چطور مردم تصمیم واقعی بلند مدت می گیرند : یک مورد از آماده سازی بازنشستگی

Large variations in retirement wealth are common, with some households accumulating hundreds of thousands of dollars and others accumulating next to nothing. We examine to what extent formal planning or simple rules of thumb contribute to these differences in wealth accumulation. In particular, we investigate whether those who follow simple rules of thumb or those who come up with more complete plans accumulate more wealth than those who take an unsystematic approach. We test this empirically using a specifically designed survey about retirement preparation. We find that people who rely on a rule of thumb behave like literal planners. However, people without any systematic approach save substantially less. Our results, taken together with evidence from psychology, suggest that financial planning advice based on simple rules of thumb may be helpful for those who currently take no systematic approach.

Households differ markedly in their wealth accumulation for retirement. Attempts to understand these differences, such as Ameriks et al. (2003), Bernheim et al. (2001), Beshears et al. (2008), Laibson et al. (1998), Lusardi and Mitchell (2007), Saez (2009) and Scholz et al. (2006) have taken various approaches. People may differ in their budget constraints, information, discount rates, the functional form of their discount rates (such as hyperbolic discounting), whether they engage in planning, or their financial literacy. In this paper, we consider the role of what we will dub decision processes, i.e. the steps one takes to make a choice. Bernheim et al. (2001) suggest that the extent to which accumulated retirement wealth varies across households is difficult to reconcile with the classical life cycle model. Traditionally, economists assume that choices are the outcome of optimization over consumption in the presence of a budget constraint. Agents are assumed to identify the choice that leads to the highest satisfaction out of all feasible choices. In many domains, this optimization process is fairly straightforward, most notably for choices that do not involve long time horizons or substantial uncertainty. In contrast, in the domain of life cycle saving, literal optimization requires engaging in contingent planning and backward induction. Indeed, most economists agree that people do not literally engage in optimizing in this domain. Rather, there is consensus that people's behavior should be understood “as if” determined by an optimization process (Friedman, 1953). Interestingly, though, economists have remained surprisingly vague about what the “as-if” metaphor would mean in practice. In other words, economists have paid little attention to how real people actually arrive at the retirement savings decisions that they make. A growing literature has started to fill this gap. In particular, Ameriks et al. (2003), and Lusardi and Mitchell (2007) provide evidence that those who plan for their retirement accumulate more retirement wealth. What remains an open question is which precise aspect of planning, i.e. which decision process, leads to the positive relationship between “planning” and the accumulation of retirement wealth. Is it calculating a detailed savings plan? Or could it be that a simple rule of thumb may also work as a “plan” to boosting wealth accumulation? We define a decision process to consist of a series of steps that one takes to make a choice. In this paper, we consider three potential prototypes of decision processes people may adopt when deciding about their retirement savings. We dub the first prototype the literal planning approach. This relies on careful intertemporal budget calculations, perhaps based on financial software or expert advice. We dub the second prototype the rule-of-thumb approach. In contrast to the literal planning approach, some individuals may not actually engage in any formal planning process. Rather, they may follow some simple rules of thumb such as putting aside a fixed percentage of their monthly earnings. Finally, some individuals may not engage in any planning at all nor consciously follow any specific rule. We dub the latter the unsystematic approach. Crucially, individuals’ decision processes may differ in terms of the degree of their sophistication and therefore how costly they are to implement. Gathering detailed information for working out a careful plan may require a substantial amount of time. In contrast, copying a simple rule of thumb from a friend or simply maximizing the company match for a 401(k) plan does not require a large time investment. Importantly, the costs of working out a careful plan may differ across individuals. It may require little effort for someone with substantial mathematical or accounting skills. On the other hand, it may be very burdensome for people with low planning skills and to those with a high disutility from thinking about economic issues. Decision processes should be distinguished from preferences over consumption profiles. Intertemporal preferences determine the optimal choice of lifetime consumption profiles. However, finding out the choice that maximizes intertemporal preferences may be costly due to bounded rationality. As a result, individuals may make choices that differ from the preference-maximizing ones. Individuals may use simplified planning or rules of thumb as a procedure to come up with a savings choice. Individuals with identical preferences over intertemporal consumption streams may easily make different choices if they differ in terms of cognitive abilities or the propensity to plan, and hence in their decision making process. In this paper, we investigate how the three prototypes of decision processes, literal planning, a rule of thumb, and the unsystematic approach, contribute to differences in wealth accumulation. Bernheim et al. (2001) posit that observed variation in wealth at retirement is likely to be caused by deviations from “rational, farsighted optimization” and that differences are more consistent with a rule of thumb. However, previous research has not had access to data that could directly measure whether individuals follow a plan based on farsighted optimization or instead follow a simple rule of thumb or even exhibit an unsystematic approach to life cycle saving. For our analysis, we use a novel data set that has been collected for the purpose of this study. These data allow us to categorize individuals according to the three prototypes mentioned above by means of specifically designed survey questions. We first examine whether all of the three prototypes can actually be observed. We investigate their relative frequency as well as the determinants that lead individuals to adopt either prototype. The main topic of the paper is then to explore whether all three prototypes of decision processes lead to comparable savings outcomes or whether they are associated with systematic differences in outcomes. Our data come from a detailed survey module on decision behavior that we fielded with the American Life Panel at the RAND Corporation. Our module consists of questions on how individuals proceed when making their retirement savings decisions, as well as questions on choice outcomes and individual characteristics. In contrast to our survey module, traditional economic data sets typically contain only the latter two. We present two main findings. First, the planning and rule-of-thumb approaches are associated with substantially higher retirement wealth accumulation than the unsystematic approach. Second, we do not find any statistically significant difference between the outcomes for planners and rule-of-thumb savers. Thus, rule-of-thumb types behave as if they were planners. In Section 2, we discuss evidence from psychology of how even a randomly assigned rule of thumb can change behavior. In light of this evidence, our results suggests that a rule of thumb may be an effective device for retirement saving for individuals who find working out a careful plan too demanding. Our paper is closely related to the papers of Ameriks et al. (2003) and Lusardi and Mitchell (2007). Both investigate the role of planning for wealth accumulation and find that planning does indeed have an economically significant effect on wealth accumulation. A crucial difference between these two papers and our own is our aim to investigate multiple decision processes rather than just “planning.” In particular, we address the question whether adopting a simple rule of thumb leads to a similar amount of wealth accumulation as working out a careful plan. This question is important from a policy perspective; advice in the form of a simple rule of thumb is by definition simpler than a plan and may be an effective way to help those who do not have a sophisticated plan. Overall, our finding is that those who do not work out any plan but simply rely on a rule of thumb behave in a similar way as proper planners. In contrast, those following an unsystematic approach save substantially less. This suggests that treating both rule-of-thumb and unsystematic-approach people just as one single “non-planning” category as in the analysis of Ameriks et al. (2003) and Lusardi and Mitchell (2007) may not reveal the full picture. A further important difference between the work of Ameriks et al. (2003) and ours relates to differing samples. Their work is based on a highly skilled sample of TIAA-CREF participants. In contrast, our sample comes from the American Life Panel and is more representative for the overall population. Rodepeter and Winter (2003) provide an in-depth discussions of the literature on rule-of-thumb behavior in the domain of life cycle saving and simulate the outcomes resulting from various rules of thumb. While many authors have considered the use of a rule of thumb for savings,1 to our knowledge we are the first to measure their prevalence or their effect on wealth accumulation using micro-data. The rest of the paper is organized as follows. Section 2 outlines the psychology of planning and wealth accumulation and spells out our main hypothesis. Section 3 presents our data and outlines the definitions of decision making types used for analyzing the data. Section 4 investigates the relationship between types and retirement wealth accumulation. Section 5 considers several robustness checks. Section 6 concludes.

