اقتصاد سیاسی تأمین اجتماعی: یک نظرسنجی

This paper surveys the literature on the political economy of social security. We review models that address the following questions: (i) Why do social security programs that transfer resources from young and middle-aged workers to the elderly exist? (ii) What are the economic and political interactions between social security systems and other redistributive programs of the welfare state? (iii) How does political sustainability shape social security systems in a dynamic economic and demographic environment, and which social security reforms are politically feasible? We characterize this literature along two lines: economic factors and political institutions. We then assess the empirical relevance of the models by comparing their implications to stylized social security facts.

All developed countries have a social security system. An overwhelming majority of these systems is unfunded (or pay-as-you-go). The systems are financed through a payroll tax levied on current workers' labor income, and provide a pension benefit to retirees, i.e., former workers, who have reached retirement age and have exited the labor market with an entitlement to an old-age pension. The size of these unfunded social security systems has increased over the last few decades. In 1995, they absorbed 4.5% of GDP in the US, 13% in Italy, 16.5% in France, and over 20% in Sweden. There has been a wide body of literature on the economics of social security (see, among many others, Diamond, 1996 and Gramlich, 1996). However, decisions on social security policy, from the initial design of the system to later modifications, go beyond economic theory into the realm of politics. Contributions to political campaigns, lobbying, votes of vested interest groups influence career-oriented policy-makers to sacrifice economic principles for political objectives. This paper surveys the main contributions in the literature of political economy models of social security. We first review models that analyze the institution of the unfunded social security systems and their development into the most widespread instrument of social insurance. The challenge common to this line of research is to understand why there exist social security programs that transfer resources from young and middle-aged workers to the elderly. We highlight the empirical implications of the stylized models, and compare them to the social security “facts”, that emerge from the empirical literature. Many of the elderly political economy models used unidimensional voting models, and thus at most accounted for a single social security feature, typically the size of the system. This is a significant limitation, since social security systems are known to have several features.2 We then turn to a recent body of literature that applies multidimensional voting models to the analysis of social security systems. These models examine the interactions between social security and other redistributive programs of the welfare state, or among the different characteristics of the social security system. An insight of the multidimensional models is to recognize that the different programs of the welfare state may be economic and political complements or substitutes. These aspects are crucial in determining the size and composition of the welfare state. Finally, we consider the theoretical contributions to the more heated topic of the current social security debate. We review the literature that studies the response of the current social security systems to changes in the economic and demographic scenario, and the models of social security reforms. In particular, we focus on the following questions: How does political sustainability shape the social security system in a dynamic economic and demographic environment? Which social security reforms would be politically feasible? We characterize the political economy models of social security by political institutions and economic factors. Agents' individual preferences over the social security system can be aggregated through several political mechanisms. The political institutions encountered in the literature in the case of a one-dimensional issue space can be classified in three broad groups: majority voting, veto-power or constitutional rules, and interest-group models. There is a fundamental difference between interest-group models and models of voting (majority voting or veto-power). The former focus more on the political process, which may allow a powerful minority, the elderly, to carry through an intergenerational redistribution policy. In the latter models, social security arises if there are sufficient economic reasons for at least a majority of the electorate to support the system. Thus, voting models can be classified according to which economic factors lead young and/or middle aged agents to favor positive levels of social security. We identify five major economic reasons: (i) dynamic inefficiency (Aaron, 1966), (ii) a reduced time horizon in evaluating the social security program (Browning, 1975), (iii) the crowding-out of the aggregate savings by the social security system (Cooley and Soares, 1999a), (iv) the within-cohort redistribution element shared by many social security schemes (Tabellini, 2000), and (v) an altruistic motive (Hansson and Stuart, 1989). The combination of these two lines of classification—economic factors and political institutions—provides a natural way to catalogue the different models. Table 1 summarizes how the models address our first question: why do social security programs exist that transfer resources from young and middle-aged to retirees? Although voting models (majority voting and veto-power) may display several of the economic features, this characterization does not apply to interest-group models, which rely instead on the political power of the old.This survey complements Breyer (1994a), who reviewed the previous literature on the political economy of social security. We concentrate on the most recent body of contributions, and also adopt a broader view of the literature that includes models of policy reforms as well as multidimensional voting models of social security. Persson and Tabellini (2000) also provide a comprehensive treatment of the political economics of intergenerational transfers. They however have mainly a didactical purpose and focus on specific economic aspects to explain the rise of social security as an equilibrium outcome of a majoritarian election. They do not survey the literature. The paper proceeds as follows: Section 2 introduces a general economic environment in order to examine the economic elements in the voting models (majority voting and veto-power). Section 3 analyzes the three different political arrangements encountered in the literature. Section 4 reviews the empirical facts on social security and compares the implications of the models to the facts. In 5 and 6, we discuss, respectively, models of welfare state determination and models of social security reforms. Section 7 concludes.

