SmartETF

Defining ETF Risk: Does It Pass the "Smell" Test?

Matt Hougan wrote about this very thing on June 8th and lists the exact same information plus 'tax risk' and 'counterparty risk' (in the case of commodity and leverage funds). I think you are missing the two most important risks: 'Expected Shortfall risk' and 'volatility risk'.
Expected Shortfall is the extra (fat-tail) loss that is ignored using a normal distribution. By converting to a 'Stable' (logarithmic) distribution you can actually see the true risk of a frequency distribution. In other words, it is a Value-at-Risk (VaR) model that better describes the tails of a distribution. With VaR, with may think you stand to lose 3% of the portfolio value on a given day, one percent of the time (at a 99% VaR). With conditional expected shortfall (or conditional VaR) the actual loss 1% of the time may actually be 6%; like what happened this past February.
Volatility Risk is the extra risk you assume by investing in less diversified asset classes. This is a big deal with ETFs. The cause of this problem stems from the sudden interest in ETFs and the need for ETF manufacturers to gobble up their stake in the ETF real-estate game. As the land-grab for ETF shelf space continues so does the increase in volatility. The first ETFs were broad-based market indices, like the S&P 500. The next wave of ETFs was the industry sectors (health care, financials, basic materials, etc.). Because they are less diversified the risk on one industry, in terms of volatility (measured in standard deviation) is 1.3 to 8.6 times the volatility of the S&P 500. Having seized the industry sector space the ETF manufacturers went to the sub-sector frontier to build their niche (such as bio-tech); and henceforth more risk. Not to be out done, competing manufactures launched inverse funds and leveraged funds; again, more risk. Only since June of last year has the risk in new ETF’s subsided with the introduction of fixed income, real estate and some commodity ETF’s. The largest risk in managing a portfolio of ETF’s is in choosing the proper fund universe; then comes the ardent task of fundamental research and asset allocation.

Expected Shortfall is the extra (fat-tail) loss that is ignored using a normal distribution. By converting to a 'Stable' (logrithmic) distribution you can actually see the ture risk of a frequency distribution. In other words, it is a Value-at-Risk (VaR) model that better describes the tails of a distribution. With VaR, with may think you stand to lose 3% of the portfolio value on a given day, one percent of the time (at a 99% VaR). With conditional expected shortfall (or conditional VaR) the actual loss 1% of the time may actually be 6%; like what happened this past February.

Volatility Risk is the extra risk you assume by investing in less diversified asset classes. This is a big deal with ETFs. The cause of this problem stems from the sudden interest in ETFs and the need for ETF manufacturers to gobble up their stake in the ETF real-estate game.