as those observed in fashion. Further, advertising and even the mere exposure to a particular good, is able to alter tastes.  

2.1 Decline in fertility 

      To explain the observed decline in fertility we go by the synthesis theory of Easterlin which seems to provide a viable, broad framework. The other, highly mathematical modeling approach, is taken by the human capital school of economics, started by Becker. We shall touch upon certain issues arising from that school, in due course. We begin by discussing the complex puzzle of human fertility. 

2.1.1  A basic neo-classical framework 

      In the standard neo-classical framework, we keep the utility function fixed and analyze the change in incentives, through an analysis of factors affecting the supply and demand for children.  

Classification of the determinants of fertility 

      The earliest classification of factors affecting fertility was made by Kingsley Davis and Judith Blake (1956). Bongaarts (1978) called those factors which have a direct and immediate (biological) bearing on fertility as proximate determinants of fertility. These are combinations of biological and behavioral characteristics that determine fertility, and include24 (a) entry into marriage or sexual union, (b) contraceptive prevalence and effectiveness, (c) induced abortion, (d) postpartum infecundability, (e) the frequencey of intercourse, (g) the waiting time to conception, and (h) intrauterine mortality.25   

      These determinants are to be distinguished from the basic or “true” determinants which refer to the “behind the scenes” determinants, such as education, urbanization, cultural factors, psychological and environmental factors. The old age security motive would be one such “true” determinant. It is believed that apart from the proximate determinants - some of which have changed over time, it is the basic determinants that throw more light on variations in fertility in pre-transition societies.  

      Easterlin has demonstrated the mechanism through which these determinants operate (from Chapter 15, on Modernization and Fertility, of a book by Easterlin): 

Figure 2: Modernization and Fertility: Evolving Approaches 

The supply-demand, economic-social-biological twin “Synthesis”  

      Easterlin (1978)26 created a channel through which the basic determinants operate. In his model, Cd represents demand for children and Cs the supply of children, then simply speaking, given certain preferences, 

      A situation like Cs - Cd  represents an excess supply of children over the number demanded. If the present value of the psychic costs (over a lifetime) of this excess supply are greater than the psychic costs of fertility regulation, i.e., ≥ RC, then parents will prefer to go in for fertility control. 

      Therefore, the chain of linkages as visualized by this model is: 

Figure 3: The chain of linkage of the synthesis model 

Step 1 

Reduction in demand  +    Increase in  → Increased demand for 

for children    supply of children   contraception

(economic/social)  (mortality revolution) 

Step 2 

Increased demand for  + Lower costs of  → Rapid decline in   

contraception   contraception   fertility

                            (family planning) 

      This model explains somewhat more than this. There is an intermediate step where there is in fact an increase in fertility. This is attributable to the fact that with the onset of the mortality revolution parents have lesser disease, and are able to have more children. But soon enough this “hump” disappears as parents find they have an excess supply of children.  

      Easterlin tried to operationalize this framework in Easterlin and Crimmins (1981).28 The World Fertility Survey utilized this theoretical framework.29 This synthesis model - also called the Pennsylvania school model, was formalized for the first time in a utility maximizing framework in Easterlin et al. (1980). Unfortunately, as Sanderson (1980) pointed out, the formal model led to rather ambiguous results. We therefore do not consider this model in greater detail here, but the basic thrust of the model would be kept in mind. 

      The question which naturally arises is: which of these sides, the supply side, or the demand side, is more significant in terms of explaining the changes observed in fertility over the past two hundred years?  

      We begin by an analysis of the changes that might have taken place in the supply side in the past one hundred years or so (and in the more recent past in the developing nations). 

Demand factors: [Desired family size, Cd]  

      Since the supply side explanation, taken in isolation, leads to lower declines in fertility than those actually observed, there must have been changes taking place in the demand for children.  

      The costs of fertility control can be psychic as well as economic. The psychic costs are influenced by the culture or tradition of the society. It is virtually impossible to quantify these costs. It can be said, however, that with increasing education, the psychological barriers against contraception are breaking down and these costs are perhaps decreasing. The other costs are the economic costs of contraceptive devices/ procedures. As technology has improved and the supply of contraceptives become heavily subsidized - particularly in developing nations,  the economic cost of fertility control has tended to decline. It is here that the family planning programs have played a very important role. The results of the famous Matlab experiment (Koenig, et al 1992) showed how a reduction of RC can reduce fertility. If there is a large unmet need in a society, then RC might be barrier facing the decline in fertility. In such cases (as explored by the Matlab experiment), the causality of Figure 3 can be turned around as in Figure 4: 

Figure 4: Significance of the costs of contraception: the Matlab viewpoint 

      While the slow growth rates of population in the agricultural epoch do not quite justify this argument, there is a widely held view which believes that there has always been a latent demand for contraception in the human species, which was not fulfilled till the arrival of modern contraception. The moment human beings were brought in the presence of  essentially costless modern contraceptives, they began to use them. The argument would sound plausible only if it gave sufficient reasons for the existence of such an unmet need. In the Matlab experiment, it was perhaps the intervention of the nurses and doctors in health and other education, that brought about awareness of alternative lifestyles with lesser number of children, and hence created the unmet need which was found in the Matlab area of Bangladesh.  

      Economists are generally uncomfortable with this argument of widespread unmet needs. Leibenstein (1978) shows how “many of the countries that achieved their fertility decline earlier did so prior to the widespread distribution of modern contraceptive practices.”35 Also, as Pritchett (1994) has pointed out that “even costless availability of contraception would not drive down ‘unmet need’ very far, a point confirmed by the existence of  substantial ‘unmet need’ even in countries with excellent contraceptive access.” They therefore discount the role of RC, and cite evidence that its was not the lack of availability of alterative means of contraception that kept fertility to relatively high levels but the higher demand for children. One would tend to support this view. The essential causes of the fertility decline have been changes in the demand and supply side. While definitely facilitating fertility decline, modern contraception cannot be cited as a major cause of this decline.  

      It is also possible that there are other effects of family planning publicity, for example, a change in the utility function. There is reason to believe that a common villager is not merely a wealth-maximizing individual, but gives ‘honor’ a considerable weight too.36 If honor is included in the utility function, we could have a situation where the promotion of family planning programs by respected individuals can cerate a bandwagon effect in favor of low fertility. In Costa Rica, since the catholic norms gives greater ‘honor’ or respect to having less abortions, there is less incentive to reduce the number of children, whereas in Thailand, such constraints presumably do not bind, and once a bandwagon has been created in favor of two children in a family, people would be forced to abide by the social norm since they care about the respect and reputation they command in society.  

Extension of the Synthesis framework 

      The neo-classical synthesis framework of Easterlin thus possesses the ability to accommodate various aspects of change in supply and demand. On the other hand, its limitations are that it is too general, and therefore its predictive power is limited, apart from pointing out the general direction of trends. There are aspects of decision making which are better modeled using a strict household model on the lines of Becker and Cigno, etc. Further, it does not take into account the macro and community level variables appropriately.  

      Therefore, two extensions of this framework have been considered in the literature: 

      A diagrammatic illustration of the flow of causality among relevant variables in the extended versions of the supply-demand synthesis is given below (from Nugent and Anker, 1990).  

3.1.2 Household economics approach 

      Becker is well known for pioneering the economics of the household, in which decisions of the household were placed in the neo-classical framework. The major approaches used here to explain fertility decline are: (a) quality vs. quanity, (b) altruism. We take a brief look at these approaches here: 

a) Quality vs. quantity 

      This incorporates a substitution between quality and quantity in the determining of fertility. In his “An Economic Analysis of Fertility”, 1960, Becker assumes that parents have preferences regarding both the number and educational level of their children, where the educational level is affected by the amount of time and other resources that parents spend on their children. Investments in children’s human capital may then be derived as a function of income and prices. As wages rise, parents increase their investments in human capital, combined with a decrease in the number of children. Becker uses this theory to explain, for example, the historical decline in fertility in industrialized countries, as well as the variations in fertility among different countries and between urban and rural areas.  

      This approach fails completely in my view, because it does not show why there has been a sudden incorporation in the utility function of mankind from quantity to quality. It does not also explain why there has been a significant decline in fertility in societies where there has been no major increase in income or in the opportunity cost of time  

b) Altruism 

      Economists have been aware of a strange twist in human “self-interest.” Very often, people seem to behave against their own self interest (not in a delinquent sense, as in suicide), but out of goodwill for others: “How selfish soever man may be supposed, there are evidently some principles in his nature, which interest him in the fortune of others, and render their happiness necessary to him, though he derives nothing from it except the pleasure of seeing it” (Adam Smith, in The Theory of Moral Sentiments). 

      Barro (1974) used the concept of altruism to prove the Ricardean equivalence. According to him, “if people are altruistic towards their children, (dynastic utility maximizers), an increase in public expenditure or cut in taxation financed by increased public borrowing will have no real effect other than by distorting marginal incentives. Realizing that present government generosity will have to be paid for by their descendents, people will in fact privately transfer to future generations an amount of resources matching the debt accumulated by the government” (Cigno, 1992).  

      This result has been critiqued by Wilasdin (1990) who showed that this result of neutrality of government debt does not hold if parents can choose how many children to have. “Given convex preferences, altrusitic parents will respond to the government’s increased borrowing partly by transferring more wealth to each of their children, and partly by having fewer children; government debt is not neutral” (Cigno, 1992). In other words, government debt can change the economy through changing N, the population. 

      Logan and Spitze (1995) point out that “older people tend consistently to be least likely to adopt the ‘pro-elderly’ position… Altruism, not self-interest, seems to govern the attitudes of the older generation in this sample.” What Logan and Spitze found is that elderly people are altruistic toward their young. This is precisely what we would expect to find on the basis of evolutionary biology.  

2.2 Lower limit on fertility 

      We have seen above that it is relatively easy to model the decline in fertility that has taken place in the past century. The Easterlin synthesis seems to capture a wider range of factors, but does not seem to do quite as good a job at predicting a lower limit to fertility. As Schultz (1981: 235) points out, we need a mechanism that will yield a long-term stable equilibrium population. According to the basic models which consider the asset value of children, “If technical change and the accumulation of physical capital continue to increase labor productivity and wages, and encourage further per capita investments in human capital, the price of time should continue to increase in the future and contribute further to the reduction in fertility. Consequently, the price-of-time hypothesis that drives the demand theory of fertility does not prescribe how in a closed economy a new long-run population-economic equilibrium is achieved.” He seems to be quite pessimistic about this, and wants the state to ultimately subsidize the production of children.  

      In other words, if incomes → infinity, urbanization → 100%, female education and status → the maximum possible, etc., then we might have that number of children born → zero, leading to the collapse of civilization. The corner solution problem has been highlighted in Cigno (1991), and Schultz (1981). Cigno, while discussing the predictive power of the model which considers children as assets, states (1991: 153): "the growth of the financial sector (including in that the social security system, as well as banks, private insurance, and the stock exchange) tend to coincide, in the development of an economy, with a sharp fall in fertility, the break-up of the extended family networks, and a widespread reluctance on the part of the middle-aged to accept responsibility for the maintenance of elderly relatives.” Then he goes on to mention that in such situations, “The fact that fertility does not actually fall to zero, even for couples who make little or no contribution to the welfare of elderly parents (and must, therefore, realistically expect the same treatment from their own children), suggests … that the demand for children is not entirely derived from the demand for old-age consumption. In other words, we need to assume that, for some couples at least, children enter the utility function of their parents in some form in order to explain why children are raised even when they are not required for old-age security purposes.”  

      There are three ways to incorporate a lower limit to fertility in economic theory: two ways are endogenous (or supposedly so), and one is exogenous.  

a) Endogenous positive lower bound: 

b) Exogenous positive lower bound: 

      According to my limited understanding of the two reasons postulated above, these so-called endogenous reasons are in fact exogenous reasons. For, we have no explanation for why the utility function for consumption of children is so shaped, nor why altruism exists in the first place. While the first explanation can only succeed by working backwards from data (i.e., parametrizing the concavity of the function in a way as to “fit” the data) the second explanation is actually exogenous, relying on our evolution to give us such a utility function. It is possible to explain the existence of altruism as an economic solution to a repeated game, as demonstrated by Axelrod (1984), but I suspect that such solutions are extremely tenuous, and susceptible to collapse given changes in technology, mobility of agents, etc.  

