Essays on inflation expectations

Abstract

“Expectations about macroeconomic variables play an important role in economic theory and policy-making. Household inflation expectations are key to understanding household consumption and investment decisions, and hence these ultimately impact the monetary policies. Yet how these expectations are formed, and how best to model this process, remains an open question (see Bernanke (2007), Bachmann et al. (2015), Coibion and Gorodnichenko (2015)).” In this thesis we undertake the analysis of quarterly household inflation expectations survey data collected by the Reserve Bank of India (RBI) to address issues related to quantification of ordinal qualitative responses, testing validity of rational expectations assumptions, modeling the formation of household inflation expectations, modeling the disagreement of household expectations, and finally, joint modeling of household inflation perception and expectation. The thesis comprises four essays. In Essay 1, we provide a method to quantify the ordinal qualitative responses on household inflation expectations. Our empirical results show that the simpler methods viz., balance statistics (BS) and Pesaran’s regression methods (PR) perform significantly better than the more complex Carlson-Parkin (CP) method. We proposed an extension of BS and PR methods with time-varying weights. The performances of the methods with time-varying weights and thresholds are found to be significantly better than the methods assuming constancy of weights and thresholds. In Essay 2, using the household expectations data we test the rational expectation hypothesis (REH) of economic agents by using a battery of statistical tests. The empirical results lead to the rejection of the REH. The inflation expectations are found to be either inconsistent or biased or both. Next, we consider a general model for mean inflation expectations formation which includes the adaptive expectations model, and the sticky information model (Mankiw and Reis (2002), Mankiw and Reis (2001), Carroll (2003)) as special cases besides others. We estimate the model using household expectations data and interpret the results. 7 Disagreement is a common phenomenon in the inflation expectations of the households in survey data (Mankiw et al., 2004). In Essay 3, we explored the links between disagreement and other aggregate macroeconomic variables. Additionally, we discussed the role of the inflation regime in disagreement. In Essay 4, we employ unit level household responses to identify the factors that determine inflation expectations. We use bivariate ordered probit model (see Sajaia (2008)) to model the household inflation perception and the household inflation expectation jointly. Our results suggest that perception is a key determinant of household expectation formation besides other idiosyncratic and macroeconomic factors. To summarize, this thesis furnishes a comprehensive empirical analysis of household inflation expectations data collected by RBI to address some important macroeconomic questions. To the best of our knowledge, there is very little evidence of empirical research being undertaken on household inflation expectations using RBI data.