March 2002
Return Engagement
|
"The gates of hell are open night and day; Smooth the descent, and easy is the way: But to return, and view the cheerful skies, In this the task and mighty labor lies. " |
| -- Virgil, Aeneid |
UR QUANT PORTFOLIOS HAVE BEEN AROUND for over a year and half. While that's not a long time, it's still long enough to be able to discern some trends. Specifically, this is true in regard to Portfolio 3 since it's reformulated every two months.
A look at the bi-monthly changes is the starting point. These are listed along with the current portfolio on the Stocks of P3 page.
The first thing that jumps out at you is that over such a relatively short period of time, there have been a tremendous number of changes. In a volatile market, this may be expected, but is there a trend in the changes?
Sole Survivors
P3 holds the top 30 stocks from our quantitative model. At the close of 2001, only six stocks (Amgen, Medimmune, Mercury Interactive, Network Appliance, PayChex, and Yahoo) were charter members. The remaining 80% were the result of turnover. One might conclude that at least for this model, these six were core holdings.They certainly aren't a diverse group representing only three of the ten S&P sectors. Could that imply that the model works better for these sectors -- Healthcare, Industrials, and Technology -- than the other seven?
|
||||||||||||||||||||||||||||||||||||||||||
In an effort to answer both questions, we checked how a portfolio made of these core holdings measured up against the overall portfolio. Interestingly enough, their return both from P3’s inception (July 1, 2000) and for calendar year 2001 was almost exactly that of the entire portfolio. In essence, these six stocks yielded the return of P3 while the other 24 simply cancelled one another out.
That was unexpected, especially since it’s the other 24 stocks that are constantly coming and going. The point of reformulating the portfolio every two months is to allow it to adapt to market conditions. If all of the return is explained by the 6 core holdings all of the effort of reformulation goes for naught.
Déja Vu
And there’s a funny thing we found about those 24 other holdings: The same stocks keep coming and going.Though December 31, 2001 there were ten reformulations of the portfolio. During that time, there were 36 stocks that came and went. Of the 36, there were 8 that entered the portfolio, left, and came back again. It’s not supposed to work that way.
Stocks leave P3 when they are no longer in screen’s top 40. In order to enter, they must be in the top 30. We designed the model this way to prevent excessive turnover. Evidently we were too stringent in limiting existing holdings to the top 40. Perhaps the top 50 or 60 would have worked better.
But that wouldn’t necessarily solve the problem either. Stocks appear to move around pretty freely.
All Over the Place
Consider Palm. It originally entered P3 on October 16, 2000. At that point it was the highest rated stock not yet in the portfolio. On June 15, 2001, it removed when it fell to 54. In just two months it had climbed back to being the highest rated stock not yet in the portfolio. By October 15, 2001 it was again purged as it had fallen to 434 out of 500 stocks.Increasing the cutoff point might have kept Palm in the portfolio in June 2001, but nothing would have saved it four months later when it had fallen into the bottom quintile.
Palm hasn’t been the only stock to bounce around. After spending six months in the portfolio, Apple Computer was removed at 349. Broadvision was in P3 for four months before being removed in April 2001 after falling to 482. Most amazing was General Mills which was removed after four months, having fallen to 498 out of 500.
As volatile as these stocks are, others may be even more so since we only track stocks that move into and out of the portfolio. Those that don’t quite make it may be even more volatile just as those that do enter may have been in the lowest quintile at the prior reformulation
So What Does It All Mean?
Well-designed models shouldn’t be so volatile. If the basis for their regression is valid, it should measure fundamental features that don’t change so drastically in two short months.|
| |||||
|
The fact that the same stocks come, go, and come back again may simply mean the cutoff to be removed is too low. But the fact that stocks fall from the top 30 to the bottom of the lowest quintile is of much greater concern. It may undermine the fundamental factors selected for use in the model. If they yield such volatile results, they may not be capturing the key elements of return which was the goal of the model.
The existence of core holdings is more intriguing. The fact that they represent the vast majority of the overall portfolio return suggests that P3’s regression model can be focused on only a few sectors to capture the overall target return.
Unfortunately, while the core holdings do a good job of mirroring the overall portfolio, they fall just as short of the S&P 500 as P3 itself. Nevertheless, this may suggest a starting point for future models: Rather than trying to explain the benchmark’s performance by attempting to isolate fundamental factors across all sectors, it may be more appropriate to concentrate in only a few.
One of the good things about quantitative modeling is the fact that even when models fail, they can still be beneficial in pointing you in the right direction for the next model. Maybe P3 just did that.
Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
|||||||||||
| Homepage | Return to Top |
