That’s the provocative title of a recent article in Smart Money. The article makes a controversial claim:
The Journal of Indexes gives academic treatment to bland investments, and so might not seem a likely source of hot controversy. The latest issue, however, is packed with it-and has greatly annoyed mutual fund titan Vanguard. A report therein gives new support for the claim that most index investors are unknowingly missing out on a large portion of the returns that their passive approach ought to provide.
Are investors really missing out on a large part of the passive return? The article implies that a variety of alternative weighting methods like efficiency weighting, volatility weighting, equal weighting, and fundamental weighting provide better returns than traditional capitalization weighting. ETFs are even available for some of these alternative methods, most notably equal weighting (RSP) and fundamental weighting (PRF). Over the eleven-year period cited in the article, all of them have better returns than capitalization weighting.
Source: Smart Money/The Journal of Indexes
Here’s what drives me crazy: when alternative weighting methodologies are discussed, there is always one method omitted. That method is relative strength. Perhaps it is no surprise that over the eleven-year period cited in the article, relative strength weighting outperformed all of the other methods. (And I didn’t get to cherry pick the time period-the article made that choice. If the blowout year of 2010 was included in the results, relative strength would have an even larger advantage over its rivals.) It’s also nice to note that there is a relative strength weighted index available, the Technical Leaders Index (PDP). Here’s what the same chart looks like if relative strength weighting is included.
Source: Journal of Indexes, Dorsey Wright
Why is relative strength weighting always left out? If I were cynical, I might say that relative strength is intentionally disregarded so that alternative methodologies do not have to show their comparative performance. But since I am in a charitable mood, I think the reason it is often ignored is because it is too simple. Yes, too simple.
It does not require manipulation of a massive fundamental database. It does not require equations and a mainframe computer to calculate a covariance matrix. It does not require a CFA, MBA, or PhD. (Think how much money you could save on grad school!) Instead, it requires a pocket calculator or spreadsheet and one of the many methods for measuring relative strength. We are partial to our proprietary measurements, but lots of methods work just fine. What does it say about your complicated alternative indexing method when it can be outperformed by something a middle-school student could learn to calculate?
There is one big advantage to capitalization weighting: it can be implemented in nearly infinite size. Along with theoretical reasons (“owning the market portfolio” in Modern Portfolio Theory), that may well be one reason why institutions, and Vanguard, believe it is the way to go. On the other hand, it is not really a stretch to believe that there are alternative indexing methodologies that could be designed for better performance. We think that relative strength has been demonstrated to be the simplest and most robust way to build the better mousetrap.
PDP has not outperformed the equal weight RSP or the other equal weight indices. It also hasn’t outperformed the QQQQ. http://finance.yahoo.com/echarts?s=PDP+Interactive#chart11:symbol=pdp;range=my;compare=^rut+qqqq+rsp;indicator=volume;charttype=line;crosshair=on;ohlcvalues=0;logscale=on;source=undefined
PDP had a very large loss and hasn’t yet recovered or reached a new high since it’s inception nearly 4 years ago. In addition to the equal weight RSP an QQQQ outperforming it, so did iShares S&P MidCap 400 Growth (IJK). Even the bond ETF’s like LQD has achieved a much better return. The price of LQD has matched it over the period, but when you add in the interest it was much better. LQD didn’t have the 70% loss.
http://finance.yahoo.com/echarts?s=PDP+Interactive#chart13:symbol=pdp;range=my;compare=lqd;indicator=volume;charttype=line;crosshair=on;ohlcvalues=0;logscale=on;source=undefined
PDP has outperformed both RSP (equal weight) and QQQQ (cap weight) over the trailing twelve months. As an indexing method, it also outperformed S&P equal-weight over the 11 years cited in the article. (The article looked at cap weighted S&P 500, but not NASDAQ. QQQQ has been helped massively in recent years by its 20% weighting in Apple Computer. If Apple stumbles, so will this cap-weighted index.)
When I looked at price returns in all of the overlapping 5-year periods where there is available data (2000-2004, 2001-2005, 2002-2006, etc.), here is what I found:
Return rank RSP PDP QQQQ
00-04 1 2 3
01-05 2 1 3
02-06 2 1 3
03-07 3 1 2
04-08 2 1 3
05-09 2 3 1
05-10 2 3 1
The average rank for PDP was 1.7, for RSP was 2.0, and for QQQQ was 2.3. In other words, PDP did better than the other two indexes for more of those periods than not. I think what we are really looking at is a situation where the performance is very sensitive to the measurement period. All of the indexes were the best for some periods-and all were the worst for other periods. Clearly, it will take a very, very long time of live performance to draw a firm conclusion.
However, if you look at academic back data on different weighting methods (going back 70 years in some cases), you would conclude that the likely long-term order of finish will be 1)relative-strength weighting, 2)equal weighting, and 3)capitalization weighting. Bond returns like LQD would be in a distant 4th place. They’ve been good lately, but not great over the longer term.
Equal weight vs Market Cap. The study goes back 11 years and the S&P 500 EW from 2000 has beaten the S&P 500 MCW a lot. My ratio moves from 0.8 up to the current 1.52 in 11 years. However the bull market slice from 1994 up to 2000 was market cap driven with the SPW/SPX ratio from 1.1. down to 0.8. I would say that there is no absolute style for passive investment, the perforamce of EW vs MC is subject to trends. The RS apprach, instead, is and remains a long-term winner/outperformer.
[...] has no mention of relative strength weighting. In a previous post, I discussed how relative strength weighting may well outperform both of these alternative methods over a long period of time. The new Wall [...]