The Marginal Propensity to Consume in Heterogeneous Agent Models

Abstract

What model features and calibration strategies yield a large average marginal propensity to consume (MPC) in heterogeneous agent models? Through a systematic investigation of models with different preferences, dimensions of ex-ante heterogeneity, income processes, and asset structures, we show that the most important factor is the share and type of hand-to-mouth households. One-asset models either feature a trade-off between a high average MPC and a realistic level of aggregate wealth or generate an excessively polarized wealth distribution that vastly understates the wealth held by households in the middle of the distribution. Two-asset models that include both liquid and illiquid assets can resolve this tension with a large enough gap between liquid and illiquid returns. We discuss how such return differential can be justified from the perspective of theory and data.

Heterogeneous Agent Trade

“…The scaling here is done (again) to deliver a balanced-growth-like property of the model so a households debt capacity is invariant to a country's autarky level of income. The precise number of 50 percent seems reasonable to me, but as a check, I show that the marginal propensities to consume and how they vary across households are consistent with the evidence of Kaplan and Violante (2022).…”

“…This procedure works well, and thus the model is able to match spatial distribution of economic activity in the data just as well as standard, constant elasticity gravity models. In addition, I ensure that the model replicates salient facts regarding household-level expenditure patterns (Borusyak and Jaravel (2021)), elasticities (Auer et al (2022)), and marginal propensities to consume (Kaplan and Violante (2022)).…”

Did marginal propensities to consume change with the housing boom and bust?

“…Our findings of higher MPCs, especially for low liquidity homeowners, are immediately suggestive of the deterioration in housing wealth combined with liquidity constraints making consumption more sensitive to transitory income shocks following the housing bust. A question, though, is how much these results actually reflect a movement along household consumption functions due a change in wealth distributions versus being driven by a shift in consumption functions for other reasons (see Kaplan & Violante, 2022). We argue that the change in MPCs does not actually appear to be related to a change in wealth distributions but rather is consistent with tighter borrowing constraints given less credit availability for low liquidity homeowners.…”

How Do Households Respond to Income Shocks?

“…In this literature, Hall and Mishkin (1982) and Altonji and Siow (1987) represent seminal early contributions, and the initial body of work is discussed comprehensively in Deaton (1992). Jappelli and Pistaferri (2010a) and Jappelli and Pistaferri (2017) are comprehensive surveys of the subsequent empirical literature, and Kaplan and Violante (2022) summarize the corresponding predictions of standard incomplete markets models concerning this question. How strongly consumption responds to income shocks of a given persistence is the central question of this literature.…”