| Summary: | 碩士 === 國立東華大學 === 經濟學系 === 98 === This paper runs the monthly data from 1985:1 to 2009:7, and try to understand the behavior of the disaggregate Korean real exchange rates (RERs). The four different disaggregate price levels are the decompositions of consumer price index (CPI) from sub-price index level to product level. In order to detect the possible structural changes in the data series, the LM unit-root test with level shifts (0-2 breaks) are also used. Our empirical results show that most series accept the unit roots hypotheses by univariate unit root tests or by univariate LM unit root tests with breaks. It implies that the Korean
disaggregate RERs might be persistent if they are examined by traditional unit root methodologies. However, we found all of the disaggregate RERs are stationary processes and decrease half-life estimate successfully by panel unit
root tests without structural breaks. Comparison with the panel LM unit root test with structural breaks under correction aggregation bias, only part of the disaggregation RERs can be rejected unit root at least 10% significant level, and the convergence rate is similar. As our panel LM unit root test with structural breaks is calculated by individual LM test of t-test statistics
and AR(1) coefficient and then total average. Thereby, the measurement is the individual persistence rather than persistence of sectoral. Consequently, the influence on aggregation bias becomes unobvious. Although the half-lives
for all Korean disaggregate RERs are slowly, but still less than the consensus views of 3-5 years (Rogoff, 1996). In conclusion, the most disaggregation price
level data provide the strongest evidence to support the PPP between Korea and Japan.
|
|---|
