Why do we invest in cap-weighted indices?

By Felix Salmon
July 8, 2010
paper (PDF) revisiting the question of why so many of us invest in capitalization-weighted stock indices. It turns out that the theory behind our behavior is pretty weak: first of all, you have to believe in the capital asset pricing model, or CAPM. The CAPM includes lots of assumptions which don't hold in the real world: that all investors have the same risk appetite, for instance; that they all have the same investment horizon; that they can short securities freely; that they pay no taxes or transaction costs; and that all assets can easily be traded, including assets such as human capital and real estate.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Felix Goltz and Véronique Le Sourd have an interesting paper (PDF) revisiting the question of why so many of us invest in capitalization-weighted stock indices. It turns out that the theory behind our behavior is pretty weak: first of all, you have to believe in the capital asset pricing model, or CAPM. The CAPM includes lots of assumptions which don’t hold in the real world: that all investors have the same risk appetite, for instance; that they all have the same investment horizon; that they can short securities freely; that they pay no taxes or transaction costs; and that all assets can easily be traded, including assets such as human capital and real estate.

If you make all those assumptions, then the CAPM says that the market portfolio is efficient — the market portfolio being, essentially, everything in the world: stocks, bonds, real estate, commodities, human capital, art, social security benefits, automobiles, everything. But the problem, of course, is that cap-weighted indices do a pretty bad job even of reflecting the performance of the stock market as a whole, let alone all global assets. (The world’s assets have grown a lot in the past decade; the S&P 500, not so much.)

The authors conclude:

In view of these arguments, it seems that financial theory alone does not justify the current practice of cap-weighting. In fact, from a theoretical perspective, cap-weighted stock market indices seem to offer no particular advantage.

So why are cap-weighted indices so popular? One reason might be that indices seem to outperform a simple buy-and-hold strategy. And another is that they’re easy to understand and most of them have been around for a long time. Still, it’s worth noting that the most famous stock index in the world, the Dow, isn’t an index at all, and it certainly isn’t cap-weighted. (Although with the Dow, you would have been better off just holding the original 30 stocks than following the vicissitudes of DJIA itself.)

The main reason for buying cap-weighted indices, I think, is that it’s easy and it’s cheap. (That’s probably the main reason not to buy cap-weighted indices, too, since anything easy and cheap is likely to get crowded.) Of all the assets in the world, stocks are the easiest to invest in and the easiest way to invest in stocks is to simply buy the index. I suspect there are many portfolios which do a better job of simply “investing in the world” than the S&P 500 does. But once you take into account the costs of putting them together, it’s probably not worth it.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
10 comments so far

Easy in 2 ways: don’t have to follow 20 or 30 seperate issues and with 1 thing you get rid of all non Beta volatility. Last is not relevant in CAPM but of course for individual real investors not unimportant.

Not agree with the crowded point. You make it sound like a negative thing. Crowded means just more (size and/or quantity) will come. Assuming that it is iso investment in normal shares. And crowded likely means better and cheaper as it creates more competition. Only negative point could be that non index gets less popular, but with all new products this doesnot look to me a real danger.

Posted by Rikh | Report as abusive

Of course, if one is unhappy with market cap indices, they are a whole range of fundamental indices. The likes of GWA and RAFI offer a pretty comprehensive choice. Not all have products based on them, and many of those products are not as easily accessible as a tracker mutual fund or ETF – and cost more. And that assumes one can get past the fairly black box nature of those research outfits. (btw – fundamental indices are based on things like earnings, assets etc… essentially accounting numbers rather that market cap; the idea is that it avoids bubbles (backtesting (warning) shows they avoid the internet buble) as it focuses on the “real” economic performance rather than the mood swings of investors…)
What is interesting though, is that the last couple years performance of fundamental indices has been pretty dismal. They were overweight in financials as those companies were reporting really strong figures. Yet, those figures turned out to be pretty illusory when set in the context of the losses that followed. The problem is that accounting is not an exact science, it’s a moving set of rules, and different in various countries. All things equal, $1 of earnings from say china, UK or argentina should be treated equally (ie same rating) by investors – yet, it doesn’t happen. Fundamental indices try to address that, but clearly the accounting in those 3 countries will be different, hence that $1 will probablty turn out to be quite a different figure.

