Is “Not Trading” Informative? Evidence from Corporate Insiders’ Portfolios (2022)
Luke DeVault, Scott Cederburg, Kainan Wang
Financial Analysts Journal 78(1), 79-100, URL/SSRN
Transactions of insiders are usually a useful source of information when evaluating a stock. Insiders typically have a good understanding of the underlying business and buys are therefore often considered as positive signal. On the other hand, insider sales are not necessarily negative. There are many non-informative reasons to cash out. Maybe the insider needs some cash for personal expenditures or just wants to diversify his assets. This week’s AGNOSTIC Paper challenges this asymmetry and creatively shows that even those transactions convey important information…
Everything that follows is only my summary of the original paper. So unless indicated otherwise, all tables and charts belong to the authors of the paper and I am just quoting them. The authors deserve full credit for creating this material, so please always cite the original source.
Setup and Idea
The paper starts with a review about the literature on insider transactions for equities. There are specific legal definitions of “insiders” but generally, they are shareholders with better access to information than others (officers, board members, large shareholders). Unsurprisingly, the literature mostly finds that the transactions of those investors tend to be profitable. Replicating their trades tends to be a good strategy and many data providers offer the required data. Most of these studies, however, focus primarily on insider buys.
With respect to insider sales, most of the existing literature doesn’t find profitable patterns. As I mentioned in the introduction above, this not surprising. There are no reasons to buy a stock if you don’t see an appealing opportunity. On the other hand, there are many reasons to sell a stock beside the current attractiveness. For example, an insider might just need some cash for personal expenditures.
This week’s authors find a novel and creative approach to still exploit such insider sales. The idea is actually quite simple. There are investors who are insiders at more than one company.For example, a CEO who also sits on the board of another company. So those people hold a portfolio of insider-positions and the authors call them portfolio insiders.
Now suppose that such a portfolio insider sells a stock. By itself, observing this transaction is typically non-informative. But with information about the portfolio insider’s other holdings, it becomes interesting. Why? Because if you have five stocks in a portfolio and (for whatever reason) need to sell one, which one will you choose? Of course, the one with the worst outlook. For that reason, a sale is actually a good sign for the remaining stocks in the insider’s portfolio. If you must sell, then at least try to sell the (expected) losers and keep the winners. In my opinion, this idea is very logical and the authors also find it in the data…
Data and Methodology
To test their hypothesis, the authors obtain data on insider transactions from Thomson Reuters (a.k.a. Refinitv). Stock returns and fundamentals are from CRSP and Compustat, respectively.Both are high-quality databases, so there shouldn’t be any issues here. The sample ranges from 1992 to 2020 and covers all US stocks with insider transactions at one of the major US exchanges.
In the next step, the authors identify portfolio insiders as investors with more than one reported insider holding. While this sounds very straight forward, it is still just an estimate. For example, “normal” holdings of the portfolio insiders remain unknown. In addition to that, events like resignation or retirement of insiders can lead to a sudden stop of data availability even though the insider may have not liquidated her position. The estimate is therefore certainly not perfect, but the best the authors can do with the available data.
The authors consider four types of (portfolio) insiders. Directors, officers, large or institutional investors, and “others”. The following table provides some statistics about their portfolios. I want to highlight two observations. First, there is considerable heterogeneity. The largest insider portfolio consists of 71 stocks whereas the smallest one consists of just 2 (the minimum threshold for being a portfolio insider). The median is also 2, so a lot of portfolio insiders are indeed very focused on few companies. Second, from 139,684 insiders in total, only 8,580 are portfolio insiders. This corresponds to about 6% and is actually not that much.
Since large institutional shareholders are not the typical insider in the sense of “access to better information”, the authors exclude this group from the main analyses. But they consider them in a separate section of the paper and use them as robustness check for there findings.
Important Results and Takeaways
“Not sold” stocks from insider portfolios outperformed
The first part is all about the hypothesis that portfolio insiders tend to sell their weakest portfolio companies. As mentioned above, a sale of a portfolio insider should therefore be a positive signal for her “not sold” portfolio stocks. “Not selling” may be not such a strong commitment as buying, but if you have the choice and actively decide to keep a stock, that also expresses an opinion.
To test this idea, the authors conduct an event study and report the average performance of “not sold” and “not bought” companies after a reported transaction of portfolio insiders. The following table summarizes the results. Panel A reports excess returns with respect to the value-weighted CRSP market index, Panel B with respect to a portfolio of companies with similar market cap (size-adjusted), and Panel C with respect to a portfolio of stocks with similar size, value and momentum characteristics.
As it is most relevant in practice, I mainly focus on the excess-returns above the market in Panel A. The results are substantial! After portfolio insiders reported a transaction, their remaining “not sold” portfolio companies outperform the market on average by 0.77% over 30 days, by 1.67% over 60 days, and by 2.84% over 150 days. The results are statistically significant and suggest very attractive performance! In addition to that, the results support the author’s hypothesis. Insiders seem to know what they are doing and if they have a reason to sell, they at least keep their best holdings when they have the choice.
A portfolio of “not sold” stocks easily beat the US market
In the next part, the authors test whether other (“outsider”) investors can exploit this pattern. The event study above has already shown that it would have been profitable to buy an equal-weighted portfolio of “not sold” stocks after a reported insider sale. Trading such event studies, however, is difficult for most investors and the authors therefore present another strategy.
For each month, they identify all “not sold” stocks from insider transactions. They also carefully incorporate all deadlines and time-lags of the relevant filings to avoid any look-ahead bias. They also remove stocks with conflicting signals. For example, if Apple is “not sold” by insider A but sold by insider B, they just exclude the stock from the portfolio.I think this is reasonable. How should they decide which insider is more likely to be right? The following chart shows the cumulative performance of the equal-weighted “not sold” portfolios for monthly and annual rebalancing frequencies after estimated trading costs.
