Abstract
The misidentification of monetary policy shocks typically produces the price, and the exchange rate puzzles in empirical studies. Numerous identification approaches have been introduced in the VAR literature, and these approaches have been applied to address anomalies in various countries. This study examines how well these identification approaches have worked in resolving price and exchange rate puzzles under state dominance in the financial sector in a highly dollarized and shallow financial market environment of Central Asian Economies (CAEs). First, a baseline Vector Autoregression (VAR) model with short-run restrictions is estimated using Cholesky factorization, with anomalous results observed for both prices and exchange rates. A sign-restricted VAR model is then applied with minimum restrictions, producing results consistent with conventional views on monetary policy transmission. Moreover, monetary policy shock is identified imposing the theorybased zero contemporaneous restrictions. The results show that the exchange rate puzzles disappear in three countries once identifications were modified relying on current economic theories. Also, the price and exchange rate puzzles in Kyrgyzstan and Tajikistan were fully solved, but the price puzzle in Kazakhstan was only resolved for a single monetary policy regime. Hence, puzzling behaviors appear to be the result of misidentification rather than the structural characteristics of CAEs.