In this paper we have addressed whether the adoption of different decision making processes leads to systematically different levels of accumulated retirement wealth. Interest in this question arises for several reasons. First, previous research has found that variables related to the classical life cycle model do not explain much of the observed variation in retirement wealth. Second, knowing more about individual decision processes in the realm of retirement savings has practical consequences for a good design of a pension system. Third, addressing this question relates to a fundamental methodological issue in economics: whether or not decision processes are relevant for decision outcomes. Famously, Friedman (1953) suggested that decision processes are irrelevant. He argued that, even if some (or the majority of) people did not literally optimize by means of a careful planning process, their behavior would nevertheless look “as if” they optimized or had made a careful plan. While we cannot be sure that our planners are literally optimizing, they are calculating retirement needs and savings rates necessary to achieve those needs. At some fundamental level, this is similar in spirit to solving the life-cycle model. Our results allow us to conclude that rule-of-thumb types behave “as if” they were planners in that they accumulate a similar amount of wealth. This finding is novel and may be very welcome since, for a substantial part of the population, understanding the logic of a sophisticated plan may be rather difficult. On the other hand, we find that planner and rule-of-thumb types accumulate substantially more retirement wealth than the remaining category that we have dubbed unsystematic. Thus, decision processes do matter. The unsystematic type category counts for about one third of our sample. Hence, the behavior of a substantial fraction of the population does not look “as if” it is derived from a careful plan. Because of this, variation in decision behavior does indeed contribute to explaining part of the variation in observed retirement wealth. Building on the research of Ameriks et al. (2003) and Lusardi and Mitchell (2007), our results show that non-planners are not a homogenous group. Half of the non-planners are in fact rule-of-thumb types whose outcomes in terms of wealth accumulation are similar to planners. Our paper also has implications for the discussion about the validity of the life cycle model. The existence of unsystematic types and their much lower level of savings adds to the literature that refutes the life cycle model, while the behavior of rule-of-thumb types supports it. Thus, our research points to the importance of considering multiple models of behavior to describe subsets of the population. Our results suggest two main implications for policy making. First, it is important to be aware that people use different processes to make retirement savings decisions. In particular, unsystematic types may react to changes in pension policy differently from the other types. Economists should try to account for heterogeneous decision types in their theoretical models. This is all the more an issue given that all three decision making categories are substantial in size. Second, our results show that simple rules of thumb may be a powerful device for accumulating retirement wealth. We cannot make any strong conclusions that rules of thumb have a causal influence on wealth based on our data. However, there is a substantial body of evidence from psychology on the causal effect of such rules on behavior related to the achievement of long-term goals such as health behavior. This evidence suggests that at least part of the positive association between rules of thumb and retirement wealth may reflect a causal effect from a rule of thumb on wealth. If this is indeed the case, then it may be very helpful to spread information about simple rules of thumb that lead to an adequate level of retirement wealth for a range of representative cases. This may lead to a spread of adequate levels of retirement saving among the population by means of a non-intrusive policy measure.