The Feasibility Of Alpha Hedging Strategies Based On Financial Fraud Detection

Under the current economic situation, lots of money in the market can not find suitable investment products, asset allocation chaos occurs, causing certain types of investment products in the market to be overheating. Stocks, bonds and real estate markets are good examples to explain it.In order to find a suitable investment portfolio in such economic environment and to gain higher return on investment, I got the inspiration from the muddy water company. That is building the alpha-arbitrage strategies, which is based on financial fraud detection. The alpha of those listed companies becomes extremely different after the declaration of the reports, which means that the strategies can work quite well under theoretical conditions.By extrapolating models to screen listed companies, I found that Random Forest classification model work significantly better than the other models. Sales rate, cost margin of one year age, interest coverage ratio have a high level of interpretation.Finally, I tested the policy performance in bull, bear and concussion market. Simulation results indicate that strategy in bear market gain the most, the cumulative gain is up to 34% during 120 calendar days, while in bull and concussion market, the monthly income can reach about 5%. However, because of the lack of a good method to hedge the system risk, resulting in yields at a premium in bear market and yields underrated in bull market.