Featured Post

Was it constitutional for Proposition 124 to replace PSPRS' permanent benefit increases with a capped 2% COLA?

In this blog I and multiple commenters have broached the subject of the suspect constitutionality of PSPRS' replacement of the old perma...

Saturday, August 8, 2015

How PSPRS digs itself deeper into a hole without even trying

For me the math and methodologies involved in pension accounting get so complicated and arcane, that it seems more like art than science so I am grateful whenever I see something that provides a simple explanation of a pension accounting concept.  An example is this small piece, David Crane explains the ramifications of CalPERS' 2.4% return for the past year, that appeared on the Pension Tsunami website.  In it, Mr. Crane's explains how an underfunded pension, like PSPRS, is in much worse shape than it appears at first glance:
Recently newspapers have reported that CalPERS earned 2.4% over the last twelve months and contrasted that return with its 7.5% assumed rate of return. But what those newspapers have not reported is that CalPERS needs to earn much more than 7.5% per annum for its unfunded liability not to grow.
This is because (i) under US public pension fund accounting, liabilities grow at the assumed rate of return and (ii) currently, liabilities exceed assets. That means assets have to grow faster than the assumed rate of return in order to keep up with liabilities.
As a simplified example, let’s say a public pension fund has a 77% funding ratio, which means that it has assets equal to 77% of liabilities. For greater simplicity, let’s say it has assets of \$77 and liabilities of \$100, and therefore an unfunded liability of \$23 (100-77). Because of item (i) above, liabilities grow 7.5% per annum. That means liabilities that today equal \$100 will in one year equal \$107.50. For the unfunded liability not to be larger than \$23 at that time, that means assets have to grow from \$77 to \$84.50 (107.50-84.50 = 23). That means that the pension fund needs to earn 9.7% (\$77 times 1.097 = 84.50). Anything less and the unfunded liability will grow.
This is why it’s so hard for US public pension funds to catch up once they fall behind. See this relevant article from The Economist.
If we use Mr. Crane's same calculations, how much does PSPRS, which was only funded at 50.4% on June 30, 2014, need to earn in a year to keep its unfunced liability from growing?  At the 50.4% funded ratio (assets divided by liabilities), PSPRS has only \$50.40 in assets to pay off the \$100.00 it owes to its members, leaving an unfunded liability of \$49.60.  The \$100.00 liability will grow the next year to \$107.85 at its current expected rate of return (ERR) of 7.85%.  In order for the unfunded liability to stay at \$49.60, PSPRS' assets would need to grow from \$50.40 to \$58.25 (107.85-58.25 = 49.60), which translates to an annual interest rate of 15.56% (50.40*1.1557 = 58.25).

If PSPRS were to earn the 7.85%, the 50.4% funded ratio would not change since both its assets and its liability would go up proportionately, but the actual dollar amount of the unfunded portion would increase from \$49.60 to \$53.49 [the new assets would be (50.40*1.0785 = 54.36) and the new unfunded amount would be (107.85-54.36 = 53.49)].  We can see why returns that exceed the ERR are so important when you have a deficit.

It appears that PSPRS will not reach its ERR for the 2015 fiscal year that just ended.  I estimate that PSPRS will earn somewhere between 4.5% and 5.0%.  If PSPRS earned only 5.0% in the last fiscal year, its liability will again grow to \$107.85, but its assets will only grow to \$52.92.  This will mean that its unfunded portion will grow from \$49.60 to \$54.93, and the funded ratio will only be 49.07%, a decrease of 1.33%.

This is an oversimplification of how PSPRS calculates its funded ratio because PSPRS uses a seven-year smoothing period to spread out gains and losses more evenly.  If we do not factor in the negative effects of the Hall lawsuit and a lowering of the ERR to 7.5%, PSPRS should actually see a positive movement in its funded ratio over the next two years because the extraordinarily large losses incurred during the Great Recession should fall out of the seven-year smoothing period.  Regardless, Mr. Crane's calculations are still very enlightening as to how critical market returns are to an underfunded pension.

The interesting thing to look at here is how COLA's or permanent benefit increases (PBI) under the excess earnings model affect the funded ratio.  Remember that at the current funded ratio of 50.4%, PSPRS would need to earn 15.56% to keep its total unfunded liability from increasing.  Remember also that with the excess earning model, PSPRS must pay half of any earnings over 9.0% into a special fund that can only be used to pay COLA's.  This means that PSPRS would actually have to earn a whopping 22.12% in order for its liability to remain the same.  The 6.56% over the 9.0% COLA threshold would need to be doubled in order to maintain the 15.56% rate since half of anything over 9.0% would be placed in the COLA fund.

If PSPRS actually did earn the 15.56% in a year, its unfunded liability would still grow because 3.28% of the gain over 9.0% would go into the COLA fund.  Using our earlier numbers, the liability would still grow from \$100 to \$107.85.  However, the assets would grow from \$50.40 to \$56.59 (50.40*1.1228), and \$1.65 (50.40*0.0328) would go into the COLA fund.  This means that the original unfunded liability would increase about \$1.65 from \$49.60 to \$51.26 (107.85-56.59).  So even if PSPRS earns 15.56% its unfunded liability will still grow larger when the excess earnings COLA model is in place.

Now if we multiply the figures we have been using by \$100 million to bring them more in line with PSPRS' actual assets and liabilities, we can see how much money we are actually talking about.  A
\$4.96 billion unfunded liability would grow by about \$165 million to \$5.126 billion, despite the fact that PSPRS earned an outstanding 15.56% return on its investments.  The other \$165 million would be used to pay a COLA of up to 4% of the average normal retirement, which would increase the total liability and further widen the gap between assets and liabilities.  This is exactly what you don't want to happen when you are already in the hole.