| |
 January 2002 Apples to Apples
But what what we do always point out is the portfolio's growth bias. Both were created using regressions of fundamental characteristics to the S&P 500's performance over the decade of the '90s (see sidebar below). Unfortunately, right around the time we created these models, value resumed its dominance of growth. Leadership in the S&P 500 reverted to value stocks and our new models quickly fell behind the index. As long as value continues to outperform, P3 and P4 will continue to underperform. No apologies there -- they're de facto growth portfolios. Growth and value come in and out of vogue so over a longer measurement period, they may pick up some of their lost ground. Obviously they haven't tracked the S&P 500 very well over the short-term so as a proxy for that index, the models have failed. But once you realize they're growth portfolios, the more interesting question is how do they stack up against growth indexes? The Growth BenchmarksThe first step is to find the appropriate growth benchmarks. Most major indexes are broken down into growth and value components, providing plenty of alternatives.
P3 and P4 are based on the S&P 500 so by definition, are large cap portfolios. Because of this, we only considered large-cap growth indexes. The S&P 500 Barra Growth index is the logical starting point. This benchmark is drawn from the stocks of the S&P 500. Its creators divide the main index into growth and value, basing the division on fundamental factors such as price-to-book. The style indexes are reconstituted on a quarterly basis. Similarly, the Russell 1000 large cap index is also divided into two style indexes. Again fundamental characteristics are the basis of the division. Like all Russell indexes, the Russell 1000 Growth Index is only reconstituted once a year in July. There are other large cap growth benchmarks such as Wilshire's, but information about their fundamentals isn't as freely available. Given that, we decided to stick with the S&P and Russell measures.
Chart 1 compares the benchmarks to P3 and P4 since the latter's July 1, 2000 inception. The underperformance relative to the S&P 500 has been well documented (see Work in Progress Archives), so that's no surprise. Despite being reconstituted every two months, P3 has lagged more than P4. P3 looks a little better when compared to the S&P Barra Growth and Russell 1000 Growth, yet still suffers from its underperformance in February and March of 2001. It also lagged both before and after the events surrounding September 11th. The nearby table shows its performance as well as that of the benchmarks for specific periods through November 30, 2001. P4 fared better. Up until September 2001, it actually outperformed the S&P Barra Growth and Russell 1000 Growth. Since then, it's fallen back to their levels. Most notably, in the July 1 - November 30th period, it's trailed the benchmarks by double digits. This is significant for P4 since like the Russell indexes, it is only reconstituted once a year on July 1. It's quite possible this portfolio was whipsawed last year by taking on a more value-oriented cast when reformulated only to see growth reaccelerate after September 11th. Going InsideTo check this out we compared the portfolios' fundamental characteristics to those of the benchmarks. Information was more limited here, so we focused on the S&P 500 Barra Growth.
Table 2 shows growth rates and price ratios for each. P3 is clearly the most aggressive of all. Its growth rates are significantly above the benchmark's as are its valuation measures. This is understandable since its stocks are selected based only on their performance (growth) characteristics. P4's growth characteristics are also well over those of the index, yet its valuation measures are more in line. That's probably a result of its sector weighting. When reconstituted, P4 is designed to mirror the sector weightings of the S&P 500. With representation in all sectors, P4 must hold a stocks in traditional value sectors as well as growth. Not only would this affect its valuation, it could also explain the relative underperformance to the growth indexes in the July - November period in 2001. The Weighting is the Hardest PartBear in mind that even P4 is designed to reflect the sector weighting of the S&P 500, not the S&P 500 Barra Growth Index. Given that, it wouldn't be surprising to find its weights to be much different from the index.
On the other hand, P3 being much more fluid should look more like the growth indexes. Since it's reformulated every two months, it should more closely reflect the sector performance of the best performing growth stocks. Also, since there's no explicit sector weighting requirement, it's never forced to hold stocks in traditional value sectors. To test this, we compared the sector weightings (as of November 2001) vs. those of the the S&P 500 Barra Growth Index. The results are shown on Chart 2. As expected, P3 is much closer to the growth index than P4. Six of the ten sectors are essentially market weighted. The major overweights are Consumer Discretionary (a cyclical sector) and Healthcare, a traditional growth sector. Industrials, a traditional cyclical value sector, is the the major underweight. This is what you'd expect. What you wouldn't expect is that Technology -- the poster child for growth in the '90s -- is the second greatest underweight. This is probably a function of the dramatic tech selloff of 2000 - 2001. It's worth noting that when reformulated in December 2001, Technology's weighting almost doubled to over 30% of the portfolio. This would actually move Technology to an overweight position. Again, this is understandable in the context of tech's 40% plus gain since the September 21st low. P4 looks pretty much how you'd expect. Overall, it's much further from the growth index, with only three sectors within 3% of the index weights.
The overweighted sectors tend to be value while the underweighted ones are growth. As we've noted before, this is a function of the sector weighting requirement relative to the S&P 500. Unlike P3, P4 is forced to hold stocks from the traditional value sectors. It Grows on YouWhich brings us back to the performance issue. P3 acts more like a growth portfolio but is somewhat backward-looking. It's more frequent reformulation allows it to more quickly reflect the trends in growth, yet it is somewhat momentum driven. When trends shift as they did in late 2000 and early 2001, it gets left behind and takes too long to adjust to make up lost ground. On the other hand, if growth were to resume its dominance, P3 would offer tremendous opportunity.P4 is less growth-oriented given its sector weighting requirement. When growth falls out of favor, P4's value holdings -- no matter how slight -- help keep it from suffering the sharp losses of P3. Nevertheless, in a growth dominated market, P4 would fall behind P3 as well as the growth benchmarks. Over the past year and a half, both portfolios have compared more favorably with growth measures than the overall S&P 500. Given the growth-bias inherent in the data from which they were created, this isn't unreasonable. Even so, they were not supposed to be growth indexes, they were supposed to mirror the S&P 500. From a statistical standpoint, this is an important point. P3 and P4 may ultimately prove to be excellent growth models, but they weren't meant to be. Even if they ended up with perfect correlation to growth benchmarks, it would not be a success for the model, just an interesting coincidence. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
INCE THEIR INCEPTION EIGHTEEN MONTHS ago, every mention of our quant portfolios (P3 and P4) has sounded somewhat apologetic. Sure, the performance could be better, but the whole market is down which explains a lot of the problem. We don't apologize for that.
