Research

Job Market Paper

• Constrained-Optimal Social Security Reform with a Debt Ceiling and Learning by Doing

  Abstract

  Once the Social Security trust fund is depleted, all bets are off. The most likely scenario is what I call the baseline or no-reform case: an initial transition of increasing government transfers to finance the Social Security deficit, which would increase national debt up to a ceiling, followed by a combination of higher payroll taxes and lower benefits. Can a different adjustment path mitigate this burden without making any generation worse off on average than under no reform? I study this transition in a general-equilibrium overlapping-generations model calibrated to the United States in 2010, with learning by doing and a public-debt ceiling of 150% of GDP. The baseline economy finances scheduled benefits through borrowing until the debt ceiling binds and benefit reductions follow. I construct a reform that lowers the payroll tax to 3.5% by 2200 and reduces benefits more relative to baseline, without average welfare losses for any birth cohort. Cohort-specific payroll-tax relief during working life and benefit supplements for low-income individuals late in retirement protect transition cohorts. A consumption tax finances both under the same debt ceiling, at a present-value cost of 2.2% of 2026 GDP. Learning by doing is essential to this result: without it, the gains from lower payroll taxation are insufficient to finance the cohort-specific payroll tax relief and benefit supplements. Complete phase-out produces larger long-run output gains but imposes welfare losses on many cohorts alive today. Retaining a smaller program, as implied by my reform, yields higher social welfare than phase-out when the government counts only cohorts alive today or weights future cohorts lightly through dynastic altruism.

Working Papers

• The Intergenerational Welfare Cost of Social Security Insolvency

  Abstract

  This paper asks which generations bear the welfare cost of Social Security insolvency. The trust fund is expected to be depleted in 2033, but each generation meets it at a different age, with different years of benefits at stake and different means of adjusting. In a life-cycle model with endogenous saving, labor supply, human capital, and claiming, I quantify how the expected duration of reduced payments shifts that welfare incidence. The main result is that the cohort losing the most unpaid dollars is not the cohort losing the most welfare. Households that expect the shortfall self-insure by saving, working, and human-capital investment, but the cost of doing so depends on age. The youngest cohorts spread the adjustment over a career, whereas mid-career cohorts must compress it into the years before retirement and end up worse off than if the shortfall had been a surprise. Duration matters as well. When the shortfall is brief, the burden lands mainly on retirees, and when it is prolonged, the burden shifts to the working generations behind them.

  Paper (PDF)

• Fertility Inertia under Ambiguity Aversion

  Abstract

  High family-policy spending often produces only small changes in fertility. I study this problem in a Barro–Becker model. Parents value descendants’ utility, dislike risk in its total value, and assess family plans by their worst-case value. They are uncertain about how well public news predicts children’s outcomes. Information inertia creates an interval of bad news in which welfare falls but locally optimal fertility, investment per child, and consumption remain unchanged. Reliable subsidies can still raise fertility. An explicit policy solution discounts uncertain benefits by the lowest possible delivery rate. Thus, unclear delivery requires larger promises for the same fertility response. More effective support raises fertility and shifts and widens the inertia interval. Less uncertainty about signal informativeness narrows it for a fixed family plan. This suggests two policy dimensions: generosity and legibility. The exact inertia result is limited to maxmin preferences, continuous fertility, and a local normal representation.

  Paper (PDF)

Work in Progress

• Optimal Immigration Policy under Private Information: A Macroeconomic Approach

• Insurance and Learning about Self-Protection: Evidence from Conservation Reserve Program (with Hyewon Lee)