We study how early withdrawal policies in retirement savings balance consumption smoothing and commitment. Using Danish administrative data, we document that early withdrawals increase sharply after life events entailing financial hardship. However, using a natural experiment, we also show that individuals respond strongly to an exogenous reduction in early withdrawal penalties, unrelated to periods of financial hardship. We use these empirical patterns to discipline a quantitative life-cycle model featuring a liquid savings account and a pension account subject to early withdrawal penalties, addressing both liquidity and commitment needs. We show that policies combining penalty-free access in adverse states with positive penalties otherwise dominate uniform schemes, improving both consumption smoothing and retirement saving outcomes.

This paper develops a calibrated theory of economic decision complexity. For any decision problem, the framework delivers a measure of its complexity with no free parameters, making it highly portable across settings. We illustrate its use in central economic tasks, including dealing with taxes, forecasting, consumption, intertemporal financial planning, and the choice of financial products with opaque features.

The microfoundation builds on a macro-style production framework that treats the mind as a cognitive economy. Complexity is defined as the inverse productivity of thought: it increases with the number of importance-weighted components and decreases when outcomes hinge on a small set of dominant factors. More complex problems generate larger errors and reduced sensitivity to underlying parameters. We complement the theory with an experimental methodology for measuring subjective perceptions of complexity, and an experiment that confirms its main predictions. The measure provides a portable formula for predicting ex ante and quantifying ex post decision complexity across economic domains, potentially useful both theoretical and empirical applications.

We study disparities in retirement savings opportunities using data on 401(k) plan fees from 2011 to 2021. Despite industry-wide fee declines, we find large and persistent geographic variation in plan costs. These disparities correlate with local demographics and remain after controlling for plan and sponsor characteristics. Using ERISA lawsuits and large firm entries as quasi-exogenous shocks, we show that external legal and competitive pressures causally reduce fees. Our findings indicate that plan design reflects not only participant or firm characteristics but also frictions in local financial intermediation. Participants' asset allocations further exacerbate cost differences without offsetting performance benefits.