Did Pennsylvania take a wrong turn with portable alpha?

On November 25, the Pennsylvania State Employees Retirement System (SERS) announced its Q3 results.  Public pensions across the US issued similar press releases detailing the Q3 carnage.  But what makes this pension plan different is its widely publicized use of portable alpha (see our April 2008 post.)  As a result, the media has been quick to associate the fund’s losses with the “aggressive“, “exotic” and “unusual” investment strategy.  To be sure, it appears that portable alpha cost the plan in Q3.  But is that an indictment of the portable alpha architecture per se?  We’re not so sure.

The Wall Street Journal reports that:

“The blowup is yet another example of the wide-ranging damage caused by sophisticated investment strategies peddled to pension funds and other institutional investors when the stock market was soaring.”

Pennsylvania SERS reportedly lost $1.5 billion on swaps used to gain market exposure.  These swaps allowed the plan to get the exposure it wanted without having to allocate as much capital (i.e. it only had to post margin on the swaps).  It then invested the savings in “absolute return” funds (funds of hedge funds in this case).  According to media reports and the plan’s own press release, it appears that it lost on both the swaps AND the absolute return funds in Q3.

This leaves the plan in the uncomfortable position of having to explain how it lost money on “derivatives investments” – as if the swaps were a stand-alone investment strategy or an out-right market call.  But they were not.  The loss on the swaps is analogous to the market losses experienced by any other plan that invested in equity markets the usual (fully-funded) way.  The real trouble seems to have resulted from a surprise in the performance of the “absolute return” portion of the portfolio, not the swaps.  As a spokesman for the plan told P&I,

“The losses were not because of the use of swaps per se, but because (underlying) equity markets are down.”

In theory, the absolute return investments should have had a very low (ideally a zero) market correlation and a modestly positive long-run return.  Thus, the combination of the market-tracking swaps and the absolute return funds should have produced returns equal to the market plus a bit of icing.

But many hedge fund strategies – including the funds of funds used in this particular portfolio – seemed to mimic equity markets in Q3.  In other words, they seem to have contained a hidden market exposure that was only revealed in times of distress.  The result was that Penn SERS seemed to have an excess exposure to markets (100% exposure, for example, in the swaps and an additional, x% exposure via the supposedly “uncorrelated” hedge funds of funds).

But hedge funds’ surprisingly large correlation to equity markets in Q3 does not, on its own, repudiate portable alpha as a portfolio construction technique.  As we mentioned after August 2007’s quant drawdown, a zero correlation does not mean that the fund and the market are immune to concurrent drawdowns.  While a perfectly stable monthly return (i.e. cash) has a low (zero) market correlation, so can a more volatile hedge fund.  The hedge fund, in theory, produces intermittent monthly drawdowns that are simply unrelated to those of the market.  The key is that “unrelated” means they can easily be negative when the market is also negative.

Here’s an example of two randomly generated return streams with a zero correlation over 30 months.  Both return streams have an 11bps average monthly return and have very similar standard deviations.  Note that despite their zero correlation, their signs are the same half the time.

In four of the thirty months, both funds have a significantly negative return.  Yet such a combination remains a viable portable alpha solution since, over the long run, you never really know what one fund will do based on the returns of the other.

This is a simple illustration only.  The actual “absolute return” funds used by SERS may have had a significant positive market correlation over many months or years.  If so, then one might expect the amount invested in the market-replicating swap to account for this.  (We have no more information than what has been reported in the media and by SERS.)

The Massachusetts Pension Reserve Investment Trust (PRIT), another fund reporting a recent loss in its portable alpha program, pins the blame on the investments selected for the program, not on the concept itself.  Its Executive Director told Dow Jones last week:

“Over the short period of time [that it’s been used], it’s not performed well…But we think that’s more a reflection of market conditions. We’re not thinking of scrapping portable alpha.”

And the WSJ reports that Penn SERS’ portable alpha program remains in the black since inception:

“[Pennsylvania SERS] officials say that, even with that setback, the strategy has generated $500 million in cumulative above-market returns.”

So the bottom line seems to be that portable alpha strategies, like 130/30 strategies, are only as good as the raw materials used to construct them.  The reported “failure” of SERS and PRIT programs may offer some important lessons, but they will likely not end the debate over portable alpha.

Be Sociable, Share!

Leave A Reply

← London Notebook: HF regulations in 1609, VIX to the rescue, and no premature demise for clones Fee-conomics →