We elicit the intertemporal Marginal Propensity to Consume (iMPC) based on hypothetical different-size lottery winnings through questions in the 2023-24 Italian Survey of Consumer Expectations (ISCE). Survey respondents were asked to allocate three hypothetical lottery winning amounts (€1,000, €10,000. and €50,000) between consumption and saving in both the year following the survey and over the longer term. The iMPC for a €1,000 win declines from 26% in the first year to about 1% five years after the shock. Larger win amounts have a smaller impact in the first year and a larger impact in the long run. The iMPC for a €10,000 (€50,000) prize declines from 19% (15%) in the first year to 2.5% (4%) in year five. Regardless of the size of the shock, the iMPC shows a weak negative relation to the cash-on-hand amount and a negative relation to income risk. We show that calibrated simulations of incomplete market models with borrowing constraints, income risk, and household heterogeneity are broadly consistent with these empirical findings. 

We study the transmission and distributional effects of monetary policy in an environment where consumption-saving choices reflect both precautionary motives and life-cycle considerations. Age emerges as a key state variable that links multiple dimensions of heterogeneity: young households tend to have low wealth, high marginal propensities to consume, and strongly procyclical hours. In our quantitative model that matches these facts, monetary policy operates primarily by stimulating investment, boosting labor demand for young workers who consume most of their additional income. Wealthy retirees are affected by the repricing of financial assets and persistently low future returns. However, the effect on the consumption and welfare of most retirees is small because they hold little financial wealth. 

This paper studies household consumption behavior when saving motives differ due to preference or income risk heterogeneity. I show that self-insurance operates along two margins: a standard liquidity margin determined by the level of current resources and an intertemporal margin governed by target savings. Stronger saving motives imply higher target savings, which tilt consumption toward the future. This forward-looking behavior lowers marginal propensities to consume (MPCs) independently of current resources. Canonical incomplete-markets models confound these margins because saving motives and resources co-move. I provide conditions under standard preferences and rational expectations that overturn this mapping. When saving motives and saving capacity (such as income) are negatively related, higher-income households can be wealthier yet exhibit higher MPCs because their target savings are lower. Cross-sectional MPC–resource profiles can therefore be non-monotonic due to type sorting along the resource distribution. Empirical and quantitative applications to income-risk heterogeneity support the mechanism and highlight a new perspective in which affluent households shape aggregate demand beyond standard borrowing-constraint channels.

We study the role of heterogeneity in the revenues of individual firms for euro area macroeconomic dynamics. To this end, we specify two models: a standard aggregate vector autoregressive model (VAR) and an ``heterogeneous VAR'' (HVAR). The VAR model includes only aggregate data, while the HVAR model also incorporates the feedback loop between firms' revenue distribution and aggregate variables. Our results demonstrate that the behavior of firms' revenue distribution plays a significant role in explaining the dynamics of key euro area macroeconomic variables.

We study whether firm-level heterogeneity helps explain U.S. macroeconomic fluctuations in response to aggregate shocks. Using quarterly Compustat and CRSP data from 1986 to 2025, we construct two revenue-based statistics inspired by the  Melitz (2003) model: the average firm and the marginal near-default firm. These statistics summarize key features of the firm distribution. We augment a Bayesian VAR with these measures and compare its performance to a standard aggregate VAR and to a functional VAR that incorporates the full cross-sectional distribution of firm revenues.  We find that firm-level heterogeneity contains information not captured by aggregate variables. Including the two statistics allows the VAR to closely replicate the impulse responses obtained using the functional VAR and improves out-of-sample forecast accuracy. These findings are robust to a replication using UK data.