The effect of historical ties is captured through another dummy variable. This historical tie is
assumed to be reflected in the official languages of the trading partners dummy (DLAN). It
takes value one if both partners share a common language or zero if they do not. The primary
interest of this study is the artificial trade barriers. These variables cannot be easily quantified,
hence we used dummy variables to capture their effects. In southern Africa, three major
25
regional integration efforts can be identified. Thus we introduced two sets of three dummy
variables for the Southern African countries and another set for the preferential trade body to
which some of the control or normal trade group of countries belongs. In all, we have eight
dummy variables. The first set of dummy variables equals one if the importing countries are
members of SADC, MERCOSUR (Common Market of Southern American), ASEAN
(Association of East Asian Nations), and NAFTA (North American Free Trade Agreement),
respectively, otherwise they are equal to zero. This is to test whether trade barriers in each of
the groups are significantly different from what obtains in other countries in the sample. The
second set of dummy variables equals one if both partners are members of SADC,
MERCUSOR, ASEAN, and NAFTA, respectively, otherwise it is assigned zero. This is to test
whether the bodies have any significant effect on trade flows of their respective members.
These variables for SADC are of interest to us in this study. They are expected to be negative
or insignificant prior to the integration period. The value of SADC dummy after integration
will depend on effectiveness of the particular regional body.
In summary, the gravity model for the period prior to integration is specified as: