(With apologies to readers of the original version of this post, I later discovered a serious mistake in the chart settings on which it was based. Here is the revised study, with corrected charts and a somewhat different set of conclusions)
As investors, we’re always looking for a little edge in the markets. One of my central tenets, not only for investing but in all of life, is that differences at the margin are important – it’s that little extra that often distinguishes the successful from the mediocre. By continually making positive incremental improvements, by going the extra distance, we can achieve long term success.
With all of that said, here is an interesting visual exercise in seeking a better alternative to the plain vanilla S&P 500 Index EFT, a core holding of so many portfolios. The most established of these is State Streets SPY and, like the index itself, is capitalization weighted. Articles on various alternatives to capitalization weighting abound; Rob Arnott, founder of Research Affiliates, is perhaps one of the better known proponents of alternative weighting.
Invesco sponsors several PowerShares ETFs based on the Research Affiliates Fundamental Index, with the FTSE RAFI 1000 (PRF) being the most comparable to SPY. Another fundamentally oriented ETF, and one of my personal favorite alternatives, is Vanguard’s Dividend Appreciation (VIG), a Morningstar 5 Star rated offering which is based on the sponsor’s customized index, which selects “high quality” stocks primarily from among the US large cap universe. A different type of alternative, employing the simple approach of equally weighting all of the S&P 500 stocks, is used by the Rydex S&P Equal Weight ETF (RSP).
Our interest here, as always, is to test the theories by looking at what the market tells us. For this exercise, we compare performance over the last five years, a period which has seen a market crash, recovery, and consolidation phase, to see which approach has produced the best results over the full cycle. In all three charts, we use VIG as the basis for comparison to the other alternatives, simply because it is my preferred ETF. Readers can easily replicate this with any ETFs of their choosing.
First, we can see that Vanguard’s dividend growers (dotted line) have outperformed the cap weighted ETF since the onset of the financial crisis. Looking at the composition of the two funds, it was fairly clear to see that VIG benefited from a smaller allocation to large cap financials:
(click on charts to enlarge)
When compared to the RAFI approach, which also beat cap weighting, we have more mixed results, but RAFI generally seems to have outperformed slightly in rising market phases, while VIG had better downside performance:
Finally, the simple equal weighting approach seems to have provided results very similar to RAFI:
You can see that PRF and RSP track nearly identically:
What can we take away from this rudimentary study? Aside, of course, from the fact that we haven’t established anything definitive here…but an investor has to make the best call with the available information. Here are my conclusions:
1. The critics of indexes based on market capitalization are probably on to something. Is the drag 2-4% per annum, as Arnott’s research suggests? Hard to say, but the thesis appears to hold up through this market cycle.
2. Criticism of stock picking – and rule based ” fundamental indexes” like Vanguard’s and RAFI are a form of stock picking, albeit perhaps a “weak” form – may not be as valid, if the outperformance of the Vanguard ETF’s approach is durable. This is somewhat ironic as the fund comes from the firm that champions indexing. Of course, one could make the argument that anything other than buying the total market is a form of stock picking, but you get the point.
3. It is probably a healthy sign for the market that cap weighting isn’t producing the best performance, in that it tells us by definition that money isn’t flowing disproportionately into a relatively small group of mega caps.
Now we should add one more note. This exercise looked only at share price performance, and in the real world, of course, we have to account for dividends, reinvestment and any associated costs, and fund fees.
Dividends and costs matter, and their cumulative effect is greater the longer the holding period, so these investors with different holding periods will draw different conclusions from this study. Those who trade into and out of the market with shorter cycles might prefer the slightly higher beta RSP or PRF, while VIG may be more attractive for a long term buy & hold portfolio.
Dividend yields as of this writing (data from Morningstar) are:
Fund fees as of this writing are: