Aim 1: Empirically quantify the impact of automatic enrollment retirement policies on retirement savings and total savings. Aim 2: Test the hypothesis that the perceived costs of retirement saving prevent workers from saving for retirement. Aim 3: Estimate the size of the upfront costs. Aim 4: Explore alternative optimal default policies
Numerous models have been proposed to explain why people offered access to pension plans tend to adhere to defaults when included as part of a retirement system design (Carroll et al. 2009; Bernheim et al. 2015; Goldin and Reck, 2017). Our work seeks to explore how people respond to alternative default policies, and also how default polices could be designed to achieve social optima. While it would be useful to explain the underlying mechanism driving default behavior, in general it is necessary to specify how individual utility enters into the social welfare function in order to characterize optimal default policies. Yet if behavioral factors generate observed default behavior, it is often unclear how to account for such psychological factors in the social welfare function. Another consideration is that many theories can account for default behavior, making it challenging to differentiate their predictions in empirical data, particularly when people are heterogeneous. Instead of proposing models to explain the underlying reasons for defaults, this project introduces a generalized sufficient statistic model to characterize individual saving and consumption decisions in the presence of a default. The goal is to better understand why people often fail to save for retirement on their own, yet tend to default into retirement plans when offered them at their place of employment.
Our project seeks to evaluate how UI differentially affects older versus younger workers’ labor supply and the impact of UI extensions (and their removal). Our results will improve understanding of how people with heterogeneous characteristics respond differently to policy changes. If the proposed aims are achieved, the results can provide guidance to policymakers on designing more effective UI policies in future recessions.
Prior research has shown that unemployment insurance (UI) reduces labor supply among younger workers, but it does not seem to do so for older workers. A potential explanation for the difference may be that the older population (age 40-60) faces consumption commitments in the form of goods involving transaction costs, so they are constrained from changing their consumption patterns frequently. The objective of this project is to investigate the effects of consumption commitments on labor supply and welfare. Our hypothesis is that older workers with consumption commitments respond less to policy changes in unemployment insurance than younger workers. In other words, moral hazard induced by unemployment insurance is modest for older workers. As a result, a deviation from optimal consumption caused by adjustment cost can reduce older workers’ welfare. For this reason, social welfare gains are potentially feasible by redistributing unemployment benefits from younger workers to older workers to maintain consumption commitments without sacrificing too much non-commitment consumption. We will use the expiration of unemployment insurance extension (EUC08) in December 2013 and data from the CPS and SIPP to examine whether older and younger workers responded differentially.
Retirement benefits under the US Social Security system are currently provided as a lifelong benefit stream, though some workers might be willing to trade a portion of their annuity streams un exchange for lump sum amount. In a previous theoretical paper we explored whether allowing people to receive a lump sum as a payment for delayed retirement rather than as an addition to their lifetime Social Security benefits might induce them to work longer on a voluntary basis. This proposal outlines out intention to implement an experiment using the American life Panel (ALP), where we will seek to assess whether people might be willing to work longer in exchange for receiving their delayed retirement credit under Social Security as a lump sum payment instead of as an annuity. We will also evaluate what the price would be for deferring retirement.
This project will analyze how changing family status over the life cycle influences optimal portfolio choice, for stocks, bonds, life contingent-assets (life annuities and term life insurance). Prior models have not focused in much detail on the way in which changes in marital status and numbers of children might influence such investors behaviors. Nevertheless, in the presence of differential labor income profiles for men and women, there is expected to be an increased demand for term life insurance, joint survivor or single life annuities, since labor income cannot be adjusted easily in case of a spousal death. In addition, children affect parental decisions by requiring consumption as well as time inputs. In the case of a divorce, assets (including pension claims) must be split by the partners. We will analyze the effect of changing family status over the life cycle on optimal retirement accumulations and decumulations behavior taking into the account current and prospective future US Social Security and taxes.
Millions of unorganized and informal sector workers in the developing world are excluded from formal pension and social security systems. Old-age economic security is a pertinent problem for such population groups and providing adequate and secure income flows in the future is a formidable challenge. The problem is aggravated by demographic transitions associated with significant increases in life expectancy and changing social structures like the breakdown of the traditional extended family system, making today’s workers vulnerable to unmitigated longevity risks, uncertain health costs and poverty in their post retirement period. Further, with underdeveloped annuity markets and poor financial literacy, workers face considerable challenges in retirement planning decision-making. Of late, a market for micropensions -- defined contribution pensions for the informal sector workers – has emerged. To determine how such long-term saving products might help solve the problem of old-age income security for informal sector workers, we require an improved understanding of the behavioral, economic, and institutional barriers to participation. Our study will examine the emerging micropensions market in India, to better understand the economics and key institutional aspects of defined contribution retirement benefit schemes currently on offer for informal sector workers. Our work will lead to additional research and policy analysis, as well as experimental analysis, of the potential for micro-pensions to enhance retirement security in developing countries.
