구매력 평가설과 이탈요인에 관한 소고 : Studies on purchasing power parity and factor of deviation
- 발행기관 서강대학교 대학원
- 지도교수 정재식
- 발행년도 2006
- 학위수여년월 200608
- 학위명 박사
- 학과 및 전공 경제
- 식별자(기타) 000000103188
- 본문언어 한국어
초록/요약
Most general equilibrium models of open economies impose the Purchasing Power Parity(PPP) as a long run equilibrium condition. However, the literature has casted doubt on its empirical validity. The main goal of this paper is to examine the hypothesis of long run purchasing power parity among 19 OECD countries and 10 Asian countries under the current float using the non-stationary panel approach and semi-parametric approach. In non-stationary panel approach, I apply two newly developed panel unit root tests that explicitly deals with the numeraire effect that causes cross sectional dependence. We find robust evidence that supports the PPP hypothesis, In most cases, unit root tests reject the possibility of a non-stationarity in real exchange rates. Especially, the price index plays key role in determining whether purchasing power parity holds or not. The real exchange rates measured by wholesale price index ratio can reject unit root hypothesis with test procedure which is considered cross sectional dependence, but real exchange rate based upon the ratio of consumer price index cannot reject unit root hypothesis. Secondly, panel unit root test which is considered cross sectional dependence give us more clear economic implication than panel unit root test which is not considered cross sectional dependence. The latter finding is dependent on the combination of both price index and base currency, but the former is not. When panel is classified according to common factor of deviation from mean reversion level, the result undergoes significantly various influence such as the exchange rate arrangements, its volatility, inflation and GDP growth rate per captia. but is not influenced by inflation differential and the reliance on trade. The other strand approach which is to assess the validity of Purchasing Power Parity, I adopt two semi-parametric analysis tools to estimate fractional differencing parameter using Fourier Analysis and Wavelet Analysis. In the Fourier analysis, I find that we get the inconsistent results which are varied the combination price index and base currency. But the wavelet analysis gives us more consistent results than Fourier analysis. Because, we can find 3 real exchange rates are stationary, and 12 real exchange rates from 26 sample have long memory. These finding is virtually invariant to the use of price index and base currency. In addition to these results, some real exchange rates are influenced by the use of either price index or base currency. six countries affected by use of price index, three countries affected by use of base currency. On the basis of our findings we conclude that there is substantive evidence for purchasing power parity although not to the same extent as for every currency.
more