      The attempt to force endogeniety on the model in this artificial manner therefore appears to be highly contrived. Easterlin et al. (1980) did argue that there is a substantial biological, unavoidable, component to fertility  (as interpreted by Sanderson, 1980). Yet, there was no formal modeling of this important characteristic in their model.  

      It is therefore much simpler, in my view, given our state of knowledge of evolution, to allow the forces of Nature to ensure to this. From the fact that the human species has survived so far, it can be deduced that there is an ‘instinct’ to survive and to pass on one’s genes to the future. This ‘instinct’ or hardwiring has evolved as a powerful component of human nature over millions of years, and there is no reason to believe that it will suddenly stop functioning, or become overpowered, merely due to economic reasons. Mankind in fact, generally speaking, from the study of demographic history, has always had a tendency to reach an equilibrium where it just reproduced itself. This required it (a) first, to reproduce itself, i.e., to have children, and (b) to not overdo it. We can see tendencies toward both these things in the current time; only, the period is too short, and the overshooting that takes place due to boundedness of human nature, seems to be obfuscating the observation of such a response.  

      In any case, I do not believe that we can reasonably attribute a lower bound to human fertility through economic reasoning. The forces of nature are too strong to be limited by mere economic reasoning. I feel more comfortable, therefore, with an exogenous explanation of the lower bound of human fertility, than with an ‘endogenous’ economic explanation. Therefore, we can much simplify this matter, at least for purposes of this study (which is more interested in the decline, rather than in the lower limit of fertility), by postulating an exogenous evolutionary factor (EEF), which ensures that the number of children per household will, on the average, be equal to or greater than one (n ≥ 1). This factor is supported by other facts, too, which have not been explained in the economic models on altruism: 

a. There are no general markets for children. If we were to gain utility from children in general, including utility from investment in them, then it would make sense to have markets for children.  

b. Disproportionate expense on lost children: Not only does a market for children not exist, but also people strongly desire and are willing to spend huge amounts to “recover” their lost children, or to get custody of their children even after a household has been dissolved after divorce.  

c. Desire to give children a start in life better than one’s own: As Poulson (1994: 176) points out, “What we observe in most societies is that parents want to give their children a start in life at least equal to or better than their own.41 In some cases this may mean investing in their child’s education, but in other cases this may be accomplished by transferring land or other forms of wealth to the child at various stages of the life cycle.” Parents often undergo inordinate and seemingly irrational labor to promote the development of their children.  

      Further, if one asks any parent, he or she will never be willing in any way to place a monetary value on a child, and if asked persistently, will state that the value of the child has been long recovered. That does not mean that economic calculations do not come into the picture in determining fertility, but that once a child has been born, the economic calculations decline in importance, and ‘quality’ considerations come to the forefront. 

      The EEF helps us drive fertility away from corner solutions, as well as explain why everyone will produce their own children for all times to come, even if economic logic goes completely against children. We need only hypothesize that n ≥ 1. That does not preclude the collapse of human civilization, but it is hoped that it allows for other possible reasons for having a child, to play a positive role in determining fertility. EEF, as postulated above, is actually very weak. The intention is to merely support an equilibrium of perfect reproduction. It does not in any way drive women to have 15 children. Therefore, it is soon overpowered by economic factors, and allows the ‘rest’ of the fertility to be determined endogenously.  

      To postulate an EEF almost sounds heretical. As Bergstrom (1996) has stated, “It is easy to convince most economists that economic analysis would greatly enrich other academic disciplines, but economists are surprisingly reluctant to believe that reading anthropology, biology, history, psychology, or sociology is important for doing good economic analysis.” This work “has great potential to enrich our understanding of economic relations within families.” For example, it is possible to hypothesize that “human preferences were shaped by natural selection.” This leads us to genetically programmed utility functions. “Much as economists postulate that individuals maximize utility, biologists postulate that individuals maximize fitness.” 

      In an even more biologically driven approach, Berstrom (1996)postulates that “natural selection is for utility functions rather than for hard-wired actions,” in other words, it is assumed here that human beings are moving away from the ‘instinct’ programmed into animals toward getting utility in a rational maximizing manner. This avoids the problem of ‘programming’ implied by the EEF. But I am sceptical about such arguments because they imply that if children given sufficient disutility, then a corner solution could occur again, leading to a collapse of the human civilization. Bergstrom (1996) then mentions a semi-Kantian utility function that is half-way between selfishness and the Kantian ethic. According to the Kantian ethic, “parents do what they do for their children not because they like it (an ‘altruistic’ parent gets more utility feeding his child than feeding himself) but because they think it is right.”42 The semi-Kantian maxim is, “Act toward your siblings as you would if you believed that with probabilty one-half, your sibling would copy your action.”  

      One’s confidence in rejecting altruism as a valid cause for having children is supported by Cigno (1992). He finds that the predictions of the models of altruism do not match observed data. “Recent economic explanations of changes in fertility behavior have focused on the effects of labor-market-related incentives. The present paper draws attention to another set of incentives, those connected with the transfer of resources over time.  The theoretical implications of intergenerational altruism as a possible motive for having children and making transfers to them are considered, and contrasted with those arising from the competing hypothesis that such actions are motivated by old-age-security considerations.  From a comparison of these theoretical predictions with the findings of a number of empirical studies, it would appear that self-interested concern for one’s old age, rather than any great love for future members of one’s dynasty, is or has been so far the dominant force driving fertility and intergenerational transfers worldwide” (from the abstract of Cigno, 1992). 

      Before closing this topic, it is necessary to admit that the EEF sounds like human beings are at least in part “slaves to their hormones,” which is a major criticism of sociobiology and evolutionary psychology. I think that to deny this would be more irrational than admitting that we are at least partly determined by our nature as primates.43  

2.3 Link between Old Age Security Motive and Fertility 

      Having set the general framework for the paper, we explore in detail the role of the OASH in determining the demand for children.44 We will look both at the theoretical and at the empirical literature. The area of fertility is very complex; it is not quite easy to distinguish relevant theories from the irrelevant ones. We should also realize at the outset the limitations of considering human fertility in purely mathematical terms. Alfred Marshall said in 1890, in his Principles of Economics admitted that most economic phenomena “do not lend themselves easily to mathematical expression.”  

      The approach taken here is to consider most of the issues related to the OASH and fertility, and to try to illustrate the issues with theories and empirical literature. 

What is old age? 

      The word “old age” can only be precisely defined in the context of a particular study. In specific situations, the word might refer to the age of retirement, though most people at the beginning of their retirement might not agree that they are old. Old age is also a function of a persons’ health and state of mind.  

Old age security vs. pension 

      Dutta and Nugent (1984) define the old-age security motive as the need for “protection or insurance against inability to earn one’s own support because of disability.” They add that the “value of insurance against disability and old age should not … be construed as being limited to those with either the misfortune of being disabled and unable to work, or the fortune of having nothing to do.” Therefore it is quite possible for an old aged person to be employed and receiving wages, but still feel more secure as a result of the availability of his children. Further, this is not quite the same as the pension motive, which deals only with the earnings component in old age. Old age security is concerned with the risk factors, primarily. Even a weathy person (who presumably has adequate

‘pension’) might need a son for support: “the fact that wealthy men are more secure than poor ones should not necessarily be taken to imply that sons were not particularly important sources of support” (ibid). In this context, it is not merely in the developing nations that parents desire to be close to their children for old age security: “On the contrary, several sociological studies, notably in the United States, the United Kingdom, and Denmark, have shown that most old people have at least one child living nearby, often only a few minutes’ traveling time away” (Britannica Online).  

      It would thus appear that old age security is a stronger motive than the pension motive, since it includes non-monetary factors.  

      Gomez and Nugent (1995) state that “[b]y a considerable margin the old age security motive for fertility remains the least studied and least understood motive for fertility. Over the last two decades have appeared a variety of theoretical models claiming various positions on the motive’s relative importance, ranging all the way from ‘it cannot possibly be important since in present value terms there will always be better substitutes than children for providing support or security for old age’ (Lindert, 1980, 1983) to the ‘safety first’ argument which argues that children are a unique insurance substitute in rural areas of developing countries which lack both credit and insurance markets (Leibenstein, 1957 1975; Cain, 1981, 1986a, 1986b). The more realistic of the models would seem to allow for the possibility of an old age security motive, though by no means inferring that it would be the primary one.” 

2.3.1 Models which incorporate only the pension motive (children as assets)  

      The underlying neoclassical model here is that observed fertility outcome for any couple is the solution to their life-cycle optimization problem. In such models, children are viewed as assets. Parents are thought to evaluate the costs and benefits from children against the ‘market’ or other opportunity costs. The expected net present value of the marginal child determines the demand for children.  

a) Children might be valuable even before leaving the household: 

      It is possible that children in pre-transition societies provide a net positive value even before becoming adults and leaving the house. Nerlove (1991)  considers the important role in household production, as in the family farm, provided by children, in valuing children. According to this theory, children might have positive value even before they leave home.  

      Bulatao and Lee (1983) reviewed various studies on net child costs and concluded that average child (boys and girls combined) would not provide net flows to parents before leaving the parental residence. Thus children are not net producers while still with their parents. 

b) Children are valuable in pre-transition societies over a lifetime: 

      Becker (1960) argued that “… it is possible that in the mid-nineteenth century children were a net producer’s good, providing rather than using income.” Caldwell (1976) hypothesized that the average return from a child were positive in pre-transition societies over the life-cycle, and became negative in post-transition societies. According to Caldwell (1976, 1982), there is a transitional point in this cost-benefit analysis before which fertility is high and subsequent to which it is low. In the Caldwell thesis, the direction of the net wealth flows between parents and children plays and important role in fertility behavior. The demographic transition, according to him, hinges on the direction of intergenerational wealth flows. According to him, the change in the direction of the net flows from children to parents [when the children are used for labor, etc.] to parents to children [when the parents have to invest in the children in order to make them competitive in the market place], explains the transition from high to low fertility. “However, inasmuch as Caldwell has not been able to explain the turning point in the direction of these net flows, even the considerable interest and controversy that this thesis has generated has not been sufficient to allow its integration into the supply-demand framework” (Nugent and Anker, 1990). We also see that empirically, this hypothesis is extremely suspect. 

      Cain (1977) in a study in Bangladesh,45 Caldwell (1982) in a study in Nigeria, Dow et al (1994) in a study in Kenya found that there were positive average returns to children in these nations. However, there are a few criticisms of these views: “Caldwell’s (1982) own analysis of transfers in Nigeria is primarily based on qualitative data and hardly sufficient to provide a reliable quantitative answer on the net flows between parents and children” (the Ivory Coast article). Also, it has been felt that while the analysis of Dow et al is provoking, “their data are also subjective and incapable of determining net economic flows” (the Ivory Coast article). 

      On the other hand, various studies have shown that the net asset value of a child is actually negative. Mueller (1976)’s findings about the value of children in peasant agriculture violate the Caldwell hypothesis. She analyzed detailed life-cycle data of developing countries and found that children are net financial burdens on parents in peasant societies when survivorship is included and a zero discount rate is used. Lee and Miller (1994), while studying intergenerational flows in the 1980s in the US, confirmed that at in the post-transitional societies, the present value of children is negative. Paul Turke (1989) carried out field studies in the Micronesian islands of Ifaluk and Yap and concluded that children tend to be a net economic burden on their parents. Studies of hunter-gather tribes (the Ache in Paraguay, the Piro of Peru, and the Masiguenga of Peru), carried out by Kaplan (1994) also reject the Caldwell hypothesis. Thus evidence seems to be increasingly pointing to the NPV of children being negative in both pre- and post-transition societies. Or, at the best, as in the case a study of USA by Lee, the return was seen to be very low - between -1 to 1 percent.46 

      Various studies have tried to measure the economic value of children. Ahn (1995) estimates the gender- and age-specific values of children using a dynamic programming model. Other studies that have carried out important analyses of intergenerational transfers, include Altonji, Hayashi and  Kotlikoff (I992, 1995), and McGarry and Schoeni, Cox, on developed nations. There have been studies by Lucas and Stark (1985), Ravallion and Dearden (1988),  and Hoddinott (1992) on developing nations, and Lee, Parish and Willis (1994) on the study of nations with strong norms of filial loyalty 

      “Overall, the evidence seems strongly consistent with the evolutionary view as expressed by Kaplan. Over the course of  a lifetime, resources tend mainly to flow from the old to the young and not the other way around” (Bergstrom, 1996).  

