Elsevier

Journal of Public Economics

Volume 91, Issue 10, November 2007, Pages 1944-1966
Journal of Public Economics

The choice between an annuity and a lump sum: Results from Swiss pension funds

https://doi.org/10.1016/j.jpubeco.2007.09.003Get rights and content

Abstract

This paper uses unique micro data from Swiss employer-based pension plans to study the annuitization decision at retirement. The administrative nature of our data, though limited with respect to individual background characteristics, allows us to analyze real choices over large retirement balances, rather than subjectively reported intentions to annuitize. We find a strong and robust impact of a utility-based measure of the annuity's value (computed within a life-cycle framework) on individual annuitization rates. Low accumulation of retirement assets is strongly associated with the choice of the lump sum, presumably due to the availability of means-tested social assistance. The sponsor's default option, in most cases the annuity, is also found to be highly influential in the decision to annuitize.

Introduction

An annuity is the only contract that guarantees income right up to the point of death. Nonetheless, international numbers show that only a small minority of individuals voluntarily purchases an annuity. A number of explanations have been put forward to explain this so-called Annuity Puzzle. First, the price of the annuity might be too high as a consequence of load factors, which in turn may be due to administrative costs, not fully competitive markets, and information asymmetries. Second, the desire to annuitize may be weakened by bequest motives, and precautionary savings to cover spending boosts (like health expenditures). Third, intra-family risk sharing may act as a partial substitute for the insurance implied by an annuity. In the presence of such motives, the existence of a first pillar that provides a basic income stream further diminishes the demand for the annuity. Notwithstanding all of these factors, it is not easy to reconcile the low annuitization rate with the much higher predicted rate even under more realistic conditions.

Despite the importance of the issue and the increasing involvement of the private sector in pension provision, there is surprisingly little empirical evidence on why individuals (do not) annuitize the accumulated pension capital at retirement. To the best of our knowledge there are only two contributions that empirically study the annuitization choice. The first one is presented by Hurd et al. (1998), who analyze pension cash-outs in the US using Health and Retirement Survey (HRS) data. The second contribution is provided by Brown (2001), who applies simulation based models of utility valuation to HRS data. In both studies the preference for lump sum pension settlements turns out to be a very persistent empirical result.

The main reason for the limited amount of empirical analyses of the pay-out phase seems to be a lack of reliable data. Without a sufficient level of voluntary annuitization, many potential data sources are of little use. Private annuity contractors and pension sponsors, moreover, are often reluctant to disclose individual choices. Funded mandatory pension schemes, which may act as an alternative source of data, are still immature in most countries, and consequently cannot alleviate the scarcity of data yet.

This paper contributes to the existing empirical literature in that it directly examines individual decisions to annuitize or to cash out pension wealth. Our analysis is based on unique micro administrative records from ten Swiss pension sponsors. It includes real choices over the use of substantial amounts of retirement savings (approximately $400,000 on average), as a consequence of the fact that in Switzerland occupational pensions constitute a second mandatory pillar and account for roughly half of retirement income on average.

The Swiss data has many advantages over alternative data sources. First, it includes actual choices with respect to mandatory insurance and not just the intention to annuitize in voluntary plans. Second, it provides reliable and complete information on pension plan details, which is not usually present in survey data. This allows us to price the annuity with respect to age, gender and marital status. Third, it is field data and generally involves large stakes. We can therefore expect individuals to spend more time and effort in thinking about their choice than in a laboratory setting. Fourth, although the data is not entirely representative for Switzerland, it exhibits similar average stakes, annuitization rates and retirement patterns. This is partly due to the fact that the entire workforce of a company is covered in a given period. As individuals do not have a choice between different pension providers in Switzerland (apart from the fact that they may choose their employer), the selection bias is thus presumably smaller than in comparable studies. Nonetheless, the data suffers from a number of shortcomings, notably a lack of information on non-pension wealth and limited individual information.

To account for the fact that an annuity also provides insurance against outliving one's assets, we compute for each individual the utility-value of the annuity in a life-cycle framework for different levels of risk aversion. This so-called annuity equivalent wealth, a measure developed by Brown and Poterba (2000), takes into account that insurance requirements differ by gender, marital status, and the level of pre-existing annuities.

We show that, in line with other empirical studies, the annuity equivalent wealth is the most important determinant of the annuitization decision. We also find sizeable differences among companies: Individuals often choose the standard option offered by the company or seem to follow their peers. Moreover, our analysis demonstrates that small stocks of old age capital are much more likely to be withdrawn as a lump sum. This may be due to higher rates of time preference. More importantly, for low levels of capital, it may be optimal to spend down the resources to qualify for means-tested social assistance.

The paper is organized as follows. Section 2 provides a brief description of the Swiss pension system, with an emphasis on the characteristics relevant for the objective of our analysis. The potential determinants of the demand of alternative pay-out options upon retirement are analyzed in Section 3. The data are presented and discussed in Section 4. Section 5 reports the results from several empirical specifications, and Section 6 concludes.

Section snippets

Background information on the Swiss (occupational) pension system

Switzerland's pension system has two main pillars, a publicly financed pay-as-you-go scheme and a mandatory fully-funded occupational pension scheme.1 Fig. 1 gives a first overview of the benefit structure for a typical individual. Its components will be explained in greater detail below. The first pillar

Potential determinants of the demand for different pay-out options

This section briefly summarizes the potential determinants postulated in the theoretical literature as to whether to annuitize retirement wealth or to cash it out. These factors, which are discussed in the light of the data at hand, will then be used (jointly with other individual background characteristics) in the empirical analysis.