Browning (1975) proposed that the large size of social security systems in democracies is due to political factors related to the age of the median voter. Since this insight, there has been a subsequent literature on political economy models of social security. What can be learned from this literature? Initial contributions have focused on the nature of the “social contract” that supports the social security system. Browning (1975) assumed an extreme form of commitment over future policies in the form of once-and-for-all decisions. Later models have shown that, in absence of a commitment device, social security can arise as the outcome of an implicit, unwritten contract among successive generations of individuals. A social contract defines a system of rewards and punishments that induces young voters to transfer resources to retirees, because they expect to be rewarded with a corresponding transfer in their old age. Interest-group models, on the other hand, have highlighted the role of political pressure exerted by the elderly to obtain a pension transfer. In the last few decades, there has been a dramatic increase in the social security expenditure, which in most industrialized countries has been related to the contemporaneous population aging process. As a response, a large literature has looked at financial sustainability of social security systems under the expected demographic dynamics (see Auerbach and Kotlikoff, 1987). The political economy literature has examined the demographic issue from a different perspective. Voting models have shown that an aging population has two opposite effects on social security: it increases the dependency ratio, thereby reducing the profitability of the system, and it raises the median voter's age, which creates more support for a larger social security system. Theoretical and simulated models have related the impressive increase in the size of the system to the aging process, and have considered how current systems can be expected to react to future demographic dynamics (see 4 and 6). Demographics alone, however, is not sufficient to explain the entire magnitude of the rise in the social security expenditure. Mulligan and Sala-i-Martin (1999a) identify a missing element in the political power of the old, created by the retirement policies that require, or induce, the old to exit the labor market in order to receive a pension. But why has this political power increased over the years? The reason for the rise in the social security expenditure of the last few decades remains an open question. Literature that analyzes social security in a wider, multidimensional context, which we reviewed in Section 5, will perhaps be able to make some progress in this direction. Future studies in this literature should also address the following questions: Is the increase in social security expenditure related to the retirement policy? Or to the rise in other welfare programs, mainly targeted to the old, such as health care? Or to some intrinsic feature of the development process that leads richer countries to have larger systems? The normative message of the literature on financial sustainability has been that reform is required to keep the system in balance, either through a structural change, for example from the PAYG to a mixed or unfunded system, or a “parametric” reform, for example, a rise in the retirement age, or a change in the formula used to calculate the pension benefits. Since decisions on social security reforms, just like any other social security policies, are in the realm of politics, political feasibility is important. Which reform would obtain the support of a simple (or qualified) majority of voters? The few contributions along these lines have not been able to provide general results, and have focused on calibrated models of the US social security system. In these cases, they have provided a negative answer. Although there may be gains from moving to an unfunded, or mixed system, no social security reform would obtain the required majority, due to transition costs that would have to be imposed on the current (voting) generations. Does this represent a description of the reality? Or is it a limitation that these political economy models of social security have not been able to identify a politically feasible reform? Some reforms have indeed been implemented around the world, mainly in countries in which the existing social security systems were relatively small. In other countries, like the US, despite the political debate, no reform has taken place yet. In fact, the calibrated studies of the US social security system suggest that no reform is politically feasible. Should a reform be implemented in the future, the existing models would prove to have very little explanatory power.