Blog - Investing Notes

April 27, 2007 - Update: A Weighted Guru Groupthink

Suppose we take the current intermediate-term (notionally 3-9 month) stock market outlooks of the experts covered in Guru Grades and weight these outlooks according to current calculated guru forecast accuracies. Would the current accuracy-weighted aggregate opinion be bullish or bearish? Here's another update on weighting guru stock market outlooks.

As background, the following chart shows the distribution of gurus according to the accuracy of their forecasts for overall stock market direction, with accuracies segmented into intervals of 5%. The average (equally-weighted) accuracy of all 42 tracked gurus is 49%. Given the small sample size, the underlying distribution may be normal, even though the empirical distribution is lumpy. The approach outlined below reflects weighting the outlooks of gurus on the right (left) side of the distribution positively (negatively), with the magnitude of the weighting increasing with distance from the center.

The weighted group outlook development method is:

1. Applying judgment, identify gurus with current intermediate-term outlooks. We find 34.
2. Assign +1, 0 or -1 to each of these 34 gurus for bullish, neutral and bearish outlooks, respectively.
3. Weight each guru's outlook by the factor [Guru's Accuracy - 50%]. If a guru's accuracy is above (below) 50%, the weighting factor is positive (negative). For example, a guru with a poor forecast accuracy of 36% would have a weighting factor of 36% - 50% = -14%. If that guru were bullish (bearish), his contribution to the overall outlook would be negative (positive).
4. Sum all of the weighted guru outlooks.
5. Normalize the result such that the maximum (minimum) possible group outlook based on current guru accuracies is 100% (-100%).

The output of these calculations is +59%, with a contribution of +32% from gurus with historical accuracies greater than 50% and +27% from gurus with historical accuracies less than 50%. (The more accurate subgroup is net bullish, and the less accurate subgroup is net bearish.)

For comparison, we weight the 34 gurus equally and find a net +12% outlook, quite a bit lower than the accuracy-weighted outlook and, in fact, not far from neutral.

In summary, the accuracy-weighted intermediate-term group outlook of the gurus we track remains bullish and has become more so over the past three months. The more accurate half is bullish, and the less accurate half is bearish.

For past measurements of guru groupthink, see our blog entries of 10/25/06 and 1/24/07, when the groupthink outlooks were +31% and +37%, respectively. The principal change since the last measurement is that the more accurate gurus have become more bullish.

For other research on sentiment indicators, see Blog Synthesis: Sentimental Journey.