The household demand model follows the tradition of neoclassical economics with rational maximising agents. It was devised in the 1960s and 1970s with contributions from Becker, Willis and Easterlin. The model posits a couple as being rational agents who wish to maximise their utility function which incorporates pleasure from children (their role as a consumption good), subject to their income constraint which includes time and thus allows us to examine the opportunity cost of having children on time.
Factors making it difficult to model demography:
- Parents are both the demanders and suppliers of children, which makes it difficult to observe prices of child services. Furthermore a lot of childbearing and rearing services are non-market based, and so it is difficult to impute prices/costs on this.
- Culture plays a large role in fertility, not only in looking at contraception but in how long a parent’s responsibility over the child lasts, and the degree to which they look after the child (how much they invest in quality).
- There is a large degree of uncertainty in fertility: parents can only attempt to have children (and hope that they fall pregnant), and even then there is the probability of a miscarriage, or infant mortality.
The utility function is given as U = U(N,Q,S) where N is the number of children, Q is the quality of each child and S is an aggregate basket of other commodities which the couple derive utility from (Willis). Our production function is given as C = NQ where C is child services and is the product of the number and quality of a child. This means that we can increase child services (the utility derived from having children) by either increasing the number of children, the quality of children, or both. Our budget constraint is given as I = NQπc + NPn + QPq + Sπs where pi represents the shadow cost (the non-market price reflecting opportunity cost) and I is the family’s full income. As a result, by maximising utility subject to our constraint, we get derived demand for children as a function of I, πc, πs, Pn and Pq.
For simplicity we assume that the quality of each child is constant, so no favouritism can occur, although psychologists have cast doubt on this assumption as it is normally the case that the first born is given greater attention and as a result has higher quality. Moreover, the issue of son-preference in many developing countries means that males have greater quality than females, one explanation is because they are expected to continue the family name, and will receive the inheritance. The model has taken a one-period static approach, assumes that both the male and female in the relationship are making the fertility decision, and so have equal weighting within the utility function and income constraint. In reality the assumption that a household is composed of a male and female may not accurately described developed country societies, where there has been a rise in the number of single parents – often the mother – meaning that one party bears much more cost than the other, whilst the other party may still be able to derive utility from seeing the child. Additionally, it ignores the situation of adoption and minorities such as LGBT. The model may describe more accurately developing countries’ relationship, where they are more traditional, but this is still not the case in all societies, particularly where polygamy exists. For example, Guyer observed that in the face of deteriorating economic opportunities for some African women, they decided to have children fathered by many different men in order to increase their “lateral links” with them, and perhaps offset failing economic circumstances, something which isn’t incorporated into our model.
Moreover, it is unlikely that both men and women in the relationship will have equal power, and equal weight, it is much more likely that the man is more dominant and so it may be more appropriate to use some form of Nash bargaining theory to evaluate the utility of each side of the couple, which would further complicate matters. If we were to do this then we might assume that fertility would be higher than in the equal weighting situation, because men have more dominance and face less costs from having more children – both in terms of bearing and rearing – then they are likely to want to demand more children than the woman would desire.
The model also makes the drastic assumption of perfect information, which is unlikely in reality, as couples may have had no experience in dealing with children before, and so don’t know the full costs of constantly being responsible for a child, nor do they know exactly the level of utility they will gain from having a child, as this has never previously been observed and their expected future income is uncertain. This critique may perhaps be more apt for describing developed countries, as it is likely that females have some experience of looking after children, such as younger relatives or other relations, in developing countries (Cain). Furthermore, the probability of conceiving, and then having a successful pregnancy, along with the child surviving are all unlikely to be unknown and add a degree of uncertainty to the analysis. We are also fundamentally assuming that it is biologically possible for a woman to maximise fertility, this might not happen for other reasons such as religion or culture.
Child services are a normal good and as such as incomes rise parents demand more child services, and thus either a greater number of children or greater quality. So a pure wealth effect – which doesn’t affect wages or other market prices so doesn’t affect the optimising behaviour of individuals – would result in higher demand for child services (but not necessarily in a greater number of children, it may instead be for a greater quality of children). Rising wages increase the value of time and thus increase the opportunity cost of having more children. This means that the shadow price of a child would increase. The price of children rises more relative to the price of less time-intensive satisfactions because having more children increases a women’s time costs disproportionately.
Quantity Quality Trade-off – the demand for quality is much more elastic to income than the elasticity of demand for quantity. This means that an increasing income is likely to increase the demand for quantity more than the demand for quality. Hence, as incomes rise, parents may substitute quantity for quality, so the substitution effect has a negative relationship with fertility, whilst the income effect is positively related to fertility. This is empirically observed in developed countries whereby greater incomes have led to lower levels of fertility as parents opt to invest more in their children, rather than produce more.
