I've been at the Pershing INSITE Conference today in Hollywood, Florida, moderating a panel on Advanced Strategies In ETFs.

It was a good panel, with Kelly Gallagher of SSgA, Bob Holderith of ProShares, and Dan Waldron of First Trust.

Before the conference started, I thought I might have to give a speech at my panel. In preparation, I roughed out a speech on the risks of buying ETFs and what investors should look out for when they are researching funds.

As it turned out, I played more of a traditional moderator role, and the speech stayed on the shelf.

Below, I've posted my rough notes on the speech. The talk is incomplete and unpolished - I stopped before I got around to writing the introduction. Still, it lays out some good, basic information about researching ETFs, so I thought I'd share it here.

---Risk In ETFs---

Label Risk

The first risk investors face in the ETF market is what I call "label risk." Basically, ETFs don't always hold the kinds of stocks you expect.

Take the iShares MSCI Pacific Ex-Japan fund. On the surface, it sounds like a great way to get access to the fast-growing Asian tiger economies. But if you look under the hood, the fund is 67% invested in Australia and New Zealand, with just a smattering of exposure to Hong Kong, Singapore and other markets.

At the same time, there are three different BRIC (Brazil, Russia, India and China) ETFs on the market today. But you can't just go out and buy one without thinking. Depending on which you choose, you could be buying 53% exposure to Brazil (EEB) ... or 28% (BIK); you could be getting 5% exposure to Russia (EEB), 20% exposure (BKF) or 25% exposure (BIK). Those kinds of differences matter ... a lot ... and can huge impacts on performance.

Tracking Error Risk

One you understand the index you're buying, you have to make sure the fund tracks that index well.

For the most part, ETFs do track well ... but not always.

Consider what happened with emerging markets in 2007. There are two ETFs that track the MSCI Emerging Markets Index: one from iShares (EEM) and one from Vanguard (VWO). The two funds take different approaches to tracking this index: Vanguard uses what's called full replication, meaning the fund buys every stock in the index in exactly the right weights. The iShares fund takes a different approach, called optimization. Essentially, iShares uses computer algorithms to select a subset of the broader index it believes will track the index as a whole. At one point in 2007, the iShares fund held just 250 of the 750 stocks in the broader index.

Unfortunately, with all the volatility in emerging markets, the iShares fund had trouble tracking its benchmark. While Vanguard's fund stayed within a percent or so of the index, iShares fell behind by 7% over the first six months of 2007 alone. It since recovered, but in general, funds that optimize have greater risk of tracking error than funds that fully replicate their index.

Spreads Risk

The third risk to look at is trading risk, or more specifically, spreads. Since ETFs are bought and sold like stocks, spreads are a real issue.

ETF issuers for years have said that the liquidity of the ETF doesn't matter; what matters is the liquidity of the underlying shares. But a recent paper I wrote for ETFR showed this wasn't true. It looked at average bid/ask spreads on every ETF in the industry and found a direct and profound correlation between fund size, trading liquidity and average spreads. Most ETFs had spreads of less than 10 basis points, but 14 ETFs had spreads greater than 50 basis points. When you think of paying that kind of spread on a round-trip sale, it really adds up.

Tax Risk

The fourth risk I want to highlight is tax risk. For the most part, ETFs are more tax-efficient than competing mutual funds. But there are exceptions.

For one, ETFs tend to have higher nonqualified dividends than some mutual funds, so it pays to check the percentage of QDI a fund has had in the past before buying.

But a bigger problem comes from new asset classes. One of the great things about ETFs is that they have opened up new markets to investors for the first time - commodities, currencies, and more. But those new asset classes aren't taxed like equities.

For instance, the SPDR Gold Fund - GLD - is the 5^th-largest ETF in the world. It's one of the cheapest and most liquid ways to buy gold. Thousands - maybe millions - have used it.

But here's a question: If you hold GLD for two years, and it appreciates 50%, what do you pay in taxes when you sell? The 15% long-term capital games tax? Nope - 28%, because the IRS considers gold a "collectible" and charges all collectibles at 28% regardless of how long you hold the funds.

Currencies and commodities have different but equally important tax quirks. The further you get from traditional asset classes, the more complex the tax issues become.

Counterparty Risk

I'll close with just one more: counterparty risk.

For the most part, when you buy an ETF, it's like buying a traditional mutual fund. You own a direct stake in the stocks or bonds that make up the fund.

But there are a few exceptions: ProShares and competing levered funds, as well as some commodity funds; so there is some slight counterparty risk there.

And exchange-traded notes - or ETNs - are something else entirely. ETNs are promissory notes underwritten by a bank. If that bank goes under, you end up as a creditor looking for pennies on the dollar. Of course, the likelihood of an underwriting bank going under is small - Barclays, for instance, probably isn't going to disappear tomorrow. And since ETNs can be redeemed from the issuer on either a daily or weekly basis, and are traded on the stock exchange intraday, you're unlikely to be caught out if you're paying attention. But still, strange things can happen. Just ask the folks from Bear Stearns.