Can Dynamic Panel Data Explain the Finance-Growth Link? An Empirical Likelihood Approach

The short run effect of the financial intermediary development on economic growth is ana- lyzed using an unbalanced panel of 77 countries covering 35 years. Empirical Likelihood (EL) estimation is used and compared to more conventional GMM methods that weight moment conditions equally over the sample. However, if a part of the data is associated with only weak instruments, GMM estimators are subject to considerable small sample bias. EL appropriately re-weights the moment restrictions to deal with that problem. Using EL, we obtain more robust estimates of the effect of financial intermediation on economic growth than GMM.