| Ryan Pitylak |
|
Does Aid Starve the Developing World? By: Ryan Pitylak 05/03/2006 University of Texas-Austin
1. Introduction The effect of aid on GDP is
important because many developing countries receive foreign aid. It is important to understand whether foreign
aid negatively or positively impacts the lives of the people living in the
developing country, and GDP is one measure of this impact. My approach is designed to look at the
difference between the countries that are extremely underdeveloped and those
that are just less developed. My
hypothesis is that the countries that are the most underdeveloped need aid so
much that the aid will actually boost their GDP levels. My hypothesis is extended to include the
notion that the less developed countries that receive aid actually receive
lower GDP levels because the aid actually diverts incentives, and people do not
increase their marginal productivity to labor, and therefore GDP drops for that
reason. For the most underdeveloped
countries, I assume that the support infrastructure to gain higher marginal
productivity levels do not exist, and therefore aid doesn’t change incentives
that much. This intuition comes from
looking at some of the least developed countries, like those in Sub-Saharan
Africa. To caveat this point, the people
in these least developed countries could raise their marginal productivity
levels, as argued by William Easterly in The
Elusive Quest for Growth, but changes in these countries need to occur
before major change to these workers lives can happen. 2. Analytical Approach and Data My analysis is about the effect of
aid on GDP, holding several other determinants of GDP constant. The information came from the World Bank
Indicators database that was available through the The dependent
variable is GDP Per Capita, which is GDP per capita according to current
international dollars of the observation, because the change in real GDP per
capita is good measure of the growth, or decline, of a country. The independent variables in this regression
are FDI, which is net foreign direct investment of the observation, year, which
is the year of the observation, Inflation, which is the inflation of the
observation, health, which is a measure of both private and public health
expenditure, and different measures of foreign aid. FDI is held constant because it is presumed
that FDI is a measure of the desirability to invest in the country. Countries that are more desirable to invest
in long-term capital investments, such as FDI, typically have sounder policies
and institutions. Year is held constant
because it is presumed that the situation in each country changes from year to
year according to effects that are specific to that year, and not specific to
our main thesis about what is the effect of foreign aid on GDP. Health is held constant because it is
presumed that countries that spend more money on health have a better economy,
and therefore GDP will be inflated too high for good countries unless I hold
health constant. Multiple aid variables
are regressed in the different tests. My
initial hypothesis is that the effect of aid on GDP lags by four years. That seemed like an appropriate amount of
time for the aid money to be distributed and the effects of the aid to be felt on
GDP levels. Now to explain the
different the different measures of aid: Aid is a measure of the effect of aid
on the same year’s GDP. Aid Lag 1 is the
measure of the effect of foreign aid received in the previous year on GDP. Aid Lag 2 is the measure of the effect of
foreign aid received two years ago on GDP.
Aid Lag 3, Aid Lag 4, and Aid Lag 5 all follow the same format. Multiple ways of looking at Aid, ranging from
the effect of the current year’s foreign aid on GDP to the effect of foreign
aid from five years ago on GDP, were important to isolate the actual effect of
foreign aid on GDP. Regressions were run
separately for the least developed countries and the less developed
countries. The least developed countries
included Table 0a. Lowest Countries Variables
Table 0b.
I found it odd that
the aid variable increases with each lagged year for the low
countries, but this was verified through manual inspection of the data. Table 1. Regression of
the Determinants of GDP in the lowest developed countries between
1997-2001
The Log ( Aid ) is statistically significant in regression 1b, but
there might be first order correlation.
This might also seem statistically significant because of the fixed
effects of each country. Table 2. Regression of
the Determinants of GDP in the low developed countries between
1997-2001
Aid was not
significant in OLS regression, but the log ( aid ) was
significant. The problem with this is
that the significance may come from variations by year or variations by
country. To resolve these conflicts, I
ran an Xtreg in Stata that
adjusted for the fixed effects of each country, and I also added dummies for each
year. The base year was 1997. Regression 3b below shows that Log ( Aid ) is still statistically significant.
|