About the research
A large body of literature (see Caldwell 1976) suggested that the decline of fertility observed in the last centuries may be an effect of the progressive deterioration of family ties, which reduces the expected old-age economic support from children.
Moreover, some authors (see Caucutt et al. 2007) explain the introduction and expansion of pension systems as the effect of the transition from a rural world, where land is the main instrument of old-age security, to an urban one, where pension systems substitute land as source of economic support for the elderly.
However, the decline of fertility reduces the profitability of pay-as-you-go pension systems, therefore the political support toward such systems.
The thesis suggested a voting model that incorporates these effects in order to investigate the relationship between family ties and political support toward pension systems. The model and some numerical simulations showed three possible development patterns (monotonically increasing, monotonically decreasing and inverted-U), all of them concave.
The empirical study tried to assess which is the dominant relationship using data from several countries at different development stage.
Methodology
The theoretical model and some numerical simulations suggested three hypotheses to be tested:
- the weakening of family ties implies a continuous increase of the pension expenditure
- the weakening of family ties implies a continuous decrease of the pension expenditure
- the weakening of family ties implies an initial increase of the pension expenditure, followed by a decrease of the expenditure
Assuming that in the city, family ties are weaker and fertility costs are higher than in the country, the author estimated the linear regression model, which explained the ratio of pension contribution to GDP with the urban rate. Additional control variables in the several specifications estimated were the per-capita GDP (based on PPP), the proportion of old-age population, the total fertility rate, the Gini index (as a measure of income heterogeneity) and the fiscal incidence (as a measure of the size of the public intervention).
In order to check which pattern was prevalent, the regression included a quadratic term for urban rate: as the relationship was expected to be concave, this last coefficient was expected to be negative and a test performed on the coefficients of the linear and quadratic terms would determine which was the prevalent relationship in the data.
Both cross-section and panel estimations were performed. As the fertility rate might be considered endogenous in the model, instrumental variable method was employed as well, using child mortality rate and the female education level as instruments.
Results/Outcome
All the performed regressions supported the inverted-U pattern, suggesting that the weakening of family ties after initially inducing the introduction of pension systems (as a substitute for fertility in old-age economic support), eventually erodes their profitability, reducing the political support towards them.
The estimates suggested that the pattern becomes decreasing for urban rates higher than 50-70%.
If fertility rates continue their rapid decline, as in the past years, pay-as-you-go pension systems may not be sufficiently profitable for them to be supported by the majority of the voters, hence being destined to disappear or be replaced by the funded ones. Some remedies have already been suggested by Cigno and Werding (2007): raising child benefits (already present in many countries) or linking pension benefits to individual fertility may restore private incentives for raising children, making pension systems fertility-neutral.
Impact
References
Caldwell, J. C. (1976), ‘Toward A Restatement of Demographic Transition Theory’, Population and Development Review 2(3/4), 321-366.
Caucutt, E. M.; Cooley, T. F. & Guner, N. (2007), ‘The Farm, the City, and the Emergence of Social Security’, NBER working paper 12854.
Cigno, A. & Werding, M. (2007), Children and Pensions, MIT Press, Cambridge (MA).
Publications
The full paper can be found at: http://mpra.ub.uni-muenchen.de/20961/