We've moved well beyond that. There's been much discussion of sector funds and other exotic ETF categories such as currency funds, commodity funds and so forth. The simple one-step decision is gone even for those who are inclined to stick with a traditional style, such as value. Moreover, there can be a conspicuous performance penalty for making the wrong decisions.
We evaluated performance data for 18 domestic ETFs that specifically identified themselves as pursuing the "value" investing style and for which there exists performance data stretching back to the beginning of 2006. Table A shows the price-change differentials between the top and bottom performers over various time periods.
Table A - % price changes
|Last 5 days||3.06||0.42||2.64|
|Last 20 days||4.11||0.88||3.22|
|Last 60 days||10.57||4.47||6.10|
|Last 120 days||12.59||4.80||7.79|
|Year to date||16.19||6.20||10.00|
The variations do not stem from any oddities that produce extreme readings one way or the other. Table B compares, for each time period, the average of the top five performers with the bottom five performers.
Table B - % price changes
|Top 5||Bottom 5||Difference
(top 5-bottom 5)
|Last 5 days||2.24||1.36||0.88|
|Last 20 days||3.01||1.75||1.27|
|Last 60 days||9.71||6.07||3.64|
|Last 120 days||10.86||6.29||4.57|
|Year to date||13.89||7.54||6.35|
At first glance, the 0.88 percent five-day performance gap between the top five and bottom five may appear slim. But if it persists over the course of a year, it would compound to 57.7 percent. That didn't come close to happening in 2006, but even the 6.35 percent differential we actually see would be noteworthy for someone who expected to get by with a reasonably passive value strategy.
Some of this likely reflects the way different ETF sponsors define "value." They all claim to passively follow their respective value index, but each has its own way of designing the index.
Generally speaking, all work with generally recognized concepts such as price/earnings, price/sales, price/book value, price/cash flow and dividend-related ratios. But there are differences in how they use the data. Not all sponsors will use all ratios. Also, different sponsors will have different definitions for each item (price/earnings, for instance, can be based on historic results or estimated earnings), different importance weightings assigned to each variable used, different cutoff points that determine eligibility for the value category, different ways of narrowing the universe down to an appropriate number of stocks, and different ways of weighting the stocks that are included.
But despite the variety of specific index-building techniques, when real-world ETF performance results are tabulated, much, seems to turn on the size category used as a starting point. Most value ETFs directly identify themselves as being geared toward a particular size category (large-cap value, mid-cap value, or small-cap value), or do so indirectly by associating themselves with a market index having known size characteristics, such as the large-cap orientation of the S&P 500.
Differences in size characteristics go a long way toward giving us a handle on performance differences between value ETFs.
Table C compares price changes in the S&P 500 with those of the Russell 2000.
Table C - % price changes
|S&P 500||Russell 2000||Difference
|Last 5 days||0.54||1.98||-1.44|
|Last 20 days||2.01||2.84||-0.83|
|Last 60 days||7.27||10.92||-3.65|
|Last 120 days||8.83||7.61||1.22|
|Year to date||11.61||16.61||-5.00|
Table D compares average performance between large-cap and small-cap value ETFs.
Table D - % price changes
|Large value||Small value||Difference
|Last 5 days||0.70||2.56||-1.86|
|Last 20 days||2.07||2.88||-0.81|
|Last 60 days||6.37||9.28||-2.92|
|Last 120 days||10.89||8.69||2.20|
|Year to date||14.28||10.22||4.06|
We would not expect Tables C and D to be identical. The association of S&P 500 and Russell 2000 with large cap and small cap respectively constitute general, rather than rigid, qualities. Also, both indexes include stocks that are not necessarily consistent with the value style of investing. (The same sponsors offer large-cap growth and small-cap growth funds too.) Viewed in that light, it's interesting to see how performance differentials within the value ETF groups wind up resembling those between the indexes when viewed over shorter time frames.
Statistically speaking, we won't claim we have seen enough to prove that differences within the value group depend on differing size orientations. But we do see enough to indicate that value-ETF investors should give thought to the issue of large versus small. That alone may not be sufficient to produce optimal results, but it seems to offer a greater chance to succeed than would be the case with passive assumptions that value is value and that all such funds are alike.
We're not still not out of the woods.
It's one thing to choose between small-cap versus large-cap value ETFs. It's more challenging to identify the right category at the right time. Still, there's something to be said for trying because it can make a noticeable difference.
Table E shows the performance ranks (from 18 being best to 1 being worst) of the top value ETFs over the past 60 days and also shows what their respective year-to-date ranks were. Notice the sometimes-stark differences between the two sets of ranks.
