Summary: | It is important to understand the effect of Small and medium enterprises (hereinaftercalled SME) on economic development and to find out whether International Organizations’ policy recommendations are “Kicking Away the Ladder” policies or not. Therefore, in this study, we seek to evaluate the causal effect of SME’s on growth of 11 different sectors (as suggested by IMF) of economy. To this end, we started with a simple regression model to check whether there is a positive association between SMEs and economic growth at all. The toy regression model we employed was based on the regression of sectoral per capita GDP on the SME growth variable (measured by the labor employed by SMEs in each sector). Furthermore, we checked the robustness of such positive association in different environments determined by the fixed and random effects assumptions and interestingly, we obtained statistically significant positive association no matter which assumption about fixed and random effects were made.The data used in this study are obtained from the website of the State Statistical Committee of the Republic of Azerbaijan (“The State Statistical Committee of the Republic of Azerbaijan,” 2019). We did initial data cleaning and transformations to make the data serve our purposes and these manipulations are given in the paper. The interesting side of the paper is it uses panel data to reveal the relationship of SMEs and growth.To the best of our knowledge, no paper has employed macroeconomic panel data to assess the effect of SMEs on each sector of a single country so far. To this end, we use Arellano-Bond estimator which is also called Generalized Method of Moments Instrumental Variable estimator. The essence of this methodology is that current values of the dependent variable cannot have a causal impact on the past values of the endogenous regressors (Arellano and Bond, 1991).Relying on this fact, we construct instruments for possible endogenous SME growth variable using its own past values. After filtering out the variation (noise) in SME growth variable in current period which is not associated with its own lagged values we hope that there will not be any signal left in the predicted values of SME growth variable which can be affected by current period GDP contemporaneously. After doing so, we believe that any positive regression coefficient obtained is purely due to causality flowing from SME growth to GDP and not vice versa. Moreover, we use three different fixed effects model for robustness check purposes and find that the relationship between SME growth variable and GDP in each sector stays positive and statistically significant which makes our results more convincing.
|