Universa Investments published a research paper titled Pension Funds Should Never Rely on Correlation. The paper is short but it is dense and I was not sure why the title pushed focus onto pension funds as I think the general premise is an interesting question for any portfolio construction process.
I have always been very big on blending together holdings with different correlation attributes as a way to manage volatility. The one line that I think summarizes the entire paper came toward the end: “Would a real pension fund manager simply allocate based on a point estimate of correlation and maintain the allocation indefinitely? As preposterous as that may sound, it has been the actual practice at some funds to subscribe to the 60/40 approach to portfolio construction.”
Basically, the same allocation can’t be optimal for all times which has been my approach from the start. I’ve written countless posts about initiating defensive action when the S&P 500 breaches its 200-day moving average (DMA). Taking defensive action is changing your asset allocation. If you buy an inverse fund you are changing, reducing, your net equity exposure. With perfect foreknowledge you’d be 100% equities on the way up and 100% T-bills on the way down. With no foreknowledge you’re doing the best you can.
If you’ve been reading my posts for a while then you might have heard me talk about holding gold because it has the historical tendency to have a low to negative correlation to equities far more often than not. It’s not absolute but it is often enough for me to believe it is effective. Just this past week I saw an article about gold having done poorly for some number of years. Well, yes because equities have rocketed higher in a manner that is hard to fully understand.
The reason gold isn’t my only low/negatively correlated holding is to protect against the possibility that gold goes down the next time that equities do. We have a little bit in managed futures. Similar story in that it tends to not look like equities. I have unyielding faith that gold and managed futures will work far more often than not but in any given downturn they may not.
To the criticism from Universa that optimal allocations cannot be static, I agree on two levels. Correlation analysis is backward looking. While the correlation of gold versus equities is a very good bet to be low to negative in a very reliable fashion, the correlation of healthcare stocks versus tech stocks and their different demand elasticities, is going to be far less reliable and not a dynamic I’m likely to build defense around.
The other level where I agree about static allocations is more of an in-practice manner which is if you buy an inverse fund for defense whether based on something technical or a hunch (hunches are not ideal of course) and the market does actually go down, your inverse fund goes up thus hedging more of the portfolio the further the market goes down. That is a dynamic hedging strategy, it’s not a static anything.
Tying in a recent post, does this sound simple to you? The answer depends on the end user. I have often said a couple of broad index funds in the right mix with an adequate savings rate is absolutely valid and can get the job done but not ideal for how I view things. Managing a portfolio with regard to changing the allocation in the manner I described above has become second nature but the goal isn’t necessarily long term outperformance. The goals are more about trying to prevent client panic and trying to have as little disruption as possible on client withdrawals.
If you are managing your own portfolio, you should do whatever is second nature to you with the added caveat that you always keep learning more. I would not say you absolutely should be doing things differently than you did ten years ago but you should know a lot more than you did ten years ago and that might lead to a few small changes.