Still, if you’re a large pension fund with buying power, you can get a financial institution to create a custom product very cheaply following fundamental rules – that ought to reduce your porfolio overall risk.

Posted by fxtrader14 | Report as abusive

In an efficient market, more money would flow to undervalued stocks than to highly-valued stocks. Yet cap-weighted index investing forces the reverse. In the middle of the dot.com boom, the indices were shoveling new cash towards the technology sector for the simple reason that they were selling at high prices.

About 8% of the shares in each S&P500 company is held by index funds, and much more is held by portfolio managers who are too fearful of under-performing the index to deviate far from that core. This serves to increase market volatility by effectively decreasing the float.

Posted by TFF | Report as abusive

Why do people buy cap-weight index ETFs? Retail investors buy them because that is the product available, if you want cheap, diversified exposure. Institutional investors want cap-weighting because that is how their performance is measured.

Why does the market supply mostly cap-weight? Well, the institutional market is important. Also, it is easier to supply the product in volume because the liquidity you need is well-correlated to cap weight.

In the age of cheap trading, it is within the reach of many individual investors to diversify themselves by buying a handful of individual stocks in equal weights – you get almost all the benefits of 500 names with 20 or even 12 – and rebalance somewhere between yearly and quarterly. But this requires discipline; once you touch individual names, it is hard to resist the temptation to start trading.

Posted by Greycap | Report as abusive

A correction: CAPM does not assume all investors have the same risk appetite. It assumes investors about standard deviation of returns as a measure of risk and return as a measure of reward (mean-variance optimal portfolios). But it allows for variation of risk appetite.

Posted by ayaga | Report as abusive

CAPM has nothing to do with it.

If markets are efficient and you’re not meaningfully different than the price-setting investor, cap weighting is ideal.

If markets are not efficient, it’s not clear what works, but cap weighting is cheap, tax efficient and diversified (by economics, not companies).

Fundamental indexing is a proxy for overweighthing small and value, which Fama French teach us is taking more risk for the hope of higher return.

Posted by jabberwocky | Report as abusive

Belief in CAPM is sufficient, not necessary, to imply that this is optimal. CAPM doesn’t usually posit that all investors are equally risk-averse, though it does posit that everyone’s conception of “risk” is entirely captured by standard deviation of returns. It doesn’t require that one be able to short stocks, though it does require that the less risk-averse be able to borrow, to leverage up. I’m not clear on what you could mean by “the performance of the stock market as a whole” that would differ from the internal rate of return, i.e. the return on a market-cap weighted total market index, including dividends. (I’d also like to see you explain the distinction between an index and an average.)

Posted by dWj | Report as abusive

We index when we have no opinion. As noted above, CAPM is totally unnecessary to come to the conclusion that if we cannot add value, then we simply take a market exposure.

The better question is why we do not attempt to proxy for the large swaths of the market for which there are no easily traded instruments.

Posted by wcw | Report as abusive

An equally weighted portfolio containing small and micro-cap stocks would in some senses be a much more crowded trade then the market weighted one as billions of dollars raced into small companies shares. That would drive up the prices of those shares quite a bit without any commensurate changes in their earning forecasts.

That doesn’t mean any one investor shouldn’t hold the equally weighted portfolio but does mean that everyone shouldn’t do it.

Posted by OneEyedMan | Report as abusive

Cap-weighted indexes exist for one simple reason. They represent the average return for a dollar invested in a asset class or subclass.

Such a strategy has the advantage of being scalable: i.e. if the strategy was pursued to its fullest, all securities in the cap weighted index would be owned 100% by the index fundholders. No individual security in a cap weighted index is disproportionately affected by the creation or liquidation of index units. With non-cap-weighted indexes, securities that are scarce relative to their index weights get disproportionately affected by the creation or liquidation of index units.

Can you do better than a cap weighted index? Sure, but others must do worse for you to do better. Everyone as a group can’t do better; they will earn the cap-weighted index return less fees.

Posted by DavidMerkel | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/