Archive for category Budget
Last week, former PERA Executive Director Miller Hudson penned an op-ed for the Denver Post, arguing that PERA’s situation has improved to the point where we need not worry about it, and that no further tinkering with it is necessary (“There is no need for panicky ‘fixes’ to PERA“). Unfortunately for the taxpayers of Colorado, Mr. Hudson’s comforting conclusions are belied by some uncomfortable facts.
Let’s begin with where Mr. Hudson places the blame for the current funding problems. He identifies one of them correctly – overly generous benefits that amount to promises that cannot be kept, except at great expense. He is also correct that the dot-com bubble was fool’s gold for the legislature, which led it to create the overly-generous benefits.
But PERA’s portfolio managers (who predate Mr. Hudson’s tenure as Executive Director), allowed the fund’s investments to become dangerously overweight in volatile stocks, in effect letting their winning bets ride. When the dot-com bubble burst, so did PERA’s funded ratio, and it continued to decline throughout the decade, recovering only slightly in the mid-00s:
This chart also shows the folly of relying on long-term returns to determine a fund’s solvency. If a plan is underfunded, adding additional return may look like the way to catch up. But along with that additional reward comes additional risk and volatility. When the portfolio has a bad year, as in 2000, 2001, and 2008, it doesn’t have the option of drastically reducing its payout that year, as you or I would with our own retirement accounts. The need to pay benefits regardless of the fund’s annual return can put it in a hole that it can never recover from. PERA’s estimate of 8% may indeed be a realistic return over 30 or 40 years. But benefits need to be paid when they need to be paid, and the results of this thinking are all too obvious in the above chart.
And while the legislature rarely met its Annual Required Contribution (a contribution set by Government Accounting Standards Board, and designed to ensure actuarial soundness), this shortfall was only a relatively minor factor in the fund’s increasing unfundedness. According to the chart below, had the legislature made the ARC every year from 2000 on, the State and School Divisions, which comprise the overwhelming part of PERA, would only have been about $4 billion better-off last year. PERA admits to a $23 billion unfunded liability, although there is reason to believe it is much larger:
Mr. Hudson also argues that, because overall, PERA contributions account for less than 3% of public spending, the burden is light. This ignores that for many entities – school districts, in particular – PERA spending is eating up an increasing portion of their operating expenses:
This is a result of the very supplemental payments (SAEDs) that are designed to save the system from ruin. PERA is correct that the supplemental payments were envisioned as being shared between the districts and their teachers. But with many, if not most, school boards under the thumb of the teachers’ unions, they have decided to have their districts absorb the entire supplemental payments. This means that as of 2011, for four major Denver-area school districts, roughly 11% of their operating expenses were going to teacher pension plans, money that could have gone into the classroom.
Mr. Hudson tries, implicitly, to discredit those who are concerned about PERA’s fiscal condition by claiming that it is only “in recent decades” that concern has grown up around the unfunded liability. While it is true that in the past, PERA has been significantly under-funded, two conditions make that of greater concern now. First, the PERA unfunded liability is much larger now as a percentage of the state GDP, meaning that should a fix become necessary, the pain to the state’s taxpayers will be considerable greater than it has been in the past. In the 1980s and early 90s, the unfunded liability hovered around an unthreatening 2% of state GDP. That has since grown to 9%:
Second, since PERA has an unfunded liability, it means that some of its current expenses are paid for by current employees. (A fully-funded program would, by definition, have all current expenses in the bank.) The ratio of current employees to retirees has been falling for decades, as well, meaning that any increases in contributions will fall more heavily on future employees and future taxpayers:
As part of his rhetoric, Mr. Hudson contrasts the concrete – and real – improvements from SB10-001 with unnamed and undescribed “fixes” proposed by those who worry about PERA’s financial condition. This leaves the reader to imagine all sorts of horribles. Let’s look at some of the “panicky” fixes proposed in the state legislature over the last several years:
- HB13-1040: Would have calculated benefits on the basis of seven, rather than three years’ pay, making “spiking” more difficult to achieve
- SB13-055: Would have applied the same liability discount rate rules to PERA as apply to US private pensions and European public pensions
- HB12-1142: Would have given all PERA members the option to join PERA’s own defined contribution plan
- HB12-1179: Would have broadened the composition of PERA’s board to reduce conflicts of interest and increase accountability
- SB12-016: Would have given local governments the same option the state government has to make plan members pick up more of their benefit contributions in times of fiscal distress
- HB12-1250: Would have calculated health care benefits on the basis of costs, rather than employees’ salaries
- SB12-082: Would have set the PERA retirement age to that of Social Security for non-public safety members, a matter of basic fairness
- SB12-119: Would have forced PERA make adjustments until its plans could meet a 30-year amortization window, the standard for pensions
- SB12-136: Would have included PERA benefits in the state’s Biennial Compensation Report
All of these changes are designed to increase transparency, increase accountability, and decrease conflicts of interest. All of them are designed to increase fairness, and increase the likelihood that PERA retirees will be able to rely on promises made to them.
