Equity Market Structure and Trading Diversification: Insights from Panel Data, Clustering, and Machine Learning

This article aims to contribute to a relatively understudied area of financial development, namely, the internal dispersion of trading activity. The focus is not on overall financial development measures such as total market capitalization and liquidity but rather on trading diversification, defined as the proportion of trading volume contributed by firms outside the VTX, representing the top ten most frequently traded firms. The article uses data from the World Bank’s Global Financial Development Database. The sample is constructed as a balanced panel of 23 countries over the period 2002-2021, starting with a sample of 38 countries. The article uses four key explanatory variables, namely, relative size of deposit-taking banks (DBS), remittance inflows (REM), market capitalization excluding the top ten firms (MCX), and outstanding international public debt (IPU). The article uses a combination of panel econometrics, hierarchical clustering, and machine learning methods. The econometric results show that a diversified financial system structure and remittance inflows are strongly, positively related to overall and less concentrated trading activity, while bank dominance and reliance on international public debt are related to more concentrated trading activity. The clustering results show significant cross-country heterogeneity and a core-periphery structure. The machine learning results show that, using all models, equity market structure is again found to be the most important explanatory variable, with external financial flows being important as well. The article concludes that equity market structure is key to understanding internal dispersion, with important policy implications.