Two recent papers investigate the effects of aggregate uncertainty in Heterogeneous Agent New Keynesian (HANK) models, addressing a significant technical challenge in the literature. They share the core insight that household heterogeneity meaningfully interacts with aggregate uncertainty to amplify macroeconomic fluctuations.
Both papers move beyond standard linearization techniques that typically ignore or inadequately capture uncertainty effects in HANK models. Each proposes alternative approaches that allow aggregate uncertainty to have meaningful impacts on model dynamics.
Together, they demonstrate that aggregate uncertainty interacts powerfully with household heterogeneity through multiple channels – precautionary savings, portfolio substitution, and nominal rigidities – to shape macroeconomic outcomes.
Aggregate uncertainty and portfolio choice
In this paper by Cosmin Ilut, Ralph Luetticke, and Martin Schneider explore the impact of aggregate uncertainty in a two-asset HANK model with portfolio frictions. The authors model aggregate uncertainty as ambiguity using multiple priors preferences, allowing for first-order effects on utility and facilitating easier computation and estimation. They then propose an iterative estimation procedure that matches long-run moments and estimates parameters from linearized dynamics. Ilut et al. use six observables and five shock series, accounting for measurement error.
Findings:
- Impact of aggregate uncertainty: Aggregate uncertainty shocks are found to be a powerful driver of the business cycle, explaining more than 70% of cyclical variation in key macroeconomic aggregates and asset prices. This effect is stronger than in representative agent models due to portfolio substitution by rich capital owners.
- Portfolio substitution: Rich households substitute away from capital towards bonds during uncertainty shocks, while poor households save more in bonds for precautionary reasons. This leads to a decline in investment and an increase in the capital premium.
- Cautious price and wage setting: Firms and unions respond to higher uncertainty by increasing prices and wages, mitigating deflationary pressures during recessions.
- Comparison with other shocks: Aggregate uncertainty shocks have a more significant impact on business cycles compared to idiosyncratic uncertainty, TFP, and nominal shocks. The latter are more relevant for inflation and nominal interest rates.
HANK’s Response to Aggregate Uncertainty in an Estimated Business Cycle Model
Authors: Cosmin Ilut, Ralph Luetticke, Martin Schneider
From: Duke University, Universität Tübingen, Stanford University
Aggregate uncertainty and the Zero Lower Bound (ZLB)
Alessandro Lin and Marcel Peruffo, in this paper, propose a novel methodology for solving HANK models that incorporate aggregate uncertainty and the Zero Lower Bound (ZLB) on nominal interest rates. The authors combine the sequence-state Jacobian methodology from Auclert et al. (2021) with a tractable two-regimes shock structure for aggregate uncertainty. This approach allows for efficient simulation of complex household heterogeneity, aggregate occasionally binding constraints, and aggregate risk. Aggregate uncertainty is modeled using a simple two-states Markov process. This structure confines uncertainty to a finite number of paths, making the model tractable despite the curse of dimensionality. The model is calibrated to U.S. macro and micro data, focusing on output loss and inflation during the Great Recession.
Findings:
- Amplification of downturns: Aggregate uncertainty significantly amplifies economic downturns when the economy is at the ZLB. This amplification leads to a substantial increase in output loss compared to scenarios without uncertainty, more than doubling the negative impact.
- Household heterogeneity: The presence of household heterogeneity further strengthens the amplification effect of aggregate uncertainty. In a Representative Agent economy, the output loss is about 60% higher under uncertainty compared to a deterministic scenario, but this effect is even more pronounced in a HANK model.
- Forward guidance: The authors find that forward guidance on interest rates is more effective in a HANK model with aggregate uncertainty than in its deterministic counterpart. This suggests that forward guidance can mitigate the strong interactions between aggregate uncertainty and household heterogeneity.
Aggregate Uncertainty, Hank, and the ZLB
Authors: Alessandro Lin, Marcel Peruffo
From: Bank of Italy, The University of Sydney
