In a couple of previous posts, I’ve discussed the folly of using a loophole in public pension accounting to increase the discount rate by taking on added risk.  Naturally, that fear would be more justified if it turned out that public pensions were doing that.  Herewith, the evidence.

The Census Bureau does an annual survey of the state of public pensions, although they tend to take their sweet time posting the data.  (The most recent data available was for FY2008.)  Since the time when online records begin, in 1993, until 2008, you can see some obvious trends in public pension funds’ asset allocation:

The proportion in US stocks had risen almost 5 percentage points, from 34% to 39%, and the amount in safer, but really boring Treasuries has dropped from almost 20% to just under 5%.  You can see where the poor years for stocks in 2007 and 2008 took back some of their gains, but inasmuch as the fund managers should have been rebalancing, this hardly vindicates them.

The fund managers have also been risking more of their money; cash and liquid securities dropped from about 7.5% to under 3%, although that could be for a number of reasons.  The pension operations could be efficient, for instance, or the managers may have had better actuarial data at their disposal.  In what looks like a gap in the Census survey questionnaire, “Other” is up significantly, from 5% in 1994 to 13% in 2008.  Anything that’s 1/8 of your investment portfolio shouldn’t really be lumped together as “Other.”

And since its introduction as a separate category in 2002, Foreign Investments (not shown) have risen from 12% to about 15%.  That’s probably a result of both better returns and increased investment.

I also want to emphasize that this is aggregate data for State & Local pensions across the US, not just for PERA.  CalPers may well distort the data, and some plans, like DERP, are around 90% funded and suffered very little in the stock downturn in terms of fundedness.

In any event, what is clear is that the instruments best-suited to stable, long-term returns (at least up until now), US Treasuries have either been sold off or allowed to mature, with the funds being put into US stocks and some other, almost certainly riskier, categories.

There’s a lot more data over at the Census site, including some state-only data over the last 3 years that also has actuarials associated with it, so I’m hoping to make some time to dig into that, as well, but as usual, no promises.