48 / 514
Analysis of Bilateral Trade Between EU-15 and Selected Balkans and Black Sea Countries
37
n
In this study gravity model analyze the bilateral trade relations between BBSCs and EU-15 countries. A panel-
based gravity model is more desirable in order to deal with heterogeneity issues because the effects of such deter-
minants can be modeled by including country-pair individual effects (Serlenge et. Al., 2004:4). According to the
main equation we used, the size of bilateral trade is an increasing function of economic size and a decreasing
function of country distance. Therefore we argue that each Balkan and Black Sea Countries’ export flows with an
each country of EU-15 has a direct correlation with economic size and the inverse relationship with the distance
between them. As pointed in many studies, the right hand side variables of the gravity equations are related with
certain cultural, historical and geographical characteristics of the cross-section units. As we mentioned before
there are many variables that may effect on trade flows between the country-pairs; therefore we can put them in
the equation as a form of dummy variables which are common land
border
and the presence or absence of regional
trade area agreements (
RTA
). In both empirical studies of gravity model and trade of Balkan countries, a common
finding is that RTAs or FTAs appear to have uncertain effect on trade flows. According to Baier and Bergstrand
(2009:63) “
the coefficient estimates of RTA (or FTA) display extreme instability across years, and in many cases
seemingly successful economic integration agreements have negative estimated treatment effects
”. However in some
studies it is a common finding that these agreements appear to have significant positive effect on bilateral trade
flows (Begovic, 2011: 52) Thus, the final model of this study is as follows, which will be used to analyze each BBSCs
export with its EU-15 partner:
_
_
InX C In gdp e
In gdp i
In D RTA Border ij
1
2
3
4
5
ij
ij
b
b
b
b
b
f
= +
+
+
+
+
+
(5)
In this equation denotes each BBSCs’ exports to each country of EU-15,
gdp e
and
gdp_i
indicate the level of
nominal GDP in country i (each of the BBSCs) and j (each of the EU-15) in current dollars,
Dij
is the geographical
distance of each country of BBSCs from each EU-15 country and consistent with the arguments in the literature,
its coefficient should be negative.
border
is a binary variable assuming the value 1 if countries
i
and
j
share a com-
mon land border and 0 otherwise,
RTA
is another binary variable assuming the value 1 if countries
i
and
j
have a
free trade agreement and 0 otherwise; a positive indicates that the export volumes of BBSCs are higher than those
of the countries outside the considered RTAs with EU-15. Finally
ij
denotes the stochastic error term
3.2 Data
Our data contain bilateral export flows for EU-15 and selected BBSCs. The data used cover a period of twenty
years (1996-2012) and sample contains 9 Balkans and 15 EU countries. The trade data of these countries are
extracted from the UN COMTRADE database, GDP data come from the World Bank-World Development Indi-
cator (WDI).The dummy variable for having a common border refers to 19 land borders shared by the countries
included in the sample.
3.3. Results
In the first step, we use panel data to estimate the trade determinants for a particular time period. For the purpose
of this work, the standard model was regressed basing on several specifications: as pooled data by OLS, as fixed
and random effects model with time dummies. The regression results for the baseline gravity model are presented
in Table 5.