Gopi Shah Goda, Matthew R. Levy, Colleen Flaherty Manchester, Aaron Sojourner, Josh Tasoff 11 February 2020

US employees have increasing responsibility and face greater challenges in preparing for retirement savings.  Lifespans have lengthened and the costs of healthcare have greatly increased.  In addition, employers have shifted from defined-benefit retirement plans, which provide regular income to workers in retirement based on a pre-set formula, to defined contribution retirement plans, which provide incentives to accumulate savings for retirement.  This transition has shifted the risks of capital markets and uncertainty about one’s lifespan to the employee, increasing the stakes for retirement savings decisions.  

Unfortunately, saving for retirement is complex.  Knowing how to optimally save for retirement is a challenge, and rates of financial literacy are low (Lusardi and Mitchell 2014). In settings for which it is difficult to know the optimal choice, the way options are presented can have a profound effect on people’s choices.  One of the most robust findings in the retirement savings literature is the power of defaults.   Madrian and Shea (2001) find that the default saving rate – the rate of saving specified if an employee takes no action – has a highly influential effect on people’s retirement savings rate.  Specifically, they find that only 49% of new employees in their sample enroll in their employer’s 401(k) plan under the initial opt-in default.  A subsequent change of the default to auto-enrollment increases the participation rate to 86%.  Since then, research has shown that the default effect is persistent and does not crowd out other kinds of savings (Chetty et al. 2014).

What drives the ‘default effect’ in retirement savings?

However, less is known about why the default effect for retirement savings exists.  Many potential mechanisms have been suggested.  The first is inertia, which can be caused by a variety of factors. Present-biased preferences – i.e. the tendency to place more weight on the present relative to benefits in the future – can lead to constant procrastination (O’Donoghue and Rabin 1999a, 1999b, 2001, Carroll et al. 2009).  Choice overload, where a large number of options in a complex choice environment overwhelms a person, can cause them to select something simple (Iyengar and Lepper 2000).  Inertia can also be caused by the transaction costs (i.e. the hassle of paperwork) associated with making a change to one’s contribution.  Another possibility is that individuals judge losses and gains against the default, and aversion to losses makes the default appear more favourable (Tversky and Kahneman 1975).  A second mechanism is endorsement, whereby employees interpret the default as being endorsed by their employer (Butt et al. 2018).  In this case, inertia need not be present.  

READ ALSO  Welsh People Blocked From Buying "Non-Essential" Items Due To Lockdown

Understanding the mechanisms behind the default effect is important because the optimal policy guiding the framework for retirement-savings choice depends crucially on the root mechanisms.  For example, if transaction costs are the main cause, then the optimal policy may be to make paperwork easier.  In contrast, if endorsement is the main mechanism, careful consideration of employer-to-employee communications is warranted.  It is possible that multiple mechanisms are at play, with different mechanisms operating on different people.  If so, we may observe different mechanisms operating for different default levels.  

Our study        

In a recent study, we measure the association between several individual-level attributes and employees’ retirement contributions and default behaviour (Goda et al. 2019a).  During our period of analysis, the employer changed the default from opt-in (i.e. a 0% default savings rate) to automatic enrolment at 3%.  This allows us to test whether the effect of the default is associated with the same individual-level attributes under both regimes, or whether it varies.

Our study combines administrative records on employee contributions to their retirement savings plan with survey data.  The data come from employees at the US Office of Personnel Management (OPM), an agency that provides human resources, leadership, and support to most federal agencies. We analyse choices pertaining to the Thrift Savings Plan (TSP), a tax-advantaged DC plan for federal employees that provides retirement income in addition to a defined-benefit pension, controlling for a rich set of covariates relevant to contribution decisions, including salary, demographics and education.

Using experimental economics methods, we elicited four key individual-level attributes of the employees: their long-term discount rates, their present bias, their exponential-growth bias, and their financial literacy. Present bias has been posited as a driver of default behaviour for several decades (O’Donoghue and Rabin 1999b), but little direct evidence has connected individual measures of such bias to default acceptance (Brown and Previtero 2014, Blumenstock et al. 2018) and no prior studies have looked at it across different default regimes. Exponential-growth bias is the tendency to underestimate the effect of compounding interest (Stango and Zinman 2009, Levy and Tasoff 2016), and recent research has shown it to be associated with retirement savings (Goda et al. 2019b). Financial literacy has been shown to play an important role in various savings behaviours (e.g. Lusardi and Mitchell 2007). 

We evaluated how these measures predict six retirement contribution behaviours: following the default contribution rate, one’s annual contribution amount and annual contribution rate, making any contribution, contributing to maximise the employer match, and contributing the annual allowable maximum for tax-preferred saving. 

READ ALSO  A (very) short history of the idea of ending poverty

Different traits for different regimes

We find that under the initial opt-in regime, financial literacy is negatively associated with following at the default contribution rate, positively associated with the total amount of TSP contributions, and negatively associated with contributing the plan cap. Specifically, a one standard deviation increase in financial literacy is associated with a 30% (2.7 percentage point) higher probability of contributing the default rate, with contributing $684 (8%) more dollars annually, and with a 19% (2.5 percentage point) higher probability of contributing at the plan cap.  We find no significant associations with the long-run discount rate, present bias, or exponential growth bias under the opt-in regime.

