قوانین نفع امنیت اجتماعی، رشد و نابرابری

We examine the balanced growth effects of pension plans on the rate of growth and on income dispersion in a closed economy where individual decisions about education are the engine of growth. We distinguish between pay-as-you-go and fully funded pension systems and differentiate between three different benefit rules: a Beveridgean regime, a Bismarckian regime depending on one’s entire earnings history and on one’s partial earnings history. Our analysis shows that social security generally reduces the long-run growth rate and our inequality measure. Growth can only be stimulated under a fully funded scheme based on partial earnings history.

There is a large body of research on the expenditure and provision of education. For one, countries spend a significant proportion of public expenditure on this item (more than 5.6% of GDP on average in OECD countries). Furthermore, education is also considered to be not only a major contributor to growth (see e.g. Barro and Sala-i-Martin (1995), Benhabib and Spiegel (1994) and Mankiw et al. (1992) who all spell this more or less out), but also to changes in inequality. In fact, various studies have looked at the role of public expenditure on either one or in fact both of these issues (we mention only a few papers): Kaganovich and Zilcha (1999) show that there is a double benefit of public funding of education which consists in reducing income inequality and yielding a positive growth effect. Bearse et al. (1998) investigate how the degree of centralization of education funding influences income distribution. Fernandez and Rogerson (1999) examine the effect on growth and income distribution of different education financing systems and show that there may be no simple trade-off between equity and resources to financing. Glomm and Ravikumar (1992) look at the effects of education on inequality and concentrate on comparing public vis-à-vis private education in terms of levels of wealth and income inequality. This paper also investigates the effects on growth and equality where education is involved. We adopt an approach that is a little less straightforward, however at least as appealing and relevant: we investigate how growth and equality are affected by the choice of a public pension system. Over the last years, research on public pension schemes has become a major player in the field of public economics as well as of late an important issue in actual public policy. The reasons for this are fairly straightforward: public pension plans not only represent a large proportion of total GDP (OECD countries already spend more than 10% of GDP on public pensions), moreover, given the demographic change that is taking place, this proportion is forecast to increase to over 16% in the next 30 years if no action is taken. Hence, it is of no surprise that numerous implications of pension policies have already been examined in the literature (e.g. induced distortions on the labor markets, effects on saving rates and capital accumulation, and consequences for the provision of public goods, see World Bank, 1994). An important aspect of pensions is how they affect growth in the economy. Hence, from a normative point of view the implications of pension plans has found some attention. Saint-Paul (1992) investigates e.g. the case for public debt in an Arrow–Romer endogenous growth model (i.e. an ‘AK’ model) and shows that the introduction of an unfunded social security system reduces capital accumulation and hence the growth rate, such that it cannot be Pareto-improving. In such an endogenous growth setting, Marchand et al. (1996) show that the case for ascending transfers (i.e. transfers from the young to the old) is rather weak on the balanced growth path. It seems surprising that there have been few attempts to combine both aspects, education and public pension schemes, in a single model and investigate the effect of one on the other. This is not to say that education and public pensions have not been combined. Turning to theoretical models first, in an endogenous growth framework Docquier and Michel (1999) examine a similar problem to Marchand et al. (1996), yet do so in a Lucas–Uzawa setting where education is the engine of growth. They reach quite a similar result to Marchand et al. (1996) regarding the weak case for ascending net transfers. Glomm and Kaganovich (1999) investigate the growth and inequality effects of increased spending on public provision of education that comes at a cost of decreased pension outlays and show that higher spending on public education leads to higher income inequality. Turning to empirical work, Zhang and Zhang (2000) perform an analysis of how social security affects growth determinants. They do so by allowing for a simultaneous interaction between growth and social security and find in a cross-country selection that the empirical evidence indicates that social security has a positive impact on growth in such that it stimulates investment in human capital. However, to the best of our knowledge, there is no study examining in detail how different pension schemes will affect the investment in education and hence how growth and inequality changes resulting from differences in pension schemes arise. Although Zhang (1995) look at effects of social security on the steady state growth path of per capita income where investment in education is motivated by parental altruism and fertility plays a role, their analysis of pension schemes is limited due to the two-period overlapping generation setup with homogenous agents (see below for more details). Hence we investigate how growth and equality are affected by the choice of a public pension scheme. We contrast economies with different pension schemes and accordingly different incentives to invest in education. These pension schemes differ for one in the way that they are financed: they may be pay-as-you-go (PAYG) pension schemes where contributions are immediately paid out to retired agents, or they may be fully funded (FF) pension schemes where contributions are invested on capital markets until the contributing generation is retired. 1 Furthermore, we deal in our model with heterogeneous agents and can thus add a dimension by looking at the way the pension is awarded: pensions may be Beveridgean, i.e. a lump-sum transfer that is identical to all agents, or pensions may be Bismarckian, i.e. earnings related. This too will of course be an important determinant in the decision of whether to invest in education or not. Finally, as we are working in a three period overlapping generations model, we are able to add an interesting feature that has so far not been considered in the literature, yet is of real-world importance: Bismarckian pension benefits can either be based on one’s entire earnings history or only on one’s partial earnings history (as e.g. in Germany and in the US respectively). We should point out that we only focus on the balanced growth path in our model. Although a model such as ours would necessarily exhibit rich dynamics in moving from one retirement system to another, we merely wish to compare schemes that are already in place. Hence, there are no explicit dynamics which must be taken into account. The structure of the model is discussed in Section 2. Section 3 analytically presents the comparison as far as possible. We present a numerical example in Section 4 to study the long-run effects for schemes where no definite analytical conclusion is possible, compare growth and equality effects of a FF scheme as opposed to a PAYG scheme and give a brief summary of our results. Finally Section 4 concludes.

