Will Fundamental and Dividend Weighting Deliver What They Promise?

With the launch in Spring, 2006 of a new family of exchange traded funds by Wisdom Tree Investments, dividend weighting has joined the fundamental weighting approach pioneered by Research Affiliate’s Bob Arnott as a challenger to traditional market capitalization index weighting. Nobody can argue with the fact that both Wisdom Tree and Arnott have identified anomalies that in the past generated significant alpha compared to market cap weighted indexes. However, two critical questions remain unanswered at this point: what caused these anomalies to occur, and will they continue to exist in the future?

Securities markets are not perfectly efficient, and over and undervaluations occur, for complex reasons. The return generating process for a security has three main parts: (1) changes at the company level (e.g. variations in customer needs, technology, competitor offerings, cost positions, etc.); (2) investors’ forecasts for how those changes will evolve and affect the company in the future; and (3) investors’ forecasts for how they expect other investors’ perceptions and behavior to evolve.

Investors find it harder to project future results (and therefore to estimate security values) for some types of companies than for others, including those that face substantial competition, companies in fast changing industries, and companies that pay no dividends. The higher the uncertainty about a company’s return generating process, the more likely that some investors will make incorrect forecasts. Since excessive optimism and overconfidence are natural human biases, incorrect forecasts will most often lead to overvaluation.

These overvaluations often receive a further impetus from another source. There is constant variation in the percentage of investors who are using fundamental strategies (which seek to identify undervalued securities, in the belief that market prices will eventually converge to them) and the percentage who trade by following market trends (i.e. on the basis of their forecasts for how other investors will behave). Unsurprisingly, these momentum investors are usually most prevalent in the markets for securities that are hardest to accurately value. Their presence can cause overvaluations to persist, as fundamental investors are sometimes reluctant to absorb inferior returns for a long period until momentum reverses and the price of its security returns to a more rational value. For example, many fundamentally oriented fund managers believed that technology stocks, particularly after 1999, were seriously overvalued. However, few were willing act on this belief (e.g. via buying put options or shorting the stocks in question) for fear that the low short term returns that would result would cause their funds to lose assets.

As critics have noted, a market capitalization weighted index will give too much weight to securities that are overvalued, and too little to securities that are undervalued. The question is, can we design an index that reduces exposure to the factors that drive over and undervaluation?

Indexes that use so-called fundamental weighting (based on a company’s relative share of book asset value or revenues) or dividend weighting (based on a company’s share of total dividends paid) have, based on their historical performance, managed to do this in the past, and have outperformed comparable market capitalization weighted indexes. What caused this result?

It would appear that the key assumption underlying fundamentally weighted indexes is that focusing on non-market capitalization measures of size helps to avoid situations where overvaluation is most likely to occur. In the context of our valuation discussion above, this implies that larger companies are either less likely face fast changing businesses environments, and/or investors are less uncertain about how these changes will affect a larger company. This also implies less exposure to the actions of momentum investors. We suspect that all of these assumptions are, to some extent, true. Large size (as measured by revenues or assets) often implies that a company has been successful at creating a stable business ecosystem in which it can thrive. To be sure, as the U.S. automakers can attest, this stability does not last forever. However, even when it gives way to a period of rapid change, size often buys a company more time to adapt, compared to smaller companies that lack a similar resource cushion.

The assumptions that underlie the success of dividend weighting appear to be similar, and perhaps even stronger. Larger companies are usually the ones paying larger dividends, so the size arguments made above also apply here. But paying large dividends should also imply that a board of directors has confidence in the sustainability of a company’s competitive advantages, even in the face of rapid change. Hence, the presence of dividends tends to reduce investors’ uncertainty about their estimates of a security’s fair value, and thus avoid the emergence of serious overvaluations.

However, understanding why outperformance occurred in the past is not the same as believing it will happen again in the future. The disappearance in recent years of the so-called size premium is a perfect example of this. Once informed that small stocks (by market capitalization) had historically outperformed large stocks, investors bought more of them, bid up their prices, lowered their returns and mostly eliminated the expected outperformance.

The question we have to ask is whether the same thing will happen with fundamental and dividend based indexing. What is to prevent hedge funds and other smart traders (not to mention the growth of fundamentally weighted index funds) from bidding up the price of those stocks with the highest weights in a fundamental index, causing their expected excess returns to be eliminated, while having the exact opposite impact on fundamentally smaller stocks that apparently have been systematically overvalued in the past? In this sense, the very discovery of the fundamental and dividend weighting anomalies should cause market capitalization based indexes to become more efficient over time.

To argue this won’t be the case, you need to believe there are significant and durable obstacles that will prevent this arbitrage from occurring. In this age of intense competition for alpha, we don’t believe these obstacles exist. As a result, we expect fundamental and dividend based index products to continue to outperform for a time, but then quite closely match the performance of market capitalization weighted products as the historical anomalies are arbitraged away.

Download Free Section PDF