Basic model:  

      Leibenstein (1957) identified three motives for fertility: (i) the consumption motive, (ii) the production motive, and (iii) the old age support motive. Initially (Leibenstein, 1957) focused on the pension motive.  

      utility = f(security, consumption levels of both pecuniary and non-pecuniary) 

where children are assumed to be able to contribute to each type of utility. Leibensten then argued that the costs of children are usually lower than the benefits from them, particularly in LDCs, and therefore there is an incentive to have more children there than in developed nations.   

An overlapping generations model with externalities: 

      “Neher (1971) was the first to use the Samuelsonian life cycle model to represent fertility and other choices of the individual decision-maker. Moreover, he built into the model certain characteristics of developing countries such as incomplete capital markets and added a third generation, namely, dependent children. In order to focus more specifically on the old age support motive for fertility, he imposed the following simplifying assumptions:  

      “Fertility decisions are made by young adults who choose the number of children in such a way as to maximize the utility they derive from their consumption streams in working adulthood and old age. With technology fixed and two inputs, land (which is in fixed supply) and labor, there is one family size which maximizes per capita consumption. Once that size is attained, the socially optimal fertility rate is merely the replacement rate, i.e., the rate needed to maintain the optimal family size. Each young adult couple, however, has an incentive to exceed this replacement fertility rate inasmuch as the couple can look forward to receiving all the benefits of children (i.e., support in old age) while paying only a portion of their costs (those of childhood but not of their old age). As a result, the old age pension motive generates” (Nugent and Anker, 1990).  

      Neher concluded that it is the existence of an externality in the benefits and costs of children which makes the old age security motive result in excessively high fertility rates. 

A simple graphical model of the old age pension motive and fertility:   

      Nugent and Anker (1990) describe a simple graphical model which captures much of the substance of the issues in relation to old age security and fertility. This uses the cost-benefit analysis of the marginal child. 

Other complexities: 

      The issue, in my view, is not quite so simple. Institutional factors come into play here. Consider the case of Costa Rica where there is a pension system in place. Pension systems are expensive, and they are inequitable: being ‘progressive.’ Therefore, there are likely to be some people who gain from their taxes put into pension funds, and others who lose, i.e., those who could have done better by placing these savings into the market. In such a case, the person who loses from the pension programs might have an incentive to increase fertility, as illustrated in the table below.

 

      We observe that when the NPV of a marginal child is positive, there will be an incentive to increase fertility  (supporting Caldwell’s pre-transition hypothesis). But when the NPV is negative (as has been empirically found, in most cases) the loser from the pension programs might have an incentive to maintain or even increase fertility. Therefore, the total effect of a pension program are not quite obvious. 

      Given the rapid nature of change of the environment, models which incorporate such change are a valuable addition to the literature. Joseph (1979) has a model in which  “a couple compares the number of children it desires with the number of children it has and decides a spacing strategy… The decisions are reviewed periodicically and may be revised as circumstances change.” 

c) A system of pay-as-you-go pensions, requiring net transfers from young to old: 

      If we consider the case of Costa Rica, however, we would like to model a formal social security system which provides pensions. One of the early, non-intuitive models of this nature was by Samuelson (1958). He “attempted to demonstrate the viability of a formal “pay-as-you-go” system of old age support such as the social security program in the United States. He did so with the help of a two-generation life-cycle model. In such a model everyone is in one of two generations, the younger generation which constitutes the labor force and contributes to the support of the old generation, and the old generation itself, which lives entirely on the support received from the younger generation. Samuelson demonstrated that the social security system of intergenerational transfers from young to old would be viable indefinitely only if the base of the system were growing steadily as a result of growth in either productivity or population. In the absence of productivity growth, population growth would be necessary for the viability  of a satisfactory system of old age support. (Subsequently, however, Aaron (1966), Samuelson (1975a, 1975b) and Arthur and McNicoll (1977, 1978) showed that Samuelson’s conclusion holds only if the capital-labor ratio is maintained in the face of higher population growth)” (Nugent and Anker, 1990). 

      We see that pensions based on net intergenerational transfers to the old are not easily supported in equilibrium unless there is population growth in the economy. This result goes clearly against the policy prescriptions which suggest that pay-as-you-go systems of pension would reduce fertility. 

Role of pensions and social security: empirical evidence 

      “The effect of social security systems [on fertility] … is not well established. There are four kinds of data that have been brought to bear on this issue: (1) cross country data, (2) anthropological analysis of fertility and social security in certain countries, (3) survey data on people’s stated motives for having children and (4) experimental data on savings programs designed to provide incentives for reduced fertility. These data sources vary enormously in their richness and their findings” (Cochrane, 1988).  

      One should also point out that the empirical literature is clear on the fact that children are not the only investment of the peasant.  For example, Simon (1980) cites various studies of Indian peasants that show “that poor Indian farmers save very respectable proportions of their incomes - cash savings of perhaps 12% gross and 8% net. And when non-monetary saving is included - as it should be - ‘the gross savings-income ratio among rural households would rise to 20% or so.’”  This shows that even in the absence of asset markets, children are not the only asset. Unfortunately, given that farmers have limited access to banks, their cash savings usually can be expected to earn negative interest: hence, children, who are also likely to earn negative net returns, are substitutes for savings. 

3.3.2 Models which include children as insurance (old age security motive) 

      As noticed above, from empirical findings, it appears that children have are have negative net present values; this would imply that if we only consider the asset value of children, then we might very well get a corner solution of zero children rather quickly. On the other hand, if the insurance aspect is included, and the OASM allowed for, we are able to get a positive demand for children despite negative asset values.  

      It makes good sense that in subsistence societies, one of the chief reasons47 why children are demanded is because of the risk aversion of agents. If we weight the mean return from a child with the insurance provided by the child in the absence of alternative markets for insurance, we arrive at a positive demand for children. In 1968, Leibenstein had restated his initial fertility theory and “placed priority on the security motive as a result of the ongoing and insecurity-increasing tendency in LDCs of the fragmentation of the extended family” (Nugent and Gillaspy, 1984). 

i) Ensuring ‘loyalty’ through internalization of social norms 

      A good model  of old age security has not only to show that there is a need for insurance and that in the absence of markets for insurance, children will be utilized for this purpose, but it has to show the incentive structure for children to be loyal to parents when there is such a need. Different models use different procedures for enforcing the returns from children. All these methods are costly and not Pareto optimal, since they involve overcoming opportunism (As mentioned earlier, we discount any explanations that require the assumption of altruism).  

      “In a wide-ranging and illuminating essay on social norms, Elster (1989:113) suggests that: ‘Intergenerational reciprocity is … found between parents and children. Assuming that parents cannot disinherit their children, the latter have no incentive to take care of their parents in old age … Yet, most societies have a norm that you should help your parents; in return for what they thought (allied to other thoughts) to argue that this third type of answer we are considering here is of no use; that internalization of norms is the central means by which norms are in fact sustained.” 

      Dasgupta (1993) works out a simple overlapping generations model for this purpose. According to him, the “near-stationarity of both kinship lines and the circumstances facing people in traditional societies together imply that mutual insurance arrangements don’t always look like mutual insurance agreements. There can be layers of behavioral norms and rules whose compliance sustains a variety of insurance arrangements. There is nothing mysterious in such acts of reciprocity; certainly, there is no reason to invoke the idea that there is greater innate generosity and fellow-feeling among poor people in poor communities than exists among members of modern urban societies.” “… Within rural communities there is thus an integral system of mutual insurance against illness, production failure, and general bad luck.” 

      Dasgupta then asks the extremely relevant question, viz., how is it that we these contracts are enforced. The answer to this is extremely relevant in the case of old age security.  For example, how do parents enforce reciprocity from their children? Among the three possible answers that he examines, the most prominent appears to be the internalization argument, viz., that the practice of reciprocity is internalized by each of us over time through communal living, role modeling, education, and through experiencing rewards and punishments.  

      “This process begins at the earliest stage of our lives. We internalize social norms, such as that of paying our dues, keeping agreements, returning a favor; and higher-order norms, as for example frowning on people who break social norms, and so forth. To use the language we have developed earlier, the claim here is that a person’s utility function is itself a reflection of an ordering over actions in part driven by social norms. By internalizing a norm, a person makes the springs of his actions contain the norm. He therefore feels shame or guilt in violating a norm, and this prevents him from doing so, or at the very least puts a brake on his violating it unless other considerations are found by him to be overriding. … Now it is evident that people differ in the extent to which they internalize social norms. They also differ in the extent to which they are willing to trade off the dictates of norms against personal desires, other commitments, competing loyalties, and so on.   … social norms could be self-sustaining even were the socialization process ineffectual.”  

      The internalization of social norms creates a set of beliefs which are then shared by all members of the society and which can then support various sequential equilibria.  

ii) Ensuring loyalty through the feeling of guilt: 

      Becker (1993) believes that parents can create a feeling of guilt in their children to ensure provision of old age support. This can be subsumed in the concept of ‘after-life’ introduced earlier, and the study of institutions designed to promote guilt. 

iii) Ensuring loyalty through strategic bequests: 

      Goode (1963) and Bernheim, Shleifer and Summers (1985) have models wherein parents use bequests strategically to promote loyalty from children. But Cox and Jiminez (1995) reject this as being an important theory in economically backward societies since they claim that there is not much asset accumulation in any case to promote the motive.  

      I am not quite sure whether the criticism is justified. In fact, it is in poorer societies that the marginal utility of even smaller amount of assets would be higher and act as incentive to the children who need every bit of assets that they can get in order to raise their standards. It is in the societies which are seeing rapid income growth that the marginal value of the bequest of the parents would be low. Also, one of the use of property rights as developed over the ages, was perhaps to ensure this loyalty from children. 

iv) Ensuring loyalty through mutual aid: 

      Kotlikoff and Spivak (1981), Cox (1987), and Cox and Stark (1992) have models on these lines. The problem here is that infirm old parents can hardly be of any aid to children. Therefore, the quid pro quo might not support the infirm parent. 

v) Loyalty through hard-wired ‘sense of duty:’ 

      Wilson (1993) argues that we are hard-wired (relatively weakly) to perform their duties toward the society. To paraphrase Wilson (1993:113) in the context of old age security, a tiny and remote chance of being honored by one’s society is of far higher value than the immediate gain from not looking after one’s parents. He points out that utilitarian calculations can become extremely hollow when honor is at stake. He traces the development of the sense of duty to a strong attachment to one’s parents, as shown from various studies.  

      To me, this ‘hard-wiring’ seems to be only a starting point. In the absence of institutions to support this, the sense of duty may or may not be adequate to support the old in their times of need. I would expect that some children would be altruistic toward their parents by virtue of this ‘hard-wiring’ even without other threats, but it is perhaps much easier to ‘frighten’ them into being virtuous by showing them the possibility of ‘hell’ if they do not look after their parents. 

vi) Ensuring loyalty through implicit intergenerational contracts and migration: 

      Lucas and Stark (1985) have a basic model on this line. Developing this line further, Arcand et al. (1995) develop a two sector general equilibrium, three period, overlapping generations model which incorporates intra-family and inter-generational contracts. According to this model (which resembles the strategic bequest model considerably), parents set up incentive constraints for children who migrate to urban areas to ensure that they “remit,” on the lines of the principal-agent problem. “In period one, the parents present those among their offspring who decide to migrate with a ‘gift’ which allows them to migrate to the city, accumulate human capital, and consume a sufficient amount to survive. In period two, the migrant secures an urban sector job with probability one and remits to his parents in the countryside. For this arrangement to be self-enforcing however (altruism is assumed away here so as to focus on what is individually rational), it must be the case that the migrant’s parents hold a ‘sword of Damocles” over his head were he to fail to remit. Otherwise, it would not be individually rational for the offspring to remit in the second period and, knowing this, the parents would not have extended the initial payment in the first period to begin with, thus unraveling the whole migration process from back to front. The existence of the last period threat available to the parents in the case where the migrant does not remit is thus the key to the whole Lucas-Stark approach.  

