Dinâmica de convergência dos preços relativos para cidades brasileiras

Abstract

The three essays of this thesis investigate the hypothesis of Purchasing Power Parity. More specifically, three empirical exercises related to this topic are carried out. The PPP puzzle is named as one of six major puzzles in international macroeconomics and therefore the first two articles concentrate on the analysis of possible solutions for it. The first paper analyzes the effect of structural breaks on the speed of convergence in relative prices of Brazilian cities from 1991.01 to 2016.11. Three structural breaks dates are estimated for all cities: 1996.02, 2001.12 and 2010.10. These dates represents different situations in the Brazilian economy, with likely impacts on the intranational relative prices. The speed of convergence, measure by half-life, diminish approximately 77% after the introduction of structural breaks. We avoid the presence of others bias that can affect half-life estimates as Nickell and temporal aggregation bias. This evidence indicates the relevance of instabilities to compute the purchasing power parity in an intranational context as pointed out by others papers. The result is robust to the compute of half-life and estimate of structural breaks using different numeraires. IMRR (2005a) point out to the aggregation bias of price with heterogeneous adjustment dynamics as a solution to the puzzle. Thus, the second one investigates whether this result is supported for Brazilian cities data using a dynamic CCEMG estimator with half-panel jackknife bias correction to mitigate the small sample time series bias proposed by Chudik and Pesaran (2015). The results corroborate the IMRR’s (2005a) evidence, with a median half-life of 13.5 months. Finally, the third article analysis the impact of transport costs on relative price volatility for Brazilian cities. To do that, use time-varying volatility measures, which are different from those used so far. Additionally, it proposes a new proxy for transport costs. The estimations in dynamic panel and an ARDL process in a heterogeneous panel augmented of cross-section averages are not available robust evidence of the impacts of transport costs on relative price volatility.