Is the fertility decline a consequence of the growth of the welfare state? Evidence from historical data

2. What determines fertility?

Our thesis is not a new one. Rosati (1996), Ehrlich. and Lui (1998), and Boldrin, De Nardi and Jones (2005) show that fertility should be negatively correlated with e.g. social security. The intuition for these results is quite clear. In less developed old societies without the welfare state and relatively undeveloped capital markets people had few means for supporting themselves in the old age. They could try to work until the very last day, or they could invest in children who had an obligation to support their parents in the old age. Because all generations mainly lived close to each other and depended on one another, it was not so easy to default from the obligation one had for their parents. Of course, bequests generated another motive for supporting one’s parents.
We make use of reasonably extensive data sets which cover both the European and the whole World data. Above all, we focus on historical data which basically start at 1750 and end at 2000. These data are compiled for five-year intervals (before 1850, ten-year intervals) and they come from the following seven European countries: Denmark, Finland, France, Italy, Norway, Sweden and the UK. In addition, we use data from the following four Non-European countries: Australia, Canada, Japan and the US. Most comparative analyses can be made for a somewhat shorter period —1860-2000 only.
For control purposes, we use more extensive World Development Indicator data from the World Bank which come from 185 countries for the (maximum) period being 1960-2003. Finally, we use MZES (Mannheimer Zentrum fur Europäische Sozialforschung) data for 1949-1993. These data only cover pension expenditures and total social security expenditures (in relation to GDP) which — in a similar fashion as with the other data sets — are related to total fertility rates. Ideally, one might use the pension/social security expenditures for the long time period, but there is no way of compiling cross-country data for pre-WWII period. Therefore, we simply use various variables describing the size of government as an indicator of the welfare state and control the results for the post-WWII period with proper indicators.
The main purpose of our comparisons is to see how fertility is related to the growth of the welfare state (for a general exposition of this issue, see papers in European Papers on the New Welfare, 2006). Our main question is: Is it indeed true that the growth of the welfare state crowds out private investment in children?
The 19th century opened up new possibilities for the old age survival. The development of capital markets made it possible to save for the old age, and thus investment in (the quantity of) children was not necessarily needed to such a large extent. On the other hand, growth of educational opportunities made it possible to invest in the quality of children instead of the number of them2. Furthermore, the growth of manufacturing and services and the decline of agriculture diminished the need for child labour. Because child labour was not so much needed — or the use was completely abolished — investment in children (quantity of children) was no longer an appealing opportunity3.
But the most obvious reason for the decline of fertility comes from the welfare state. The government in a sense intervened in the market and provided a pension system (and also a system for health care, old people’s homes and so on). It is quite obvious that government intervention crowded out private investment, and this showed up in a smaller incentive to invest in children. Put simply, the growth of the welfare state is the reason for the decline in fertility.
But how do we know whether this is an important reason? The decline in fertility could equally well be explained by the growth of income, change of the structure of economy (say, agriculture versus manufacturing), fall in mortality or growth of education. It is not easy to answer this question, because we are really dealing with the behaviour of different generations and not only the current generations, but also the expectations of current generations on the behaviour of the future generations. Thus, annual observations on the size of the government and some other relevant variables cannot help to solve this problem.
Still we can do something to settle this issue. We can compare over countries and over reasonably long time-horizons the relationship between fertility and the background variables, and see whether any of the relationship is dominant. We have also constructed a model where we estimate the particular effect conditional on other variables.
It should be emphasized that it is not meaningless to ask what the reason for fertility decline is. If the main reason were just the growth of welfare state, we would face a difficult dilemma. Almost all attempts to boost fertility would then ultimately require more government resources, i.e. in a way an expansion of the welfare state. We could face a delicate situation, where government actions could only worsen the overall demographic situation.

2 Galor and Weil (2000) emphasize the investment in the quality (education) of children in a transition to modern growth.
3 Doepke (2004) argues that child labor regulation has been an important factor in the decline of fertility.

Pages: 1 2 3 4 5

Tags: ,