      “There are a number of social structures which may easily furnish the parents of a migrant with credible threats. These include the potential return to the village of the migrant in the final period of his life (either to retire, or as was suggested to us by a Cameroonian student, to be buried), the desire by the migrant (who is usually a younger male) to eventually wed a woman from his native village, or the eventual bequest of part of the parents’ assets to the migrant. The precise form of the threat is likely to differ across cultures: what matters is that it exists, for it is the threat which eventually sustains the entire process, if one abstracts from altruism.   

      “In order to be able to integrate their approach in a tractable manner into our growth model, we will assume that the failure to remit results in the guilty offspring being subject to a “social sanction” which causes him a loss of utility which is an increasing function of the magnitude of the initial gift he was provided with by his parents in order to allow him to migrate to the urban area.”  

Complexities in the old age security motive: 

      The old age motive has been downplayed in the literature for the following reasons (compiled in Nugent and Anker, 1990):  

      People in developing countries are thought to have short planning horizons;  

      An illustration of a critique of the relevance of the OASH is in Vlassoff and Vlassoff (1980). They “provide an excellent piece of work on the old age security arrangements and attitudes among men in India. The theoretical and anthropological studies raise several questions (1) granted that children may be an important source of old age support,  how many children are needed to ensure adequate support? (2) given that men tend to have very short periods of inactivity prior to death in many poor countries and women tend to be considerably younger than their husbands and are longer lived in most areas, are women more likely than men to put a premium on the old age security for children?, and (3) how much of the need for old age support is financial and how much is emotional and thus the desire to live with children in old age may not be a financial motive? Each of these questions are important for designing policy. Some survey data can provide insight into these factors, but relatively little work has been done” (Cochrane, 1988). Dutta and Nugent (1984) have a critique of the approach of Vlassoff and Vlassoff, and it appears that arguments against the relevance of the OASH are rather weak. 

Existence of the old age security motive: empirical evidence 

a) Value of Children Study 

      I can do no better than to quote from Cochrane (1988), who has an excellent summary of the findings from this study. “The most thorough examination of the motives for child bearing in developing countries is the Value of Children Study...  data seem to indicate that husbands and wives express the same relative importance to the need for children for old age support and this declines in about the same way as national fertility declines… In those countries where there is not a strong preference for sons (the Philippines and Thailand), the extent to which people expect old age support from sons and daughters are about equal.  In countries such as Korea and Taiwan where sons are valued much more than daughters, there is a wide difference between the extent to which husbands and wives rely on sons rather than daughters for old age support.  

      “There are several interesting questions that need to be addressed with such data: (1) what are the determinants of such attitudes towards the utility of children?, (2) whether the desire for old-age support affects family size preferences?, and (3) whether the desire for old age support affects behavior? The Value of Children Survey did not provide analysis separately for old age support but combined it with data on financial assistance from children. The study found significant negative correlations between economic motives for child bearing and parental education and income in all six of the countries in their sample at that time (Arnold et al, 1975, p. 109). A recent study of husbands and wives in Pakistan showed educational differentials in the amount of support expected from children (Sotoudeh-Zand, 1987). A study of Egyptian husbands and wives showed strong negative relationships between the extent of old age support expected from children and parental education  in both urban and rural areas.  Differences across income and expenditures groups were also sharp, but not always monotonic (Hallouda et al, 1983). The effect or reliance on old age support on desired family size is not well established. The Value of Children Survey found that among individuals in Korea, Taiwan, Hawaii and the Philippines there was a significant positive association between the economic motivation for childbearing (including old age support) and ideal family size. No significant relationship was found in Japan or Thailand. A study using Egyptian  data mentioned above found a significant effect of expectation to live with children in old age and desired family size among women in urban and rural areas. There was no significant effect for men. Nor was the extent of financial support expected from children significant for men or women (Cochrane, Khan and Osheba, 1988).  

      “There is little data on whether the old age motive for childbearing causes a couple to use contraception earlier or more frequently as would be expected if it had important effects on desired family size. Bulatao in analyzing the Value of Children data compared the attitudes of those who were high parity contraceptive users with low parity users. There was some relationship between the expectations of old age support and contraceptive use, but in some cases the differences were small and in several cases were reversed (Bulatao, 1979)” (Cochrane, 1988). 

b)  Other cross country studies 

      “The major cross country studies that have attempted to estimate the effects of pension programs on fertility are that by Holm in 1975, reanalysis of his data by Kelly et al in 1976, reanalysis by Holm himself in 1976 and a more recent effort by Entwisl and Winegarden in 1984. Holm attempted to test two hypotheses: (1) the coverage of a country’s old age or retirement, invalidity, and surviorship programs will vary inversely with that country’s subsequent fertility, and (2) The benefit level of old age, invalidity and survivorship programs will vary inversely with subsequent levels of fertility. Data (partly generated) from 67 countries showed strong zero order correlations between coverage and benefit levels and total fertility rates. …” (Cochrane, 1988). 

c) Review of empirical literature by Cigno (1992)

 

      Empirically, therefore, there seems to be a relatively strong existence of the linkage between old age security motive and fertility. Therefore, it is a topic worth exploring in greater detail. 

      We must note again, that there could be other issues involved in the OASM, such as gender issues, based on the relative importance of decision making in the household, and so on. In the basic model, on would have to try to exclude such complications, but these should be included to the extent possible, in the later models. 

2.3.3 Models that emphasize altruism 

      “Altruism is incompatible with the old-age-security hypothesis (Cigno, 1992).” We now take a brief look at models that satisfy the old age security motive through altruism. It is thought that children are altruistic toward their parents, apart from parents being altruistic toward their children, as already explored earlier. Models of single-sided altruism, while capable of explaining why there could be a lower limit on fertility, are unable to explain how the old could survive through the agricultural epoch. If altruism runs from parents to children but not the other way round, aged parents may well want but will not get any support from their grown-up children. The reason is pointed out in Cigno (1991, Chapter 9): as it only cares for the subsequent one, no generation will want to pay anything to the previous generation, because it will not wish to recover the cost of that by lending to the next generation at more than the market interest rate of interest. Therefore we shall not observe net intra-family transfers from the middle-aged to the old unless children are altruistic towards parents” (Cigno, 1992). 

      Therefore Becker and Barro (1988) include models of two-sided altruism. They argue that their “… analysis contains both the Malthusian and neoclassical models since fertility is endogenous and rates of return on investments in physical capital decline as its stock increases. The endogeneity of fertility also leads to multiple steady states: A ‘Malthusian’ undeveloped state with high birth rates and low levels of human capital, and a developed steady state with much lower fertility and abundant stocks of human and physical capital.” “In the Barro-Becker model the structure of individual preferences induces a dynastic utility function which generates the first-best rate of population growth. In other models, while the rate of population growth might not be the only one which maximizes social welfare for a particular set of intergenerational welfare weights, it is Pareto optimal. This is so since even though the given equilibrium rate of population growth may imply very low child quality and therefore welfare (the so called “repugnant conclusion” of Parit (1984)), it is the rate which maximizes the utility of the parents/ family and hence, for example, lower fertility, while it might improve the welfare of the children, would reduce the utility of parents/ family” (Baland and Robinson, 1996). 

      Zhang and Nishimura (1993) assume children’s altruism toward parents in a two-overlapping-generations model with endogenous fertility. Parents raise children because, when retired, they expect gifts from their children who are essentially a capital good. Individuals’ behavior between generations is examined by analyzing a Nash equilibrium, which is then compared with a social planner’s optimal allocation.  The pay-as-you-go public pension program is viewed as the optimal gifts from the optimal allocation when the latter is implemented.  The effect on fertility of the introduction of a capital market is also analyzed.  The validity of the old-age security hypothesis is shown to depend on the parameters of utility and cost functions. 

      Cigno (1991) shows how two-sided altruistic models are compatible with the argument that “fertility appears to fall drastically, but not to zero, with the development and widespread availability of the market (or State) based methods of providing for old age may thus be taken as additional evidence that, for some couples at least, children are a source of direct utility.” 

      To counter-argue the utility of such models, we repeat our argument that there is no reason that man evolved with altruistic notions toward the old. In fact, Hamilton (1964) and Dawkins (1976) strongly question the notion that transfers from children to parents make any evolutionary sense. For children to devote resources to parents would not be an evolutionarily stable outcome, since the parents are likely to be beyond their reproductive age.48 The fact that humans have been able to live longer than their ‘useful’ life to nature, can therefore not be attributed to an altruistic motive which suddenly developed toward the old along with the agricultural revolution. It must be an artificial construct, and hence, explained by institutional design rather than by altruism.   

2.3.4 Models that allow change in preferences  

      According to this view, preferences for old age security are changing over the course of time as a result of the influences of modernization. Bengson (1993) offers viewpoints on the changing preferences on the role of children in providing old age security. Lee (1980) reminds us that it is not necessary that only the incentives for reproduction might be changing the observed demand for children, but the very utility function might be a function of time and modernization. He uses what is called a “stock adjustment models” of fertility. Achenbaum (1993) has a model of this nature too. However, as discussed earlier, we shall not investigate models of this nature.

 

Chapter 3

Theoretical and Empirical Models to be used in this Dissertation 

      We must keep in mind what Nugent and Anker (1990) have so perspicaciously stated in their (1990) paper: "... it is possible to construct theoretical models of the relation between either old age support or old age security and fertility that gives virtually any kind of result, ranging from unimportance to paramount importance. Moreover, since differences in implication can be traced to rather specific assumptions, often of questionable realism, it is clear that the controversy can be cleared up only through carefully done empirical studies." The crucial aspect therefore is the determination of suitable empirical models to test the relationship between old age security and fertility. But it is also worthwhile to determine a suitable theoretical model, if possible. This section is in the nature of an exploration in these directions, of empirical and theoretical modeling. 

3.1 Expected relationships 

      Data is available for two categories of people: the young (who have not completed their fertility so far), and the old (who have completed their fertility. I am proposing below, based on the survey of the literature, above, the following relationships which need to be tested. At this stage, one should also mention that the proposed relationships are not uni-directional, i.e., there is simultaneity in some of the relationships. Possible econometric models will therefore be written down in due course. 

3.1.1 Desired fertility 

      From our study so far we expect desired fertility of a peasant in rural areas of developing countries to display the following relationships: 

      Directly related to 

3.1.2 Completed fertility 

      It might not be possible to get a measure of total fertility, but completed fertility can be used in its place. It can be expected to display the following relationship: 

      Directly related to 

      The data is rich in data relating to inter-generational transfers, and much other information. It is not quite clear to me at this stage, however, whether suitable proxies will be possible to create for various factors mentioned above. Nugent and Anker (1990) have proposed an extremely complex, modular approach to the testing of the relevance of the hypothesis. It will require considerable effort to convert the theory  and this approach into suitable econometric models, where variables are supported by the available data. “Researchers have applied different econometric techniques to examine the determinants of fertility. The most common estimation techniques have been single equation ordinary least squares (Schultz, 1978; Olsen, 1980; Lee and Schultz, 1981), Tobit maximum likelihood (Zhang, 1990), and sequential logit (Zhang, 1994)” (Asgary and Pagan).  

3.2 Theoretical models 

      We find that it is extremely difficult to create a neo-classical  model that shows the above, ‘desired’ relationships.  

