A0432
Title: Responses of households' expected inflation to oil prices and the exchange rate: Evidence from daily data
Authors: Etsuro Shioji - Chuo University (Japan) [presenting]
Abstract: The purpose is to examine if daily data can help predict monthly changes in the inflation expectations of households. The mixed-data sampling (MIDAS) method is utilized to study determinants of a monthly survey-based measure of household inflation expectations in Japan. Two types of higher-than-monthly frequency variables are incorporated into the analysis. The first is a group of "traditional" daily indicators, such as the exchange rate and crude oil prices. The second consists of more unconventional measures. They include a daily price index called CPINOW, which is constructed from scanner data recorded at supermarkets from all over the country and (weekly) retail gasoline prices. It is found that, although both groups of variables help explain actual CPI inflation, only the latter turns out to be significant when the dependent variable is expected inflation. This finding suggests that people's perceptions are affected predominantly by prices that they actually observe in their everyday lives at supermarkets and gasoline stations.
