Beginning in 1995 and then more broadly in 2000, the Social Security Administration (SSA) began mailing hard copy statements of estimated benefits to workers (the “Statements”; see the the SSA website). Included in the Statements were available disability insurance (DI) benefits and estimated monthly benefits at certain retirement ages, along with other information. The SSA stopped sending the Statements in 2011 for budget reasons but restarted the program in 2014.
How did the Statements affect worker behavior?
Dr. Philip Armour of the RAND Corporation discussed the findings of his research on this topic in a webinar yesterday, examining how older workers’ (ages 50-64) behaviors were influenced by the Statements delivered between 1995 and 2011. Dr. Armour’s findings are summarized below:
DI applications increased: The DI application rate for older workers with significant health issues doubled. Note that financial information regarding DI was an actual available benefit, versus a projection.
Employment effect: Older workers began working LESS after receiving the Statement (about 110 hours less per year on average). The post-Statement effect was influenced by number of hours worked per week:
Those working under 10 hours per week had an average INCREASE in hours worked per year: 470 more hours worked after receiving the Statement on average.
Those working 40 hours per week or more had an average DECREASE in hours worked per year after receiving the Statement: 500 hours per year on average.
Projected benefits at certain retirement ages (e.g., 62, 67, 70) were just that: projections of future monthly benefits presuming continued work, as opposed to benefit amounts already accrued based on past work. The study demonstrated an apparent misunderstanding regarding this fact.
Are there implications for companies who provide financial information to employees or clients?
Information regarding estimated future income and current benefits available can have an impact on behavior. Therefore, large-scale, personalized, and financially-related communications should be scrutinized prior to delivery. This is true for financial institutions who create (or change) personalized reports AND companies who allow the information to be delivered to employees (e.g., in 401K statements). The impact on hours worked found in this study is significant.
There were some income and wealth effects found in Dr. Armour’s study, as well (and as expected): higher income, higher net worth individuals demonstrated a smaller decrease in hours worked in general, but there are many outstanding questions, too (do the statements affect savings behavior? How do other characteristics, including age, impact the findings?) The “restart” of the program may influence workers, once again, and will allow for more opportunities to examine the impact.