Rosenzweig and Wolpin find evidence that twins are associated with a decline in the education levels of other children in the family, this means that a higher (unexpected) quantity leads to reduced quality and demonstrates the trade-off.
Duesenberry and Blake show that the substitution effect is greater than the income effect, reasonably assuming that parents wouldn’t choose child quality independently of their own living standards, and hence rising incomes leads to lower fertility.
If we remember that our budget constraint is given as I = NQπc + NPn + QPq + Sπs, then we can analyse the empirical result that fertility rates have been rising in Scandinavian countries which provide subsidised education. We might expect that subsidised education would just mean a greater quality for children, but it actually means that the price of quantity for a given quality and number of children has fallen (due to public provision), hence Pn has fallen, so demand for quantity rises even though quality is high. We can extend this to realise that poor countries subsidising education, and other state-provided services, will lower Pn which could lead to population growth.
Determinants – Non proximate determinants affect proximate determinants which in turn affects fertility levels. Non proximate determinants include women/man education, occupation, income, infant mortality, politics, family planning services, household structure, extended family, son-preference, female autonomy, religion, urban-rural residence and demand for children as a consumer/producer/investment good. These affect the proximate determinants of natural fertility, nuptiality, abortion and contraceptive choice (Iyer).
Demand for Children – so far we have only examined the demand for children as consumption goods which assumes that parents are altruistic and wish to have children just for enjoyment and to see them grow up. There are also other benefits arising from the role of children as production goods, investment goods, and a form of insurance.
Firstly the children can be productive both in the labour market – generating an income which the parents can use to supplement their own income with – and indirectly by doing household chores, collecting fuel and water, or looking after other members of the family, thus freeing the time of the woman, who can instead engage in income labour activities. Cain finds that in a typical Bangladeshi village (Char Gopalpur), that a male becomes a net producer by the age of 12 (and is no longer a burden on the family), compensates for their own cumulative consumption by the age of 15 and additionally the cumulative consumption of a sister by the age of 22. Children of both sexes begin their economically useful lives around age 6, performing such tasks as gathering fuel, fetching water, carrying messages, and caring for younger children. Male children begin agricultural work at around the age of 11, but the activities they conduct depend on their physical strength and stamina. On average both male and female adults work around 9 hours a day with the bulk of a man’s time spent on productive work and the bulk of a woman’s time spent on housework. Children of both sexes work long hours at early ages. Children aged 4-6 years old work approximately one-fifth as long as adults. Their work time increases to one-half an adult workday by ages 7-9, to three-quarters by ages 10-12, and at age 13 and above, children work, on average, as long as or longer than adults. There is also quite a substantial difference in child work between classes. Those with more assets have children doing more productive work at a younger age, perhaps because they have the assets to permit this (Cain). This shows that children may be quite an important source of income, and may therefore be an important decision in why couples have children. This factor goes some way in explaining the son preference: due to their higher earning potential. One could extend this hypothesis of children being demanded for their ability to earn an income to developed countries, where children themselves aren’t allowed to work until adulthood, by considering the effect of the welfare system where some are compensated (e.g. single mothers) for having children; this therefore acts as an incentive to increase fertility. Drago et al examine the Baby Bonus paid in Australia (AUS$3,000) which has led to a “modest” increase in fertility with this mainly affecting, second, or higher-order, births. This reflects that variable costs decline in the number of children and so the bonus can be used to mitigate the fixed costs. This matches findings by Milligan in Canada. Whilst not the first attempt at increasing fertility in developed countries, this scheme was the first to be offered without other conditions (i.e. dependent on household income or number of other children). Because this is a fixed value, we would expect it to influence lower-income couples more than higher-income couples.
Secondly, children can act as an investment good because parents expend resources into their upbringing at a young age (they invest), and this investment then means that the child grows and can get a job earning an income – a fraction of which may be remitted to the family – and it might be expected that when the parents get old the child then looks after the parent (pays the investment back). Consequently, this would imply that fertility is higher when interest rates are higher, because children are seen as a more attractive investment opportunity. However, in reality it might be questionable whether this plays much of a role in fertility decisions, considering the greater access to credit markets today, and the development of this through technology (e.g. Kenyan farmers accessing banking on their phone) and the use of microfinance to offer finance to small operators. These factors act to lower interest rates and reduce investment opportunities arising from having children. Furthermore, this might imply that the low interest rates following the Great Recession of 2007 should have resulted in a fertility decline which may have been exacerbated by falling incomes. The development of the welfare system in many countries means that a couple no longer has to rely on their children to look after them in their old age and so this cultural trait is becoming unnecessary. Moreover, the welfare system means that a given couple do not have to have children themselves, as the children of other couples will be contributing to the welfare system which looks after this couple in old age.