After the change to automatic enrollment at a 3% default contribution rate, the relationship between individual characteristics and savings behaviour changes considerably.  Instead of financial literacy driving the relationship, present bias appears to do so.  We find that present bias is associated with a greater likelihood of contributing at the default rate, with more passive behaviour (i.e. making no active decision), with lower contributions, and with a lower likelihood of contributing at the plan cap.  Specifically, a one standard deviation increase in present bias is associated with a 30% (4.5 percentage point) higher probability of contributing at the default rate, with a 28% (3.3 percentage point) higher probability of passive choice, with contributing $369 (7%) less annually, and a with 9% (1.2 percentage point) lower probability of contributing at the plan cap.  We find no significant associations with the long-run discount rate, exponential growth bias, or financial literacy under the auto-enrollment regime.

A causal interpretation of the evidence suggests that there is no single individual-level trait that leads to default behaviour for retirement savings.  Rather, the features of the plan interact with individual traits to determine contribution behaviour.  Lack of financial literacy leads to default behaviour when the default is 0%, but when the default is 3%, present bias leads to default behaviour.  


These findings are consistent with a framework where individuals weigh the perceived costs and benefits of aligning with the default.  When the default is to not enroll (i.e. a 0% contribution rate), those with low financial literacy do not see the benefits of enrolling as being as high as the costs. Procrastinators who are financial literate, on the other hand, are able to overcome their procrastination due to the larger perceived benefit of switching from the default.  When the default is a 3% contribution rate, those with low financial literacy are contributing more, but procrastinators do not get around to changing from 3% given the up-front costs of switching and likely smaller perceived gain between the default rate of 3% and their optimal contribution rate.

READ ALSO  Tropical Storm Zeta To Intensify Into Hurricane Leading To More Gulf Coast Chaos

Our findings suggest that developing ways to mitigate present bias and improve financial literacy may change saving outcomes and ultimately improve welfare. Importantly, our results suggest that interventions may need to vary based on the terms of the default.


Blumenstock, J, M Callen and T Ghani (2018), “Why do defaults affect behavior? Experimental evidence from Afghanistan”, American Economic Review 108(10): 2868-2901.

Brown, J and A Previtero (2014), “Procrastination, present-biased preferences, and financial behaviors”, working paper.

Butt, A, M A Donald, F D Foster, S Thorp and G J Warren (2018), “One size fits all? Tailoring retirement plan defaults”, Journal of Economic Behavior and Organanization 145: 546–566.

Carroll, G D, J J Choi, D Laibson, B C Madrian and A Mertrick (2009), “Optimal defaults and active decisions”, Quarterly Journal of Economics 124(4): 1639–1674.

Chetty, R, J Friedman, S Leth-Petersen, T Nielsen and T Olsen (2014), “Active vs. passive decisions and crowd-out in retirement savings accounts: evidence from Denmark”, Quarterly Journal of Economics 129(3): 1141–1219.

Goda, G S, M R Levy, C F Manchester, A Sojourner and J Tasoff (2019a), ”Who is a passive saver under opt-in and auto-enrollment?”, Journal of Economic Behavior & Organization.

Goda, G S, M R Levy, C F Manchester, A Sojourner and J Tasoff (2019b), “Predicting retirement savings using survey measures of exponential-growth bias and present bias”, Economic Inquiry 57(3): 1636–1658.

Iyengar, S S and M R Lepper (2000), “When choice is demotivating: can one desire too much of a good thing?”, Journal of Personality and Social Psychology 79(6): 995.

Levy, M R and J Tasoff (2016), “Exponential growth bias and lifecycle consumption”, Journal of the European Economic Association 14(3).

Lusardi, A, O Mitchell (2007), “Financial literacy and retirement preparedness: evidence and implications for financial education”, Business Economics 42(1): 35–44.

Lusardi, A, O Mitchell (2014), “The economic importance of financial literacy: theory and evidence”, Journal of Economic Literature 52(1): 5–44.

Madrian, B C and D F Shea (2001), “The power of suggestion: inertia in 401(k) participation and savings behaviour”, Quarterly Journal of Economics 116(4): 1149–1525.

O’Donoghue, T and M Rabin, M (1999a), “Doing it now or later”, American Economic Review 89(1): 103–124.

O’Donoghue, T and M Rabin, M (1999b), “Procrastination in preparing for retirement”, in H Aaron (ed.), Behavioral Dimensions of Retirement Economics, Brookings Institution Press and Russell Sage Foundation, pp. 125–156.

O’Donoghue, T and M Rabin, M (2001), “Choice and procrastination”, Quarterly Journal of Economics 116(1): 121–160.

Stango, V and J Zinman (2009), “Exponential growth bias and household finance”, Journal of Finance 64(6): 2807–2849.

Tversky, A and D Kahneman (1975), “Judgment under uncertainty: Heuristics and biases”, in Utility, Probability, and Human Decision Making, Springer, pp. 141–162.