The Insider Trading Anomaly: How to Make 7% More Than Index Funds

Oct.31.10 | About: SPDR S&P (SPY)

The insider trading anomaly has been one of the most profitable investment strategies in the past 50 years. In general, information is a valuable commodity. In this case, the insiders have a great deal of information, more so than do any outsiders (journalists, bloggers, etc.) or stock analysts covering the company.But just having knowledge isn’t what gives the insiders their advantage.

The financial advantage comes from the prowess to derive valuation estimates based on that information. Sure, sometimes insiders make mistakes; they make the wrong assumptions or they can’t predict the effect of macro factors. But the probability that this happens and they come to the wrong conclusion is much smaller when more than one insider, acting independently, come to the same conclusion: buy, buy, buy.

In other words, consensus criteria reigns king. When there is consensus among insiders (remember, who are acting independently), returns to insider trading are much higher.

Back in 1968, Professors James H. Lorie and Victor Niederhoffer had one of the earliest academic papers on profitability of insider trading. It had a relatively proper methodology. Around the same time that economist Eugene Fama was preaching his efficient markets theory, Lorie and Niederhoffer showed that under intensive trading criteria (number of buyers is at least two more than the number of sellers or vice versa), insiders will outperform the market during the next six months.

That’s nothing to laugh at.

The studies that followed had the same purpose of testing strong and/or semi-strong forms of efficient market hypothesis, each time employing an improved methodology. Wharton professor Jeffrey Jaffe’s 1974 study didn’t find abnormal returns from the zero investment portfolio that was long in companies with more insider purchases, and short in companies with more insider sales in any given month.

Reducing the sample to only large transactions (at least $20,000) also didn’t change his conclusion. But when Jaffe used an intensive trading criteria – where there needed to be at least 3 buyers from each company to be included in the long portfolio and 3 sellers from each company to be included in the short portfolio – he found abnormal returns of 5.07% in the first 8 months following the transactions.

When he included each company in the portfolio 2 months after the transaction, he still found abnormal returns of 4.84% in the first 8 months. His conclusion was that outsiders can also reap abnormal profits by imitating insiders.

University of Michigan’s Joseph Finnerty’s 1976 paper didn’t use any intensive trading criteria, and yet concluded that insider purchases have abnormal returns of 4.1% in the first 6 months. It also concluded that insider sales have abnormal returns of -2.4% in the first 6 months. Finnerty found that half of these returns are realized within the first month. That made him refute the strong form of market efficiency, but not the semi-strong form.

In 1986 University of Michigan professor Nejat Seyhun used a better methodology to measure abornormal returns. He incorporated the “size effect”. In his sample, insiders in small firms had predominantly more purchases than sales.

Insider Monkey also observes the same pattern in our dataset: 60% of transactions are made by insiders in small firms, even though in dollar terms this corresponds to only 16% of all transactions. Seyhun found abnormal returns of 4.3% for the first 300 days for firms with more insider purchases than sales, and -2.2% in the same period for firms with more insider sales than purchases. He noted that employing intensive trading criteria yielded similar results.

Studies conducted until around this time were based on random subsets of insider transactions . Remember, they didn’t have comprehensive databases or powerful computers like those that are available now. More recently, studies utilize the entire insider transactions dataset and remove any remaining doubts about the profitability of insider trading.

In 2001, University of Illinois at Urbana Champaigne’s Josef Lakonishok and Inmoo Lee corrected for size and B/M effects by using a different trading criterion. They too concluded that the difference between strong buy and strong sell portfolios, which excluded transactions of large shareholders, is 4.8% for the first year after the transactions. They didn’t find any abnormal returns for large shareholders.

Two years later, Harvard University’s Leslie A. Jeng and Richard Zeckhauser, and Yale University’s Andrew Metrick calculated insiders’ actual returns for their purchases and sales. They didn’t use any sort of screening criteria such as consensus, or “intensive purchases” that are used by other researchers. They also started calculating returns as soon as an insider made a transaction, not when the transaction become public. They found that insider sales are not profitable but insider purchases are extremely so. In raw returns, insider purchases beat the market returns by 11.2% per year.

And this return is not risk adjusted. By comparison, when a mutual fund announces they beat the market by more than 10%, they’re aren’t adjusting for risk either. The risk adjusted excess returns for the Harvard professors’ study are 52 basis points per month (or 6.4% annualized) using Carhart’s four factor model (Risk adjusted returns are 8.5% if you use CAPM to adjust for risk).

Achieving 6.4% annualized returns without utilizing a consensus criteria is insane. And unbelievably, they found that purchases in small firms don’t earn significantly higher returns than do purchases in large firms.

It’s been empirically shown that insider trading has beaten the market indexes over the past 50 years. Academic studies have been publicizing this for 40 years. Yet, the markets aren’t efficient enough yet to eliminate these abnormal returns. Insider Monkey, believes that it’s still possible to beat market indexes and achieve abnormal returns in the long run by utilizing insider trading data. Markets aren’t as efficient as you think.

Disclosure: Author is long SPY