【Application】Stock Selection Factors Research: Combining Insider Ownership and Momentum Factors
In recent years, as the stock prices of popular AI companies continue to reach new highs, investors are increasingly focused not only on these companies’ operational status but also on the trading behavior of their insiders. Company insiders have more information compared to external investors, giving them an informational advantage when trading the company’s stock. To prevent illegal insider trading and enhance information fairness in the securities market, financial regulatory authorities in various countries have established strict laws and require companies to disclose insider ownership data regularly . This is intended to reduce the impact of information asymmetry on investors. This paper examines the relationship between insider trading activities and stock returns, specifically studying insider trading and momentum phenomena.
Research Hypothesis: Insider Ownership and Stock Returns
Previous research indicates that insider trading can generate significant positive returns. Companies with higher insider ownership tend to have more concentrated stock holdings, leading to insufficient circulating shares for investors to reflect company information, thereby increasing the potential for information asymmetry in their stocks. Moreover, researches have shown that insider ownership data should not be solely relied upon as buying or selling signals; external investors need further screening to effectively profit from this information.
Therefore, this paper hypothesizes that stocks with high insider holding ratios or positive monthly growth rates in insider ownership are more likely to exhibit price continuity in the short term. We speculate that a momentum strategy combined with insider ownership data can form effective composite factors, thereby generating significant positive returns. Consequently, the following two hypotheses are proposed:
(1) Companies with higher insider holding ratios have more severe information asymmetry in their stocks, allowing momentum factors to earn significant positive returns.
(2) Stocks with higher monthly growth rates in insider ownership are more likely to see insiders using their informational advantage in trading. After external investors absorb and mimic this information, short-term price continuity is expected, allowing momentum factors to generate significant positive returns.
Data
(1) Data source and sample period
Insider holding ratios: Fundamental data
Ex-dividend price data and trading annotation data for stock: TEJ Market databank
Period: 2013/01–2024/03/20
For more information on forward-looking bias, please refer to “What is Forward-Looking Bias?”
(1) Factor and IC/IR testing
- Insider Holding Ratio (Insider): This is calculated by dividing the shares held by directors, managers, and major shareholders (owning more than 10%) by the total outstanding shares of the company for each stock.
- Monthly Growth Rate in Insider Ownership (Insider Chg): This is the change in insider holding ratio compared to the previous month.
- Momentum Factor (MOM1y): This is calculated by dividing the current adjusted closing price by the maximum adjusted closing price over the past 252 trading days.
- Information Coefficient (IC): This measures the correlation between the factor value at time t and the stock return at time t+1, ranging from -1 to 1. An IC closer to 1 indicates a strong positive predictive ability, while an IC closer to -1 indicates a strong negative predictive ability. Empirically, an IC value of 0.05 or higher is considered an effective factor.
- Information Ratio (IR): This is the average IC value over a period divided by the standard deviation of IC over the same period, measuring the stability of a factor’s predictive ability. Empirically, an absolute IR value of 0.5 or higher is required.
In addition to calculating the IC/IR values of individual factors, composite factors for the two hypotheses are also tested. For the Insider Momentum (Insider Mom) composite factor, samples with zero or missing insider holding ratios in period t, as well as samples lacking momentum factor values, are excluded. Each stock’s factor values in period t are ranked, and the average of these rankings forms the composite factor value for that period. For the Insider Change Momentum (Insider Chg Mom) composite factor, samples with zero or missing monthly growth rates in insider ownership and lacking momentum factor values in period t are excluded. Each stock’s factor values are ranked, and the average of these rankings forms the composite factor value for that period.
(2) Factor Portfolio Backtesting
In the single-factor portfolios, we calculate the performance of portfolios using either the one-year high momentum factor or the insider ownership factor separately. Stocks are divided into three quantiles based on their factor values: the top 30%, middle 40%, and bottom 30%.
For the composite factor portfolios, insider ownership-related factors are first divided into the top 30%, middle 40%, and bottom 30%. Within each of these groups, stocks are further divided into three quantiles based on the one-year high momentum factor: top 30%, middle 40%, and bottom 30%. This results in nine sub-portfolios for each composite factor. All groups use equal-weighted allocations and are rebalanced on the next trading day after the monthly insider ownership data is updated.
