Does the Relative Size of an Insider's Trade Correlate with Over- or Under-Performance of Such Trade?

The question we are trying to answer today is this:

Does the size of a corporate insider's trade relative to the size of his or her past trades in the same company's stock correlate with the return achieved by such trade? Put another way, do abnormally large purchases by insiders earn, on average, higher returns than smaller purchases by the same insider?

What Constitutes A Large Trade?

A number of academic studies suggest that larger purchases do correlate with larger returns. However, it is important to note how each study defines large trades. Below are the two common ways that the line of demarcation between "large" and "small" trades have been defined in a number of different studies, as well as the potential shortcomings of such approaches.

Total number of shares purchased - Using this method, the total number of shares purchased serve as the proxy to segment trades into groups. Trades with a very high number of shares - perhaps 100,000 - might be rated as large, while trades with relatively small numbers of share - perhaps 2,000 - might be classified as small.

Shortcomings. The obvious shortcoming of this approach is that it does not factor in the price of each share. An insider who buys 1 million shares of a stock trading at less than $1 per share should not be classified as larger than a 2,000 share purchase of a stock trading at $500 per share.

This approach is appealing, as the number of shares is readily available and more easily accessible than share prices. But it introduces a lot of noise into the data (i.e., misclassified trades), making it difficult to trust the findings of any study purporting to find a statistically significant relationship between trade size and return.
Dollar value of shares purchased - Using this approach, the total purchase price paid by the insider - meaning the total dollar value of the trade - is used. The study will then define large and small trades based on a dollar value threshold for the trade. This approach is, in my view, superior to the first approach, as it includes share price and is therefore more reliably able to classify trades by their true size.

Shortcomings. The obvious shortcoming is that not all insiders are equally wealthy; in fact, their wealth diverges a great deal. Hence, while a $1 million purchase might be large for one insider, it might be small for a billionaire founder who already owns billions of dollars worth of the company's stock. As you can imagine, the level of conviction held by an insider for whom his or her $1 million trade represents a substantial portion of such insider's total net worth is probably significantly higher than the level of conviction for the billionaire who also buys $1 million of stock. But this approach does not differentiate between the two and, hence, still lacks some precision about the likely level of conviction behind insiders' purchases.

Can We Measure the Size of the Trade Against the Insider's Estimated Wealth?

In a perfect world, we would be able to create a database that accurately approximates the net worth of each Section 16 insider and compare trade sizes to that estimate. In such case, we would be able to state with accuracy what percentage of the insider's net worth was being invested in the company in any particular trade.

While this sounds fanciful, at least a rough approximation can be generated for a significant portion of corporate insiders who serve as officers of the company. Because the annual compensation paid to the CEO, CFO, and three highest compensated officers thereafter is disclosed in issuers' filings, one can construct a database of such information. This would cover the "income" side of a net worth estimate. To cover the "asset" side of the estimate, one could use aggregate share holdings as a proxy.

Having tried to implement this approach in the past, I have found that it has numerous shortcomings. Among them are the following:

Not all insiders are required to have their incomes reported.
No directors are required to have their incomes reported.
Using share ownership as a proxy for net worth is highly inaccurate.

For these reasons, I do not believe that this approach is an improvement over the approaches discussed previously. Hence, it does not seem possible to me to measure the size of a trade against an insider's estimated net worth with any accuracy, and so I have abandoned this approach.

A Potential Solution - Measure the Size of the Trade Against the Size of the Same Insider's Previous Trades

Under this approach, which is examined in greater detail here, we compare the size of any particular insider purchase against the size of all of their previous trades. We can then see whether there is a relationship between trade size and returns. In essence, we are here testing whether trade size can be used as a proxy for the level of conviction behind an insider's trade, where abnormally large trades reflect abnormally large conviction in the future return prospects of the insider's stock.

To implement this approach, I aggregated all of the purchases by insiders in my insider trading database (which is regularly updated from the SEC's EDGAR database). I then calculated the total dollar value of all purchases made by an insider on a particular day. Finally, I calculated a z-score specific to each insider's purchases. Hence, a low z-score indicates a relatively normal sized trade for such insider, whereas a very high z-score would indicate a rare outlier for such insider in which the insider made a far larger purchase than normal.

