An application of partial least squares to the construction of the Social Institutions and Gender index (SIGI) and the Corruption Perception Index (CPI)

In this paper the Social Institutions and Gender Index (SIGI) is constructed with Principal Component Analysis (PCA) and Partial Least Squares (PLS). Using the SIGI, we test the effects of social institutions related to gender inequality on several development outcomes, such as female education, fertility, child mortality and corruption, controlling for relevant determinants. As the measure of corruption we use the Corruption Perception Index (CPI), considering alternative weighting procedures using PCA and PLS. We find that gender inequality in social institutions has significant effect on fertility and corruption regardless of the weighting procedure, while for female education and child mortality only the SIGI based on PLS generates significant results.