We recently acquired a database of 177,000 insider buys and sells over the last couple of years.  Naturally, we dove right in to the process of divvying up the data for analysis.  We'll throw some preliminary tidbits to our viewers right now, but first we should make it clear that we are not at liberty to share the actual data.

First, we examined only the "buys".  Better than 90% of insider transactions actually fall into the "sell" category.  That might seem odd, but one must remember that a large chunk of insider holdings result from stock grants and the exercise of options and warrants, so it's quite possible for insiders to be burdened with stock that they never intended to acquire in the first place.  Thus sales can easily outnumber buys.  A quick look at the data proves that buys, in general, are fairly strong predictors of nice future gains, but sales are quite weak as predictors of future losses.  In the broadest terms, a "buy" signals yearlong gains that exceed the S&P 500 by better than 8%, while the average stock undergoing insider selling underperforms the S&P 500 by an insignificant 1%.  The 8% figure is very much in line with the results of longer-term studies conducted by others.

What exactly did we try to predict?  Rather than predicting the 12 month performance that follows an insider purchase, we looked at the last 11 months.  In other words, we excluded the first month.  That may seem odd, but this allows us to ask, "does the performance of a stock in the first month following insider buying have any bearing on the next 11 months?".

Going straight to the most eye-popping of results, we found that the price of a stock that is being purchased by insiders is the strongest predictor of future gains.  The cheapest 20% of such stocks averaged 18% gains above the S&P, the cheapest 10% averaged 33% gains, and the cheapest 4% averaged 51% gains.  Rather outrageous, but we did double-check for reverse splits.  Of course, we're only looking at the last couple years of data.

One objection to the above results might be that the market simply favored cheap stocks over the period of time examined.  There's some truth to that.  However, the cheapest 20%, 10%, and 4% of stocks undergoing insider selling outperformed the S&P 500 by only 5%, 8%, and 13%.

Conversely, expensive stocks paralleled or even under-performed the S&P 500. 

Stocks with nice pops shortly after the insider buying had occurred (on a scale of 3 days to one month) tended to continue moving upward.  However, a large monthlong loss (8% or more) also predicted decent gains.  What you don't want, it would appear, is a stock that stagnates in the midst of all this insider activity. 

What would you like to see before an insider purchases, gains or losses?  The data was fairly clear...you'd like to see losses.  When insiders pile on following large % losses, they're telling the investment world that they believe the stock has been unfairly punished. 

What is better, a stock that has large numbers of SEC filings detailing a long trail of purchases, or just a few such filings?  According to our data, you don't want to see long lists of purchases.  Stocks with such long lists (73 or more SEC filings over the course of the data) actually tended to underperform the indices.  You might rationalize the result this way:  frequent trades stink of an attempt to play the market in some way (e.g. tweaking the bid/ask spread to grab an extra .5%), whereas less frequent trading may signal a stronger commitment.

If the stock does have a long list of buyers, the latecomers to the party (those with the most recent purchases) will tend to do the worst.

How about the actual $ figures involved in the trades?  Do large trades tend to serve as better predictors than small trades?  Here, moderation seems to predict gains.  The trades should involve neither huge sums of money, nor exceedingly petty sums.  Trades on the order of $10,000 to $30,000 work nicely.  Again, there may be some rationale here:  huge purchasers may be more interested in controlling the company than in garnering large % gains in the upcoming months.  And small-time purchases (we saw numerous trades of precisely 1 share!) may be more related to depositing a goody in Junior's Christmas stocking than anything else. 

The figures above, relating to the size and number of transactions, don't really surprise us.  We don't find any of it to be particularly counterintuitive.  Searching the net, however, it does seem that a number of websites that specialize in reporting insider data make the a priori assumption that big and numerous trades are more bullish than smaller, less frequent trades.

