Tuesday, July 19, 2011

Cap-Weighted vs. Fundamental Indexing

Investors have their choice of index method based on either traditional capitalization-weighted funds or fundamentally-weighted funds. The terms used in this battle may be somewhat intimidating, but the ideas are simple enough. The more difficult question to answer is which indexing method will produce higher returns.

Traditional indexes are cap-weighted. For a stock index, this means that each stock is held in the index in proportion to the total value of the company. So if the market says that company A is worth $10 billion, and company B is worth $1 billion, the index will have 10 times as many dollars worth of company A stock as company B stock.

Continuing this example, if a cap-weighted basket of stocks contains $10,000 of company A stock, it will contain $1000 of company B stock. The number of shares may not differ by a factor of 10, though. If A shares trade for $10, then we have 1000 A shares, and if B shares trade for $100, we have only 10 B shares.

The fundamental indexing crowd points to examples like Nortel to say that cap-weighting is crazy. The more overvalued Nortel became, the more dollars worth of Nortel stock was included in the index. This means that wildly overvalued companies tend to dominate index portfolios.

To combat this problem, fundamental indexing came to the rescue. Instead of relying on a company's current stock price, fundamental indexes use a company's fundamental factors such as revenues and earnings to arrive at a value for the company. This fundamental value is used to choose the weight of the stock in the fundamental index. Nortel's share of the index would not have grown so large in a fundamental index because it's revenue and earnings were low compared to the market's assessment of Nortel's total value.

One disadvantage of fundamental indexes is that they require more trading. With cap-weighting, if a company's stock doubles in price, it's share of the index roughly doubles and there is no need to buy or sell shares to maintain the cap-weighting. With fundamental indexing, if the share price doubles without any change to the company's fundamentals, the index must sell off about half of the shares. Similarly, the index must buy more shares when their prices drop. This need for adjustments means that fundamental index funds have higher trading costs than cap-weighted index funds.

If the stock markets do a decent job of valuing stocks, then cap-weighting should win out due to the lower transaction costs. However, if the markets regularly create some high-fliers like Nortel that are destined to crash, then fundamental indexing is likely to benefit by more than its added transaction costs.

I don't see any way to predict which method will work better going forward. Some back-testing studies indicate that fundamental indexing would have worked well in the past. I tend to be sceptical of any investing strategy that increases costs; so I've made my personal bet on cap-weighted indexes, but I have no strong feeling about which will win over the next few decades.

14 comments:

  1. In general I follow the logic of the fundamental indexes, and am inclined to agree. But the eternal passive investing question remains: I can control the fees, but not the returns. Will the returns beat out the increased fees?

    For CRQ vs XIC, there's about a half a percent difference in MER. If that was all there was to it, I might be tempted to go with CRQ on the theory that the fundamental index has the potential to outperform by more than 0.5%. However, it seems to be counterproductive in terms of diversification: financials are already a big part of the Canadian market, almost 30% of XIC. But in CRQ they're over 45%!!!

    It just seems to trade one problem for another, so I guess I'll stick with cap-weighted as well (though I'm with e-series, so I don't have much choice there).

    ReplyDelete
  2. @Potato: I didn't realize that financials make up 45% of CRQ. That has looked good over the recent past, but who knows if it will look good in another 10 years.

    ReplyDelete
  3. I have the same opinion as you - I'm not sure that any extra returns (if any) will beat the increased costs.

    One benefit of a fundamental index is potentially less risk. Having an index with one stock making up 33% of the index might or might not affect returns, but it definitely makes the index riskier.

    I suspect both approaches are fine and it's the investor's saving rate and behaviour that will make the biggest difference, rather than the decision to go cap-weight vs fundamental.

    ReplyDelete
  4. @Mike: You're right that the volatility of fundamental weightings has been lower in back-testing. This has some value. For example, if fundamental weightings reduce standard deviation of returns from 20% to 19%, this boosts compound average returns by about 0.2%. This should all be factored into any well-done back-testing of return data.