This project will evaluate whether and how ambiguity or uncertainty aversion influences economic decision making. Risk is different from ambiguity, where the former refers to stochastic events with known outcome probabilities, and the latter refers to stochastic events where the probabilities are unknown. Ellsberg (1961) suggested that, on average, people are averse to uncertainty, strongly preferring risks with known probabilities over those with uncertain likelihoods. The current project will measure and assess the effects of ambiguity aversion on economic decisions using a purpose-built survey module to be implemented in the nationallyrepresentative American Life Panel. Ambiguity attitudes will be elicited by varying the degree of ambiguity presented to each respondent in an internet survey, depending on how he responds to an initial set of questions. Next, this project will evaluate whether respondents who indicate they are more averse to ambiguity are also those who plan and save more for retirement, are more financially literate, invest more in their health, learn more about their future pension benefits, invest less in stocks, and otherwise protect themselves against uncertainty.
The proposed research will study the importance of health insurance availability in the decision of whether to annuitize wealth at the time of retirement. Recent research has suggested that uninsured medical expenses among the elderly may contribute to the under-annuitization puzzle, but there is no direct evidence on this question. We propose to study how the annuitization decision of retirees may be affected by Medigap insurance coverage. Our first contribution is therefore to provide direct evidence on the causal effect of insurance coverage on annuitization. We will also explore the mechanisms by which insurance affects annuity demand by testing whether individuals with higher and more volatile expected medical costs are more responsive to insurance.
The proposed reserach will evaluate what role financial advisers can play in the context of a lifecycle portfolio choice model. Previous efforts have focused on optimal portfolio allocation patterns for a rational forward-looking consumer who must decide on her own how to allocate her accruals between stocks and bonds. The proposed paper will add several new features to the basic model to explore why and when a financial adviser might be optimally brought in to help the consumer manager her portfolio. The first innovation is to allow for portfolio adjustment costs which many people must bear, when managing their own financial wealth. This has particular impact if the worker must accumulate job-specific human capital through learning by doing; in this instance, spending time on one's own financial management imposes an opportunity cost in terms of current and future job-related human capital accumulation. We also propose to incorporate an age-related time efficiency pattern for financial decision-making, in keeping with observed empirical evidence. These two innovations are likely to make it costly for individuals to manage their own portfolios, in ways that are consistent with observed low levels of trading in workers' 401(k) accounts. Our framework will also permit us to analyze the factors that influence whether people choose to managing their own money, versus delegate his task to a professional money manager.
Compulsory annuitization is often proposed as a compelling solution under defined-contribution pension schemes to help plan participants manage their longevity risk. This paper explores the current annuity market in Singapore and discusses the pros and cons of a proposal to mandate annuitization under the Singaporean national pension system -- the Central Provident Fund (CPF). We evaluate the pricing of various annuity policies in order to assess whether plan participants might benefit from higher annuity returns per dollar premium and lower adverse selection costs under the new annuitization mandate. Using 2007 pricing and mortality data, we aim to compute the money's worth value for life annuities available to a 55- year-old retiree in the Singapore population in order to assess whether private annuity providers currently offer good value-for-money annuities. The paper will also feature cross-country comparisons, as well as sensitivity analysis by varying mortality and interest rate assumptions. In addition, we analyze the payout structure of the newly-launched government-offered life annuities and assess their money’s worth values in order to draw conclusions about how the government's entry into the annuity market may impact private providers and participants' choice.
The proposed research seeks to investigate the determinants of borrowing and repayment patterns from 401(k) pension plans. In this pilot project, we seek to assess how employee and plan characteristics affect the propensity to take these loans, the likelihood of repaying them, and how large the loans are when people take them. The economic life-cycle model predicts that younger people will be more likely to take loans as they are more credit-constrained. 401(k) loans are particularly attractive since they tend to be available through the employer and do not require a credit check; they are also available on a tax-preferred basis as the contributions are pre-tax and the loan is pre-tax. However, if the employee does not make payments according to the plan regulation, the outstanding loan balance is considered as defaulted, and is subjected to normal taxation plus 10% penalty tax. Our objective is to produce a working paper which investigates the type of people who take loans; we will also explore how large the loans are, how often people take them, and what characterizes repayment patterns. This will provide insight regarding who takes loans and will help predict default rates of these loans. This will lay the groundwork for future explorations of how 401(k) loans shape eventual retirement wealth.
