Page 97 - Azerbaijan State University of Economics

Fadai Mardanli Mehman, Vildan Zahidkizi Rizayeva: Do Remittances Compensate for the
Labor Market Gaps Created by Emigration?

depends on who will migrate‚ how remittances will be used‚ and whether productive
employment can be created in the country of origin based on migration income.

Pooled fixed-effects regression
Table 4 shows the pooled regression analysis across the entire sample. Dummy
variables for Moldova‚ Nepal and Tajikistan are included in the estimation‚ with
Kyrgyzstan acting as the reference category. This model asks whether variation of
remittances within a country is associated with variation of unemployment‚ holding
fixed country effects.

Table 4: Pooled regression with country fixed effects
Variable Coefficient (B) Std. Error t-statistic Significance
Constant (baseline for KGZ) 5.782 0.429 13.476 p < 0.001
Remittances (% of GDP) -0.0071 0.015 -0.464 p = 0.644
Dummy for Moldova -0.253 0.357 -0.708 p = 0.481
Dummy for Nepal 4.997 0.358 13.957 p < 0.001
Dummy for Tajikistan 6.045 0.374 16.142 p < 0.001
Model R² = 0.877, Adj R² = 0.870, F(4,71) = 126.6, Sig F < 0.001
Note: Reference category for dummies is Kyrgyzstan; thus the constant 5.782 represents

The unemployment level in Kyrgyzstan when remittances and interactions are at 0.
The dummies indicate what the intercept is for each country relative to Kyrgyzstan.
The remittances coefficient is the same for all countries in this fixed-effects model.

The coefficients for the country dummies for Nepal and Tajikistan in the last column
of Table 4 are both meaningful‚ which implies that rates of unemployment in these
two countries are much higher than Kyrgyz unemployment (the baseline). The
constant term recorded in the same column indicates that Kyrgyz unemployment was
around 5.8%‚ without remittances. The Moldova dummy of -0.253 (not important)
indicates that but for remittances Moldova should have a natural unemployment rate
0.25 points lower than Kyrgyzstan's. Thus Moldova and Kyrgyzstan have similar
unemployment rates. The dummy for Nepal is 4.997‚ indicating a natural unemployment
rate 5 percentage points higher than Kyrgyzstan (Kyrgyzstan ~5%‚ Nepal ~10%). So the
coefficient for Tajikistan's dummy is +6.045‚ about 6 points higher than Kyrgyzstan's
base level (consistent with KGZ ~5% TJK ~11%). Regardless of remittance levels‚ labor
underutilization is projected to be higher in Tajikistan and Nepal than in Kyrgyzstan and
Moldova‚ due to structural characteristics in these labor markets.

The most relevant result is the one for Remittances‚ B=-0.0071 and p=0.644. This
suggests that once average cross-country differences are excluded remittances do not
appear to have any systematic statistical effect in the unemployment of these
countries‚ either positive or negative. In other words‚ controlling for cross-country

97