It is easy to get confused about statistical correlation and the behavior of inverse funds as is found in this article recently published at SeekingAlpha. The author applauds the availability of inverse ETF's and proposes that these will enable retail investors to reduce risk through a process of diversification

One of the simplest ways to do this is through the use of non-correlated investments that zig when everything else zags. In the old days, that meant having exotic futures accounts or taking positions opposite to the markets entirely, using margin accounts to sell individual stocks short - one stock at a time.

Inverse funds, as their name implies, go up in value when the markets go down. There are plenty of choices to consider, with everything from the S&P 500 to specific sectors available in the mix.

Unfortunately the author seems to blow his argument apart about the usefulness of inverse funds as part of a strategy to seek out uncorrelated asset classes in a diversified portfolio when he also says later on:

Of course, if the markets go up, the reverse is true and these things can lose money in a real hurry, so one can’t just pile in indiscriminately.

If you are long an inverse index fund and the market goes up your returns will be highly correlated with the index fund - it's called negative correlation.

The take-away is that it's harder to immunize a portfolio from system risk than it may appear. Correlation based strategies are also hazardous for many other reasons which I have discussed in *Long/Short Market Dynamics. *

The most insidious reason is that during periods of liquidity crises the co-variance of asset classes rises dramatically towards unity i.e. they all go down together.