It is telling that each of these changes – every last one – has been opposed by PERA and its allies in the public employees unions here in Colorado.
And it’s enough to make you wonder who’s really panicking.
Defenders of PERA often argue that while the liabilities have been under-funded in the past, it is only now that PERA’s critics have begun to worry about the matter. The implied message is that the complaints are political, rather than financial. Here’s why this isn’t the case:
Yes, Colorado’s economy has grown, but the PERA liability has grown faster. While through the 80s and most of the 90s, the unfunded liability hovered around 2% of the state’s GDP, since 2000, it has grown to 9%. Of course, during the good economic years, it declined somewhat, and it may well decline a little again this year, as PERA’s returns are expected to be around 12% on its portfolio. But sooner or later, we will hit a cyclical recession, and even as the economy shrinks, PERA’s unfunded promises will continue to accumulate.
Did you recognize the faulty presumptions in PERA’s spirited defense of defined benefit plans?
You have been given a false choice about why defined benefits plans are better than defined contribution plans.
In a recent EdNews Colorado Voices column, Colorado PERA Executive Director Greg Smith avers that PERA’s existing defined benefit structure best serves both the teachers and the taxpayers of Colorado. He was responding to a report by the National Council on Teacher Quality that leads the reader to support reforms to move away from the existing scheme and toward a defined contribution plan. Smith’s claims are wrong about the advantages of defined benefit plans in general, and PERA’s actuarial soundness in particular.
Smith cites a National Institute on Retirement Security (NIRS) report that claims three advantages for defined benefit plans over defined contribution plans:
- Less error in the amount saved for retirement,
- Less need to rebalance and re-allocate assets over time, and
- Better returns, largely as a result of lower transaction costs.
Each advantage turns out not to be dependent on having a defined benefit plan, but on having a professionally-managed, aggregated plan. The same advantages would accrue to a similar defined contribution plan that was also aggregated and professionally-managed.
PERA already has such an option, PERA Plus. It’s organized as a three-part 457(b) / 401(k) / Defined Contribution option. Like any set of diversified retirement offerings, it includes a variety of funds with different investment goals. For our discussion, the most relevant set of funds are those with target retirement dates. PERA has nine of these, with target dates every five years from 2015 to 2055, and an Income Fund designed to provide current income for current retirees.
Over time, as the target date for each fund approaches, that individual fund reallocates its assets into more conservative investments, before maturing and merging into the Income Fund. While each individual fund “ages,” all the funds collectively are maintaining a proper average. Taken together, they continue to represent the aggregate ages and target retirement dates of the entire set of members, the very source of the first two alleged advantages. The third, that of lower transaction costs, is completely independent of how liabilities are calculated.
There is no inherent reason why the assets of a DB plan should earn a higher return than those of an identically-invested DC plan. The only mandatory difference is that the defined benefit plan beneficiary has a share only in the specific benefits to be paid – the fund’s liabilities. By comparison, the owner of a defined contribution plan has a property right in the assets. Therefore, while a defined contribution plan is, by definition, always fully-funded, a defined benefit plan may have to seek additional funding, or trim back on its promises, in order to remain so.
The danger of unrealistic promises
It is therefore imperative that the promises being made to future retirees be realistic. All the more so if the promised benefits are being used to attract and retain qualified or exceptional teachers. Unfortunately, it is far from certain that PERA can afford the promises it is making, given its current funding levels. Recent legislative reforms (Senate Bill 10-001 in particular), while welcome and substantial, simply do not close the gap.
By PERA’s most recent published calculations, its unfunded liabilities remain at a staggering $26 billion, and its overall funded level is well below 60 percent, on a par with the chronically ill Illinois public pensions. In fact, a recent study by Barry Poulson suggests that PERA could be in the worst shape of any statewide plan in the country.
Let’s give credit where credit is due. PERA’s adoption of a 401(k)-like portability is indeed commendable. But if it’s designed to mimic the properties of a 401(k), it can hardly then provide an advantage over one.
