Testing PPP theory using black market exchange rate data has gained vast popularity in recent times. It is claimed that black market exchange rate data more often support PPP than official market data. In this study, to assess both the long run stability of exchange rate and the short run dynamics, we employed Pooled Mean Group (PMG) Estimation developed by Pesaran et al. (1999), which best addresses heterogeneous dynamic panel data. We have applied PMG procedure on eight groups of countries based of different factors like income, economic region, economic conditions etc. The dataset contained different time periods for different groups. The model that has been used under this study is the one that has been used by Bahmani-Oskooee and Goswami (2005). The results producred are, however, found to contradict the most recent studies. We have found a very weak and insignificant support for PPP using both black market and official market exchange rate data. The results are also very sensitive to different lag selection procedures.