The theory that we are testing is whether an insider makes higher returns relative to the S&P 500 when he or she makes very large, outlier-type of purchases. If this is the case, then it means that perhaps a retail investor could target these types of trades and then mirror them to earn outsized returns.

An Example of What A Correlation between Trade Size and Total Return Would Look Like

Before I get into the data, take a look at the chart below, which illustrates what we are looking for (but on a large scale). This chart, from a nameless insider, shows the over- or under-performance by the insider relative to the S&P 500 twelve (12) months after the trade. Each bubble below represents a day that the insider purchased stock in his company, and the relative size of each bubble indicates the dollar value of each trade relative to the average size of all trades by the insider.

The grey horizontal line indicates "zero" alpha, where alpha is defined as the relative over- or under-performance of the company's stock vs. the S&P 500 over the twelve (12) months following a purchase. We see here that the insider made two very large purchases (relative to all other purchases by the insider) in 2009, near the bottom of the financial crisis. We also see that those purchases outperformed the S&P 500 by about 25% over the next year - certainly a very impressive return and one that substantially beat the S&P 500 even when the S&P 500 itself also rose substantially over the same time frame.

But this is only one datapoint out of thousands and thousands, so we can't really draw any broad conclusions without more testing. If there truly is a relationship between relative size of a trade and return on such trade relative to the market, then we would expect to see a similar relationship as seen in the chart above across all insiders on average. But is that the case? What does the data show?

Comparing Relative Trade Size to Trade "Alpha"

The chart below shows the alpha (i.e., over- or under-performance relative to the S&P 500) earned by all of the officer insiders for which I have data. Here's how to read the chart:

The X-axis reflects the z-score for every single day in which an officer insider purchased his or her own company's shares. Recall that these z-scores are measured against only other trades by such insider, helping to ensure that "big" trades are truly big trades for the specific insider.
The Y-axis reflects the total excess return of the insider's stock over the twelve (12) months following the purchase. The scale of the Y-axis is such that a value of 1 indicates a 100% outperformance relative to the S&P 500.
As you can see, there are a number of outliers on the chart where the insider's stock vastly outperforms the market by 5-10x. While the chart makes this look like a common occurrence, it is actually quite rare. The chart itself depicts more than 100,000 data points, but you cannot tell that because the vast majority of dots are stacked on top of one another. Hence, the dots right around the zero horizontal line have many hundreds or thousands of dots stacked on top of one another, while the dots at 10-40x are single dots. The odds of these returns is actually only about 2,000-to-1, which the chart does not depict well.
Note, also, that ALL returns quoted on this site reflect the return to a retail investor who mirrors such trades, and not to the insider himself. This is an important assumption, as it reflects the profitability to you, the retail investor, whereas most academic articles measure them on the basis of the return to the insider.

So what does the chart show? It shows that the vast majority of truly outsized returns occur when an insider does not make an extremely large purchase (again, as measured by the insider's past purchases). We know this because the very high returns are all clustered around a zero z-score on the x-axis.

Does this chart prove that relative trade size does not matter? No. But it is some evidence that the relative size of an insider's purchases relative to their past purchases is not an accurate level of trade conviction.

To analyze the question further, here are some stats on the 12-month returns for the data analyzed above. We'll use these again later in the analysis.

Average 12-month alpha: 13.1%
Median 12-month alpha: -2.1%
25th percentile alpha: -21.4%
75th percentile alpha: 21.5%

As the data shows, insider purchases underperform the S&P 500 more than 50% of the time. Yet, the average return on all such trades still substantially outperforms the market (13.1% outperformance), due to the fact that a small portion of total trades earn enormous returns. This brings the average up substantially.

Examining ONLY True Outlier Purchases

But what if we analyzed only truly large, outlier purchases - purchases where the z-score is 3 or above. After all, these trades reflect only very abnormally large purchases by insiders. Presumably these insiders would not have invested so much of their money in the purchase if they didn't really believe that their company's stock would perform well, right? Would this subset of the data above reflect higher trade conviction and, hence, better total returns and market outperformance?

Take a look at the chart below. It is the same as the chart above, except that it only analyzes trades with z-scores of 3 or greater.

We have now screened the original data set, which was in excess of 100,000, down to only about 800 total trades. This is consistent with what we are trying to screen for - true outliers. This is also consistent with the fact that these returns generally follow a normal distribution, as, for normally distributed data, we would expect approximately 0.1% of all data to fall at or above 3 standard deviations (and we find about 0.08% here).

