RDP 2025-04: HANK and the Transmission of Shocks to Demand and Supply 2. Literature

The HANK literature has rapidly evolved in the past several years. This literature builds off of several decades of earlier work which weakened the complete markets hypothesis to consider a variety of questions in environments with heterogeneous outcomes. Early papers in this area, such as Bewley (1977), Aiyagari (1994) and Krusell and Smith (1998), laid the computational groundwork for solving them and studied foundational properties. Later work built on the computational tools (such as Reiter (2009)), and looked at a variety of applications (such as Heathcote (2005)). As computational methods improved, attention turned to the transmission of aggregate shocks in these models, including those due to monetary policy.

An influential piece of work which was one of the first to add New Keynesian price rigidities in a heterogeneous agent model was Kaplan, Moll and Violante (2018). Arguing that the proportion of households who are wealthy but hold exclusively illiquid assets is significant, the authors show that these wealthy hand-to-mouth households have a consumption response to interest rate changes that primarily arises from the impact of the change on their wage income (a general equilibrium effect). Simulating the HANK model by way of an MIT shock, they find that this effect outweighs the traditional intertemporal substitution effect in the model.

Prior to this work, Guvenen et al (2015) examined high quality administrative tax data from the Internal Revenue Service of the United States. Among other things, this work estimated several moments of the distribution of income changes. Kaplan et al (2018) then used these moments to calibrate the process for income in their model.

Alongside this work, Ahn et al (2018) introduced a continuous time variant of the method in Reiter (2009) for solving heterogeneous agent models. This method solves nonlinearly for the stationary equilibrium of the model, linearises the dynamic equilibrium conditions around the steady state values, and then solves the linear system. Ahn et al show that this method can solve baseline heterogeneous agent models much more efficiently than older methods, when the number of individual state variables is relatively small. As noted in the introduction, this is the solution method applied in this paper.

More recent developments in the broader HANK literature have explored the implications of heterogeneity for the exchange rate channel of monetary policy in small open economies (e.g. Auclert et al 2021) and the role of financial frictions and credit constraints in the HANK framework (e.g. Faccini et al 2024). The former paper finds that heterogeneity amplifies the extent to which a lower exchange rate reduces household real incomes and hence spending. This results in the potential for a currency depreciation to be contractionary, making expansionary monetary policy less expansionary. The latter paper finds that elevated credit spreads reduce consumption of indebted households in Danish data, resulting in a countercyclical propensity to consume. The authors calibrate a HANK model with financial frictions to match these observations, and conclude that regulatory measures may be stabilising at the aggregate level but volatility inducing at the household level. These are just two examples of questions being examined in a rapidly expanding body of literature.

As well, this paper is related to other papers exploring the differences between shock transmission in heterogeneous agent models and representative agent models. Kaplan and Violante (2018) show that responses to unforeseen shocks may differ or be similar between HANK and RANK, depending on the nature of the shock. They show that in response to a discount factor shock, the models have similar responses with similar underlying mechanisms, while in response to a technology shock they have similar aggregate responses but with different underlying mechanisms. They also show that monetary policy shocks generate different aggregate responses in HANK versus RANK. Finally, Kaplan and Violante highlight the fact that certain macroeconomic questions can only be addressed with heterogeneous agents, such as those on how a given shock affects individuals across the distribution.

Auclert, Rognlie and Straub (2020) study the transmission of monetary policy and sources of business cycle fluctuations in a HANK model. They find that investment plays a larger role in HANK relative to RANK. In particular, they find that if investment could not adjust in response to a policy shock, then the cumulative output response would be only 20 per cent of that when investment does adjust, in contrast to 60 per cent in RANK. They also find that matching a realistic distribution of marginal propensity to consumes (MPCs) in the model allows for procyclical co-movement of consumption and investment in response to an investment shock, offering a resolution of a puzzle dating back to Barro and King (1984), wherein RANK models require total productivity shocks for procyclicality.

While most of the HANK literature finds that household heterogeneity amplifies the effects of monetary policy, Cantore et al (2023) find a dampening effect on policy through the household labour supply channel. They provide some empirical evidence that following a contractionary monetary policy shock, lower income households tend to increase their hours worked. Building this added worker effect into a model they show that it can dampen the effects of monetary policy.

Guo et al (2023) measure the extent to which consumers face liquidity, savings, or borrowing constraints in data from 20 European countries. They consider whether the income and wealth distributions influence the effectiveness of tax and spending polices. The authors find that tax multipliers are amplified by a higher percentage of wealthy hand-to-mouth households, while spending multipliers are amplified by a higher percentage of poor hand-to-mouth households. They also find that a high proportion of wealthy hand-to-mouth households enhances monetary policy effectiveness, but a high proportion of households that are constrained in terms of liquidity, savings, and borrowings weakens it.

Overall, the literature on HANK models and supply and demand shocks highlights the importance of heterogeneity in understanding macroeconomic outcomes. In the existing body of research, HANK models have been employed extensively to analyse various economies around the globe. These models, with their inherent ability to capture economic realities more accurately due to their consideration of individual heterogeneity, have significantly contributed to our understanding of diverse economic phenomena.

Despite HANK models being adopted by academics and policy institutions internationally, there's a dearth of studies focusing on the Australian economy. This paper seeks to fill this gap in the literature by calibrating a HANK model to Australia and examining the implications for shock transmission relative to a representative agent framework.