Correlating Risky Assets

Asset allocation is tough, because the correlations are not stable.  Here’s an example: in the 90s, at many conferences that I went to, I was told that one of the smartest moves you could make was to invest heavily in every new class of Asset Backed Security [ABS] created, because they all tighten in yield spread terms after issuance, leading to price gains.

I didn’t believe it then, and that was a good thing, because the most exotic of ABS classes got whacked in the financial crisis.  As it was was, I had already seen debacles in Franchise Loan ABS (spit, spit), and Manufactured Housing (post-1997 vintage).  At a conference for Life Insurance, I was a skunk at the party in 2006, as one ignorant presenter suggested that AAA structured assets never went bad.  History already taught us better, and as I tried to say to the then-CEO of Principal Financial as he was exiting the conference, he needed to look at the mezzanine and subordinated structured product in his company.  Free consulting, but but worth more than the consensus.  As far as I can tell, he didn’t listen.  For many reasons the stock price is lower today.

I have many other tales where in fixed income (bonds), everyone “followed the leader,” which worked in the short run, but failed in the long run.  The point is that investor behavior correlates asset classes.  There may be underlying economic differences, such as owning a natural gas producer and utility that uses natural gas, but most of those differences get erased as most investors seek portfolios immune from factors of secular change.

So as new asset or sub-asset classes are introduced, in the short-run they are uncorrelated, and likely rally, because few own them.  But after the rally, many now own it, and the future correlations are high because so many own it.  The correlations ultimately depend on two things: the underlying economics, and investor behavior.  Investor behavior is the dominant aspect of pricing.

I don’t think there is a lot of diversification in most risky asset classes from an economic standpoint.  Does it matter whether a business is public or private?   I think the answer is no.

What that means in the present environment is that there is a gap between business risk, and those that finance business risk.  In other words, there is a difference between investment grade bonds, and risk assets.  That’s the negative correlation in this market.  Do you want diversification?  Buy some ETFs that invest in long high investment grade debt.  You will not get any effective diversification out of buying different classes of risky assets.  Those are already owned by those that compete with you.

Promises to pay from sound entities that can be relied upon in the future behave very differently than risky assets.  In your asset allocation, to the degree that you need real diversification, look at that as the critical distinction.  All other distinctions are secondary at best.


  • Doug says:


    In the end, there are only three asset classes: senior claims on cash-flow (bonds); claims on operating cash-flow (real-estate); and residual claims on cash-flow (equities). I see commodities as a specialized form of real-estate, as commodity ownership is a claim on feedstocks.

    That’s where your primary diversification comes from. Since most people (unless they are young) are long residential real-estate–and levered, to boot–an equity/long-bond balanced portfolio is a reasonable allocation.

    I would say there are some gains to be made from geographic and industry diversification, on both stocks and bonds: government bonds, corporate bonds, commercial mortgage bonds; energy stocks, insurance stocks, consumer stocks; stocks and bonds from the developed world and the developing world. These diversifications mean little on a one-off basis, compared to the asset-class gains, but they are meaningful when taken in sum.

    As to asset-backed bonds in the ’90s, my Dad always taught me never to buy the first year’s production of a new automobile model. So I missed the Nissan Mac 1988s. But I bought the GMAC 90s, the Prime 93s, the Greentree 95s, the Metris 98s, but always the short deals. I never went for the David Bowie bonds–after the 144a premium, there wasn’t much economic return left.

    When Greentree blew up, I knew that the new issue perfect game had been blown, and it was time to change the pitcher and move up in the cashflow seniority. But I still got caught by DVI (fraud) and GMACM-03s, because I bought 5 year bonds (that are still out there). Who knew that small loan sizes meant it wasn’t worth collecting. It turned out that smaller loans were less secure, because they were uneconomical to collect. Better to have a small piece of a pool of big loans–a case for CMBS!

    In the late ’90s it was still true that no losses had been recorded on subordinate ABS, so no wonder the herd rushed into senior ABS and super-senior CDOs in the 00s. Now, I think opportunistic scavengers have the best pickings, but it’s hard to live on leftovers.

1 Trackback