 November 2003 Mid Cap Preference
Unless you want to game the index by going outside of it -- and many "successful" mutual fund managers do just that -- you need to find a manageable way to mirror and exceed the index. That's what we did when we created Portfolio 5. Its benchmark is the S&P 1500, the combination of the S&P 500 (large caps), S&P 400 (mid caps), and the S&P 600 (small caps). Picking and choosing among 1500 stocks and then
To do this, we divided the "equity market" into nine different segments as represented by the Morningstar Stylebox. In essence it splits the benchmark by capitalization (small, mid, and large caps) and then each of those pieces into style (value, core, and growth). Rather than relying on individual stock selections, we instead chose to utilize exchange traded funds representing the nine different categories. As a result, the entire S&P 1500 could be approached through a combination of two S&P Depository Receipts (SPDRs) and seven Barclay American iShare ETFs. Our optimization model told us which to use and to what extent making the quarterly rebalancing an easy task. Indeed, this process is so simple (and cost efficient) it's something individual investors can do at home. The Road Never TakenSo far this approach has been successful. Chart 1 shows the relative strength of Portfolio 5 to both the S&P 1500 and the better known S&P 500. Except for an eight month period from October 2002 to June 2003, P5 has outperformed both benchmarks. By the end of September, it led the S&P 1500 by almost 2% and the 500 by nearly 4%.So what's working? Why has this model succeeded where others have struggled?
With only nine potential holdings and those mirroring indexes, it's logical to assume P5's performance is really a function of its asset allocation. In other words, for this portfolio, any value added does not depend on individual security selection but rather on the mix of capitalization and style. Although Portfolio 5's official start date was January 1, 2002, we have backtest data going all the way back to January 1, 1994. For our purposes here, we considered all 116 months of data extending from 1994 through August 2003. Although the pre-2002 information is index data from Ibbotson Associates, it's use is appropriate given that P5's ETFs so closely follow the benchmark indexes. One of the first things we noticed was that the indexes themselves were almost never used in the model. In setting up the Stylebox, we used the S&P 500, 400, and 600 to represent the "blend" approach for each capitalization class. At times when a specific capitalization was leading the market, it would be reasonable to think both value and growth could be performing well, and in that case the blended index rather than one style or the other would be preferred.
That might be a reasonable assumption, but it really hasn't played out over the 10-year test period. As you'll notice from Chart 2, the S&P 500 and 400 were never utilized, while the Small Cap 600 was employed for only six months or 6% of the time. If you reduce the measurement period by ten quarters to begin with October 1997, even the Small Cap 600 isn't utilized. This is illustrated on Chart 3. Clearly style was -- and is -- important. This even held true in the late 1990s when large caps dominated. In essence, it wasn't large caps that outperformed, but rather large cap growth. Value investors remember this all too well. Middle ManagementGiven that style is important, another thing that jumps out from Chart 2 is the reliance on mid cap growth. Of all the individual styles, at 63% this is the most heavily used one. This is especially remarkable since it occurs in a decade dominated by large caps. Indeed, at the capitalization level, large caps do win out 95%-86%, yet the dominance is not as great as would be expected. Again if you consider the period October 1997-August 2003 as shown on Chart 3 even this trend is reversed. There mid cap growth is employed 88% of the time and mid caps dominate large caps 108%-96%. Perhaps even more noteworthy, this period still includes the late 1990s when large cap growth outperformed all other styles yet mid cap growth is still preferred. As you may have suspected, the October 1997-August 2003 was not randomly selected: October 1997 marks the beginning of twenty-five uninterrupted quarters of mid cap representation in P5. That's right, mid cap value, mid cap growth, or both have been in the mix for six straight years. That's quite a preference for mid caps.
Have mid caps quietly outperformed all other styles and capitalizations? Considering the quarters of outperformance over the entire measurement period it's actually quite the contrary. Chart 4 shows the percentage of the time each style was the top performer. Again notice that the indexes themselves aren't even shown since they were never the top performer. This stands to reason since they are a blend of the value and growth styles of their capitalization. Unless both styles had the exact same return, one would outperform the other and therefore the index as well. (This again is why the model rarely utilizes the "blend" indexes relying instead on the specific styles as well as capitalizations.) More to the point however, notice that the mid cap styles were not the most frequent outperformers, but rather along with large cap value were the most infrequent. On a quarter-by quarter basis, they were the least likely to beat the other styles.
If mid caps don't pound out a lot of homeruns, they must be consistently hit singles. And indeed, they do. Chart 5 shows the average monthly returns for each style (again, "blend" indexes excluded) for the period January 1994-August 2003. Not surprisingly, mid cap growth and value are the leaders. This isn't too hard to reconcile with the data from Chart 4. While mid caps aren't frequent outperformers, they are consistent. To put this in context, consider large cap growth and mid cap growth. According to Charts 2 and 3, they are the most heavily used while Chart 4 shows large cap growth as the top performer 19.8% of the time while mid cap growth is best only 15.5%. Yet Chart 5 shows mid cap growth outperforming large cap growth by over .2% per month. The difference lies in consistency. Large cap growth was the top performer in 23 out of the 116 months observed, but it was also the poorest performer in 30. Mid cap growth was only tops for 18 months but was worst in just 15. As is often the case, consistency pays off in the long-run. Mid Cap MysterySo it appears P5 has correctly focused on mid caps not only since it's official inception in 2002, but all the way back to the beginning of the backtest. This has obviously been successful for this model, but should investors gravitate towards mid caps as well?
First off, many will argue that mid caps as a class don't even exist. No one questions that there are stocks of companies with market capitalizations between $2-7 billion, but their existence as a separate asset class and investing style has been questioned. Indeed, one could reasonably argue that mid caps are arbitrarily carved out of the broader market and are no more than small large caps or large small caps. The mid cap capitalization range is really one of transition. As small cap companies grow and garner a following, they pass through this range on the way to becoming large cap stocks. While large cap companies can and do stumble losing large cap status, once viewed as large caps most retain that perception. For example, do you think of Black and Decker or Toys 'R' Us as mid caps? By capitalization they are, and at the low end at that. Nevertheless they're still in the S&P 500 and most investors think of them as large caps. We would suggest this favors "mid caps" as a style since the S&P 400 (or any mid cap benchmark) contains more upwardly mobile stocks than falling large caps. All else being equal, this should lead to consistent, positive relative performance. P5's underlying asset allocation model acknowledges this by its emphasis on the mid cap style. So far it's paid off. The short-lived periods when it's fallen behind the broad market measures occurred when small caps outperformed both large and mid caps. No surprise there. As the data from Charts 2 and 3 indicate, mid caps closely follow large caps when the latter lead the market, so underperformance is only likely when small caps lead the way as was the case from October 2002 to June 2003. Is this an ongoing formula for success? P5 may ultimately tell us. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
||||||||||||||||||||||||||||||||||||
S YOU CAN TELL FROM THE PERFORMANCE of most mutual funds, beating a broad market measure is a formidable task. The broader the index the harder it is, simply because it's usually not practical to own all or even the majority of the target index. Even if you could, owning the index will never allow you to beat the index.
