Marta Cota
Review of Finance, Volume 29, Issue 4, July 2025, Pages 1105–1136, https://doi.org/10.1093/rof/rfaf021
Biased income expectations shape retirement savings in ways that traditional policy models often overlook. In this paper, I outline a mechanism under which low-income workers, typically pessimistic about future earnings, delay retirement contributions, prioritizing liquidity as a buffer against uncertainty. High-income workers, conversely, overestimate income growth and postpone saving under the assumption of future financial flexibility. These expectation-driven behaviors contribute to persistent disparities in retirement preparedness.
Using data from the Michigan Survey of Consumers (MSC), I establish three key empirical facts: (1) Income forecast errors vary systematically by income level—low earners tend to be overly pessimistic, while high earners are persistently optimistic. (2) Workers adjust expectations based on recent earnings shocks, reinforcing extrapolative beliefs. (3) All income groups overestimate future income uncertainty, strengthening precautionary savings motives. These biases reinforce saving in liquid accounts at the expense of illiquid retirement savings plans.
A lifecycle model integrating these biases with 401(k) plan structures reveals that expectation distortions can explain observed delays in retirement savings, particularly among low-income workers. Policy simulations show that automatic enrollment increases retirement savings by 4.8% on average, with a 10% gain for the lowest-income quartile. However, contributions under auto-enrollment often decline over time relative to active enrollment.
These findings suggest that dynamic policy mechanisms—such as auto-escalation—are necessary to address expectation-driven distortions. By embedding behavioral insights into retirement policy design, we can foster greater financial security and reduce long-term retirement inequality.