Job Market Paper

Estimating the Elasticity of Intertemporal Substitution with Disaggregated Consumption Data

This paper estimates the elasticity of intertemporal substitution (EIS) of consumption using the Nielsen Consumer Panel. The Nielsen Consumer Panel is built from transactional data that follows households in the United States and their grocery purchases from 2004 to 2014. Because of the transactional nature of the dataset, there is a low source of measurement error in consumption, and aggregation bias can be minimized. Due to changes in the economy during this timeframe, the data is examined for structural breaks. The data suggests evidence for two structural changes in the U.S. economy leading to three regimes. The first regime, 2004 to 2006, was a period of economic expansion, while the second regime, 2007 and 2008, was a period of recession. Lastly, during the third regime, 2009 to 2014, the economy exhibited quantitative easing. The EIS is estimated for each regime using linearized Epstein-Zin preferences and by the use of fixed effects and instrumental variables. The data exhibits no evidence of weak instruments. In order to estimate the EIS, consumption is aggregated weekly, and consumption growth is measured over a four-week time period in order to match four-week Treasury bills. This study adds to the literature by examining individual EIS during different periods of economic activity. With a more complete dataset that has less measurement error and aggregation bias than the existing literature, this study gives evidence of a small and negative EIS during a period of expansion, a small and positive EIS during a period of recession, and a large and positive EIS during quantitative easing.

Work in Progress

Estimating the Quantile Elasticity of Intertemporal Substitution with Instrumental Variables Quantile Regression

This paper estimates the quantile elasticity of intertemporal substitution (QEIS) of consumption using instrumental variables quantile regression. In this paper, agents have a quantile utility preference instead of standard expected utility. This allows for the capture of heterogeneity along the conditional distribution of agents. The estimator is a feasible estimator based on smoothed sample moments. In order to minimize measurement error and aggregation bias, the QEIS is estimated using the Nielsen Consumer Panel, which is built from a transactional dataset. Early results show heterogeneity of the QEIS along the quantiles of the conditional distribution.