When you buy an index fund, it’s natural to expect that your fund will earn basically the same return as the index.

In a perfect world, you should expect an index fund to earn the index returns minus the cost of the index fund.  If an index fund charges 0.1 percent annually and an index earns 10 percent, you should expect to earn 9.9 percent.

There are other reasons why index funds don’t match their indexes other than cost, but for the most part, the differences don’t add up to much and only professional investors get very persnickety about these disparities.

One glaring exception to this general rule relates to funds that invest overseas, and as funny as this sounds, the index is wrong.  Well, maybe not wrong, but not the most accurate reflection of reality.

Let’s look at a fund that is not on our Approved List as an example.  The Vanguard FTSE All-World ex-US ETF (ticker symbol: VEU) tracks developed and emerging markets stocks outside of the US using an index developed by FTSE.

So far this year through June 30th, the market price for this ETF was up 5.13 percent.  The corresponding benchmark for this fund was up 4.61 percent over the same time period.  Awesome, right?  Who would think that an index fund could beat its benchmark by 0.42 percent in six months?

While it’s great to enjoy outperformance, that’s actually bad news for an index fund because it’s much better than the paradigm I laid out in the perfect world scenario and if something can be different enough to win like that, it can lose as much over a different time period.

In fact, that’s exactly what happened over the last three years ending on June 30th.  The benchmark returned 10.00 percent, but the fund’s return was only 9.63 percent, a difference of -0.37 percent, more than the 0.14 percent expense ratio would imply.  So is Vanguard doing a bad job managing this fund?  Not at all, the difference between fund returns and benchmark returns is a function of time zones.

Imagine an index that tracks nothing but Japanese banks.  When those banks stop trading in Japan (which is 3AM CST), the index calculates the performance of the Japanese bank stocks and the index price is set at 1,000.

Now assume that when markets start trading in the US, there is some bad news that sends bank stocks down.  US bank stocks fall by five percent and while British and European markets are open (until 10:30 AM CST), those stocks fall five percent too.

Now imagine that you own a US-based index fund that tracks Japanese banks.  Do you think the price for that fund will remain unchanged since the index closed at 3 AM the night before, or do you think that traders will assume that Japanese bank stocks will fall by about five percent and trade the index fund lower?

Of course, traders will assume that Japanese bank stocks will fall by five percent and the index fund will fall five percent.  But the index didn’t change because it only looks at prices in Japan when those banks are trading; the index doesn’t track the index fund price in the US.

So now, when statements go out, it will look like the fund underperformed the index by five percent.

My five percent example is a little extreme, but the differences can be larger than you think.  In 2009, the MSCI EAFE gained 31.78 percent, but the largest ETF tracking this index only earned 26.92.

That’s a massive discrepancy of -4.86 percent and it turns out that this time-zone problem accounts for about 70 percent of the difference (there are good reasons for the other 30 percent, but that’s for another day).

The good news is that over time, these problems disappear because sometimes the index wins and sometimes the funds win; but, on average, they even out.

MSCI has created a series of indexes to deal with this problem and year-to-date through May 31, the difference between the straight index and their adjusted index was 0.80 percent.  Over the five years ending at the same time, the difference falls to 0.01 percent.  If you’re interested in more details, you can click here for some nice data that illustrates the issue.

We can see the same thing in the Vanguard ETF that I referenced above.  Of the 0.59 year-to-date out-performance of the fund over the benchmark, you can attribute almost all of the difference to the time zone problem.  For the three year numbers, half of the underperformance can be explained by this issue, leaving the other half to the expense ratio – exactly what you would expect in a perfect world.

Ultimately, though, he key thing to remember is that foreign indexes and foreign index funds can deviate from each other quite a bit over the short run, but converge over the long run.

We’re investigating whether we should use the new MSCI indexes since it would clear up a lot of questions when the time zones work against us (no one says anything when they are in our favor), but it won’t be a perfect fix because we use FTSE indexes for a lot of our overseas exposure, which can create other, also legitimate differences (here is an example from last November).

Benchmarks are useful and important tools, but like everything else, you have to know what you are really looking at to figure out what they are really telling you.

Remember, you can only invest in an index fund and not directly in the index.