Table E - Value-ETF performance ranks
|PowerShares Dynamic Large Cap Value (NYSEARCA:PWV)||18||7|
|iShares Morningstar Large Value (NYSEARCA:JKF)||17||11|
|streetTRACKS DJ Wilshire Large Cap Value (ELV)||16||5|
|Vanguard Value ETF (NYSEARCA:VTV)||15||6|
|iShares Russell 1000 Value (NYSEARCA:IWD)||14||9|
|iShares Russell 3000 Value (IWW)||13||4|
|iShares S&P 500 Value Index (NYSEARCA:IVE)||12||3|
|iShares Russell 2000 Value (NYSEARCA:IWN)||11||10|
|iShares Russell Midcap Value Index (NYSEARCA:IWS)||10||2|
|Vanguard Small-Cap Value (NYSEARCA:VBR)||9||8|
|iShares Morningstar Mid Value (NYSEARCA:JKI)||8||12|
|PowerShares Dynamic Small Cap Value (PWY)||7||18|
|iShares Morningstar Small Value (NYSEARCA:JKL)||6||16|
|iShares S&P SmallCap 600 Value (NYSEARCA:IJS)||5||1|
|streetTRACKS DJ Wilshire Small Cap Value (DSV)||4||13|
|PowerShares Dynamic Mid Cap Value (PWP)||3||14|
|iShares S&P MidCap 400 Value Index (NYSEARCA:IJJ)||2||15|
|streetTRACKS DJ Wilshire Mid Cap Value (EMV)||1||17|
One obvious solution is to rotate among ETFs depending on whether large- or small-cap is performing better at a particular point in time.
But that may prove unsatisfying to many. For one thing, it's not always easy to identify the existence of a trend until we're well into it. For another, active trading is what many seek to avoid when they turn to ETFs.
The other solution would involve buying and holding an ETF portfolio that is properly diversified within the value category. This need not be complex. Table F shows the results of a portfolio consisting of equal investments in iShares Morningstar Large Value (JKF), PowerShares Dynamic Mid Cap Value (PWP), and Vanguard Small-Cap Value (VBR) the top ranked ETF in each category (based on the average of all the successive five-day ranks we computed for each fund since the start of 2006).
|% price changes||Portfolio rank
|Last 5 days||1.38||1.63||10|
|Last 20 days||1.91||2.40||3|
|Last 60 days||7.25||7.77||9|
|Last 120 days||9.26||8.80||10|
|Year to date||11.10||10.91||10|
The portfolio isn't a world-beater. But for most periods, it avoids bottom-of-the-barrel status and winds up near the middle of the pack. That's respectable and reasonably in line with what a passive value investor might seek to achieve within this category.
But we can go one step better.
We analyzed our three category-leading funds under what quantitative portfolio strategists would call the efficient frontier. This is a portfolio-building technique that allocates assets in such a way as to result in the lowest level of risk (defined as variability, more specifically standard deviation) for a desired level of expected return. It's based not only on the return and variability characteristics of each asset, but also on how assets tend to move in relation to one another (correlations).
This model suggested there was no real benefit to owning a mid-cap fund. It recommended that we allocate 47.3 percent of funds to the small-cap selection and 52.7 percent to the large-cap position. Table G shows the results of this "optimized" portfolio.
|% price changes||Portfolio rank
|Last 5 days||1.48||1.63||10|
|Last 20 days||2.41||2.40||12|
|Last 60 days||6.78||7.77||6|
|Last 120 days||10.39||8.80||16|
|Year to date||12.86||10.91||14|
This, of course, is based on data collected over a period in which small caps performed better. Mindful of the ever-present warning that past performance does not assure future outcomes, we need to be mindful of the possibility that large caps may perform better on balance in the future. But having decided to diversify between large and small strategy orientations, the specific percentages do not shape up as being all that critical. Table H shows that a simple 50-50 weighting will produce nearly identical results.
|% price changes||Portfolio rank
|Last 5 days||1.54||1.63||10|
|Last 20 days||2.43||2.40||12|
|Last 60 days||6.91||7.77||6|
|Last 120 days||10.33||8.80||16|
|Year to date||12.75||10.91||14|
We still don't have a silver bullet. It's possible that large caps or small caps could outperform so often and by such a large amount next year as to reduce the appeal of a 50-50 weighting. But absent an ability to see into the future, or to develop an unusually strong subjective conviction one way or the other, it appears that a 50-50 strategy has the potential to meet the expectations of one who turns to ETFs in the hope of implementing a passive value strategy.
It's not necessarily critical to choose these specific large- and small-cap offerings. Others could achieve top sub-category ranks if different computational methods are used. The key, here, is to recognize the sort of sub-category diversification that is likely to be necessary to give ETFs a reasonable chance of satisfying the sort of passive expectations that originally gave rise to this asset class.
At the time of publication, Marc H. Gerstein did own not shares of any of the aforementioned funds.