[Source:] The S&P 500 is an index that weights 500 companies in order of their market capitalizations – if the market values company A more than company B, then company A is weighted more heavily in the index. But this isn’t a good investment strategy. Maybe company B earns more money, or maybe the weighting differential in the S&P 500 doesn’t reflect the earnings differential.

The same can be said about revenues or dividends. ...

... In short, the S&P 500 and many other indexes reflect what the market is doing, but they don’t make for good investments. Therefore ... leave funds like SPY to the traders.

Instead, I want to look at funds that weight their holdings based on other metrics, such as earnings, sales, or dividends. Fortunately there are several funds that are geared toward this kind of thinking. In fact, there are two companies that offer such ETFs ...

The first is RevenueShares. A RevenueShares fund is going to weight its holdings by revenues, not by market capitalization. ...

The second is WisdomTree ... This company offers many different kinds of funds, but two strategies in particular include dividend-weighted funds and earnings-weighted funds.

The expense ratios for RevenueShares and WisdomTree ETFs are higher. So is the bolded true and sound advice? Do other methods of weighting holdings consistently outperform?

2 Answers 2


This is a really interesting question and something a lot of work is being done to understand. I'm going to look at the closely related question "Do non market-cap etf weighting methods consistently outperform once you take into account their investment biases?"

Let's use revenue weighting as a reason why investment biases are so important. In revenue weighting, you would own almost no fast-growing tech companies as they generally have little revenue. This sounds great if we are talking about say Pets.com in the late 90s but you also would miss most of the rise of Google.

To believe in these ETFs consistently outperform (adjusted for risk) you would have to have a strong reason to believe that earnings, sales, or dividends are a better predictor of company value than market value. Market analysts include the above three metrics and many more when pricing stocks so out-performance using only one of the above metrics seems unlikely.

There is one caveat to this and that is value and small cap stocks have been shown to give slightly better risk-adjusted returns in the very long run (see Fama/French) and many of these alternative weighting methods will have a value or small cap bias. First, it is unclear if this out-performance will continue now that it is more widely known. Second, even if you believe this will continue you can more easily and cheaply get this bias though value/small-cap etfs than these weighting schemes.

In the end, the only thing that is perfectly clear is that higher fee investments will generally under-perform.


What makes the S&P 500 Index funds have low expenses ratios is that the amount of money spent by the mutual fund company on research is Zero. Standard and Poors does the research, the fund companies piggy back on that research.

Another big factor in keeping expenses low is the fact the companies that makeup the index don't change very often. If the fund frequently changes the makeup of the investments not only can that trigger tax considerations but there are costs associated with all those transactions.

The question is: do the other algorithms overcome the additional costs due to higher expenses and taxes.

You must log in to answer this question.