      We begin with the model of old age pension and fertility proposed by Nerlove et al (1987:120), and then, after exploring its weaknesses, propose a modified, basic model with uncertainty. Thereafter, I show other models that can be developed on these lines to incorporate the complexities observed in nature. I then try to write a few basic empirical models which will test these theoretical models. Finally, I list some of the other complexities we need to incorporate into empirical modeling to take into account the complex inter-relationships that are observed in reality.  

3.2.1    A basic model of children as capital assets (pension motive) 

“A simple model of the Old Age Security Hypothesis (Nerlove, 1987:120)” 

      For convenience, I am merely reproducing verbatim below the model from Nerlove et al, only modifying a few notations for purposes of uniformity with other models to be written down in due course, and adding a few expository points at various places. 

      “Let parents live for two periods during which they consume c1 in the first period and c2 in the second. Utility is assumed to be a function of only c1 and c2, i.e., u = u (c1, c2). All income is assumed to be produced by labor alone, and parents and children are assumed to receive an endowment per capita of y1 and y2, respectively, in the two periods, measured in units of consumption. Parents are assumed to earn nothing in period two and subsist on transfers from their own children and returns from prior investment. Thus, total income received by the household in the first period is y1, and in the second, is ny2, where n is the number of children per family. Children are assumed to consume x1 in the first period of life and x2 in the second period of life when they are productive. For the moment, we consider x1 and x2 to be exogenously given (at conventional substistence levels). The difference between parents’ consumption plus childbearing costs in the first period represents savings, s. 

                  y1 = c1 + s + nx1      (1) 

In this model, savings represent only a transfer via investment from period one to period two, and no borrowing from the future is possible. As before, suppose that such investment returns R units of consumption of period two for every unit of consumption foregone in period one.  

      Each child consumes only x2 (≤ y2). Thus, parents’ budget constraint in the second period is 

            ny2 + Rs = c2 + nx2  c2, s, n ≥ 0     (2)

      

 Suppose first that there is no capital market, so that s = 0, by definition, and children became the sole means of transferring consumption from the present to the future. In this case one can solve for n from (1) and (2) to get 

                        (a) n = y1 - c1 

                                       x1     (3) 

                        (b) n =      c2  _

                                      y2 - x2 

      Equations (3a) and (3b) can also be used to solve for c1 and c2 as functions of n: 

                        (a)  c1 = y1 - nx1

                                                                  (4) 

                        (b) c2 = n (y2 - x2) 

      One can view (4) as defining parametrically (via n) the consumption possibility frontier of the parents in c1-c2 space. One can also substitute (3a) into (4b) to get the direct relationship between c1 and c2: 

      This consumption possibility frontier is depicted in Figure 5.1.1. 

      Parents choose the point on the consumption possibility frontier (5) that maximizes their utility function u(c1, c2), a point (c1*, c2*) in Figure 5.1.1. Once they find the optimal consumption bundle (c1*, c2*), the optimal number of children n* is determined from (3a) or (3b). 

      Observe that an increase in the cost of children, x1, reduces the slope of the budget line in Figure 9.1 and the intercept with the vertical axis, leaving unchanged the intercept with the horizontal axis. Therefore, if c2 is not a Giffen good, then c1 falls and, by (3b), n also falls. However, the effect of a decrease in the return from investment in children, (y2 - x2) on the number of children is ambiguous. Such a change has the same effect on the budget line as before (making c2 more expensive in relation to c1) and again, c2 must fall if it not a Giffen good. However, whether n falls or rises depends on whether the decrease in c2 is proportionally higher or lower than the decrease in (y2 - x2); see (3b) since the return to their investment through children falls, families may need to invest more (i.e., have more children) even if they are content with consuming less in the future. In contrast, an increase in the parents’ endowment, y1, has a pure income effect. The consumption possibility frontier shifts upward without any change in its slope; c2 increases if it is a normal good and thus, by (3b), the number of children n, must also rise.  

      In general, families differ in the amounts of endowment parents have (y1) and can expect their children to have (y2). Both affect the number of children desired, but differences in the latter affect the rate of return on investment in children, so that, if an alternative means of transferring present to future consumption, such as capital markets, is available, the total number of children as well as their distribution among families may change.  

      If a capital market exists, in the sense of an alternative means (to children) for transferring present to future consumption, s may be strictly positive for some families. In this case, (1) and (2) may be consolidated by substituting for s in (2) from (1): 

                                          (6) 

with an added requirement that s = y1 - c1 - nx1 ≥ 0. The expression in large brackets is the net present value of having a child. It consists of a return (y2 - x2) in the future, which is worth only (y2 - x2) / R in the present, and a cost, x1, in the present. Clearly, since n does not enter the utility function, a family will have children only if  

                                                            (7) 

Thus, in the presence of a capital market, those families for whom x1 or x2 is sufficiently high will have no children and will transfer present to future consumption via the capital market. Families for whom the expected endowments of their children are sufficiently low may also choose to have no children. Those families for whom the rate of return on investment in children is sufficiently high will not save at all: they will hit the constrant s = y1 - c1 - nx1 ≥ 0 as they try to increase n. Consequently, they will be subject to exactly the same constraint as in the absence of a  capital market, i.e., equation (3), and will demand the same number of children. Since some families will have no children, total population must be lower than without a capital market. This analysis is the essence of the old age security hypothesis” [end of reproduction]. 

Implications of this model 

This basic model can show some interesting phenomenon, such as the effect of change in y1, x1, and y2-x2,49 etc., which might make ‘intuitive’ sense. But because the model fails to deliver, in the general equilibrium context, we restrict analysis of such implications. 

Critque of this model 

(a)  The budget constraint for the second period does not appear to treat the expense incurred on the n2 grandchildren. However, that objection can be overcome if we assume that x2 represents this expense, somehow.  The main objection is that the returns from children are assumed in this model to be guaranteed returns, while this might not hold. 

(b) The opening up of capital markets is seen to push fertility down to zero in this model. This is never observed, even in the advanced countries with relatively advanced capital markets. 

(c) As Nerlove et al show further, in their book, this model fails to show a decline in over-all income in the presence of capital markets in the general equilibrium framework. That is essentially because those whose return from children is high, will borrow from those who do not want children, until the returns from children and assets will equalize. 

Basically, therefore, treating children merely as a form of assets does not show why there should be any affect on fertility at all. 
 
 
 
 

      Unfortunately, I have not yet been able to find out models in the literature which incorporate children as insurance. Therefore I have tried to work out such models. The work done in this direction so far is summarized below. 

Model 3.2.2.1  A basic model of children as risky assets 

      Here I extend the Nerlove model to incorporate children as risky assets. I assume the CRRA form of VNM expected utility function since we assume that agents wish to smooth their consumption across time.50 The function is assumed to be separable in the two periods and have all ‘nice’ properties enabling us to ignore the second derivatives. The household is modeled as a single unitary entity with utility function where there are two states of nature in period two, with the first state representing the probability of returns being received from the children, with a probability of λ and the other representing the probability of default from the children. Then we can write the agent’s problem as  

      Max  

subject to  c1 = b + y - nαc1  -  a  - p

                         

            c21 = a (1+r) + n (P - B) 

            c22 = a (1+ r) 

            n ≥  1     (EEF, discussed earlier) 

where  b  = bequest received by agent from his/ her parents

      B  = bequest given to children subject to their taking care of the agent

                  in agent’s old age

      y  = agent’s income in youth (there is no income in old age)

      p = ‘pension,’ or consumption good given to agent’s parents

      P =  ‘pension’ received from children

      r = capital market return

      β = discount factor

      θ = parameter representing risk aversion of the agent 

      The problem gives rise to two FOCs, w.r.t. a and n, but there is no explicit solution to these simultaneous equations. In any case, one does not expect this model to predict the entire set of relationships discussed earlier, since it does not quite take into account the insurance value of children. 

Model 3.2.2.2  A basic model of children as risky assets and insurance 

      Instead of having two states of nature in the second period, we now have four states as follows: 

      Agent’s income  Return from children  

      Low (say, 0)   Yes 

      Low    No  (agent starves)

      High (say, y2)   Yes

      High    No 

      Now, we see that there is a state in which, in the absence of capital markets, and even with a negative NPV, children could provide for the agent in old age when the agent has zero income. It is the risk aversion of the agent, operating through this state, that drives the old age security hypothesis. Accordingly the model can easily be revised to incorporate four states of nature in the second period. 

      We observe in the above case that if the agent’s low income status is non-zero (say, as in the developed nations), then the old age security hypothesis will eliminate the need for children completely, if the expected return from children is negative. Therefore, we now postulate the existence of another argument in the utility function, called ω which represents the utility an agent gets from ‘old age care,’ a component that is not directly related to consumption, and which is assumed to be best provided through one’s own children. This will enable the model to generate a demand for children even if the children are risky, and the agent has a relatively high income in the two states of nature at time t = 2.  

      A question would naturally arise: from where does this argument in the utility function arise? We argue, as shown in Section 1.3 on the evolution of the old age security motive, that it arises from the fact that human beings now expect to live well into their old age when they are aware they shall be physically restricted and might even be physically handicapped: it is not just consumption that they are concerned about, in that situation, but basic things functioning and capability51 (Sen, 1992). Even if the person has the ability to purchase medicine (consumption of health), it might not be possible to put that medicine into one’s body, unaided.  

      We have to answer a basic question here: why is it that ω is best provided by one’s own child, and not, say, by outsiders. It is here that we make use of the Hamilton kin rule (Hamilton, 1964) according to which, one ‘cares’ about people in the proportion of the genetic relationship we have with them. We summarize the coefficients of relationship between kin below: 

 

      This seems to be at the source of the ‘weak hardwiring for sense of duty’ (Wilson, 1993) that we discussed earlier, and it is this that parents are able to make use of to instill a ‘sense of guilt’ (Becker, 1993) in their children if the children do not care for them. While siblings are a good alternative source of ω as per this argument, one knows that (a)  siblings have their own families to ‘look after’ and (b) siblings have the same expected life span; hence it is improbable that they will be around when one is infirm. The operation of this rule is clearly observed in day-to-day life, and we do find that ‘outsiders,’ even if they are well trained nurses, are not expected to provide the kind of care that family members are expected to provide. This does not rule out the moral hazard or ‘default’ problem from children, but it does show why one’s own children are strongly preferred to ‘non-kin’ nurses in old age homes, as providers of care.  

      Accordingly, the above model is modified as follows. The utility function now looks like this: . We will have the four states of nature as above. Additionally, in each of these states, there is the further uncertainty about receiving old age care (which is distinct from pension). Therefore, it is possible that children might pay a ‘pension’ P but not provide ω, while there is also a possibility that children might not pay any pension but provide ω.  

      If income and insurance risks were reduced as a result of economic growth and old age home were to become perfect substitutes for children as providers of old age care, then this model would also drive down fertility to n = 1. However, there is sufficient reason to suppose that old age homes will always remain imperfect substitutes for children. 
 
 
 

b) We can add human capital investment through α, thus making P a function of α.  

Numerical simulation 

      As can be seen, with increasing complexity, the model will become more and more intractable. However, it should be possible to use various software (such as GAMS) to carry out nonlinear programming and to arrive at numerical solutions for plausible parameter values and thus to plot the relationships between the variables and fertility. It is hoped that a basic theoretical model will be possible to work out on these lines which will support most of the relationships hypothesized in Section 3.1. 

Advantages of this model 

3.3 Complexities 

      It is easy to see that the above discussion has not even scratched the surface of the complexity that exists in the inter-relationships between variables. For example, we have not been able to capture the dynamically changing decision when expectations both of longevity and of incomes of children get revised upward (as in a society like Thailand). We have not been able to capture the relative risk aversion of the male/ female partners to the fertility decision. The dynamic changes caused in almost all variables are not caputured in this model. There is also the complexity of moral hazard added on when there are too many children. It is possible that some children might free ride on the care being provided by their siblings. And so on.  

      However, this basic theoretical model can be taken as a starting point, and it should be possible to arrive at results, by adding sufficient complexity, closer to what is observed in reality.