The chart directly shows the strong excess returns from the event study. Both “not sold” portfolios strongly outperformed the US market between January 1996 and October 2020. It is especially interesting that the “slower” strategy with annual rebalancing performed better. This is because of trading costs. Before costs, the monthly strategy was much better and achieved a wealth multiple of more than 35. Trading costs, however, massively reduce performance to a final after-cost wealth multiple of less than 20. For this particular strategy, it is therefore better to trade less frequently and give up some of the expected before-cost outperformance to achieve higher after-cost outperformance. Note that this is obviously not true for all strategies. For example, strategies like momentum or short-term reversal live from high turnover and tend to perform worse with slower rebalancing.
The fact that the strategy also worked with slow rebalancing makes it even more interesting as basically every investor can use it.As I have learned more and more from my Wikifolios, frequently trading large portfolios in practice is much harder than doing backtests… It also suggests that insiders seem to trade on longer-term and probably fundamental information. In my opinion, the strategy is very interesting because it is derived from a clear and logical hypothesis and apparently quite easy to implement.
Finally, a brief note for everyone who wants more performance statistics. The authors of course also show that “not sold” stocks generate significant alpha versus common factor models. Depending on the parameters, they estimate monthly alphas of about 30-40 basis points. Notably, these alphas also survive trading costs, especially for the strategies with slower rebalancing.
“Not sold” stocks with momentum are even better
The authors also explore the interaction of the “not sold” effect with other return-predictors by means of the price-momentum factor. For this purpose, they rank companies by their past 12-month returns and define the third with the highest returns as “momentum stocks”.This is the classic academic definition of momentum. See for example, Kenneth French’s website. Next, they combine the two strategies and present an equal-weighted portfolio of “not sold” stocks that also have high momentum. Since annual rebalancing is too slow for momentum, the authors shift to quarterly rebalancing for this backtest.Momentum tends to have fairly high turnover and worked best with fast rebalancing (before costs!). This is also a common critique of the factor and has been disentangled by two of the world’s leading momentum-managers, AQR and Robeco.
The chart speaks for itself. The “Combination” portfolio generates substantial outperformance and achieves a wealth multiple of 36.7 after trading costs. That’s a compounded annual return of 15.6% versus 8.9% for the US stock market over the same period. The “Combination” portfolio also generates significant alpha versus common factor models. Notably, the alpha is even positive when controlling for the momentum factor. This suggests that considering “not sold” stocks of insiders improves the momentum strategy beyond its traditional definition.
Corporate insiders know more than institutional investors
As I mentioned earlier, the previous results were limited to directors, officers, and “other” insiders. In the final part of the paper, the authors address this issue and repeat some of the analysis with insider transactions from large institutional investors. Interestingly, the outperformance of “not sold” stocks disappears for this group. This is in-line with the authors’ ex-ante expectations as directors and officers tend to have better information (and influence) on companies than probably most financial institutions. So this also serves as an robustness check for the results above and suggests that transactions of “real” insiders do convey important information.
Conclusions and Further Ideas
I really like the idea and strategy of the paper because it is based on a very clear fundamental idea. When insiders (for whatever reason) need to sell a stock and have a portfolio to choose from, they obviously try to sell the stock with the worst outlook and keep the winners. You need to think a little bit around the corner to interpret such an insider-sale as a positive signal for the remaining portfolio companies. But given the historical outperformance of “not sold” stocks, this effort was definitely worth it!
The strategy is also quite easy to implement in practice. Data on insider transactions are readily available from data-providers and you don’t need fast rebalancing or sophisticated trading to earn the premium. This brings me to my key question. Who is on the other side of this trade, i.e. why did the portfolio of “not sold” stocks perform so well? Of course, the most evident explanation is that insiders have better information that is not reflected in the market price. But why does it take so long (with annual rebalancing, up to one year) to incorporate this information into market prices?
That is of course the old question of empirical asset pricing. Is there an unknown risk for which the outperformance compensates, are there any other frictions to lock-in the profit, or is it indeed a market inefficiency where enough investors just don’t understand what’s happening? To be honest, I am not yet sure for this one. The only thing I would rule out are frictions. Data on insider transactions are readily available, annual rebalancing is not a problem, and the strategies also survive trading costs. But regarding the other two, I can only speculate. Given the results, however, it seems smart to follow corporate insiders event though we don’t exactly know why…
- AgPa #56: The Equity Risk Premium of Small Businesses
- AgPa #55: Backtests in the Age of Machine Learning
- AgPa #54: Transitory Inflation
- AgPa #53: Investing in Interesting Times
This content is for educational and informational purposes only and no substitute for professional or financial advice. The use of any information on this website is solely on your own risk and I do not take responsibility or liability for any damages that may occur. The views expressed on this website are solely my own and do not necessarily reflect the views of any organisation I am associated with. Income- or benefit-generating links are marked with a star (*). Please also read the Disclaimer.
|1||For example, a CEO who also sits on the board of another company.|
|2||Both are high-quality databases, so there shouldn’t be any issues here.|
|3||I think this is reasonable. How should they decide which insider is more likely to be right?|
|4||Note that this is obviously not true for all strategies. For example, strategies like momentum or short-term reversal live from high turnover and tend to perform worse with slower rebalancing.|
|5||As I have learned more and more from my Wikifolios, frequently trading large portfolios in practice is much harder than doing backtests…|
|6||This is the classic academic definition of momentum. See for example, Kenneth French’s website.|
|7||Momentum tends to have fairly high turnover and worked best with fast rebalancing (before costs!). This is also a common critique of the factor and has been disentangled by two of the world’s leading momentum-managers, AQR and Robeco.|