Exchange-traded funds (ETFs) are mutual funds that trade on the markets like stocks. Investors like them because of the easy way they supposedly offer returns that closely match their underlying indexes -- but that assumption could be wrong. It turns out that some ETFs have been delivering returns significantly different than expected.

Citing a Morgan Stanley study, Dow Jones recently reported that ETFs had an average tracking error of 1.25 percentage points, more than double 2008's 0.52-percentage-point average.

It seems that most ETFs don't completely mimic their underlying indexes -- for reasons that admittedly make sense. You'd assume that any fund intending to mimic the S&P 500 would buy all 500 companies, but some don't. The same goes for many other ETFs and their respective indexes.

The table below may help explain why:

Company

Rank by Size in the S&P 500

Weight in the S&P 500

Market Cap

ExxonMobil (NYSE: XOM)

1

3.08%

$309 billion

Microsoft (Nasdaq: MSFT)

2

2.18%

$249 billion

Apple (Nasdaq: AAPL)

3

1.87%

$191 billion

Procter & Gamble (NYSE: PG)

4

1.84%

$185 billion

Eastman Kodak (NYSE: EK)

497

0.02%

$1.6 billion

New York Times (NYSE: NYT)

498

0.01%

$1.7 billion

Meredith

499

0.01%

$1.5 billion

Titanium Metals (NYSE: TIE)

500

0.01%

$2.3 billion

Sources: Yahoo! Finance, indexArb.com. As of March 3.

As you can see, scores of companies, even powerful performers, make up only tiny fractions of the index. Fund managers can easily decide to leave some of them out, reaping the benefits of lower commission costs by not having to buy and sell them.

Big and small differences
When managed effectively, ETFs that own just a subset of their indexes can perform well, closely mirroring their supposed sources. But in recent years, some have diverged widely. According to Dow Jones:

  • In 2009, the iShares MSCI Emerging Markets Index ETF fell short of its benchmark by 6.7 percentage points.
  • Some new bond funds have posted even worse showings. The SPDR Barclays Capital High Yield Bond ETF returned 50.5%, 13 percentage points less than its index.
  • Sometimes, the discrepancies actually work in investors' favor. The Vanguard Telecom Services ETF returned 29.6%, versus just a 12.6% return for its benchmark.

In the past few years, the market has swung widely, which may have amplified otherwise small differences. In addition, over the long haul these discrepancies may even out, resulting in largely similar average returns between ETFs and their underlying indexes.

Despite these differences, you shouldn't dismiss ETFs out of hand. They're still a great way to gain easy entry into otherwise tricky sectors, including attractive foreign markets. Still, it's good to remember the potential for divergent performance between an ETF and its index. Keep an eye on your ETFs' performance to make sure they're delivering no less than they've promised.

Should ETFs be allowed to leave out some index holdings? Is that a savvy move, or a dirty trick? Sound off in the comment box below.

A free exchange of further Foolishness: