Household Uncertainty, Risk Attitude and Implications for Macro Economy

Informal interest in the consumption savings decision under uncertainty dates as back as to writings of Alfred Marshall. The 2007-2008 Financial Crisis brought about a renewed focus on the spending and savings decision of household and the quantitative strength of the precautionary saving channel in response to time varying uncertainty.

This dissertation focuses on direction and strength of households precautionary saving motive and its implications for modelling household preferences and disciplining macroeconomic models. It tests empirically for precautionary saving motive and combines these estimates with simulation-based distance minimizing methods to jointly estimate two key structural parameters for recursive preferences- coefficient of relative risk aversion (RRA) and intertemporal elasticity of substitution (IES). RRA turns out to be small suggesting mild precautionary motive among U.S. households. Further, loosening the link between households preferences across time and risk seem more data consistent approach to model household preferences. The estimated structural parameters provide micro-data consistent estimates for calibrating macroeconomic models and shed light on a variety of macroeconomic phenomena like the precautionary saving channel, the inter-temporal channel and the propagation and amplification of aggregate uncertainty shocks in DSGE models and the modelling of household preferences. Overall, precautionary motive does not seem to be a possible channel for propagation and amplification of aggregate uncertainty shocks.

Chapter 1 focuses on investigating the response of household’s planned spending to time varying self-reported wage uncertainty and tests for the precautionary saving motive among U.S. households in growth rates. It uses novel self-reported density forecast of household ex-ante wage growth, expected inflation and point prediction of ex-ante household spending growth in a region-time fixed effects regression framework. The smaller estimates on wage growth uncertainty and modest estimates on expected inflation imply that households spending is more sensitive to interest rates than it is to uncertainty. Overall, the precautionary saving motive among U.S. households seems weak.

Chapter 2 uses the estimated coefficients from chapter 1 and combines it with Indirect inference, a simulation-based technique to jointly estimate structural parameters RRA and IES for recursive preferences. It also studies the implications of disciplining DSGE models of aggregate uncertainty with microdata consistent estimates of RRA and IES. The estimates of RRA and IES suggest that using CRRA preferences where one parameter is inverse of the other can be restrictive. Further, smaller estimates of risk aversion casts doubt on the ability of precautionary saving channel to propagate and amplify aggregate uncertainty shocks in a way that is consistent with the observed co-movements during the Financial Crisis of 2007-2008.

Chapter 3 uses dispersion in consumer confidence as an empirical measure of time varying household uncertainty and studies its consequences for the macroeconomy. It uses cross-sectional dispersion in consumer confidence over time as a measure of consumer confidence uncertainty or 'disagreement' among consumers. Consumer confidence has prolonged expansionary effects on economic activity even when accounting for effects of dispersion in consumer confidence. These responses are robust to ordering in the structural VAR. The quantitative responses of consumption and output to consumer confidence shocks are higher than the shocks to dispersion in consumer confidence.