Thai Inflation Dynamics: A View from Micro CPI Data
Abstract
This paper examines the patterns of price adjustment at the micro level in order to further our understanding of price rigidity at the aggregate level. We highlight 5 stylized facts: 1) Prices change infrequently with a mean duration of approximately 4 to 7 months between price changes; 2) Price decreases are common accounting for roughly 45 percent of all price changes; 3) Price changes, both increases and decreases, are sizable compared to the prevailing in ation rate; 4) The size of price changes covaries strongly with the rate of in ation, whereas the fraction of items changing prices does not; and 5) There is signicant dispersion in price levels as well as in the synchronicity of price changes across geographical regions. Based on a dynamic factor model, we also utilize prices at the disaggregated level to perform an in ation decomposition to understand the underlying driving factors of in ation. The key ndings are: 1) Prices at the micro level are driven mainly by idiosyncratic shocks but these shocks become less important for CPI in ation at the aggregate level; 2) Pure in ation which drives long-term price movements in Thailand is responsible for approximately 10 percent of overall price movements; 3) More than half of all within-quarter uctuations can be classied as relative price changes in response to aggregate shocks; 4) The short-run in ation-output tradeoff which appears weak in aggregate data becomes much stronger once volatile idiosyncratic price changes are removed.