In the US, most employees offered defined contribution (DC) pension plans are provided a wide variety of investment options to chose from, when allocating their retirement saving portfolios. This study will investigate what happens when workers are offered a new type of investment option, namely, Life Cycle (LC) funds, which have been introduced in the last five years by pension plan sponsors.
The introduction of lifecycle funds into 401(k) plans offers a rich environment in which to assess workers’ portfolio allocation decisions. Consistent with behavioral models, employer design decisions strongly influence lifecycle adoption behavior while fundamentally altering portfolio characteristics, both in the cross-section and longitudinally. Yet there are also elements of rational choice by new employees, as well as choice constrained by information costs among workers with low literacy characteristics. We conclude that recent legislation encouraging riskier 401(k) portfolios will modify investment patterns, with the rate of change varying according to whether behavioral or rational elements dominate in a given setting.
To evaluate the effect of Chile’s pension system rules and regulations on individuals’ contribution patterns.
The main goal of this project is to evaluate the effect of Chile’s pension system rules and regulations on individuals’ contribution patterns. The few empirical studies on the Chilean Pension System have been limited to the use of aggregate and macro data. This project’s contribution is to analyze pension contribution patterns under the Chilean AFP system using micro data and state-of-the-art modeling methods. Previous researchers employing dynamic models to analyze retirement decisions have used US data which may not reflect behavior of workers covered under an individual accounts defined contribution retirement system. Accordingly, we will use the 2002 and 2004 rounds of the Chilean Encuesta de Prevision Social or Social Protection Survey (EPS), containing socio-demographic data, labor market data, and information regarding the pension system at the individual level, and link it to administrative records on monthly individual account movements and monthly wages. These linkages will allow the longitudinal analysis of contribution decisions and labor force participation decisions. We will then use regression analysis under different specifications to identify the variables that affect contribution patterns. In addition we will work to develop a dynamic behavioral model of individual decision-making about labor force participation and monthly contributions taking into account the fact that individuals working in different labor sectors face different contribution rules. Finally we propose to develop ways to estimate the parameters of this behavioral model using the simulated maximum likelihood method. Our particular interest will be to use the results to assess the possible impact on contribution patterns of changes in pension system returns and fees, as well as requirements for receiving the minimum pension.
To measure the efficiency shortfall patterns in pension plan menu offerings and to characterize these patterns according to employee and plan characteristics.
The proposed research seeks to investigate the adequacy and characteristics of investment choices offered by 401(k) plans. When constrained or inappropriate investment menus are offered to participants, they may experience decreases in returns as compared to a benchmark efficient portfolio. The goal, therefore, is to produce a working paper which measures the efficiency shortfall patterns in pension plan menu offerings, and to characterize these according to employee and plan characteristics. We do this so as to gain insight regarding the efficiency of investment menus offered in 401(k) plans, so that in future work we can further explore the investment efficiency of participants given the investment opportunities and the corresponding loss from participant investment choices which depart from the efficient ones.
To evaluate what Chileans do and do not know about their retirement system after two decades of the new program, and, to the extent that current participants and potential participants prove under- or mis-informed, to learn what aspects of the plan seem particularly difficult to fathom and what might be done to correct the lack of information and/or misinformation.
Microeconomic research generally assumes that workers are able to determine and follow optimal saving and retirement paths, and that to this end they have all the information necessary regarding the pension plan rules covering them. For instance, labor supply and saving outcomes at older ages are conventionally modeled by economists as depending on specific Social Security benefit and tax incentives that impart important notches and kinks in workers’ lifetime budget constraints. Most of these studies report finding quite small empirical behavioral elasticities, which may in fact be accurate measures of key behavioral parameters. On the other hand, these small elasticities could instead be the result of workers’ failure to understand how their pension systems actually work. This proposal seeks to evaluate worker knowledge of their pension systems by drawing on an invaluable new microeconomic survey of households linked with administrative records on actual accruals and benefit entitlements. In particular we will draw on a Chilean household survey known as the Encuesta de Protección Social (EPS). This survey, collected in both 2002 and 2004, queried respondents regarding demographic and household makeup, education and training, and most importantly for our purposes, it included a wide range of questions about peoples’ pension expectations and knowledge. To these files, we can link data from administrative records that we will use to compare participants’ self-reports with administrative information. We propose to use the project funding to prepare an exploratory report on initial findings, and to begin developing a range of econometric models of financial literacy using alternative identification strategies, including tests of the role of education in forming pension expectations.