The data, sample restrictions and limitations

We use data collected at the individual level from ten Swiss companies, both public and private, active in several areas of the economy.10 The novel aspect of our data is that it is not survey data, but

Specifications and empirical results

This section presents the results of a two-limit Tobit analysis applied to the pooled sample. Unlike traditional regression coefficients, the Tobit coefficients cannot be directly interpreted as estimates of the magnitude of the marginal effects of changes in the explanatory variables on the expected value of the dependent variable. In a Tobit equation, estimated coefficients include the influence of the regressor on both the intensity of annuitization and the probability to annuitize.

Conclusions

To the best of our knowledge, our paper is the first to provide empirical evidence on real rather than voluntary annuitization decisions. Our study is based on a unique data set of individual decisions made with regard to the Swiss second pillar. Several novel aspects are incorporated. First, the analysis is based on administrative records made available by ten Swiss employer-based pension funds. The limited flexibility of the administrative data is compensated for by the fact that it allows to

References (20)

  • J. Brown

    Private pensions, mortality risk, and the decision to annuitize

    Journal of Public Economics

    (2001)
  • R. Thaler

    Some empirical evidence on dynamic inconsistency

    Economic Letters

    (1981)
  • G. Ainslie et al.

    The motives of the will

  • R. Barro

    Are government bonds net wealth?

    Journal of Political Economy

    (1974)
  • G. Becker

    A theory of social interactions

    Journal of Political Economy

    (1974)
  • D. Bernheim

    How strong are bequest motives? Evidence based on estimates of the demand for life insurance and annuities

    Journal of Political Economy

    (1991)
  • J. Brown et al.

    Joint life annuities and the demand for annuities by married couples

    Journal of Risk and Insurance

    (2000)
  • M. Bütler

    Mandated annuities in Switzerland

  • T. Davidoff et al.

    Annuities and individual welfare

    American Economic Review

    (2005)
  • A. Finkelstein et al.

    Selection effects in the market for individual annuities: new evidence from the United Kingdom

    Economic Journal

    (2002)
There are more references available in the full text version of this article.

Cited by (103)

  • Evidence of behavioural life-cycle features in spending patterns after retirement

    2022, Journal of the Economics of Ageing
    Citation Excerpt :

    Other studies focus on the decision to annuitize pension wealth. While this would be optimal from a standard life-cycle perspective (e.g., Davidoff et al., 2005), individuals who are given the choice between annuities or receiving a lump sum at retirement, often tend to choose a lump sum — in the literature this is commonly referred to as the annuitization-puzzle (e.g., Bütler and Teppa, 2007). Insights from behavioural economics can be helpful to understand how individuals decumulate their pension wealth.

  • An investigation of time preferences, life expectancy, and annuity versus lump sum choices: Can smoking harm long-term saving decisions?

    2020, Journal of Economic Behavior and Organization
    Citation Excerpt :

    Variables such as gender, marital status, children, and attitudes such as risk aversion and time preferences are discussed with no definitive conclusions (Agnew et al., 2008; Bütler and Teppa, 2007; Cappelletti et al., 2013; Chalmers and Reuter, 2012). Some papers highlight other personal characteristics in an attempt to explain the puzzle, such as medical conditions (Sinclair and Smetters, 2004; Turra and Mitchell, 2008), prior existence of a social security annuity (Benartzi et al., 2011; Chalmers and Reuter, 2012), and the subjective discount rate (Bütler and Teppa, 2007; Warner and Pleeter, 2001). Other papers reveal several psychological and irrational behavioral barriers to annuitization, such as complexity of the decision (Brown et al., 2013, 2017), default biases (Agnew et al., 2008; Bütler and Teppa, 2007), difficulty in making irreversible decisions (Brown and Warshawsky, 2004), different biases that arise from framing (Benartzi et al., 2011; Beshears et al., 2014; Goldstein et al., 2016), difficulty parting with accumulated money (Benartzi et al., 2011), mental accounting (Benartzi et al., 2011), availability errors (Hu and Scott, 2007), ambiguity regarding self-rated life expectancy (Smith et al., 2001), and the belief that annuities have a “smell of death” (Statman, 2017).

View all citing articles on Scopus

This research was supported by the Swiss National Science Foundation (1214-67875.02), the “Bureau de l'Egalité des Chances” (Université de Lausanne, 2656958), and the RTN Project “Financing Retirement in Europe” (HPRN-CT-2001-00225). We are most grateful to the pension funds that provided us with data and other valuable information on the working of the funds. Martin Ruesch did a great job programming the code to compute utility measures of the annuity's value. Tess Faessler and Sharon Bochsler supported us in proof-reading the text and designing tables, respectively. We received extremely helpful comments from the participants of 2006 Tapes conference in Uppsala, in particular our two discussants Joshua Rauh and Vincenzo Galasso. The paper greatly benefited from very constructive comments of two anonymous referees and the editor. We would also like to thank Rob Alessie, Axel Börsch-Supan, Olivia Huguenin, Arie Kapteyn, Joachim Winter, as well as seminar and conference participants at various institutions and locations for their many useful suggestions and comments. Any errors are our responsibility.

View full text