The RAFI Fundamental Index approach sought to solve this conundrum in a shockingly simple and intuitive manner. Suppose we weighted our index by some other gauge than price. We would no longer be forced to ride up with the most popular and beloved stocks, and the exorbitant expectations that come with them. Arnott, Hsu, and Moore (2005) proposed using other measures of economic size, like sales, cash flow, book value, and dividends paid, and then rebalancing once per year. They showed how this and other non-cap-weighted indices plugged the “2% leak” in the cap-weighted boat, by breaking the link between price and portfolio weight. But unlike other measures—crude ones like equal-weighting back then and the opaque and overly complicated “quant in drag” techniques today—the use of economically meaningful measures preserves virtually all the desirable attributes of cap-weighted indices, including broad economic representation, large capacity, low turnover, and ease of governance.
Learning from Valid Critiques?
Hmm…an index that stands to deliver 2% in long-term outperformance while preserving nearly all of the implementation advantages of cap weighting? Sounding too good to be true, fundamentally weighted indexing attracted immediate skepticism from Jack Bogle, Burt Malkiel, Cliff Asness…and John West. Remember, I was a consultant at the time! My clients expected me to poke holes in money managers’ latest snake oil remedies. And I cared deeply about my clients’ success.
Frankly, many of my concerns were similar to those expressed by Messrs. Bogle, Malkiel, and Asness, unsurprising given the deep respect I held and continue to hold for all three. Jack Bogle correctly and fairly pointed out “these are hypothetical returns for the underlying indexes that don’t take into account fees, costs, and taxes” (Lim, 2007). The operative word in my mind at the time was hypothetical. We consultants had an old joke: There is no such thing as a bad backtest. They never see the light of day. I relied on my intuition. Avoiding the big bubbles like Cisco circa 2000 (or the small bubbles like Krispy Kreme Donuts circa 2003) that beset a price-weighted approach seemed promising even without simulated results. As for transaction costs, “Fundamental Indexation” demonstrated that turnover was likely to be low. Furthermore, the stocks of big companies tend to be traded in volume. As a company grows in economic importance, its target weight naturally rises, and so does its liquidity.
I was more interested in the tie-in with value. Burt Malkiel stated, “fundamental indices have done very well over the past six years because value stocks and small cap stocks have done well. Will they do well over the next six years? I’m not so sure” (Floyd, 2007). I ran the numbers, and throughout 2005 I explored the value and size issue with Research Affiliates. Jason Hsu walked me through the time-varying nature of the RAFI Fundamental Index style tilts. I became more and more aware that noisy stock prices create opportunities to rebalance. The benefits of rebalancing across asset classes are universally acknowledged, but it took the RAFI methodology to illustrate the power of rebalancing within equity markets. Rebalancing entails selling what has done well lately and buying what has done poorly. So when value stocks do particularly well relative to their fundamental size, the RAFI Fundamental Index method trims value and adds to recently lagging growth stocks.
It is precisely due to this dynamic exposure that the RAFI Fundamental Index strategy tends to win more in value markets than it gives back in growth markets. I concluded this was the driving force behind the strategy’s 1.5% annualized premium over the Russell 1000 Value Index. I also concluded that the size bias of the RAFI Fundamental Index portfolio was overstated. True, the RAFI portfolio had about half the weighted average market capitalization of the S&P 500 at the peak of the TMT bubble. But that was less of a bet against large companies and more of a bet against high-priced tech stocks.
Much of the rest of the debate in 2005 and 2006 centered on semantics. Was it an index? Personally, I didn’t much care what people called it. In most industries, customers celebrate innovations that deliver some combination of better performance and lower costs, rather than getting tripped up in arguments over nomenclature. Recall the two implementation choices at the time—active management and cap-weighted index funds. I concluded it was a better investment portfolio than cap weighting and it was clearly cheaper than top-quartile active managers (for those with the chutzpah to claim they can pick them).
Was it new or, as Cliff Asness (2006) suggested, just a cleverly repackaged form of value investing? Well, rebalancing—the driver of the RAFI Fundamental Index dynamic value and other tilts—is by definition a value-oriented activity, and it was identified well before Fama and French. Ben Graham (2005, p. 42) in 1949 intimated this in The Intelligent Investor by explaining, “Basically, price fluctuations have only one significant meaning for the true investor. They provide him with an opportunity to buy wisely when prices fall sharply and to sell wisely when they advance a great deal.”
So what was new? Putting this kind of a value-oriented approach on a rules-based index chassis—that’s what. Whether you call it an index or not, the portfolio construction methodology has critical implications. It places downward pressure on fees and upward pressure on transparency. How can that be bad?! My questions were answered. I believed this strategy was something every institutional client should examine. I figured the best way to make that happen was to join Research Affiliates as a “RAFI missionary.”
The Numbahs Please
“You need a full market cycle to evaluate an investment strategy.” How many times have you heard a manager or consultant intone this dictum? Well, it would be hard to argue that we haven’t seen a full cycle since my revelation at 35,000 feet. We’ve seen a bull market from 2005 through October 2007, the sharpest bear market since the Great Depression, and a six-year bull market. Large and small companies alike have rotated leaders. As we’ll explore in greater depth shortly, value won early, but growth stocks have had a nearly continuous run since mid-2007. We’ve also seen sector leadership shift across the economy. Seven different sectors, from health care to utilities to financials, have been calendar-year winners.
So how has the RAFI Fundamental Index methodology done in its first 10 years? What have we learned as we’ve migrated from the Lake Wobegon world of backtesting?
First, we have cumulative annualized performance of 9.4% from December 1, 2005, to December 31, 2014.7 How does this compare to the two implementation options of the time? Active management has had a rough go of it with a return of 7.1% for the median Lipper Large-Cap Core mutual fund, compared with 7.9% for the S&P 500. So that’s an annualized excess return of 2.3% and 1.5% above the median active manager and the cap-weighted index, respectively, for our first 10 years. If you had invested $10,000 in the FTSE RAFI US 1000 Index in December 2005, your balance would have grown to approximately $22,640. This is $2,660 more than if you had invested in a fund that tracks the S&P 500 and $3,960 more than if you had invested with the median active manager. To be sure, the two index results are before costs, but the costs in both cases would have been modest.
Nonetheless, the value-added returns did fall short of the hypothetical excess returns found in the original research. Why? This brings us to the value criticism. Burt Malkiel was right. Value stocks had done very well prior to the publication of “Fundamental Indexation,” whether one used the previous five years’ returns (dominated by the unwind of the tech bubble) or the longer stretch of 35 years (as far back as the Russell 1000 Value data were available). And the RAFI Fundamental Index strategy had a near universally acknowledged value tilt, sometimes big and sometimes small.
In Figure 2, I show the excess returns of the RAFI portfolio over the S&P 500 along with the approximate value and size premiums that were commercially available. The first set of bars on the left shows the results that this former consultant would have looked at in 2005. The backtested RAFI portfolio produced a 2.3% excess return from 1979 to November 30, 2005. Meanwhile, value stocks, as represented by the Russell 1000 Value, produced an excess return of 0.9%. The size premium, as measured by the S&P 500 minus the Russell 2000 Index, was negative, confirming that the RAFI Fundamental Index small-cap bias was a red herring.