سیستم های تامین اجتماعی و تنزیل شبه هذلولی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|24177||2007||16 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : European Economic Review, Volume 51, Issue 5, July 2007, Pages 1247–1262
Hyperbolic discounting has become a common assumption for modeling bounded rationality with respect to individual savings decisions. We examine the effects of hyperbolic discounting on the comparison of alternative social security systems. We show that this form of bounded rationality breaks the equivalence between funded and pay-as-you-go (PAYG) systems established in Sheshinski and Weiss [Sheshinski, E., Weiss, Y., 1981. Uncertainty and optimal social security. Quarterly Journal of Economics 95, 189–206]. Intergenerational transfers within a PAYG economy are usually secured by the social security system and independent of longevity, whereas this is not the case for the funded economy. The savings level under hyperbolic discounting is lower than under exponential discounting [Laibson et al., 1998], but the ratio between the savings level under hyperbolic discounting within a funded economy and a PAYG economy depends on the effectiveness of the commitment devices. It is shown that if individuals are hyperbolic discounters, then in a PAYG economy any change in the mandated level of intergenerational transfers is neutralized by individuals’ voluntary bequests. This does not apply to a funded system.
Time inconsistency of individual preferences and the notion of dynamic preferences or with evolving “selves” is a well-known and studied phenomenon (Pollak, 1968; Phelps and Pollak, 1968; Peleg and Yaari, 1973; Hammond, 1976; Thaler and Shefrin, 1981; Schelling, 1984). The prevalent tool for modeling time inconsistency of individual preferences in the last decades of the twentieth century was the hyperbolic discounting function (Ainslie, 1992; Lowenstein and Drazen, 1992; Laibson et al., 1998; Laibson, 2001), and such modeling can be traced back to the mid-20th century (Strotz, 1956; Chung and Herrnstein, 1961). The hyperbolic discounting function was applied to the study of undersaving and the sharp reduction in consumption of old people (Laibson, 2001), to the early retirement pattern of workers (Laibson et al., 1998; Diamond and Kőszegi, 2003), and to job search behavior (Paserman and Della-Vigna, 2000). The hyperbolic or quasi-hyperbolic discounting hypothesis (henceforth, HDH1), as an empirical finding, was based on experiments performed on humans and animals that the researchers interpreted as supporting this hypothesis (for a survey, see Laibson et al., 1998). However, as an empirical fact, HDH is controversial (Rubinstein, 2001 and Rubinstein, 2003; Read, 2001; Besharov and Coffey, 2003). From a theoretical point of view, hyperbolic discounting raises several problems. For instance, time inconsistency reflects irrationality, or at least bounded rationality, especially in its naïve version. However, in certain circumstances of uncertainty, hyperbolic discounting can be reconciled with rationality and does not necessarily generate time inconsistency and reversal of preferences (Weitzman, 1998; Azfar, 1999 and Azfar, 2002; Dasgupta and Maskin, 2003). Our purpose in this paper is neither to decide on the empirical controversy nor to find any further theoretical justifications for assuming hyperbolic discounting by rational individuals. Rather, taking HDH for granted and leaving aside the empirical controversy, we explore its implications for optimal social security systems. The main result of this paper is that the equivalence between an optimal pay-as-you-go intergenerational transfers system (henceforth a PAYG system) and an optimal funded pension system, established in Sheshinski and Weiss (1981, henceforth SW), does not hold in general under hyperbolic discounting. The model we analyze in this paper is an extension of the SW model. In order to facilitate the analysis of the hyperbolic discounting effect, we have added a third period to the standard two periods’ model. We show that the equivalence between funded and PAYG pension systems, established in SW, holds only under exponential discounting (which is a special case of the hyperbolic discounting). A social security system provides perfect insurance under PAYG structure, whereas in a funded economy bequests are random and depend on realized longevity. The correlation between bequests and longevity in a funded economy depends on the elasticity of the parents’ marginal utility from bequests. A similar result is obtained for savings. The savings level is identical in the two economies only if effective commitment devices exist in the PAYG economy. We also analyze the effect of government intervention in private decisions (i.e., introducing a compulsory pension scheme) and show that under hyperbolic discounting, such intervention has different consequences on the two types of economies. The theoretical equivalence of PAYG and funded social security systems in various theoretical models was generally ignored in pension reforms carried out in western countries. The common direction of all these reforms is a shift from a PAYG system to a funded system. However, the logic of this shift and its costs are controversial. It has been argued that one social security method has no advantages over the other (Orszag and Stiglitz, 1999). Our results imply that the ability to rank these systems holds only under the limiting assumption of exponential discounting. If people are hyperbolic discounters then each pension method has its own unique advantages and disadvantages. In the next section, we briefly describe the hyperbolic discounting function and the time inconsistency that it causes in individual decisions. Next, we present the familiar SW model modified for the hyperbolic discounting analysis using a three-period version. The following sections include a number of propositions and a discussion about their implications for optimally designed pension systems. The final section is a discussion and conclusions.
نتیجه گیری انگلیسی
The global trend during the last decades of the twentieth century was a shift from PAYG and defined benefits (DB) pension systems to funded and defined contributions (DC) systems. The rationale for this shift is controversial (see Orszag and Stiglitz, 1999). However, as Sheshinski and Weiss (1981) show, assuming standard exponential discounting provides the positive question of how this shift affects the real variables of the economy with a strict and clear answer. The two types of social security are equivalent and, therefore, as long as the pension systems are optimally designed, no real change is expected. In this paper, we show that if people assign a higher weight to current period consumption relative to future periods as well as future generations consumption, this answer is no longer valid. While a PAYG system operates within the economy very similarly no matter what discounting function the individual actually uses, provided that there are effective commitment devices, the effect of the funded system and its equivalence to the PAYG system depends on the discounting function and on View the MathML sourceε(B^i). In an economy with hyperbolic discounters and effective commitment devices, a PAYG system might be preferable to a funded system, because it neutralizes bequests from the randomness of longevity. However, this neutrality is conditional on the refraining of the government from any attempt of intervention in the pension market. Effective governmental intervention in a PAYG economy with hyperbolic discounters reduces it to the funded equilibrium, but with random bequests and without the advantages of the funded system. As we emphasized above, we have not attempted to solve the controversy surrounding hyperbolic discounting, especially not the empirical aspect. However, since it is unreasonable to assume that the features of a utility function change by the shift from a PAYG system to a funded system4; if future evidence shows that savings and bequeathing patterns have changed because of the shift from one type of pension system to another, this may also serve as an empirical piece of evidence in favor of hyperbolic discounting.