New York – We have looked at the relationship between performance and batting average on several occasions. However, the analysis was limited to our estimates of weighted average batting averages. In this article, we separate the buy recommendations from the sell recommendations and assess the relationship between performance and batting average for each portfolio. The results are interesting.
The buy batting average is the number of buy recommendations that produced positive returns divided by the total number of buy recommendations, expressed as a percent. The sell batting average is the number of sell recommendations that produced positive returns (i.e. price declines) divided by the total number of sell recommendations, expressed as a percent.
The batting average is an important indicator of predictive ability, but it does not necessarily translate into return. For example, a research provider could have a very high batting average where all of the stocks increased by a very small amount. Alternatively, a research provider could have a very low batting average and produce high returns, if several stocks went through the roof over the period of study. So the question remains, is performance affected by batting averages at all?
The following looks at this question using buy recommendations and sell recommendations. A concern, particularly with the sell portfolio, is that a research provider could have as little as one stock in the sell category and that stock could produce very large returns. To avoid this type of bias, we considered only firms that had more than 100 recommendations over the one year timeframe.
We used correlation and regression analysis to assess the strength of the relationship between buy performance and the buy batting average. In addition, we looked at the average return of each research provider per correct recommendation. The table below shows the top five research providers, who produced the highest returns per correct recommendation.
Top Providers Measured in Basis Points per Correct BUY Recommendation
|Basis Points per Correct
|B Riley & Co.
In a more formal analysis of the relationship between performance and batting average, we estimated the correlation between performance and batting average and regressed performance on batting average. The correlation between performance and batting average was found to be 0.0155 for the buy portfolio, suggesting that no relationship is likely.
In the regression analysis, the R-squared was 0.0807, the slope coefficient was slightly negative and the t-score of the slope term indicated that the batting average has no impact on the performance of the research provider.
An identical approach was taken to assess whether the sell portfolio returns were affected by the sell batting average.
Top Providers Measured in Basis Points per Correct SELL Recommendation
|Basis Points per Correct
|RBC Capital Markets
|Bank Of America
The correlation between performance and batting average was found to be -0.6809 for the sell portfolio, suggesting a negative relationship is likely. That is, correct sell recommendations lead to negative price movements and positive returns to the sell portfolio.
In the regression analysis, the R-squared was 0.4637, the slope coefficient was minus 0.6432 and the t-score of the slope term indicated that the batting average does have an impact on the performance of the research provider.
Our analysis indicates that buy batting averages have no impact on performance, but sell batting averages do have an effect on sell portfolio performance. Whether this is a result that is maintained in subsequent studies remains in question.
Note: The performance and batting average data is sourced from Investars, a research performance measurement, commission management and payment tools company. The analysis and presentation of the data has been done by Integrity Research Associates.