Earlier this week, I suggested a different way to look at the Keeneland yearling sales numbers. This is a follow-up to that post.
My equine finance professor at the University of Louisville, along with a fellow student, compared the sales results from 2009 to 2010 by using the percentage of horses offered instead of comparing day 1 to day 1, day 2 to day 2, etc as most people have been doing. Since Keeneland changed its format this year, the percentage method makes more sense. Here's an update on how the sale is progressing:
Update on Keeneland Sale Averages from Kate Ellis and Robert Losey
Scott Jagow was kind enough to let us use his blog as a sounding board on Sept. 21, where we argued that the cumulative averages for the September Keeneland Sale were overestimating what is really happening. We also predicted that the cumulative averages reported in the media would appear to drop substantially each day because of the unrealistic comparisons that result when session by session comparisons are made instead of comparing similar segments by percentage completion of the sales. We expect that when the sale is completed, the cumulative average (both ours and the reported in the media) will be up by 1-3%. If we’re right, this suggests two things: 1) the market may have bottomed out and is hopefully starting to turn up (we predict an increase in the November Sale averages because of diminished supply effects), and 2) there is a better way to compare sales from one year to the next than what the media has been reporting.
The first two columns below report our calculations for the sale. In the last column we provide the media calculation of the cumulative average as reported for days 9-11.
Sale Change Ellis/Losey Media
From Comp Calculation Calculation
In 2009 of Cumulative of Cumulative
Day Change Change
1 +49% +49%
2 +44% +46%
3 -41% -6%
4 - 7% -6%
5 +18% -2%
6 +39% +3%
7 +2% +2%
8 +29 +4%
9 -27% +2% +24.3%
10 +1% +2% +12.3%
11 -10% +2% + 9.9%
Note that the first two sessions appear to be anomalies, and they are. This is the first time that Keeneland has used night sessions for days 1 and 2, and the “super select” nature of the first two sessions resulted in a skyrocketing average for those sessions. The third session drastic drop off is also predictable. Keeneland attempted to equalize the quality of yearlings for sessions 3 through 7 this year, while in 2009 session 3 was probably the third best group of yearlings. Thus the comparisons of sessions 3 and 4 this year suffered, while the comparisons of session s 5 and 6 this year were favorable relative to similar segments last year.