A number of recent papers have investigated the growth effects of taxes in the context of neoclassical growth models. Our study also considers growth effects, however, we adopt an approach that is different from the literature in such that we focus on the implications of social security plans. Due to the setup of our model, we are able to differentiate amongst a variety of these plans: we consider the way they are financed as well as the benefit rule. Since the benefit rule may induce intragenerational and/or intergenerational transfers, the macroeconomics of social security raises distributional issues. This is the reason why our study examines how pension plans affect both growth and inequality measures. Our model depicts an endogenous growth economy where human capital accumulation is the engine of growth. We distinguish PAYG and FF pension plans and, by introducing heterogeneous agents, are also able to contrast a Beveridgean and a Bismarckian benefit rule. Finally, due to the particular three period overlapping generations structure of our model, we can distinguish two sub-cases of the Bismarckian benefit rule: benefits can either be based on a partial earnings history, meaning that not all contributions agents have made are taken into account when calculating their pension benefits, or benefits can be based on a full earnings history, i.e. all contributions are taken into account. We find that whereas all PAYG regimes are detrimental for growth but beneficial for inequality, this need not be the case for FF schemes. Although the latter are usually neutral in terms of total capital accumulation in the economy (i.e. any decrease in individuals’ savings resulting from an increase in the contribution rate to the social security system is offset by the increase in savings of the social security system) even if they are not actuarially fair at the individual level (intragenerational redistribution can be involved), we show that the growth rate of the economy may well be affected by the contribution rate as the benefit rule will affect the incentive to invest in human capital. We find that only under a Bismarckian FF rule where benefits are based on partial earnings will the growth rate be increasing in the contribution rate: by ignoring contributions made early in life when calculating the pension benefit, the cost of not investing in human capital is two-fold: not only does one have a low wage, rather more, one contributes longer, yet is not ‘rewarded’ for this. The inequality implications of FF systems also depend on the benefit rule: inequality decreases under a Beveridgean plan, stays constant under a Bismarckian plan based on full earnings and increases under a Bismarckian plan based on partial earnings history. Finally, as in the literature on tax reforms, we should note that our analysis is purely descriptive and does not involve any normative judgement. Our only purpose is to show how the choice of a particular pension plan matters in terms of the long-run growth rate and income disparities. Given that an economy already has a pension scheme, our analysis does not allow us to recommend that a different scheme should be implemented: the choice of the best pension plan would require (i) solving a long-run efficiency-equity trade-off and (ii) comparing transitional costs and gains. This could be done by selecting a complex social welfare function taking into consideration intragenerational and intergenerational disparities. However, this is a line for future research.