New CEO Watcher scoring system

CEO Watcher Update

We spent the last couple of months analyzing the ~1M insider trades in our database to identify the types of insider purchases that outperform the S&P and the insider sales with negative returns.

We then used that data to create a scoring system for all unscheduled insider trades our system finds each day.

For CEO Watcher Premium subs, we have begun including this new data in the CEO Watcher Premium Dashboard Google Sheet and will include it in the daily emails going forward.

Here is what we found:

• Copying insider purchases is most effective as a value strategy (by pairing high-signal insider purchases with quality companies trading at reasonable prices).

• The highest-scoring insider sales (that correlated with negative forward returns) are almost exclusively from expensive, unprofitable, and dilutive companies.

• Companies with above-average insider buying over their previous 90 days perform well

• Industries with above-average insider buying over the previous 90 days perform well

• Companies with above-average insider buying, that are also in industries with above-average insider buying, perform even better

Let’s dive into the data and explore the new scoring system!

-

Quick notes before we start:

• All data is from Form 4 filings between 2010 and 2024.

• As always, we filter out scheduled and low-signal trades (10b5-1 plans, tax sales, stock purchase plans, public offerings, etc.) before any analysis.

• For insider purchases, we optimize for median 1y excess returns v the S&P.

• For sales, we optimize for mean 1y returns (looking for stocks that had negative 1y returns after the insider sold).

• All returns are calculated from the open price the day after the filing was made public

• The data is PIT (point-in-time) and includes companies that went bankrupt, were acquired, were delisted, etc.

-

Net Insider Buying in Companies (last 90d)

Companies with more net buying (num buyers - num sellers) in the previous 90 days than their 3-year average significantly outperformed companies with less net buying than their typical amount.

Net Insider Buying in Industries (last 90d)

We find similar results when looking at the industry level.

At the end of each quarter, we calculated the percentage of companies in each industry with at least one unscheduled insider purchase and compared it to the median value for that industry from the last three years.

The industries with insider buying at more companies than usual had better forward performance.

Net Insider Buying in Company + Industry (last 90d)

Unsurprisingly, combining them yields even better results.

We see similar results for insider sales, though the delta is far less drastic (and the sales don’t get anywhere close to negative returns on average).

Stocks with below-average insider buying (i.e., above-average net selling) in industries with below-average insider buying underperformed those with the opposite.

CEO Watcher Scoring System

The scoring system is based on the quality of the insider trade (cluster buying, reversal, transaction size, insider title, etc.) and the company’s fundamentals (buybacks, current valuation, dilution, debt coverage, etc.).

Our three main findings were:

1. Insider buying works best as a value strategy. Profitable companies at low valuations, returning money to shareholders, with high-signal insider buying, are the most successful buying strategy.

2. The high-signal insider sales that correlated with negative returns are almost exclusively found in expensive, unprofitable, dilutive companies.

3. Insider buying at banks trading below their historic P/TBV with moderate leverage and decent (or strong) ROEs performed very well.

The two caveats are:

1. You’ll have to get most of your tech and high-beta exposure elsewhere. That is not where insider buying shines (though software is seeing quite a bit of compelling insider buying right now, and we flagged a high-signal insider purchase at Marvell at the end of last year that is up 100%).

2. The short book gets hit hard during “shitco” rallies. Shorting the High Signal Insider Sales worked very well historically EXCEPT during the post-COVID rally when unprofitable companies absolutely crushed the market (more on this later).

For insider purchases, it’s actually not too hard to find insider purchases that outperform the S&P over the following year on average.

Just avoid the lowest-scoring purchases (<= 1 point).

However, to identify insider purchases with median returns that outperform the S&P, focus on the top-scoring insider purchases in our system (>= 10 points). ~3% of insider purchases score >= 10 (typically, a handful each week).

Looking at 2010-2024, excluding 2020, our scoring system for insider sales performs very well.

