country are independent from year to year, though this assumption does not hold up for loans
given to the same country in the same year.
I drew the data for each of these variables from several sources. IDB funding was coded
at the project level from the Bank’s annual reports.
11
Additionally, the IDB funds additional
social projects each year an average of approximately $US one million as a part of special
accounts/waqf fund operations. We have not yet collected all of this data. As such, any picture
presented by this project is incomplete, though the absence of special accounts/waqf fund data
should at worst lead regression analysis to underspecify the amount of social lending which
occurs at the IDB.
Data for the other variables comes from the CIA World Factbook (2003) (IO), while
Islamic population data (ISLAM) comes from the World Christian Encyclopedia (2001). The
measure for DEM is drawn directly from POLITY IV, as mentioned above, and HEGEM is
based on membership data for the AsDB, World Bank, and IADB. Controls came from the
World Development Indicators (2001). I understand that my operationalization of these
variables may not properly capture the relationships I intend them to. The variables ISLAM and
IO and their interaction are most subject to this criticism. I would defend these
operationalizations on the grounds that I have identified a value within Islam that is so strongly
institutionalized that its required expression is an important part of what makes an individual
Muslim. Thus, we ought to see preferences for social lending come through clearly in most
Muslim groups. Environmentalism, on the other hand, is not institutionalized and weakly
11
A group of BYU students was trained to code these variables, based on coding rules derived from the PLAID and
Greenaid Projects at BYU and the College of William & Mary. A second group of students will recode this data to
check for intercoder reliability. This dataset is not yet complete, however, as we are still missing several years of
project data for the IDB.[0]
Convention