Stock Selection Using Ward Clustering and Mean Absolute Deviation Portfolio Construction on the IDX Quality30 Index

Investment in stocks offers high potential returns but also involves risks that need to be managed through diversification. This study aims to find the best stocks for constructing a Mean Absolute Deviation (MAD) portfolio by using Ward Clustering to select stocks from the IDX Quality30 index. The data includes stocks listed continuously from November 2023 to October 2024. Stocks were grouped based on financial ratios: Earnings Per Share, Price to Earnings Ratio, Debt to Equity Ratio, and Return on Equity. Ward Clustering was applied to create clusters with low internal variation by merging groups that minimize the increase in Sum of Squared Error. The Silhouette Coefficient was used to determine that five clusters were optimal. From each cluster, stocks with the highest positive expected return were chosen. The MAD model then calculated the best portfolio weights by minimizing risk measured by absolute deviation, with limits on minimum return and stock weights. The final portfolio contains four stocks: ADRO (30%), BBCA (30%), MIKA (10%), and UNTR (30%), while ACES was excluded due to its insignificant contribution to portfolio performance . The portfolio achieved an expected return of 0.00090 and risk of 0.01018. A Sharpe Ratio of 0.07224 indicates the portfolio outperforms risk-free investments, making it a suitable option for investors looking for a balanced risk-return portfolio within the IDX Quality30 index.