 

 
 
 
 
 
 
 
 
 
 

PART II 
 
 

EMPIRICAL ANALYSIS 
 

 

Chapter 4

      This chapter outlines the kind of study/ data that shall be necessary to build up in order to gain a better understanding of the societies in Costa Rica and Thailand. This will require not only understanding the demographics of these nations, but also their society and culture, especially in relation to the family. 

      The material included here has not been edited thoroughly yet. Much of it has been merely re-arranged and classified after being downloaded from various web sites on the internet.  Therefore it should be taken only as an indicator of the kind of work that needs to be done.  

The country’s setting and its people  

      The Kingdom of  Thailand, known as Siam before 1939, and situated in the west of the Indo chinese Peninsula of Southeast Asia, has remained the only nation in the region that was never colonized. Area With an area of 198,115 square miles (513,115 square km), and population, in 1993, of about 57,829,000, it is one of the middle-sized nations in Asia.  

      Thailand is divided into 4 geographical regions and 76 provinces (Changwat), 699 districts (Amphoe), 85 sub-districts (King Amphoe), 6,905 communes (Tambon), 62,994 villages and 132 municipal areas. The four primary geographic regions - Central, Northern, Northeastern, and Southern. The Northern and Northeastern are the largest in terms of area (Knodel, Chamratrithirong and Debavalya, 1987).  

Government  

      Thailand is a constitutional monarchy with a parliamentary government that, since 1932, has largely been dominated by the military. The hereditary king is head of state and commander in chief of the armed forces. Legislative power resides in a bicameral National Assembly, consisting of the 270-member Senate and the 360-member House of Representatives. Senators are appointed to six-year terms on the recommendation of the incumbent prime minister, and members of the House are directly elected to four-year terms. The head of government is the prime minister, who is required by a 1992 constitutional amendment to be an elected member of the House of Representatives. The prime minister names a Council of Ministers (cabinet) for appointment by the king. A multiparty system operates, and there is no dominant party or coalition other than the military. 

The economy 

      The process of modernization in Thailand was speeded up during the period immediately following the 1932 Revolution and particularly after World War II. Rapid socioeconomic change occurred throughout Thailand in the 1960’s with the introduction of a new mode of production: industrialization. The new type of economy helped to increase the standard of living and to modernize the country at a fast rate.  In addition to this, close contact with Westerners and western countries has brought about changes in social life style, attitudes, and values of the Thai population in modern times (Limanonda, 1995). Today, Thailand is a market economy based largely on services, light industries, and agriculture.  Agriculture accounts for about one-eighth of the gross domestic product (GDP) but employs more than half the workforce. Small landholders predominate through most of the country. Manufacturing industries account for one-fourth of the GDP but employ only one-tenth of the work force, mostly concentrated in and around Bangkok - which is the primate city. 

      Social-welfare programs are meagre but include assistance to children, families, and the old and indigent and disaster relief. Education is free and compulsory between the ages of 7 and 15.  

Religion 

      Theravada Buddhism is predominant and is accepted as the state religion in Thailand. Approximately 95 percent of Thai population are Buddhists.  At present there are approximately 30,000 Buddhist temples all over the country, over 200,000 monks and more than 100,000 novices (Wasi as cited in Phongphit, 1988). The application of the Dhamma, the Buddha’s principles, is vast in all kinds of organization in the society.

                 

a) The Family System 

                  

                  

                  

                  

b) Extent of son preference 

c) Age at marriage   

d) Inheritance pattern 

e) The Elderly

                  

Demographics 

Ethnic composition 

      Thailand’s population is relatively homogeneous, with about 80 percent of the total population being ethnic Thais of which more than 95 percent are Buddhists.  The major minority groups are the Chinese - located mostly in Bangkok and urban areas, and the Malays, who live primarily the Southern provinces. “Indigenous minorities include the hill-dwelling Karen; the Semang, who live by hunting with blowpipes and spears; and the Lawa, who are believed to be the original dwellers of the delta plain. The Khmer, Soai, and Indians make up small immigrant groups. By the mid-1980s there were some 380,000 refugees in Thailand from Myanmar, Vietnam, Laos, and Cambodia.”55  

Population Size 

      “During the twentieth century, Thailand has experienced a relatively rapid rate of population growth due to a rapid decline to mortality that preceded and then accompanied a decline in fertility.  From a population estimated to be approximately 8 million at the time of the first census in 1911, Thailand grew to approximately 54.5 million by 1990 (The total enumerated population on April 1, 1990 was 54,548,530 persons).

                  

      “During the 20 years after the declaration of a national population policy in 1970, Thailand’s population growth rate has substantially declined.  The rate has declined from about 3% during the late 1960’s to about 1.4% per annum at present. This demographic change is due mainly to more widespread contraceptive use among Thai ever married women who are in the reproductive ages, and a strong commitment by the Thai government to reduce fertility and control the size of population” (from Limanonda, 1995). 

      The sex ratio which is defined as the number of males for every 100 females in the population was 98.5. This indicates that females were slightly more than males.  

      In comparisons of the population among regions (Table 1), it is apparent that the Northeast was the most populous region with the poplation of 19,038,497 persons, followed by the Central Region (excluding the Bangkok Metropolis), the North and the South respectively. The population of the Bangkok Metropolis was 5,882,411 persons. 

      The overall population density was about 106 persons per square kilometer. The region that had the highest density was the Central Plains with 118 persons per square kilometer, followed by the Northeast, the South and the North woth a density of 11,399 and 62 persons per square kilometer respectively. The density of the Bangkok Metropolis was 3,758 persons per square kilometer. The average size of household for the whole kingdom was about 4.4 persons. The average size

between municipal and non-municipal areas was not much difference of about 4.2 persons and 4.4 persons respectively. 

      Due to population momentum, although the replacement level of fertility could be maintained, the size of the Thai population in the year 2025 may be as big as 80.9 million, an increase of 25.2 million within the period of 35 years after 1990 (Leoprapai, 1991-92).  

Fertility 

      Thailand’s total fertility rate has dropped to near replacement level in a generation; the TFR dropped from 6.3 in 1965 to 2.4 in 1989.56 57  

      “The North region was the first to show fertility decline, starting in the late 1950s. This is due in part to the scarcity of land in the region. The Northeast has lagged behind the North in fertility level, and the South has consistently had the highest level mainly due to the Muslim population. A recent study found that individual-level fertility determinants are similar across all four regions but that community-level determinants vary. Variables measuring women’s status have a large impact in the North and Central regions while those measuring population pressure have an impact in the Northeast and South. Family planning program variables have a large impact in the South (Chamratrithirong et al, 1989).” 58  

      “In general, the fertility decline in Thailand is pervasive, with little difference found by rural/ urban or educational group distributions. A recent study found that women’s status variables did have an impact on fertility at the regional level, but variables measuring children’s school status and contribution did not have an impact” (from Background Country Paper for the OAS survey: Thailand). 

      Ever-married women aged 15 years and over born an average of 3 live births. The average number of children ever born in municipal areas was 2.5 live births and 3.1 live births for non-municipal areas.  
 
 
 

Age at marriage 

      Considering the time exposed to childbearing, it is seen that the singulate mean age at marriage (SMAM) of women was about 24 years. The data also shows that women in municipal areas were married later than those in non-municipal areas. The proportion single among women in the reproductive ages of 15-49 years in municipal areas was markedly higher than that in non-municipal areas. (45.9 percent for municipal areas compared to 29.7 percent for non-municipal areas).59 

National Population Policy and Family Planning 

      Phyormyont (1992-93) has discussed the population policy of Thailand in some detail, and shows how earlier population policies tended to be policies aiming at increasing the size of population through pronatalist and public health measures. Before 1970, population growth rates were about 3%. In March, 1970, the government eventually issued the first national population policy statement “supporting family planning through a voluntary system, in order to resolve various problems concerned with the very high rate of population increase, which will constitute an important obstacle to economic and social development of the nation.”60 The first five-year family planning program was included in the health plan of the Third National Economic and Social Development Plan (1972-1976). From the Fourth National Economic and Social Development Plan (1977-1981) onwards, population policies were made more comprehensive to cover all dimensions of population: population growth, quality of population and population distribution and human settlements. Since the introduction of the national policy, the impact of contraceptive use on reducing the fertility of Thai women has been significant.   

      During 1970-1990, the service statistics, consistent with findings from various periodic surveys, indicated that the number of new acceptors entering the program has grown steadily since its inception.   

Contraception use 

      The contraceptive prevalence rate (CPR) was estimated to be about 8% during 1963-1969. “Evidence from a series of national surveys confirmed the fact that a substantial decline in fertility in Thailand is due chiefly to changing reproductive behavior within marriage, and in particular a dramatic increase in contraceptive use.  There was an attempt to estimate the contribution to the change in the TFR of Thai women during the decade 1968-1978 due to each of the four proximate variables: contraceptive use, the proportion married, induced abortion and postpartum infecundibility.  The results indicated that the increase in contraceptive prevalence made a far greater contribution to the decline in fertility than did any other variables during this entire decade (Knodel, Chamralrithirong and Debavalya, 1987)” (Limanonda, 1991). 

      Approximately 66.7 percent of currently married women aged 15-49 years were practicing contraception. Among women in ono-municipal areas, the proportion currently using was higher than their counterparts in municipal areas (68.1 percent compared to 60 percent). The most popular method was female sterilization being used by about 24.4 percent, followed by oral pill (24.2 percent), infection (9.5 percent), and IUD (4.8 percent) respectively. The proportion of currently married women aged 15-49 years reported practicing contraception was highest in the North (74.5 percent), followed by the Northeast, the Central and the South respectively. The Bangkok Metropolis the contraceptive prevalence rate was about 58.5 percent.61 

      It has been estimated that approximately 70% of Thai ever married women use one kind of contraception or another (Bennett et al., 1990). This is confirmeed by Knodel, Chamratrithirong and Debalaya 1987, by the Contraceptive Use Prevalence Survey 1987, and by the National Statistical Office 1989. 

Mortality and Life Expectancy 

      “Mortality has dropped to the lowest Level ever, at about 5-6: 1,000 population.  This reflects in the increasing life expectancy of both male and female population” (Limanonda, 1995). The average life expectancy is relatively good, at about 66 years. 

Population Distribution  

      Population distribution between municipal areas and non-municipal areas was considerably unbalanced. About 81.3 percent of the population was the inhabitants in non-municipal areas while only about 18.7 percent were those residing in municipal areas. The census results in table 8 shows that, the total population in Thailand of 54,548,530 persons, about 91.9 percent were born in the region of current residence, 7.7 percent were born in other regions, and about 0.4 percent were born in foreign

countries.62 

Table A. Major characteristics of the population by area. 

-----------------------------------------------------------------------------

        Population characteristics      Total   Municipal       Non-municipal

                                                  area              area

----------------------------------------------------------------------------- 

Age and Sex Structure 

Total population                   54,548,530  10,215,098         44,333,432

Sex ratio 1                              98.5        93.4               99.7

Percentage of population aged

under 15 years                           29.2        23.3               30.6

Percentage of population aged

15-59 years                              63.4        70.3               61.8

Percentage of population aged

60 years and over                         7.4         6.4                7.6

Dependency ratio 2                       57.7        42.2               61.8 

Fertility 

Average number of children ever born

per ever-married women aged 15 years

and over                                  3.0         2.5                3.1

Percentage of single women aged 15-49

years                                    33.3        45.9               29.7

Singulate mean age at marriage of

female (SMAM) 3                          23.5        26.5               22.6

Percentage of currently married women

aged 15-49 years practicing

contraception                            66.7        60.0               68.1 

Population Distribution 

Percent distribution of population in

municipal and non-municipal areas       100.0        18.7               81.3

Percentage of persons were born in

this region of current residence         91.9        73.9               96.1 

----------------------------------------------------------------------------- 

1. Sex ratio =  Number of male population  * 100

               Number of female population 

2. Dependency ratio =  Number of population aged under 15 years and 60 years and over  * 100

3.  Singulate mean age at marriage (SMAM) is calculated from the proportion of  single women. 

Literacy.  

