How to identify short candidates through anomalous insider selling activity

  by Dominic Finney
  Senior Analyst & Chief Technical Editor, Gradient Analytics LLC (a Sabrient Systems company)

Perhaps the most reliable shortcut to identifying a company at elevated risk of a downturn in its share price is looking at how executives and directors use their equity instruments. This might sound too simple to be predictive – something that would be quickly understood by the market and integrated into investors’ thinking on a scale that would cause the “edge” to disappear. But there are complications that have prevented that from happening, on which I will elaborate shortly. But first, let’s look at some recent examples.

Over the past two years, Gradient Analytics has published five brief “snapshot” reports based on our Equity Incentive Analytics examining signs of unusual and concerning equity use by executives and directors. The subject companies were Amarin (AMRN), United States Cellular (USM), WW International (WW, or WTW when we wrote on it), Supernus Pharma (SUPN), and Magellan Health (MGLN). All five of the reports preceded significant declines in company share price, with four of the five stocks showing double-digit declines over the ensuing three months and all of them hitting double-digit declines over six months. Read on....

The average 3-month/6-month/max drawdown returns were -18.6%/-32.4%/-38.5%. Relative outperformance (versus the benchmark S&P 500) was even better, at -19.5% over three months and -37.4% over six months.

And what’s more, these were companies that were not widely shorted, with an average short interest of only 7.5% at the time of our reports (ranging between 1.9% to 14.0%). If we look a bit further back to three years or so, Gradient also wrote on OraSure Technologies (OSUR) on 06/09/17, which painfully went against our short thesis for about seven months before hitting an inflection point and ultimately falling well below our initiation price by late 2018, and MYR Group (MYRG) on 12/02/16, which was down -22.7% six months after our initiation (and -34.0% relative to the S&P 500).

Such numbers are eye-catching, but they naturally raise the question: Why only five reports over the past two years if the analysis is so effective? And that goes to one of the complications we noted above: The overwhelming majority of Section 16 executive and director equity transaction activity – I would estimate well over 99% – conveys almost no predictive information. Or to put it another way, the difference between a company one standard deviation from the mean and another company two or even three from the mean isn’t normally worth fussing over (for most people anyway – for Gradient, however, it may influence our decision-making for broader evaluation of short or long candidates if the insider activity treads close the “line of concern,” even if it hasn’t crossed it). 

So, what’s the logical connection behind activity that we consider to be concerning and moves in share price? It’s obviously a step removed from the connections we find between earnings quality issues and share-price movements, although there might be a sequential relationship. For example, we might observe that a company has a falling ratio of accounts payable to inventory, which indicates that it might have decreased its inventory build while slowing purchases from suppliers, suggesting that management does not expect a near-term increase in demand. That’s a red flag from an earnings quality perspective. But then, if you’re an executive of the company looking at the data or model outputs that are driving those business decisions, you might also see it as a good time to initiate a new Rule 10b5-1 trading plan. Depending on the specifics, that could be inappropriate or even improper, but likely not in an enforceable way.

Note however that we are emphatically not suggesting that there is necessarily any wrongdoing by the companies we cover – and that’s not just covering our heinies. We rarely see such activity as something that an SEC investigator would consider actionable use of material nonpublic information. Insider activity that we would consider concerning is less like observing clogged arteries and seeing a heart attack coming and more like observing the secretion of glucocorticoids, which could indicate preparation for such diverse stressors as being chased by a lion or delivering an important presentation.

The specific signs we look for are varied, but we primarily focus on anomalous activity levels, SEC Rule 10b5-1 trading plans that deviate from the spirit of the rule (allowing for orderly divestiture over a long timeframe), and broad-based activity among a cluster of insiders (rather than just one).

The most common reason to dismiss activity as inactionable is that the volume is not unusual enough to be concerning (or “TJ” in our internal parlance, in honor of Tom Jones and his 1965 hit song). But once that threshold is crossed, we want to make sure that the anomalous activity isn’t coming from a single individual, as that would make it more likely that the activity is driven by a personal need for more liquidity or diversity (such as a divorce, pending retirement, new summer home in The Hamptons, etc.) than it would be if a cluster of executives and directors were engaging in it.

Next, we move on to option exercise activity. In a nutshell, we expect Section 16 filers to exercise options well before expiration, and our option timing model gives us a precise expectation for when exercises should occur (based on firm size, sector, recent share price performance, and many more factors). In some cases, our model anticipates exercises as soon as options vest, and in others, exercises are not expected until the second half of option life (i.e., the time between vesting and expiration). Options that are early relative to the model’s expectation are our biggest red flag. When a company’s share price runs up, we expect to see increased transaction activity (especially if options had been underwater), so it’s not necessarily concerning. Our option timing model helps us to navigate the curve of the nonlinear expectation function to identify anomalous activity.  

Finally, the other key category of issues we examine is SEC Rule 10b5-1 trading plans. Several years ago, we wrote an Issue Commentary that went into detail on this topic, including citing academic research and performance numbers (spoiler alert: the performance was quite good). I won’t go into it all here, but to summarize, regularly timed and sized sales made pursuant to a plan (e.g., 5,000 shares on the last day of every quarter) are not something that raises our interest. But when we see, for example, executives and directors use multiple 10b5-1 trading plans to unload a large percentage of their ownership in a short period of time (particularly if it’s soon after plan initiation), we are concerned – and even more so if such plans have apparent price triggers that were very close to the prices at the time of plan initiation (which is often but not always disclosed).

You might think that planned sales are nothing to worry about, but that’s because you probably have the mental image of the ideal plan described earlier. However, if the plan is “sell half of what I have if prices rise 5% after today,” it’s not so easily dismissed. In fact, to us, the natural response to something like that is curiosity about why a plan was used at all. And moving on from there, if sales are a small percentage of the total ownership of the sellers, we are less interested in them. Lastly (skipping some very rare issues or more minor ones), we look at historical returns following past transactions by the same individuals. It’s not common, but occasionally we come across individuals with eyebrow-raising histories of well-timed sales, which would be supplemental to our main concerns but potentially of interest.

To sum up, every once in a while there is a confluence of signals related to insider transaction activity that indicates something is up at a company, whereby a cluster of executives and directors have significantly increased their risk aversion related to the firm’s shares, and you really should want to know about it when that happens. After all, if the future is bright, why would so many individuals having privileged information want to drastically reduce their exposure to company stock?

Disclosure: At the time of this writing, the author held no positions in the securities mentioned.

Disclaimer: This newsletter is published solely for informational purposes and is not to be construed as advice or a recommendation to specific individuals. Individuals should take into account their personal financial circumstances in acting on any opinions, commentary, rankings, or stock selections provided by Sabrient Systems or its wholly owned subsidiary Gradient Analytics. Sabrient Systems makes no representations that the techniques used in its rankings or analysis will result in or guarantee profits in trading. Trading involves risk, including possible loss of principal and other losses, and past performance is no indication of future results.