    ReplyDelete
  5. I believe the folks behind fundamental indexes have identified a real problem, but not a real solution. The approach requires a formula to tell you the intrinsic value of a company more accurately than its market cap. I'm skeptical such a thing exists.

    ReplyDelete
  6. @Patrick: Good point. It isn't enough for the fundamental index supporters to criticize market-weighting. They must also demonstrate that any particular method they choose for valuing companies will lead to excess returns.

    ReplyDelete
  7. You might be interested in Bogle's take on the matter:

    http://johncbogle.com/wordpress/wp-content/uploads/2006/08/WSJ%20op-ed.pdf

    ReplyDelete
  8. @CC: Thanks for the pointer. Bogle makes a strong and convincing case for market-weighting.

    ReplyDelete
  9. Another benefit of CRQ or any Claymore ETF that is fundamentally weighted is the ability to dollar cost average on a monthly basis without paying for the trading costs. I don't know why this hasn't made bigger headlines for the average retail investor. So I'm happy to pay slightly more of an MER and not pay trading costs.

    ReplyDelete
  10. It's interesting how Bogle opens his piece by challenging the assumption that a new model can take on the dominant success story of the past. I wonder what his index funds did differently 30 years ago :)

    He is of course right that the average return will always be the cap-weighted index. That means the fundamental indexes have to get their marginal returns from active managers who underperform. If this would happens it seems like it would be because those active managers undervalue enduring value (such as profits) and favor trends with nothing behind them. Or the fundamental indexes could get some extra returns from individual investors who underperform their funds through bad timing.

    If you're telling me fundamental indexes can only succeed when other investors make short-sighted and poorly-informed moves... well I can believe that pretty quickly :) But maybe a cap-weighted value index works just as well as a fundamental index.

    ReplyDelete
  11. @Zamphir: That's a useful feature for those who like to make frequent modest-size additions to their portfolios. There is no reason why this has to be linked to fundamental indexing, but in practice it is right now.

    @Value Indexer: I can believe that there will always be bad investors, but that doesn't necessarily mean that fundamental indexers will be the ones who take their money. I guess time will tell.

    ReplyDelete
  12. Come to think of it, Bogle is somewhat wrong in that the potential for fundamental indexes to outperform cap-weighted indexes would be less in an efficient market. If you reduce exposure to an overvalued stock at the right time, that does give you above average returns at the expense of those who don't. And that can't happen if the market is inefficient.

    But that potential is only realized to the degree that the fundamental index gives you the right weightings at the right time. That comes back to the challenge of active management except it's done with a formula instead of someone's judgement. Since a formula can't change frequently it may be easier for traders to get ahead of the fundamental index funds. Nortel makes a nice example but not all mispricings are that noticeable.

    ReplyDelete
  13. Use fundamental indexes for emerging markets, out-performance is strongest here.

    ReplyDelete
  14. Bogle and Malkiel criticize fundamental indexing and its variants due to increased costs and increased taxes. Fundamental indexing and its variants tilt to small cap and value stocks. Bogle and Malkiel point out that historically such a tilt has outperformed. However, it is only relatively recently that this outperformance became known. Also, it is only relatively recently that investors had vehicles to try to tilt to small and value. Now that it is known and investors can try to take advantage of it, the possibility that the outperformance will decrease (or disappear?) is raised.

    The counterargument is that both the small and value tilts are risk premia. If they are risk premia, then they should persist. Most people seem to agree that the small premium is a risk premium. But you have to go quite small to get a premium of significance. Whether the value premium is a risk premium is hotly debated. Some say yes; others say it is a behavioral finance phenomenon. IIRC, William Bernstein tends towards the latter opinion.

    IMO, the small and value premia may persist, but I would not be surprised if they decline. If one tilts towards small and value in Canadian stocks, one decreases the diversity of one's Canadian stock exposure. IMO, the lack of diversity in Canadian stocks is a problem, even without a small or value tilt. As for foreign stocks, a tilt to value will result in increased dividends. It isn't difficult to pay 46% on foreign dividends in Canada. If you use American domiciled ETFs, your tax rate could be 61% on dividends, due to loss of the foreign tax credit.

    ReplyDelete