      It can be seen from the Table that about 93 percent of the population aged 6 years and over were literate. Male literacy was relatively higher than female literacy, 94.8 percent for males, as opposed to 91.3 percent for females. The literacy rate in municipal areas was higher than that in non-municipal areas, 96.7 percent and 92.2 percent for municipal areas and non-municipal areas respectively. It should be noted that the literacy rate were above 90 percent for all regions, except for the Northern region where the literacy rate was 88.6 percent. 

School Attendance and Educational Attainment.  

      Population aged 6-29 years who were enrolled in school at each educational level were approximately 37.2 percent of those 6-29 years of age. The proportion of those who were attending school was higher in municipal areas, about 45 percent for municipal areas and 35.4 percent for non-municipal areas. 

      With respect to educational attainment, the results reveal that only about 18.5 percent of those aged 6 years and over attained more than primary school. The findings also show that the proportion of those who attained more than primary school in municipal areas was considerably higher than that in non-municipal areas (44.4 percent for municipal areas, as opposed to 12.5 percent for non-municipal areas). Considering sex differentials in school attainment, it is apparent that the proportion of males who completed more than primary school was higher than that of female counterparts (20.9 percent for males compared to 16.3 percent for females). At the regional level, the proportion attained more than primary school level was highest in the Bangkok Metropolis (47 percent), followed by the Central, the South, the North and the Northeast respectively. 

Table B. Social characteristics of the population by sex and area. 

-----------------------------------------------------------------------------

        Social characteristics and sex  Total   Municipal       Non-municipal

                                                  area              area

----------------------------------------------------------------------------- 

Percentage of literate persons aged 6 years and over

        Total                            93.0        96.7               92.2

        Male                             94.8        97.8               94.1

        Female                           91.3        95.7               90.3 

Percentage of school attendance of the population

aged 6-29 years

        Total                            37.2        45.0               35.4

        Male                             38.0        46.9               36.0

        Female                           36.5        43.2               34.8 

Percentage of persons aged 6 years and over who

attained more than primary school

        Total                            18.5        44.4               12.5

        Male                             20.9        48.4               14.7

        Female                           16.3        40.7               10.4 

----------------------------------------------------------------------------- 

Occupation  

      31,296,805 persons 13 years of age and over were employed. The largest number of these (67.3 percent) were agricultural, animal husbandry and forest workers, fishermen and hunters. 11.2 percent were craftsmen, production workers and laborers, and 7.7 percent were sales workers. 

      When consideration is given to employed persons in municipal areas, the pattern of occupation differred from that of the whole kingdom. In municipal areas, about 25.6 percent of the total employed in municipal areas were professional, technical and related workers. However, it is found that the occupation patterns of population residing in non-municipal areas and of  employed males and females were similar to the occupational patterns in the whole kingdom. Occupation by region reveals that the patterns of occupation were similar among all regions. Most were involed in agricultural sector. The Central Region constituted the lowest proportion in agricultural sector (54.9 percent), but the proportion of other regions were above 70 percent. 

      For the Bangkok Metropolis, craftsmen, production workers and laborers ranked the highes (29 percent). Next were sales workers (19.5 percent), professional, technical and related workers (14 percent). 

      Considering economic activity of population aged 13 years and over during the week prior to the cencus date (25-31 March, 1990), it is found that 41,016,701 persons or 75.3 percent of population aged 13 years and over were economically active (inculding those who were employed, looking for work, and waiting for farm season). Table 20 depicted that about 24.7 percent were those who were non-economically active (including housewives, students, unable to work, etc.). Proportion of male who were economically active was higher than that of females (82.9 percent for males compared to 68 percent for females). With respect to regions, the percentage of those who were economically active were over 70 percent for all regions except the Bangkok Metropolis which was only 65.5 percent. 

Table C. The percentage of employed population aged 13 years and over by last year occupation, sex and area. (Occupations ranked in the order of percentages)  

------------------------------------------------------------------------------

           Major occupation group            Total    Municipal  Non-municipal

                                                        area         area

------------------------------------------------------------------------------

                Total

1. Agricultural, Animal Husbandry

   and Forest Workers, Fishermen

   and Hunters                                67.3        a          79.4

2. Craftsmen, Production Workers

   and Laborers                               11.2      25.6          8.4

3. Sales Workers                              7.7      23.1          4.7

4. Professional, Technical and                 4.2      14.0          2.3

   Related Workers                              a       10.1           a

5. Service Workers 

                Male

1. Agricultural, Animal Husbandry

   and Forest Workers, Fishermen

   and Hunters                                65.2        a          77.4

2. Craftsmen, Production Workers

   and Laborers                               12.9      29.2          9.5

3. Sales Workers                               5.9      18.0          3.4

4. Professional, Technical and

   Related Workers                             3.9      11.6          2.3

5. Transport Equipment Operators

   and Related Workers                          a       10.4           a 

                Female

1. Agricultural, Animal Husbandry

   and Forest Workers, Fishermen

   and Hunters                                69.6        a           81.5

2. Sales Workers                               9.7      29.1           6.2

3. Craftmen, Production Workers and

   Laborers                                    9.4      21.4           7.1

4. Professional, Technical and

   Related Workers                             4.6      16.9           2.4

5. Service Workers                              a       13.3            a

------------------------------------------------------------------------------ 

Symbol “a” means the lower percentage or lower ranking order than those

shown in the column. 

Age Structure 

      During the past 30 years, the many changes in Thailand’s society and economy have been closely correlated with its demographic experience. Falls in mortality from the late 1940s, particularly among infants, have brought a very high rate of population growth. A greater number of births have been seen in each decade, until very recently, and this alone will generate an increasing population of old people well into the future. Since the 1960s, there has been a rapid decline in fertility, largely due to contraceptive use. This has begun to increase the elderly population as a share of the national total. The vast dramatic decline in fertility has profound implications not only for the age structure of the Thai population, but also for the support arrangements of the ever-growing elderly population in Thailand. The ratio of adult children to older parents, which is one consequence of particular significance, will certainly decrease over time. The implication is that the fertility decline will result in fewer traditional care providers, hence restricting old age support options for future generations of elderly parents” (from “Patterns of Support Among Children to Old Parents”). 

      The population in the young ages of 0-14 year was about 29.2 percent of the total population in 1990. Those in the working ages of 15-59 years were about 63.4 percent and approximately 7.4 percent were those in the old ages of 60 years and over. Regarding regional differentials, the Northeast had the highest proportion of population in the young ages, while The Bangkok Metropolis had the lowest proportion of the young age, and the highest proportion of the working ages. The highest proportion of population of those in the old ages was found in the Central Region.  

      “Demographic changes, and particularly a rapid drop in fertility, have brought about changes in the population’s age structure.  The size of the population aged under 15 has declined from 43% in 1960 to 33% in 1990 and is expected to be as low as 27% at the end of this century.  At the same time, the size of the population in the labor force will increase from 52% in 1960 to 66% the year 2000. This is due to the high population growth rate in the past 3-4 decades.  The proportion of old people also will grow, although at a slower pace, from about 54b in 1960 to about 8% by the year 2000 (Limanonda, 1991).

                  

      “As a result of the fertility decline which is the most important determinant of age structure, the proportion of elderly people (60 and above) and the adult people (15 to 59 years) has been gradually increasing since 1970, while the younger people (0 to 14 years) has been decreasing over time. Based on these trends, the elderly people in the next 50 years will increase to a very much larger proportion than at present. In addition, another indicator of population aging is the change in the elderly dependency ratio (the number of person aged 60 and over per 100 persons 15-59 years). In Thailand, the elderly dependency ratio will also increase from 11.4 in 1990 to 16.8 in 2015 (calculated from Population Projections for Thailand 1980-2015).” (from “Patterns of Support Among Children to Old Parents”). 

      “The elderly population of Thailand is growing much faster than the population as a whole. Between 1991 and 2020 Thailand’s total population will increase by about one-third, but the population aged 55 and over is projected to more than triple, adding 11 million older men and women, almost half of whom will be 65 and older. Currently, 9.7 per cent of the population of Thailand is aged 55 and over (Table I). Thailand now has higher proportions of people in older age groups relative to other countries in the region; with the exception of Singapore (where more than 12 per cent of the population is aged 55 and over), Thailand is the “oldest” country in Eastern South Asia. Projections to the year 2000 imply that 22 per cent of the population will be aged 55 and over, making the Thai elderly one of the fastest growing older populations in Asia” (from the Parental expectation article). 

      “Another indicator of population aging is the change in support ratios - the number of ‘dependent’ persons (children under age 20 and / or adults 65 years and over) per 100 persons in the ‘productive’ ages (20 to 64 years). In 1991 the total support” (from Parental expectation article, p.90). 

      The age dependency ratio, which is the ratio of persons in the dependent ages (under 15 years and 60 years of age and over) to those in the working ages (15-59 years), was 57.7. In other words, there were about 58 persons in the dependent age for every 100 persons in the working ages. Considering the age dependency ratio among regions, the South had the highest age dependency ratio (66.9), and the Bangkok Metropolis had the lowest age dependency ratio of about 37.9. 

Empirical evidence on old age care 

1. Results from questions on support of old age from fertility surveys 

2. Traditional care of parents 

3. Sources of support in old age 

      Support from children in old age 

4. Changing patterns of support for the elderly 

4.2 Costa Rica 

      A similar study of Costa Rica will be carried out in due course to understand the background and setting of the nation. 

4.3 The Old Age Security Survey and its Preliminary Results 

      We now discuss the salient features of the Old Age Security Survey (OASS, 1991-92) as carried out in Thailand, and some of its preliminary results. A similar study will be carried out for Costa Rica (Section 7.2).  

4.3.1 Thailand: The Survey and the Data 

Geographical Area surveyed 

      Though initially, the idea was to work in the Northern and Northeastern regions,64 finally the survey was carried out in Central and Northeast Thailand (Nugent, 1990: 62). Southern Thailand was excluded to avoid complicating the study with problems of heterogeneity of the preferences arising from differences in race, religion, an cultures. A dual random sampling procedure was followed in selecting the villages and the households within this geographical area.65  

      In each of the two regions, two provinces were selected randomly. Then districts were randomly selected within each province, tambols (sub-districts) within each district and villages within each tambol. 

Region - Province - District - Tambol (sub-district) - Village. 

      Within these sample villages, households were listed to determine their eligibility for the survey, and finally, sample households were selected randomly from this list. The criteria for eligibility are discussed in Nugent (1990). 

Eligibility of households in the villages 

      Households with members falling within two age groups were considered for the survey: 

Listing of eligible households/ respondents 

      “The household  registration system at the district level does not provide up-to-date information. Births, deaths and changes of residence, for example, may not have been reported to the district office for a number of years. In addition, the relationship between various household members in the registration is often not clear. In some households there may be three or more individual families residing together, making it impossible to identify which members belonged to which families. We felt that this data source provided numerous complications affecting its reliability. Using the household registration during actual fieldwork would also cause problems as a selected respondent may not have actually been residing in that particular household for a number of years. 

      “Thus it was decided to use information from the sub-district health center, for several reasons. Information contained on the forms is up-to-date, being compiled within the year previous to the fieldwork. Secondly, it is a de-facto counting schedule. Thirdly, it is a national form available at all sub-district health centers. Finally, the health center personnel were in most cases directly involved in this data collection and thus were able to clarify certain problems we had during the listing process. 

      “From the randomly selected villages all potentially eligible respondents were listed. These respondents included eligible wives in the first form and their husbands in the second; eligible older males and eligible older females in the third and fourth forms. From these forms of all eligible respondents we randomly selected 11 persons to be actual respondents for the survey fieldwork. This above procedure was undertaken at all the 32 Sub-district health centers of eight districts in all selected provinces” (from “The old age security motive and fertility: first results from the Thai survey,” p. 9).  

Data Set and Data Description  

      An outline of the files contained in the data set is provided in Appendix I. 

