Risk and Mutual Fund Clustering in an Emerging Market: Evidence for Thailand

This study groups mutual funds using k-means clustering analysis and compares the k-means clustering process with existing clustering techniques using mutual fund data for equity funds, general fixed-income funds, and balanced open-end mutual funds rated by the Association of Investment Management Companies. Data are from January 2016 to December 2020 for 60 months and includes information on prices, risks, and investment policies. The sample for this study comprises 173 funds from 10 asset management companies with the highest net assets. The tool used for analysis is the k-means technique using a statistical package set fork= 3. The funds can be divided into three groups: Group 1 has 5 mutual funds (2.89%), Group 2 has 24 mutual funds (13.87%), and Group 3 has a total of 144 mutual funds (83.24%). In Group 1, four of the five mutual funds are equity funds with a track record of beating the market, and fund managers have good market timing skills. Moreover, the efficiency of fund grouping using thek-means technique was compared with the existing grouping with close results at 57.23%. This work provides a methodology to obtain a better categorization of mutual funds by usingk-means clustering, allowing the investors to know how mutual funds are. This categorization is very useful for improving the formulation of mutual funds, with the goal of further optimizing investment.