Working Paper: CEPR ID: DP15346
Authors: Lilia Maliar; Serguei Maliar
Abstract: We introduce a deep learning classification (DLC) method for analyzing equilibrium in discrete-continuouschoice dynamic models. As an illustration, we apply the DLC method to solve a version ofKrusell and Smith's (1998) heterogeneous-agent model with incomplete markets, borrowing constraintand indivisible labor choice. The novel feature of our analysis is that we construct discontinuous decisionfunctions that tell us when the agent switches from one employment state to another, conditional on theeconomy's state. We use deep learning not only to characterize the discrete indivisible choice but alsoto perform model reduction and to deal with multicollinearity. Our TensorFlow-based implementationof DLC is tractable in models with thousands of state variables.
Keywords: Deep Learning; Neural Network; Logistic Regression; Classification; Discrete Choice; Indivisible Labor; Intensive and Extensive Margins
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| indivisible labor (J59) | labor market dynamics (J29) |
| indivisible labor (J59) | volatility of labor relative to output (J49) |
| indivisible labor (J59) | correlation between labor and wages (J31) |
| indivisible labor (J59) | income and wealth inequalities (D31) |
| indivisible labor (J59) | sensitivity of income and wealth inequalities to variations in the coefficient of risk aversion (D31) |