4.3.2 Preliminary Evidence  

      “The evidence from the first results of the study on ‘Parental Expectation and Experience of Support from Children in Old Age and Its Relationship  with Fertility’ support the findings which were mentioned previously. It clearly indicated that sons and daughters were listed most frequently and ranked as the most important sources of support. Moreover, a majority  of respondents (more than 85 per cent) received financial help from children. Assistance when ill, companionship and help in housework are also practical help received by the old. Daughters provided most of the care. This information shows

that the elderly parents expect and receive support from their direct offspring very much more than from other sources (Archavanitkul et al., 1993). (from “Patterns of Support Among Children to Old Parents”).

Monetary support (from Patterns of Support …) 

      p.21: “Table 5 indicates that the average amount of money provided per year by the children of the old, according to the sex, age, marital status, number of living children, occupation and place of residence of children. The results seem to differ after the place of residence and occupation of children are taken into consideration. In general, the average amount was highest (3,223 baht per year) for children who stayed outside their parental village or for those involved in non-farm work (3,884 baht). The average was lowest (1,094 baht) for children who stayed in another house in village. Sons who co-resided in the parental households or stayed in another house in the village still seemed to contribute slightly more money than daughters. However, the opposite was true for children who stayed further away or who were involved in non-farm work. 

      “Additionally, for children who stayed outside the village, the average amount of money provided decreased with age of children. However, for the children who co-resided in their parents’ households or for those who stayed in the village, monetary support was highest from those aged 30-39 years an this pattern was also found among children who worked in the non-farm sector. 

      “Children who had ever been married seemed to provide more money to their parents than did single children among those who stayed in parents’ households or in another house in the village. However, for children who stayed further away, single children gave much more money in comparison to the ever-married children. The pattern of monetary support did not differ among married children and the single children who worked in farming or non-farming sectors. The single children still gave more money to their elderly parents. 

      “In addition, it can be seen that monetary support to the elderly decreases with the number of living children (grandchildren of the elderly people) among those children who stayed with their parents or those who stayed in the village. The average money provided to the parents was highest for those who had two or three children and who stayed further away. This also occurred among those children who were involved in the non-farming sector. However, the children’s current place of residence did not affect patterns of monetary support for the children who worked in the farming or non-farming sectors. 

Support in kind 

      “Table 6 displays approximate costs of support in kind given annually by children to their elderly parents, according to sex, age, marital status, number of living children, residence and occupation. In general, the provision of support in kind was highest (932 baht per year) by the children who stayed further away or among those children who worked in the non-farming sector. Daughters tend to provide more food and/ or clothes than sons irrespective of their occupation and place of residence. 

      “Provision of support in kind increased with age of children for those who lived with their parents, stayed in the village or were involved in farming. For children who stayed outside the village, the provision of food and/ or clothes seemed to peak at around the ages of 30-39 years. This pattern was also found among those who worked in the non-farming sector.  

      “Single children who stayed in the village or stayed further away were likely to provide food and/ or clothes to their parents, in comparison to ever-married children. However, the reverse holds true for children who lived in parental households. There were practically no differences in terms of patterns of support in kind between married and single children according to occupations. Married children provided more food and/ or clothes than single children, both among the children who worked in farming or in the non-farm sectors. 

      p.23: “.. the value of support in kind to the aged ... increased regularly with the number of living grandchildren, across all forms of occupation.”  

Provision of physical care 

      “Variations in provision of physical care, which is defined as a form of personal support to parents in this study, were also related to the differences in backgrounds of children (Table 7). In general, as  might be expected, the percentage of children who provided physical care to parents was highest among those who lived with their parents and decreased with increasing distance between parents and children. 92% of children who worked on farming did provided physical care, the percentage was lower for children who worked in the non-farming sector. .... (from “Parental Expectation and Experience of Support from Children in Old Age and Its relationship with Fertility”) 

Summary of findings from the younger sample 

      p.14: “Our findings reconfirm the strong cultural norm of dependence on children in old age. Sons and daughters were the most frequently named sources of support and were consistently ranked the highest in importance. We did find regional differences in expectations of children, as those in the Northeast were more likely to expect help with work, housework and financial help than those in the Central region. Most intended to live with their children in old age, with a continued preference in the Northeast for living with a daughter 

      “We found some evidence of a relationship between expected fertility size and expected support in old age. Though there was little difference in expected family size between those who listed children as first ranked source of support and those who did not, those who expected financial help from children tended to have larger expected family size than those only expecting companionship and/ or care when ill. 

      “Though expectations of support from children in old age were high, we also found evidence that children were seen as a financial burden. Given the hypothetical situation of having one more child, most expected their financial status to be worsened, and many said their financial situation would improve if they had one less child. This finding supports the notion that the quality of children was more important to our respondents than the quantity; in other words, that only a small number of children (two or three) is sufficient to assure old age security. Together these findings indicate a relationship between the perceived value of children and family size, whether this value is economic or non-economic; though the major finding is the pervasiveness of the expectation of support. 

Summary of findings on older people 

      p.23: “Our findings confirm the strong reliance on children for care and financial support. When asked about different types of support, the oldest respondents ranked sons and daughters highest in both regions. With regard to direct care, both on a regular basis and during times of illness, daughters clearly were more likely to provide care and emotional support than sons, in both regions, and this was especially true for the older women. It appears that this is partly because men are able to rely on their spouses for care. A small but significant proportion of respondents however (about 15%) did not rank their children as the first source of support. Most of those in this group stated that this was because children did not earn enough for themselves. We would conclude that aging parents did not see their adult children as irresponsible, but that they understood the economic constraints placed upon their children. 

      “We found clear regional differences in the perceived benefit of having an additional son or daughter. Northeastern residents were more likely to feel an additional child would benefit them, and particularly for the male respondents, were more likely to see the benefit of an additional son. In contrast Central residents were more likely to say an additional child would have no effect on their situation, though women were more likely to see the benefit of an additional daughter. These regional differences are particularly vivid in light of the high percentage of respondents who said their financial situation was inadequate, particularly in the Northeast. Financial hardship was also found to be related to number of living children, though the relationship was opposite in the two regions: Northeastern residents with more children were slightly less likely to say their finances were inadequate, while Central residents were more likely to have financial hardship if they had many children. Though most respondents in both regions stated that they received an adequate level of care and assistance, a not insignificant proportion also stated that they did not, particularly in the Northeast; and the figures on inadequacy of care increased with family size in both regions. 

      “A majority also reported adequacy in companionship in both regions, but again the proportion with inadequate companionship was higher in the Northeast; and it is interesting that this was related to family size ...” 

 

Chapter 5

Testing of the Models 

      This chapter will be formulated in due course, and will constitute the bulk of the work leading to the dissertation.

 

Chapter 6

Conclusion 

      At this stage, one has begun to realize that “As yet ... there is little understanding of how social, economic, political and cultural structures and institutions interact to produce the motivation for reducing fertility” (Sadik, 1991:11).  

      The dissertation arising from this proposal may not dramatically enhance our understanding of this area. However, it is hoped, that this work will improve our understanding of economic basis of fertility, and in particular help us in understanding the role of the old age security motive, and possibly throw some light on the comparative efficiency of traditional institutions versus formal institutions. 

 

REFERENCES

(to be fully verified in due course) 

Caldwell, J.C. (1982). Theory of Fertility Decline. London: Academic Press.

Weiner, M. (1991). The Child and the State in India. Princeton: Princeton University Press. 

World Bank (1995a). Support for the Aging. Development Brief 57. Aug. 1995

World Bank (1995b). Engendering economics. Development Brief 58. Aug. 95

 

Appendix I

Data Description 

      The data set is currently available in the form of the following files. I have not yet prepared the data for the analysis, but that should not be too difficult, since data stored in .por files is in SPSS portable format, and can be viewed in SPSS for Windows. 

ACOA     LOG         2,400  01-23-97 12:22p acoa.log

ACOA     SAS       102,482  01-23-97 12:22p acoa.sas 

CHILD1   POR       219,678  01-23-97 12:23p child1.por

CHILD1   SYS       249,510  01-23-97 12:23p child1.sys 
 

CHILD2   POR       249,116  01-23-97 12:24p child2.por

CHILD2   SYS        28,672  01-23-97 12:24p child2.sys 

EDS                      2  01-23-97 12:24p eds

GREENSYS SYS       981,874  01-23-97 12:31p greensys.sys 

HHOLDER  POR       157,030  01-23-97 12:31p hholder.por

HHOLDER  SYS       123,342  01-23-97 12:32p hholder.sys 

HHTABLE  POR       332,592  01-23-97 12:32p hhtable.por

HHTABLE  SYS       433,959  01-23-97 12:33p hhtable.sys 

HTH014   LOG        26,380  01-23-97 12:33p hth014.log

HTH014   LST        52,160  01-23-97 12:33p hth014.lst

HTH014   SAS        18,629  01-23-97 12:34p hth014.sas 

HUSBAND  POR       455,510  01-23-97 12:34p husband.por

HUSBAND  SYS       277,673  01-23-97 12:35p husband.sys 

JERRY1   PRO           103  01-23-97 12:35p jerry1.pro 

NCHILD   POR       219,678  01-23-97 12:35p nchild.por

NCHILD2 POR       249,116  01-23-97 12:36p nchild2.por 

OLDER1   POR       168,018  01-23-97 12:36p older1.por

OLDER1   SYS       115,364  01-23-97 12:36p older1.sys 

OLDER2   POR       171,052  01-23-97 12:37p older2.por

OLDER2   SYS       110,114  01-23-97 12:37p older2.sys 

OLDER3   POR       237,226  01-23-97 12:37p older3.por

OLDER3   SYS       142,799  01-23-97 12:38p older3.sys

OLDER4   POR       146,370  01-23-97 12:38p older4.por 

OLDER4   SYS        98,269  01-23-97 12:38p older4.sys 

SPSS     JNL             0  01-23-97 12:38p spss.jnl 

TEST     LOG         2,503  01-23-97 12:38p test.log

TEST     SAS         1,432  01-23-97 12:38p test.sas 

THAI     LOG         9,968  01-23-97 12:38p thai.log

THAI     LST       134,939  01-23-97 12:39p thai.lst

THAI    SAS         1,414  01-23-97 12:39p thai.sas 

TTT                 53,727  01-23-97 12:39p ttt

TTT      LOG        76,428  01-23-97 12:39p ttt.log

TTT      LST       192,145  01-23-97 12:39p ttt.lst 

WIFE1    POR       499,380  01-23-97 12:40p wife1.por

WIFE1    SSD     1,908,737  01-23-97 12:46p wife1.ssd

WIFE1    SYS       320,004  01-23-97 12:46p wife1.sys 

WIFE109  LOG        65,618  01-23-97 12:47p wife109.log 

WIFE1091 LOG        69,254  01-23-97 12:47p wife1091.log

WIFE1091 LST       272,079  01-23-97 12:48p wife1091.lst

WIFE1091 SAS        55,442  01-23-97 12:48p wife1091.sas 

WIFE1092 LOG        67,727  01-23-97 12:48p wife1092.log

WIFE1092 LST        63,833  01-23-97 12:48p wife1092.lst

WIFE1092 SAS        47,191  01-23-97 12:48p wife1092.sas 

WIFE1093 LOG        81,077  01-23-97 12:48p wife1093.log

WIFE1093 LST       220,211  01-23-97 12:49p wife1093.lst

WIFE1093 SAS        49,528  01-23-97 12:49p wife1093.sas 

WIFE2    POR       612,950  01-23-97 12:50p wife2.por

WIFE2    SYS       367,765  01-23-97 12:51p wife2.sys 

WIFE3    POR       578,100  01-23-97 12:52p wife3.por

WIFE3    SYS       376,165  01-23-97 12:53p wife3.sys 

WIFEHH   DAT        49,430  01-23-97 12:53p wifehh.dat

WIFEHH   POR       429,598  01-23-97 12:53p wifehh.por

WIFEHH   SYS       275,398  01-23-97 12:54p wifehh.sys 

WTH014   LOG       113,631  01-23-97 12:54p wth014.log

WTH014   LST       192,145  01-23-97 12:55p wth014.lst

WTH014   SAS        53,752  01-23-97 12:55p wth014.sas 

XXX                 19,438  01-23-97 12:55p xxx