“…7 Given the description of the trade performance and reforms during the period under investigation discussed above, it may be hypothesized that the variation in degrees of trade openness is related to economic size (as measured by output and population of the trading countries), transportation costs and the exchange rate and trade policies of the trading countries. In addition to the standard gravity model specifications applied in other studies of trade openness of the economies in transition, such as those by Havrylyshyn and Al-Atrash (1998), Elborgh-Woytek (2003) and Babetski and others (2003), this study suggests controlling for a number of other CIS-specific factors. The following log-linear specification of the gravity model has been applied to identify the determinants of trade openness: where TO i is trade openness of i country to the EU and China respectively, Y i is GDP per capita in PPP of country i; P i is population of country i; D i is distance between country i's capital city and Frankfurt or Beijing, as proxies of the vicinity to EU or Chinese markets respectively; TI i is trade liberalization index, as measured by the EBRD index of exchange and trade liberalization; SI j is similarity index between country i and its trading partners; DM im is a set of m binary dummy variables, which take the value of one if there are common borders of country i with the EU and China respectively or if country i is a member of a free trade area or WTO, and zero otherwise; t is subscript referring to time; and e it is a standard error term.…”