New Performance Ideas – More Sneaky than Geeky

New York – We have been writing on aspects of informational alpha, or synthetic returns of research providers for about 5 years now. One might think that we have exhausted our ability to analyze research provider returns in new and exciting ways, but this underestimates our geekiness.

 In the table, there are six columns of numbers. The first three columns are directly from the Investars system and relate to the absolute returns produced by the recommendations of the top research providers. The second three columns attempt to analyze the same numbers, but only for those recommendations that were directionally correct.  This is done by dividing the returns by the number of recommendations that actually moved in the right direction (utilizing the batting averages).

 But research provider recommendations always have some right and some wrong calls. Why on earth should we give them credit only for the recommendations they get right? The answer is related to the conviction of the recommendations. By necessity, recommendations are translated into a three category rating system by Investars, so that the firms can be compared to one another on an equivalent basis. However, many firms have differing recommendation structures.  For example, one firm many have strong buy, buy, neutral, sell and strong sell. As well, quantitative firms tend to rank stocks in deciles.  Obviously, the tails of the distribution, say the top and bottom deciles are more likely to move in the expected direction than are say the top and bottom 3 deciles.

 While we cannot recreate a more granular rating system, we can make an assumption, that a larger proportion of the returns from correct calls are generated by those calls in which the research provider has stronger conviction.

 Analysis of the Returns per Correct Recommendation

 By this admittedly clunky logic, the B-S column on the far right of the table represents the overall returns of the research providers’ from the calls that were in the correct direction. B. Riley & Co. was the winner by a long shot, racking up nearly 74 basis points per day in the year ended July 2009, and more than doubling the next closest provider. Ativo Research was second in line with 36 basis points return and Sandler O’Neil came in third, with a return of about 21 basis points per day.  Market Profile Theorems, Rochdale, DA Davidson and Northland Securities all produced double digit returns.

 Another interesting insight from the data is the contribution of buy and sell recommendations to the overall return. For example, B. Riley & Co. had a very strong overall return, mostly based on the sell portfolio. Ativo Research had roughly equal contributions from the buy and sell portfolios and racked up the strongest return on its buys of any firm. Market Profile Theorems’ 19 basis point per day return was dominated by the buy recommendation portfolio of the firm. Market Profile Theorems’ buy portfolio provided over 80% of the total return produced.

 Despite the obvious assumptions applied to this analysis, the results remain interesting and helpful to the analysis of overall performance and informational alpha provided by the research providers presented.

 Note: The performance data is from Investars. The analysis and conclusions are Integrity’s.