Are Insider Trades Informative?
If you google the four words above, you’ll get well over 27,000 results.
Stocks being purchased by corporate insiders tend to outperform the rest of the market, while stocks being heavily sold by insiders tend to underperform. For about fifty years this has been considered gospel in academic circles, but it has been almost impossible to make money from it in the real world. Three advisory services tracked by the Hulbert Financial Digest have based their strategies on insider behavior. Not one of the three has fared well.
Only one is still published today. The Vickers Weekly Insider Report was started during the beginning of 1993. Over the ten years ending June 30, 2014 it returned 2.40% annually while the Vanguard 500 Index Fund returned 7.66% annually and the Vanguard Small Cap Index Fund returned 10.08% annually. That performance ranks the Vickers Weekly Insider Report 93 out of 111 newsletters monitored for ten or more years by Hulbert.
During early 2012 a WW member told me that he was only buying picks that had a history of more insider buying activity than insider selling activity, and he suggested that this strategy could add value to the recommended picks. Your intrepid newsletter editor immediately wrote code to download insider activity on the 1,000 plus stocks that report increased earnings every quarter that also passed my preliminary filters. This was a time consuming PITA because the data were downloaded from NASDAQ’s website, and that website was incredibly slow and frequently unreliable, but the promise of added value kept me going.
After the increased earnings were announced, I filtered the stocks, and then downloaded their price history after 3-1/2 months had elapsed. Here are my results for seven quarters of research ending with 2014Q1.
IMPORTANT All return values shown are for three month holding periods i.e., a 7.30% return for three months is equal to a 32.56% compounded annual return.
7,672 companies passed my preliminary filters during 2012Q3 thru 2014Q1
Stocks were purchased at the closing price one trading day after the earnings were announced, and sold at the average HI-LO daily price three months later.
Transaction costs are not included in the reported returns below.
7.3% average three month return for all 7,672
13.2% average three month return for 540 that passed the WW filters
So far, very good.
How well did the stocks insiders were buying perform?
During the three months before the earnings were announced, there were 2,665 stocks that insiders were buying more than they were selling. The average three month return of these 2,665 stocks was 7.3%, showing that excessive insider buying added no value to the return of the entire population.
How well did the stocks insiders were selling perform?
During the three months before the earnings were announced, there were 3,287 stocks that insiders were selling more than they were buying. The average three month return of these 3,287 stocks was 7.5%, showing that excessive insider selling added insignificant value to the return of the entire population.
How well did the WW Picks insiders were buying perform?
During the three months before the earnings were announced, there were 189 WW Picks that insiders were buying more than they were selling. The average three month return of these 189 WW Picks was 9.0%, showing that excessive insider buying SUBTRACTED considerable value from the 13.2% return of the entire population.
How well did the WW Picks insiders were selling perform?
During the three months before the earnings were announced, there were 193 WW Picks that insiders were selling more than they were buying. The average three month return of these 193 WW Picks was 17.8%, showing that excessive insider selling ADDED considerable value to the 13.2% return of the entire population.
The results shown above were based on insider transactions that occurred three months before the earnings were announced. Results for insider transactions that occurred twelve months before the earnings were announced were consistent with the three month results.
Conclusion
The results shown above are based on 1-3/4 years of data. This is not a long enough period to prove anything, but the results are interesting, and the results are more or less consistent with insider behavior. Insiders tend to be contrarian value investors i.e., they buy stocks with poor past performance and with low Book to Market ratios, and they tend to sell stocks that performed well in the past.
Josef Lakonishok and Inmoo Lee conducted a study on more than one million insider transactions from 1975 thru 1995. They concluded that insiders were generally sellers of small growth stocks, and they concluded that insiders knew what they were doing when they sold those stocks.
That is not consistent with my results, but my results are based on a considerably shorter period than theirs. The bottom line is that almost anything can happen when one plays the investing game. If past performance were an accurate indicator of future performance, then investing would be easy, and everyone would be rich.
And that is clearly not possible.