Insiders can be divided into four groups in the SEC filings:  officers, directors, 10% or more owners, and "others".  Which group tends to have the most success?  The answer was fairly clear:  the officers.  Simply following officers in their purchases allowed one to beat the S&P by 11%.  By comparison, the "10% or more" group pretty much paralleled the indices.

Another way to look at insider data is to seek out individual insiders who have been particularly successful in their trades.  Such a study would require a longer time span than we're looking at.  In fact, we didn't identify a single insider who made a habit of buying and selling on a frequent basis.  None of these guys look anything like day-traders...given SEC rules, they couldn't daytrade even if they wanted to.

We did identify insiders who reaped huge gains nearly immediately after their purchases.  One Gary Evans over at NVAX saw a 142% gain after his initial purchase.  The tendency is to accuse the guy of corrupt practices, but then again, is it really so odd to see one trade out of several thousand jump 142% in a couple days?  We also saw supposedly prescient insiders lose 20 or 30% within three days of their initial trades.

How about individuals or groups that appear as inside buyers in more than one company?  We saw folks who held inside positions in as many as 30 different stocks.  It's almost bizarre how poorly these folks made out in their investments, however...you've got to go down the list to "Steel Partners II", who held 8 different stocks, before you find an entity that outperformed the S&P 500.  Goldman Sachs Group, with 7 different purchases, was the only multiple purchaser that seemed to excel in its decisions, averaging about 30% over the S&P 500.  Even then, the percentage gains were skewed by one very nice decision...to purchase big chunks of "Lazard Ltd." (laz).  It should also be noted that Goldman Sachs was rather weak in its sell side decisions...a number of its holdings went on to show big gains after a sale.  Of course, it's possible that Goldman Sachs used the sale to pick up even stronger equities, so perhaps we shouldn't be overly harsh in punishing insiders for seemingly poor sell-side timing.

Steel Partners II, by the way, is an interesting situation.  The group is referred to as a "hedge fund", but their strategy is not the typical hedge strategy.  These guys seek out situations where they believe a fundamentally sound company is being dragged down by poor management.  They then do what it takes to boot out the directors and management (read "hostile takeover" or "proxy battle"), and try to rebuild the company.

On a monthlong scale, did any holders of multiple companies seem to excel?  The answer, simply, is "no".  Some entities averaged as much as 16% gains in the month following a purchase, but once you examine the data, you see that you're hardly looking at a pattern of buying that's likely to repeat (e.g. the data was skewed by one very large gain over the course of a month).  What we'd like to see is a pattern of consistently outperforming the market over the course of the test, with gains in multiple stocks, and relatively few losses.  We didn't see that.

****************

As might be expected, our examination of insider selling yielded results that weren't as clearcut or dramatic as those for insider buying. Insider selling may mean nothing more than a desire for a new house.  As with our test of insider buying, we looked at the 11 month period beginning one month after the initial transaction. This allowed us to test whether the price performance in the month following a sale has any predictive value.

The most interesting result, though not unexpected, is simply that a strong monthlong performance preceding a sale is a fairly strong predictor of losses.  Here, the insiders seem to be saying that gains in their stock are overdone.

Transactions of large market value (greater than $250,000) do seem to predict greater-than-average losses in the future.

Looking at the various designations of "insiders", selling on the part of the "10% or more" holders was the strongest predictor of future losses.

Again, inside sellers are hardly immune to serious timing errors. We saw gains as high as 70% within three days of a sale, and as high as 500% within a year. We did, of course, see possible examples of insider prescience as well...losses as high as 46% within three days, and nearly 100% within a year.

The natural tendency is to assume that insiders who buy before a big spike, or sell before a major drop, are cheating. Undoubtedly this occurs to some extent. The opposite should also be considered, however...some insiders, in their zeal to be perceived as ethical, may actually make moves that are not in their financial interest. In this case, it might prove profitable to pay attention to these special insiders when they make their moves.

Copyright © 2008 MarketSynopsis.com. All rights reserved.