ETF Analytics
ETF Analytics
IndexUniverse.com

Articles

Print This Article

The Winner's Curse
By Rob Arnott, Lillian Wu

Endnotes
  1. We could have used measures of economic footprint, popularized by the growing interest in the Fundamental Index® concept, such as sales, book value, profits, dividends, buybacks, number of employees, and so forth. Market capitalization is, of course, a product of economic footprint and valuation multiples. For instance, company sales times the price/sales ratio gives us market cap. So one might argue that we’re looking at a blend of company size and company popularity. Indeed, we are. For purposes of this paper, our "top dogs" are the companies that—with very few exceptions—are both a dominant player within their business and popular enough to carry a premium multiple. If a company has the largest market cap in its sector (or country), this tacitly implies a consensus expectation that it will deliver larger profit distributions to its shareholders in future decades than any other company in its sector (or country). These companies are also expected to continue to increase their dominance.
  2. In the Tour de France bicycle race, the leader after each day’s race wears a yellow jersey the next day so that competitors can recognize the leader from a distance.
  3. Observers may sensibly suggest that a company with 51 percent market share can still double if its market doubles. Of course, growing the market helps competitors in like proportion.
  4. In countless empirical studies (e.g., see Chow et al. 2011), equal weighting tends to beat cap weighting by 1-2 percent per year. Equal- or cap-weighting will not change the basic findings in our research. It bears mention that we do not exclude the top dog from its own sector return or country return. So while some might argue that equal-weighting our benchmarks will lead to a larger shortfall, we would counter that including our top dogs in the benchmarks will have the opposite effect. In any event, the top dog effects that we explore in this article are much more powerful than the effects of benchmark construction.
  5. The 719 one-year samples were statistically independent, both cross sectionally and intertemporally. The 611 10-year samples were based on rolling 10-year results, so they are roughly equivalent to 60 independent samples.
  6. Now that Apple has taken over at the top, we now have eight U.S. national top dogs in 61 years!
  7. We also carry out additional tests on the sector top dogs for 24 developed economies; these are handled separately, because most of the 24 countries are much less diversified, with much stronger dominance by their top dogs than the G-8 primary countries that we tested. The data is "noisier," with big outliers, so we compile averages across these markets. Still, we are interested to test whether the "too big to succeed" story applies globally.
  8. This index spans the 1,000 largest-cap stocks in the US market and the 1,000 largest-cap stocks in the Developed ex-US markets, hence comprising 2,000 stocks. We refer to this list as the "top 2,000," although it’s actually the combination of two top-1,000 lists.
  9. We don’t show the sector results here. But they are impressive, albeit with considerable variability.
  10. Even relying on the 12 countries with the largest average market cap, this leads to remarkable concentration in some years. As one example, for Russia in 1996, only four companies ranked in the top 1,000 in the emerging markets, by market cap. In 1997 and 1999, only six made the cut.
  11. For purposes of this paper, we ignore the 10-year results, as there are not even two independent samples.


 

Discussion

Post a Comment
Comment
(Max. 2,000 characters)
Name:
E-mail:
Home page:

(optional)

Type in the
displayed characters:
CAPTCHA Image [ Different Image ]
Email follow-up comments to my e-mail address