The market simulation backtesting considers factors such as friction costs, which helps in more accurately evaluating the performance of factors in real-world scenarios. The disadvantage is that it involves handling numerous details during backtesting, leading to higher time costs. The settings for the backtesting are as follows:
Empirical Analysis
(1) IC/IR testing
From the figure below, it can be seen that the one-year high momentum factor has the best predictive ability, though its monthly predictive power also shows significant volatility. Regarding the insider ownership factors, both the holding ratio and the monthly growth rate perform best when focused on managerial positions. Moreover, the monthly growth rate of managerial holdings demonstrates notably higher predictive power, making it the best-performing single factor. This indicates that the monthly growth rate of managerial holdings is an effective and stable stock selection factor.
From the figure below, it is evident that combining various insider factors with the one-year high momentum factor significantly enhances their predictive ability. However, when examining the insider holding ratio factors, it appears that this enhancement comes at the cost of slightly reducing the momentum factor’s predictive power, while also addressing the high volatility associated with the momentum factor.
Among the three insider factors, the managerial holding ratio composite factor performs the best, though it also shows a significant increase in volatility. Regarding the monthly growth rate of insider ownership, combining it with the momentum factor leads to substantial improvements in predictive ability, but also results in a marked increase in volatility across all three insider ownership categories.
(2) Factor Portfolio Backtesting
The research findings indicate that among composite portfolios that simply buy stocks with high insider holding ratios and high momentum factors, the portfolio based on managerial holding ratio performs the best. This supports Hypothesis 1: stocks with higher insider holding ratios experience more severe information asymmetry, allowing the momentum factor to generate significant positive returns.
In the case of the monthly growth rate of insider ownership factors, the portfolio based on managerial ownership also performs the best among all composite factor portfolios. It represents the optimal combination that best aligns with Hypothesis 2: stocks with higher monthly growth rates in insider ownership indicate that insiders are effectively utilizing their information advantage, leading to a potential phenomenon of price continuity in the short term after external investors absorb and follow this information. Consequently, the momentum factor can earn more significant positive returns.
(3) Simulated Market Backtesting
This study further conducted simulated market backtesting on the aforementioned single-factor and composite-factor portfolios. The primary objective was to analyze the performance of these portfolios in terms of market entry and exit, and whether they could still generate significant positive returns after considering slippage costs, transaction fees, and trading constraints.
The backtesting results in the figure below demonstrate that after combining with the one-year high momentum factor, both the managerial holding ratio factor and the monthly growth rate factor show significant improvements. Annualized returns doubled, and cumulative returns increased three to four times. However, there was no significant increase in annualized volatility. In fact, it was often lower compared to the single-factor scenario. This indicates that composite factor portfolios can still achieve substantial excess returns compared to the market after accounting for slippage costs, transaction fees, and trading constraints, primarily due to the optimization of the insider holding ratio factor using the momentum factor.
Conclusion
This study analyzes whether insider holding ratio factors for directors and supervisors, major shareholders, and managers can be used to earn significant positive returns in the Taiwan stock market and how to optimize these factors further to achieve significant positive returns in backtesting. The empirical results show that factors related to managerial holding ratio significantly outperform those related to directors and major shareholders. Specifically, using the monthly growth rate of managerial ownership as a stock selection indicator and buying the top 30% of stocks in terms of this factor on the trading day following the monthly update of insider ownership data results in significant positive monthly returns.
Furthermore, this study optimizes the insider holding ratio factors using the one-year high momentum factor. It examines two hypotheses: higher holding ratios indicate more severe information asymmetry and higher monthly growth rates in holding ratios lead to more imitation and following behavior by investors. The results show that composite factor portfolios formed by combining the managerial holding ratio factors with the one-year high momentum factor can earn significantly higher returns than the Taiwan Stock Exchange Weighted Index over the same period.
Reminder: This strategy is only for reference and does not constitute any product or investment advice.
Further Reading:
Block Trade Strategy Achieves Performance Beyond The Market Index
Price Momentum Factor Strategy: The Market Favors the Strong
