Empirical studies on social capital in the United States

This paper empirically analyzes two aspects of social capital in the United States---the negative relationship between social capital and economic growth at the state level, and social capital measurement.;The effect of social capital on economic growth has been found to be robustly positive in cross-country analysis. However, the effect of social capital on the economic growth of U.S. states has been found to be negative. The first part of this dissertation tests systematically whether this negative effect is robust using the Sala-i-Martin robustness test and a quasi-experimental design effect size test. The results show that this negative effect is statistically robust. This creates a puzzle regarding what conditions make the relationship between social capital and economic growth in the U.S. a special case in the cross-country statistics. We speculate that this negative relationship can be explained by poverty traps caused by persistent socioeconomic segregations created by social capital, especially social network that is strong within groups but weak between groups.;Even though social capital has gained significant attention from researchers, no one has clearly defined social capital measure that can capture its multidimensional definition. The second part of this paper applies the single-index model developed by Stock and Watson (1993) to captures the underlying comovement of three social capital variables---trust, civic engagement, and philanthropy giving. The results are consistent with other empirical studies.;In the last part of this paper, we apply a dynamic factor analysis using a Markov-switching model---to find a composite coincident economic index that captures the overall current state of the Thai economy and probabilities of recessions that can give early signals of recessions. The empirical results show that the new coincident index represents the movement of the Thai economy's GDP growth rates more accurately than does the Bank of Thailand coincident index. Moreover, the probabilities of turning points may detect a recession more accurately than does the Bank of Thailand index. Therefore, we suggest this novel methodology as an alternative to the Bank of Thailand coincident index.