The lowest-scoring sales have the highest forward returns, while the highest-scoring sales have very negative forward returns.

The median returns chart is even cleaner.

Note: I had to remove the lowest bucket (scores 0-1) because ~90% of insider sales fall into it, which distorts the chart's bucketing.

However, if we look only at 2020, the chart does a complete 180: the lowest-quality companies with insider selling averaged nearly 100% 1y returns.

We backtested a portfolio from 2010-2024 (including COVID) that held only the highest-scoring insider sales, and it still posted slightly negative annual returns (roughly -2%).

However, COVID wiped out ~90% of the overall negative returns up to that point. Obviously, if you are a good investor, you’d manage those short positions better than our backtest (which was always 100% short the highest-scoring insider sales).

As always, the shorts need to be careful out there.

How to use the scoring system

1. Focus on insider purchases with scores >= 10. Avoid purchases with a score <= 1.

2. If looking for short opportunities, focus on scores >= 7. Avoid sales with a score <= 2.

3. Your own judgment as an investor should override the scoring system. I still own stocks with insider selling if I believe in the company (or want exposure to that theme, like AI right now). I also won’t buy a high-scoring insider purchase if I don’t like the company.

4. Remember, hot industries/themes tend to have a lot of insider selling (like AI right now). Those sales may score highly, but hot themes can run a lot further than you think (and insiders tend to be early). There is a very good chance you get your face ripped off if you try to short it. As the previous point said, use your judgment.

CEO Watcher Premium Dashboard Data

I have updated the CEO Watcher Premium Dashboard Google Sheet to include:

• Companies with the most net buying (v their 3y median) over the last 90 days

• Companies with the most net selling (v their 3y median) over the last 90 days

• Industries ranked by the amount of insider buying (v their 3y median)

• All unscheduled insider purchases from the last 90d (ranked by their score)

• All unscheduled insider sales from the last 90d (ranked by their score)

All the data in the Google Sheet is updated daily, and I’ll share screenshots in the Data Dump section of the daily emails.

The Google Sheet and Data Dump section are only available to CEO Watcher Premium subscribers.

Conclusion

The final takeaways:

• We optimize for the best returns over a 1y timeframe. CEO Watcher is not meant to be a day-trading data service.

• For insider purchases, a score >= 2 correlates with mean outperformance v the S&P. A score >= 10 correlates with mean and median 1y outperformance.

• For insider sales, a score >= 7 correlates with negative mean and median 1y returns.

• Copying insider purchases is most effective as a value strategy.

• Shorting the high-signal insider sales works incredibly well… until a “shitco” rally (like post-COVID) comes and rips your face off.

• Companies with above-average insider buying over their previous 90 days perform well

• Industries with above-average insider buying over the previous 90 days perform well

• Companies with above-average insider buying, that are also in industries with above-average insider buying, perform even better

I’ve shared the link to the CEO Watcher Premium Dashboard Google Sheet with premium subscribers in the Data Dump section at the bottom of this email (you will not see it unless you're a premium subscriber).

You can upgrade to CEO Watcher Premium at any time by visiting ceowatcher.com.

If you ever have any questions/feedback, email me at [email protected].

The rest of this email is for premium subscribers only.

CEO Watcher Premium

The rest of this email (including today’s top insider trades + all the charts and data) is for CEO Watcher Premium subscribers only. To see an example of the Premium email, click here.

Upgrade to CEO Watcher Premium at ceowatcher.com to unlock:

• The Premium daily emails

• Full access to the website, which has all insider trades back to 2009. It includes search/filters and the returns of every trade and every insider to help you find the top insider trades

• Full access to the CEO Watcher Premium Discord where I share my notes and holdings in the CEO Watcher Portfolio

• The Premium Dashboard, which has a bunch of charts/tables with insider trading data that is updated each day

If CEO Watcher Premium is not for you, I give full refunds with no questions asked.