Again, because of the clustering of dots, it is difficult to know whether this subset of data outperforms the full data set above. If purchase size is a measure of the insider's trade conviction, we would expect that this subset would outperform. But does it?

Here are some stats on the 12-month returns for the data subset analyzed above

Average 12-month alpha: 0.5% (vs. 13.1% for the full dataset)
Median 12-month alpha: -6.1% (vs. -2.1% for the full dataset)
25th percentile alpha: -24.1% (vs. -21.4% for the full dataset)
75th percentile alpha: 12.4% (vs. 21.5% for the full dataset)

This shows, rather convincingly, that extremely large, outlier purchases by market insiders actually underperform the full data set!

This is counterintuitive, but there are several reasons why this may be the case. Perhaps the most important of all is the fact that an insider needs to trade relatively frequently in order to generate a single outlier transaction with a standard deviation greater than 3 relative to such insider's other trades. Why is this the case? Consider an insider who almost NEVER trades in his or her stock. If this insider only ever makes one or two insider purchases, then such trades will not register as an extreme outlier from a z-score perspective. Hence, perhaps a better way to screen for the most actionable insider trades is to look for those made by insiders who have either never before purchased their own stock or who have only made a very limited number of purchases in the past. But that research will be covered in a future article.

Is Market Outperformance the Correct Measure to Use When Assessing An Insider's Trade?

Back to the subject at hand, is it possible that we are mismeasuring trade performance? The analysis above has measured the success of any trade based on whether it outperformed the return of the S&P 500 over the subsequent 12 months. But is it the right measure? Or would a better measure be the absolute return of the insider's stock itself? Put another way, if both the insider's stock and the S&P 500 rise by 25% over the year following the trade, the "alpha" earned by the insider's trade is zero - for the insider neither outperformed or underperformed the market. But when evaluated on an absolute basis - i.e., the insider earned 25% on the trade - it is unquestionably a success and a trade that would have been good for a retail investor to mirror.

Let's re-analyze the data above using absolute returns, rather than returns relative to the market, as the measure of trade success.

Comparing Relative Trade Size to Absolute Return on Trade

The chart below shows the absolute return (i.e., actual return, not measured against a market index) earned by all of the officer insiders for which I have data.

You can read the chart in the same way as those above, except that the Y-axis reflects the absolute return on the insider's trade (rather than the over- or under-performance relative to the S&P 500).

We don't see too much difference relative to the similar chart above. But what if we look at the summary statistics?

Average 12-month absolute return: 24.7%
Median 12-month absolute return: 7.9%
25th percentile absolute return: -13.5%
75th percentile absolute return: 35.3%

The summary stats above show that insiders do appear to know, generally, when to buy shares of their own company. This is unsurprising, obviously, but consider that the average 12-month return on any purchase by any officer insider in the data set above is almost 25% in a single year - certainly an impressive return.

Absolute Return on ONLY True Outlier Purchases

Does the same hold true for outlier purchases (i.e., those falling three standard deviations or more above the typical trades by an insider)? See the graph and summaries below.

Summary stats:

Average 12-month absolute return: 11.8% (vs. 24.7% for the full dataset)
Median 12-month absolute return: 1.2% (vs. 7.9% for the full dataset)
25th percentile absolute return: -17.9% (vs. -13.5% for the full dataset)
75th percentile absolute return: 24.2% (vs. 35.3% for the full dataset)

As can be seen from the data above, abnormally large trades yet again underperform the full data set in absolute returns. Hence, we can now conclude with relatively high confidence that abnormally large trades by officer insiders likely do not signal a higher level of conviction or belief by the insider in the future profitability of the trade. This runs counter to the wide body of literature that suggests that "bigger is better."

Do These Findings Hold for Directors As Well?

Though I have not included a discussion of the effect of relative trade size as a predictor of future returns achieved by mirroring insider purchases by directors (i.e., non-officer directors), the results largely hold true and closely mirror those described above for officers. Put another way, just as was seen in trades by officers, abnormally large trades by directors tended to materially underperform the full data set (i.e., all trades by all directors, regardless of size). Hence, there was no evidence to suggest that larger trades by directors reflected a higher level of conviction in those trades relative to others made by the insider.

