In many advanced economies, an ageing population and increasing medical costs bring about concerns regarding households’ financial welfare in retirement and government finances. One solution to these issues is to encourage older workers to work beyond historical normal retirement ages so that they are more financially prepared for their retirement and contribute to government revenues longer.
Designing policy that would be effective in extending working lives, however, requires understanding what drives current retirement patterns. Although having a part-time bridge job after leaving a career job and before full retirement is becoming more common (Maestas 2010), the majority of workers still directly transition from a full-time career job to full retirement before age 65. Is this pattern of full retirement from a career job mainly driven by workers’ lack of interest in working longer? Or is it mainly driven by firms’ reluctance to employ older workers, in particular on a part-time basis? Is the primary issue that older Americans lose interest in working at all, perhaps due to changes in physical condition or the implicit disincentives provided by institutions (e.g. pensions, Social Security, and Medicare)? Or is it instead that firms find older and/or part-time workers significantly less efficient and more costly to manage? As Rogerson and Wallenius (2013) demonstrate, it is hard to tell the relative importance of the supply- and demand-side factors just from the observed retirement patterns. Observed behaviour is an equilibrium outcome reflecting both supply and demand.
Our recent research introduces strategic survey questions that are purpose-designed to solve this identification issue (Ameriks et al. 2018). The questions are fielded as a part of the Vanguard Research Initiative (VRI), which surveys a sample of account holders at Vanguard who are at least 55 years old to understand the decision making of older Americans.1 VRI respondents are healthier, wealthier, and more educated than the general US population. Because VRI participants remain healthy in older ages, are more productive (as proxied by education and wages), and work in occupations that do not require much physical exertion, this is a population strata for which understanding the reasons behind abrupt exits from the labour force are particularly policy relevant.
The strategic survey questions, building on pioneering work by Barsky et al. (1997) and Ameriks et al. (2011) among others, ask respondents to make choices about hypothetical work opportunities. The main advantage of the strategic survey questions is that they explicitly and precisely control the demand-side factors including wage and characteristics of the job offer, allowing researchers to clearly identify respondents’ preferences.
The following is an example of the strategic survey questions that are asked in our study, in which a hypothetical job offer is available at the moment of the survey, which has a fixed work schedule and otherwise the same characteristics as the most recent job that respondents had (called ‘reference employment situation’ in the survey):2
In the scenario that follows, you will be asked to make a choice between Option A and Option B. Option A will be to immediately accept a specified employment opportunity that will be open only for a short window (say one week) and will then become unavailable. Option B is instead to pursue other possibilities including searching for another employment situation or not working.
- Option A is a new employment situation that involves a fixed work schedule. Other than this possible difference, the new employment situation matches your reference employment situation (i.e. the most recent employment situation) in terms of occupation, annual earnings, and all other characteristics.
- Option B is instead to pursue other possibilities including searching for another employment situation or not working.
Once the choice between Option A and Option B is made, it further asks for the reservation wage that makes the respondent indifferent between Option A and Option B.3
The survey responses reveal a strong willingness to work among older Americans who are currently not working. Even when the hypothetical job opportunity requires them to work exactly the same number of hours as in their previously-held reference job, about 40% of the VRI sample that are current not working report that they would accept the offer (Figure 1, blue bars). The acceptance rate is slightly higher for those who had a bridge job after leaving the career job.4 Some of them are even willing to accept a significant wage reduction to go back to work. More than 20% of non-workers are willing to take a 10% reduction in wages and more than 10% are willing to take 20% reduction in wages.
Figure 1 Would current non-workers return to work?
(a) Had no bridge job
(b) Had a bridge job
Notes: N=1,337 for Panel (a) and 434 for Panel (b). Wage reduction is calculated relative to the wage of the reference job. The blue bars represent the acceptance rates under a fixed schedule scenario, and the yellow bars represent the increase in the acceptance rates with flexibility in the work schedule.
Source: Ameriks et al. (2018).
Motivated by the recent evidence that part-time options are relatively more common among post-career bridge jobs (Maestas 2010, Rupert and Zanella 2015, and Ramnath et al. 2017), the strategic survey questions also included a scenario where the job opportunity allows respondents to choose the number of hours worked. The survey responses reveal strong preferences for a flexible work schedule among older Americans. The acceptance rate for the hypothetical job opportunity is substantially higher under a flexible work schedule than a fixed work schedule (yellow bars in Figure 1 indicate the increases in the acceptance rate in the flexible schedule compared to the fixed schedule). Perhaps the most striking finding, more than half of the current non-workers would be willing to work again if they could choose the number of hours worked and earned the same hourly wage as in their most recent job. About 40% of them would be willing to take a 10% reduction in hourly wage, and about 20% would be willing to take a 20% reduction in hourly wage, to work under a flexible schedule if other conditions were similar to their most recent job.
The finding that the current non-workers want to come back to work, with some even willing to accept a significant wage reduction if jobs were flexible, points to the importance of the demand-side factors that discourage employers from hiring older workersin explaining why they are not currently working. Though the strategic survey questions are not designed to identify constraints or preferences of firms, the worker preferences identified from the questions can be used to put bounds on the demand-side factors that are consistent with the observed behaviour. For example, the retirement model in Rogerson and Wallenius (2009), along with our estimates on the preference for working part-time, implies that the hourly wage that a competitive employer would pay for a half-time position need to be at least 22% lower than that for a full-time position to explain why a majority of the sample would suddenly transition from a full-time job to full retirement.
In summary, the strategic survey questions that are designed to estimate older Americans’ preferences for labour supply reveal that they would work longer if jobs were flexible. Thus, demand-side factors, in particular those that discourage part-time work, are likely to be important in explaining current late-in-life labour market behaviour and need to be considered in designing policies aiming to promote working longer.
Authors’ note: The views expressed in this column are those of the authors and not necessarily those of The Vanguard Group, Inc. or the Federal Reserve Board.
Ameriks, J, J Briggs, A Caplin, M Lee, MD Shapiro, C Tonetti (2018), “Older Americans would work longer if jobs were flexible”, Vanguard Research Initiative Working Paper.
Ameriks, J, A Caplin, S Laufer, S van Nieuwerburgh (2011), “The joy of giving or assisted living? Using strategic surveys to separate public care aversion from bequest motives”, Journal of Finance 66: 519–561.
Barsky, RB, FT Juster, MS Kimball, MD Shapiro (1997), “Preference parameters and behavioral heterogeneity: An experimental approach in the health and retirement studies”, Quarterly Journal of Economics 112: 537–579.
Maestas, N (2010), “Back to work: Expectations and realizations of work after retirement”, Journal of Human Resources 45: 718–748.
Ramnath, S, JB Shoven, SN Slavov (2017), “Pathways to retirement through self-employment”, NBER Working Paper 23551.
Rogerson, R, and J Wallenius (2013), “Nonconvexities, retirement, and the elasticity of labor supply”, American Economic Review 103: 1445–1462.
Rupert, P, and G Zanella (2015), “Revisiting wage, earnings, and hours profiles”, Journal of Monetary Economics 72: 114–130.
 See http://ebp-projects.isr.umich.edu/VRI/index.html for more details about the Vanguard Research Initiative.
 See Ameriks et al. (2018) for more details including exact wordings, set ups, and various scenarios of the SSQs.
 Ameriks et al. (2018) report evidence that respondents understand the hypothetical scenarios well. They also show that responses are consistent across various scenarios and also with preferences reflected in their behaviours.
 A career job is the longest or main job since age 40. A bridge job is any job held after the career job.