The Perils of Using Aggregate Data in Real Exchange Rate Estimations
28 Pages Posted: 10 Dec 2016
Date Written: September 1, 2016
The rate of mean-reversion in a country's real exchange rate (RER) is a key indicator to consider when discussing exchange rate policies in any country. In practice, this rate - known as the half-life - is commonly calculated using price aggregates, such as the consumer price index (CPI). I demonstrate that using a CPI to estimate the RER mostly yields higher persistence estimates than a RER based on the disaggregated formula suggested by PPP theory. Using a novel dataset with a daily frequency of price collection and an identical set of products from multiple countries, I find that the half-life for a RER derived from price aggregates is on average 37% higher than a RER generated from product-level information. Until now estimates suggested that the rate of mean-reversion was unreasonably slow, ranging from a minimum of 2 to 5 years. The sample used in this paper indicates that this rate is in fact typically less than 1 year. The dataset has been gathered from online sources and includes food, fuel, and electronic products.
Keywords: purchasing power parity, real exchange rate, mean-reversion, half-life, impulse response, near-integrated AR process, local-to-unity asymptotics, micro-level data
JEL Classification: E31, F31, F41, C22
Suggested Citation: Suggested Citation