This dissertation develops dynamic stochastic general equilibrium (DSGE) models in which financial frictions interact with rich household heterogeneity to study the implication of financial shocks for aggregate fluctuations.

In these essays, I study macroeconomic responses to large recessions, in environments with heterogeneous agents. In the first chapter, "Long-Term Debt, Default Risk, and Policy Transmission during Severe Recessions", I study the implications of rollover risk on firm-level investment and aggregate dynamics. A growing empirical literature suggests that the maturity risk associated with long-term debt reduces firm-level investment, particularly during recessions. I introduce discretely maturing long-term debt into a dynamic stochastic general equilibrium model where heterogeneous firms borrow subject to default risk. My model is distinguished relative to existing long-term debt models in that it captures the rollover risk arising from uncertainty about what economic conditions will be when debt matures. Moreover, my firms actively save in a short-term financial asset to help hedge against the maturity risk associated with their debt. Nonetheless, the rollover risk associated with discretely maturing long-term debt exacerbates the debt overhang problem arising in conventional long-term debt models. Thus, firms effectively face greater financial frictions, and output is on average lower. Consequently, my model predicts a larger rise in defaults and a greater decline in endogenous aggregate productivity in its response to a financial shock. Thus, its financial recessions are both deeper and longer-lived than in conventional models. I also consider a large non-financial aggregate shock, and use my model to study the efficacy of targeted stimulus policies implemented over the U.S. 2020 recession. My findings suggest that the combined effects of the Paycheck Protection Program and the expansion of quantitative easing helped stem the rise in defaults and stimulate the subsequent economic recovery. The second chapter, "The Persistence of Recessions with Incomplete Markets and Time-Varying Risk" (joint with Aubhik Khan), studies the implications of precautionary savings behavior across households on aggregate responses to crises. We study the propagation of recessions in overlapping generations economies wherein households, with uncertain lifetimes and uninsurable earnings risk, face cyclical employment risk. Business cycles are driven by persistent shocks to TFP growth and household-level employment. Increases in employment risk cause fluctuations in both the unemployment rate and in labor force participation. In this setting, we introduce elements commonly used to deliver a strong and countercyclical precautionary savings motive. Specifically, households have non-separable utility characterized by high levels of risk aversion, and a diminishing marginal productivity of investment leads to a time-varying price of capital. We find that changes in precautionary savings, following aggregate shocks, have important implications for aggregate consumption. Persistent negative shocks to TFP growth, associated with increases in risk to employment, drive large declines in consumption. This helps explain the large fall in consumption observed over the Great Recession. An empirically consistent, moderate shock to TFP growth rates implies a large and persistent fall, against trend, in aggregate consumption. Moreover, an estimated rise in households' risk of long-term non-employment reduces labor force participation and reconciles the swift recovery in TFP growth rates with a protracted decline in consumption and output.

Limited asset market participation is a well-known stylized fact and a widespread phenomenoneven in developed economies. While existing models have already examinedthe effects of social security and its reforms on welfare and inequality, little attentionhas been devoted to the role of public pensions in the context of limited asset marketparticipation. I develop a quantitative overlapping generations general equilibriummodel where heterogenous agents face a financial friction limiting access to capital markets.I examine how, in presence of the market imperfection, a public pay-as-you-gosystem affects consumption and wealth inequality and compare the results with a standardmodel that does not account for limited asset market participation. In a secondexercise, I study the implications, in terms of inequality, of an increase in the retirementage in response to a population ageing shock. I find that limited asset participation isimportant for the analysis of the impact of social security on overall inequality and oninequality within age groups.