Seminar 6: Testing Purchasing Power Parity (PPP) between India and the US; 1973-2005
Part I: Univariate analysis of non-stationary processes.
The main objective of this seminar is to conduct some basic univariate
写留学生论文analysis of non-stationary processes in Eviews. This analysis involves testing for unit roots using:
• Augmented Dickey Fuller (ADF) tests.
• Phillips-Perron (P-P) tests.
• KPSS tests for stationarity.
This analysis is applied in the context of testing the PPP relationship between India (the home country) and the US. The data for this seminar are in ppp_ind_us.wf1. This file contains monthly data for the period January 1973 to October 2005 (394 observations) relating to:
• The nominal exchange rate (s) (rupees per dollar).
• An Indian Wholesale Price Index (p).
• A US Wholesale Price Index (pstar).
The univariate analysis in this seminar is an important precursor to the multivariate analysis of PPP which we will go through in the next seminar. However in this seminar we will also carry out univariate tests of PPP based on the time series properties of the real exchange rate. In this context we will:
• Test for unit roots in the (log) real exchange rate using ADF, P-P and KPSS tests.
• Test for mean reversion in the (log) real exchange rate using the Geweke and Porter-Hudak (GPH) spectral regression test for long memory.
For the GPH test you will need an Eviews program spectrum.prg which will estimate the spectrum of a time series. The learning outcomes from this analysis will be to develop your understanding of:
• The context(s) for applying unit root tests to financial time series.
• The implementation of different unit root tests in Eviews.
• Testing for long memory in financial time series.
• Estimating the spectrum of a time series and applying the GPH test for long memory in Eviews.
1. Background and overview
PPP states that the prices of identical bundles of goods, converted to a common currency, will be the same across countries:
*SPP=
wherePand *Pare the prices of identical baskets of goods in the domestic and foreign countries respectively and is the nominal exchange rate (domestic currency price of a unit of foreign exchange). For background on the PPP relationship see Chapter 24 in Cuthbertson and Nitzsche (2004). See also Taylor (2006) for an excellent review article on the theoretical background to PPP and developments in the empirical techniques used to test the relationship. S
Taking logs and making the exchange rate the dependent variable suggests the following equation for testing PPP
ttttPPSεβββ+++=*321lnlnln
Testing for PPP entails testing1:320=−=ββH. Note that it is possible for the intercept to be non-zero under PPP due to differences in the base years used to normalize the price indices. For example, if the domestic price index is based in year x:xtxtPππ=)(, whereas the foreign price index is based in year y: **)*(ytytPππ=, and assuming PPP holds for prices based in the same year, then: **)*()()*()(,xyyttxtxttxtPSPPSPππκκ==⇒=
Accordingly, basing the indices in different years induces a constant in the PPP relationship: - so when testing PPP we’re only interested in testing that the slope coefficients are 1 and -1 respectively. )ln(,lnlnln)*()(κ−=−+=cPPcSytxtt
Testing PPP is often carried out by analyzing the time series propertie
本论文由英语论文网提供整理,提供论文代写,英语论文代写,代写论文,代写英语论文,代写留学生论文,代写英文论文,留学生论文代写相关核心关键词搜索。