As incomes rise there is a shift from household to market production, which makes children less valuable because:
- They can do less work at home and may be forced to spend time in school, Kanbargi and Kulkarni find that time spent by children in school has a significant negative effect on family size
- Infrastructure reduces the value of child labour in collecting fuel and water
- Structural change in the economy – requiring more skilled labour – reduces the market jobs for children
- Greater access to capital markets reduces the savings and investment benefit of children
- Insurance, pension and government provided welfare reduces the old-age motive of having children
- Greater geographical mobility means that there is less guarantee that a child will be around to look after you in old age
Survey data confirms that parents don’t see children as a productive asset, although may see them as security for old age. This means poverty reduction is unlikely to affect fertility, if parents don’t chose fertility based on the need for productive assets. It is unlikely to affect old age security, as this factor is based on psychological needs, rather than purely economic (Dreze and Murthi).
Iyer and Velu argue that the decision to have a child or not is an investment decision because the decision to have a child “displays the characteristics of other real options investments such as irreversibility and the necessity for flexibility”. They point out that the return on investment in children can be highly uncertain due to child mortality. Hence a couple’s decision needs to incorporate this uncertainty, and the authors extend the Beckerian model by considering uncertainty. An increase in uncertainty – changes in circumstances– will have an impact on fertility in that women will delay pregnancy, and they may even adopt measures such as permanent sterilisation (quite common in developing countries). Acting on the opposite direction is a desire to have more children to act as a form of insurance. “An options approach would predict that a reduction in uncertainty will have three effects on demographic decisions: it will decrease the desired number of children that women want, it will reduce the birth intervals between each child as women will be reluctant to exercise the option to wait, and it will reduce the age at which women decide to terminate childbearing and use a permanent contraceptive method.”
Education and labour market employment
Evidence shows than an increase in female education reduces fertility although primary schooling may be an exception in some countries, such that only secondary and higher education reduce fertility. Schultz finds that women at age 40-49 with seven or more years of schooling have 1.6-2.9 fewer births than women with no schooling. This reduction in fertility doesn’t just occur due to the increase in the value of time (better educated women are likely to earn more in the labour market, even if they don’t participate, and thus have a higher opportunity cost to their time) but because education increase knowledge of contraceptive use, informs women better about non-household options, increases the efficiency of producing high-quality children, reduces infant and child mortality (and thus higher fertility associated with this) through better education in health and nutrition, and can increase performance of status production duties: women can better manage household incomes, thus increasing their bargaining power within the household (71 percent of women who completed high school had discussed family planning with their husbands compared with 42 percent of illiterate women, Dreze and Murthi).
Furthermore, if a woman stays in education then this is likely to increase the age of their first marriage and hence reduce fertility in the sense that a woman has less available time to produce more children. Although evidence from Africa shows that girls are frequently withdrawn from school at the time they reach puberty. Additionally, matching theory tells us that higher educated men are more likely to seek higher educated women, and this may lead to a reduction in fertility.
A husband’s education may also be important in the quantity-quality trade-off and also because a higher educated man is likely to choose a higher educated wife, or one who has a skilled job (Iyer runs a regression on these variables to test their importance, if husband’s education was omitted then we would have an endogeneity issue, she finds that if a Muslim man had an additional year of primary education then this increased his wife’s fertility by 0.33 children. However a husband’s secondary education decreased fertility significantly for Muslims).
The most important implication of the household demand framework is that different sources of family income have different effects on the demand for children. The woman’s wage has the most negative (or smallest positive) effect on fertility, the man’s wage a less negative or possibly positive effect, and inherited wealth or natural resource income has the largest positive effect, because it does not embody any offsetting price effect to deter the demand for children.
Kanbargi and Kulkarni find that all children go to school in cases where the father has a minimum of ten years of schooling, reflecting his desire (and that is important in patriarchal societies) to have educated children and reflecting the quantity-quality trade-off. The presence of younger siblings seems to depress school attendance, but the presence of grandparents increases it, suggesting that without grandparents the older child is expected to look after the younger siblings instead of going to school.
If we look at the relationship between female economic activity and fertility we find that a universal negative relationship does not exist, but a clear association can be seen between fertility, female economic activity and female education. We would need female employment to be incompatible with child rearing for fertility to fall with greater female employment.