Could Trade Size Still Be Important, If Other Variables Are Also Factored In?

We should not completely write-off the importance of relative trade size in determining which trades are likely to be the most profitable to mirror. This will be covered in more depth in future posts. But it is also clear and safe to assume that trades within 1-2 standard deviations of any particular insider's average trade size should be given at least as much consideration by a retail investor as abnormally large purchases by such insider.

Other articles that may be of interest

Disclaimers

Not Investment Advice

The content on this website is for informational purposes only. You should not construe any such information or other material as legal, tax, investment, financial, or other advice. Nothing contained on this website constitutes a solicitation, recommendation, endorsement, or offer by this website, its authors or contributors, or any third party service provider to buy or sell any securities or other financial instruments in this or in in any other jurisdiction in which such solicitation or offer would be unlawful under the securities laws of such jurisdiction.

All content on this site is information of a general nature and does not address the circumstances of any particular individual or entity. Nothing on this website constitutes professional and/or financial advice, nor does any information on the website constitute a comprehensive or complete statement of the matters discussed or the law relating thereto. Neither this website nor any of its authors or contributors is a fiduciary by virtue of any person’s use of or access to the website or its content. You alone assume the sole responsibility of evaluating the merits and risks associated with the use of any information or other content on the website before making any decisions based on such information or other content. In exchange for using the website, you agree not to hold this website, its contributors and authors, its and their affiliates, or any third party service provider liable for any possible claim for damages arising from any decision you make based on information or other content made available to you through this website.

Not Legal Advice and No Attorney-Client Relationship Formed

The information provided on this website does not, and is not intended to, constitute legal advice; instead, all information, content, and materials available on this site are for general informational purposes only. Information on this website may not constitute the most up-to-date legal or other information. This website contains links to other third-party websites. Such links are only for the convenience of the reader, user or browser; this website and the author do not recommend or endorse the contents of the third-party sites.

Readers of this website should contact their attorney to obtain advice with respect to any particular legal matter. No reader, user, or browser of this site should act or refrain from acting on the basis of information on this site without first seeking legal advice from counsel in the relevant jurisdiction. Only your individual attorney and your individual investment adviser can provide assurances that the information contained herein – and your interpretation of it – is applicable or appropriate to your particular situation. Use of, and access to, this website or any of the links or resources contained within the site do not create an attorney-client relationship between the reader, user, or browser and website authors, contributors, or their respective employers.

No Representations or Warranties

The content on this posting is provided "as is" and without warranties of any kind. No representations are made that the content is error-free. You bear all risks associated with the use of this website and its content, including, without limitation, any reliance on the accuracy, completeness or usefulness of any content available on this website. This website, its authors and contributors, and its and their affiliates, employees, agents, and representatives disclaim all warranties, express or implied, including, without limitation, all warranties of title, non-infringement, accuracy, completeness, usefulness, merchantability, and fitness for a particular use, and warranties that may arise from course of dealing / performance or usage of trade.

Limitation of Liability

YOUR EXCLUSIVE REMEDY FOR DISSATISFACTION WITH THE SITE AND CONTENT IS TO STOP USING THE SITE AND CONTENT. NEITHER THIS WEBSITE, ITS AUTHORS AND CONTRIBUTORS, NOR ITS AND THEIR AFFILIATES IS NOT LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, SPECIAL OR PUNITIVE DAMAGES, UNDER ANY THEORY OF LIABILITY, INCLUDING WITHOUT LIMITATION, DAMAGES FOR LOSS OF PROFITS, USE, DATA, OR LOSS OF OTHER INTANGIBLES. IN PARTICULAR, AND WITHOUT LIMITATION, WE WILL NOT BE LIABLE FOR DAMAGES OF ANY KIND RESULTING FROM YOUR USE OF OR INABILITY TO USE THE SITE OR CONTENT.

About the Author

My name is Brad G Grounds, and I am an attorney (but see the important disclaimers above). I've been following, studying, and analyzing trades made by corporate insiders for years. I maintain an insider trading database, which contains all of the insider trades reported in the SEC's EDGAR database going back to 2003. My database is updated regularly and contains millions of transactions. Out of those transactions, I sift through for only those reflecting open market purchases or sales. These transactions - which still number in the millions - are the basis for the analysis presented on this site.

Search This Blog

Insider Trading Returns