Job Market Paper

Estimating Nonlinear Heterogeneous Agents Models with Neural Networks

Job Market Paper
Economists typically make simplifying assumptions to make the solution and estimation of their highly complex models feasible. These simplifications include approximating the true nonlinear dynamics of the model, disregarding aggregate uncertainty or assuming that all agents are identical. While relaxing these assumptions is well-known to give rise to complicated curse-of-dimensionality problems, it is often unclear how seriously these simplifications distort the dynamics and predictions of the model. We leverage the recent advancements in machine learning to develop a solution and estimation method based on neural networks that does not require these strong assumptions. We apply our method to a nonlinear Heterogeneous Agents New Keynesian (HANK) model with a zero lower bound (ZLB) constraint for the nominal interest rate to show that the method is much more efficient than existing global solution methods and that the estimation converges to the true parameter values. Further, this application sheds light on how effectively our method is capable to simultaneously deal with a large number of state variables and parameters, nonlinear dynamics, heterogeneity as well as aggregate uncertainty.
    Presented at
  • The Society for Nonlinear Dynamics and Econometrics (SNDE) Symposium (March 2022)
  • The Society for Economic Measurement (SEM) Annual Conference (August 2022)
  • The European Economic Association (EEA) and European meeting of the Econometric Society (ESEM) Conference (August 2022)
  • Conference on Non-traditional Data, Machine Learning and Natural Language Processing in Macroeconomics at Sveriges Riksbank (October 2022)
  • Midwest Macro Meeting (November 2022)
  • ASSA (January 2023)

Work in Progress

Nonlinear Phillips Curve and Inflation Risk

Work in Progress
How does a nonlinear Phillips curve affect inflation risk? Using a strategic surveys approach and micro price data, we establish that the price setting behaviour of firms depends nonlinearly on the inflation environment. In a high inflation environment, the share of firms that adjust their prices in response to expected inflation increases. We rationalize these dynamics using a quantitative macroeconomic model with a nonlinear Phillips curve. The model features a tractable heterogeneous firm setup with endogenous varying degrees of price flexibility. Solving the model with a machine learning approach, we demonstrate that, in this setting, contractionary supply shocks lead to higher inflation, which provides a new motive for the monetary policy to act preemptively.

Backpropagating Through Heterogeneous Agent Models

Work in Progress
This paper explores applications of the backpropagation algorithm on heterogeneous agent models. In addition, I clarify the connection between deep learning and dynamic structural models by showing how a standard value function iteration algorithm can be viewed as a recurrent convolutional neural network. As a result, many advances in the field of machine learning can carry over to economics. This in turn makes the solution and estimation of more complex models feasible.

Limits on Mortgage Lending

Work in Progress
This paper aims to study the impact of macroprudential limits on mortgage lending in a heterogeneous agent life-cycle model with incomplete markets, long-term mortgages, and defaults. Using data from the Household Finance and Consumption Survey, the model is calibrated for the German economy. I consider the effects of four policy instruments: loan-to-value limit, debt-to-income limit, payment-to-income limit, and maximum maturity. I find that their effect on the homeownership rate is fairly modest. Only the loan-to-value limit significantly reduces the homeownership rate among young households. At the same time, it has the most significant positive welfare effect
No matching items