| |
 May 2009 Feels Like a Long Time Some Stocks Have Been in P3 for Years
This isn't the first time we've examined this phenomenon. Early in the model's run, we considered the fact that many stocks were dropped only to shortly return. Turnover was thought to be above average, but longevity? Not so much. We returned to the subject three years later and found that there was no correlation between the time spent in the model and overall performance. But now that more time has passed, we're still left wondering if those stocks that are the longest veterans of P3 have anything in common.
As you can see from Chart 1, most stocks don't stick around very long. The portfolio is reoptimized in the middle of each even-numbered month, so the minimum stay is two months. Sixty-six of the 194 component stocks have been included four months or less, and an additional 37 hung around for 5-10 months. Most of the remaining 91 lasted 40 months or less.
But if you look all the way to the right of the chart, you'll see a handful that lasted more than 78 months. That's almost seven years of the model's entire nine year history. Is there something special about them? Do they all have some significant features in common? By the way, this is more than just idle curiosity. This is, after all, a quantitative model. It was designed with the purpose of detecting funds with the highest probability of outperforming the benchmark index over the next two months. We started with a theory about which fundamental factors were most predictive of future performance. We included nine in an initial regression analysis to isolate the most effective and the approximate weights to create a weighted factor model. But results don't always match theory, so now with nine years of results to go on, the longest running stocks may shed some insight on precisely how the model is working.
The Veterans To put this in context, think back to what was happening in the market at the turn of the century. Just a few short months after P3 launched, the tech bubble blew up with a vengeance. P3 was heavily weighted in techs and communication stocks (essentially like all investors at the time) so quickly fell into negative territory. Like the S&P 500, it's been trying to recover ever since.
Yahoo was a momentum favorite prior to the tech implosion. Many investors purchased it in a fit of "panic buying", feeling they had to buy it immediately because it was only going to get more expensive as it defied gravity. When the market turned, Yahoo had plenty of sellers and fell harder than the market. At its worst point in 2001, it had lost roughly 95% of it's June 30, 2000 value. Over the next four years, it made a comeback, but never really got close to it's initial value. More recently, falling ad demand and competition with Google have driven the stock down again. Last year's decision to decline Microsoft's merger offer has it again approaching its 2001 lows. At the other end of the spectrum is Apple. Chart 3 shows its phenomenal growth, and even though it fell 20% in the past year, it's still more than 1500% (no, that's not a typo) above its value on February 13, 2003 when it was added to the portfolio. As a cult favorite, Apple has been able to not only set trends (iTunes, iPhone, iPod), its devoted followers have been willing to pay premium prices for its products. But the model doesn't consider these elements. It only focuses on five fundamental factors. It used those same factors to hold onto all six veterans: Yahoo, NetApp (90 months), Medimmune (88 months), Apple (80 months), Xilinx (80 months), and Allergan (78 months). Besides Yahoo, Apple and Allergan are still included in the model today. (It's only fair to say that Medimmune might have been if it hadn't been purchased by AstraZeneca in 2007.)
The Performance Over the past year, returns for the five stocks (excluding Medimmune) have ranged from -49% for Yahoo to -9% for Xilinx. Keep in mind, over the same period, the benchmark S&P 500 was down 42%. As a group, these stocks did much better. Of course, only three of the five were in P3 for the entire year. In reality, it's not actual performance that keeps stocks in P3, it's specific quantitative factors. Five factors determine the constituents: Return on Invested Capital (ROC), Price to Book Ratio, Price to Cashflow Ratio, Forward P/E and Last Earnings Surprise. Of these, ROC and Last Earnings Surprise carry the most weight. Presumably returns follow these factors -- at least that's how the model was created. Currently Yahoo continues to have the best rating (33) of the five remaining stocks. (Again, Medimmune can no longer be included due to its acquisition.) Yahoo has always had a strong ROC, but interestingly enough, it's currently the lowest of the five (6.3%). Last Earnings Surprise (31%) is the highest of the five.
NetApp currently has the poorest score (158). It was removed from the portfolio in August 2007. It has a decent ROC (16.3%) but the Last Earnings Surprise wasn't one; it was 0%. That doesn't help. Forward P/Es for the group range from 17.1x (NetApp) to 38.5x (Yahoo). Price to Cashflow runs from 11x (NetApp) to 20.8x (Apple). Price to Book goes from 1.8x (Yahoo) to 4.5x (Apple). Then again, Price to Book isn't really a very good measure of these five companies. Of the five, four are Tech stocks, and two (counting Medimmune) are Healthcare. Price to Book is not a very meaningful measure for these sectors. It's more appropriate for capital intensive businesses and industries such as financials and industrials. It's no wonder then that P3 has always had a high level of Tech and Healthcare stocks in light of their above average Price to Book ratios.
Weight Watching
This isn't an issue of rapid-fire trading and market timing. Truth be told, turnover in P3 has been remarkably low, 68.8%. That figure has been even lower (31.0%) since the 2003 changes. So it's not like the model is chasing return. What it has been able to do, however, is acquire and hold stocks when they were making a good run. Look no further than Apple's chart to see that. But there's more to it than simply getting a hit now and then. P3 has had the uncanny ability (at least since 2003) to generally have one or several outperforming stocks at all times. As an example of this, compare Apple's chart to that of Allergan (Chart 4). You'll notice Allergan had a steep runup from April 2003 through the end of 2005. Apple took off in late 2005 and completed a double-top in mid-2008. In essence, Allergan handed the performance baton to Apple in 2005 and Apple took it from there. All stocks in the portfolio don't have to be major winners, you just need one or two at a time like these. P3's concentration helps support this approach. With only 30 stocks in the model, any constituent is heavily overweighted relative to the benchmark index. For example, the three current components, Yahoo, Apple, and Allergan represent 1.35%, 11.55%, and 4.63% of P3, but only 0.26%, 1.46%, and 0.20% of the S&P 500, respectively. Superior performance by any of these holdings will have a considerably greater impact on P3 than on the S&P 500. The result is a more volatile, but potentially better performing portfolio. Of course this volatility can work against the model as well. This is what happened shortly after inception when its concentrated growth issues -- particularly Technology -- took a hard fall. Just as outperforming stocks have their impact magnified in P3, so do dramatically underperforming ones. To this day, P3 is still trying to dig itself out from that rough beginning. When all's said and done, there isn't one particular factor or feature of the long-running constituents that binds them all together. Although they all come from traditional growth sectors (Technology and Healthcare), this may help explain why they were initially selected, but not why they remained so long after inclusion. Given that all five (with the glaring exception of Yahoo) did have periods of stellar performance once added to the portfolio, one might conclude the model is working properly: It's selecting stocks that will go on to beat the market. Once they're selected and moving up, it only makes sense that they would remain in the portfolio. But this conclusion may be a little strong for the supporting evidence. Yahoo has been in the model since inception, but with the exception of a few months in the beginning, it's never been a stellar performer. Nevertheless, it remains in the portfolio. Is its inclusion just a fluke or is the outperformance of the other long-timers the oddity? The data here isn't sufficient to answer this question -- at least not yet. For now, all we can conclude is that P3 is batting .800 on it's five long-term holdings. That in, and of itself, isn't a bad average. Search this site! Just enter you key word or words: Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
||||||||||||||||||||||||||||
UANTITATIVE PORTFOLIO 3 has been around for about nine years now. During that time, 194 unique stocks of the S&P 500 have appeared. Some have been merged or acquired, some have gone out of business, and others have been dropped from the index, yet at one time or another, almost 200 different stocks have visited this 30-stock portfolio.