While PERA is no longer “letting it ride,” as it did with its stock market investments of the late 90s, the 8 percent returns needed for a return to solvency come with risk. Even better-than-average returns during regular years won’t make up for prior losses in bad years, because funds must then catch up, while payments can’t be deferred.
What success SB1 does offer is predicated on both benefit reductions and payment increases. However, a court challenge to the limitation of COLAs to 2 percent has been upheld by a State Court of Appeals, and its future is uncertain at best. Should the lower courts find that limitation not to be justified, most of the immediate reduction in PERA’s unfunded liability will be wiped out.
On the contribution side, PERA plans to require supplemental increases, rising incrementally from 2 percent to 5.5 percent until 2018. School districts have been picking up the tab for these increases, rather than passing them on to the teachers themselves, as they are allowed to do. As a result, PERA now absorbs upwards of 15 percent of annual operating expenses in many large school districts, a number that is expected to rise to 20 percent as the existing plan increases for make-up contributions.
Disclosure of ties to lobbying group needed
It is also worth noting that the institute that issued the favorable DB article (NIRS) is the lobbying and public policy arm of the defined benefit public pensions, with a particularly close relationship with Colorado PERA. Smith sits on the board of directors of NIRS, as does Meredith Williams, PERA’s former executive director. Colorado PERA is both a charter member and in NIRS’s Visionary Circle, along with such other public plans as CalPERS and the Illinois Municipal Retirement Fund.
Inasmuch as NIRS is not an independent think tank, but instead is a creation of interested parties to the debate over public pensions, this relationship ought to have been disclosed.
While there is no doubt that total compensation is an important part of attracting and retaining effective teachers, those promises must be grounded in reality. Until realistic arguments are used, PERA will continue to fail not only its member teachers, but also the schools and parents it is intended to serve.
This article originally appeared in EdNewsColorado.
We’ve been here before with PERA. Sort of. Most people following the issue remember that in 2000, PERA was over 100% funded, and that its funding ratio has fallen steadily since then. What they don’t know is that back in 1974, PERA was woefully under-funded, at about the same 60% funded ratio that it is now. It took advantage of the long bull market to pull itself out of that situation:
Note that as the funded ratio rose, so too did the percentage of the portfolio allocated to common stock (both domestic and foreign) rose, as PERA basically decided to let the bets ride, rather than re-allocate to maintain the lower-risk portfolio. When the bubble burst, they ended up paying the price for having stayed too long at the fair. Now, PERA has returned to a somewhat more conservative allocation strategy, targeting 25% of its money for fixed income, and a target of 58% in stocks. Nevertheless, this is a far cry from the 45% or so in bonds that they held up until 1992 or so, and the nebulous “Alternative Investments,” which includes things like venture capital (and in which I’ve included the Lumber investments), suggests that PERA is still chasing yield there:
So this just puts us back where we were before, right? We climbed out of this hole before, we can do it again.
Not so fast. First, as noted before, PERA’s in a less aggressive portfolio now than it was in 2000. This is a good thing, since it takes out some of the volatility from its portfolio. But it also means that it probably can’t count on a run of good luck to lift it out of unfundedness the way it did last time. Also, as we’ve previously noted, the fall from grace in 2001 and 2002 wasn’t just a matter of poor returns, it was also a matter of increasing liabilities with more generous benefits. That hasn’t gone away.
And not all 60% funded ratios are created equal. Here are PERA’s inflation-adjusted, per-capita unfunded liabilities since 1974 (constant 1983 dollars):
On a per-capita basis, the overhang is about 4x what it was in 1974. So in fact, we’re in much, much worse shape than we were 40 years ago when this roller-coaster ride began.
In the discussion on public pensions, there’s been a great deal of focus on the projected rate of return. I’ve posted on what I think is PERA’s optimistic 8% here, and on the fact that that’s actually an improvement from the 8.75% that they were projecting as recently as 2002. That said, for pension estimates, their inflation assumptions matter as much as their raw return assumptions. The actuarial consequences of poor inflation estimation are too much to summarize here. But even on the basic question of returns inflation matters: the real return on an investment is the nominal return minus inflation.
Over the last 10 years, public pensions have gotten some credit for modestly reining in aggressive growth assumptions. PERA, for instance, has moved from a 8.75% growth assumption to 8%, and CalPERS has made similar adjustments. Overall, the average growth assumption has dropped slightly from 8.04% to 7.86%. But the average inflation assumption for public pensions nationally has dropped from 4.0% to 3.31%. This means that instead of decreasing the real return assumption has actually gone up from just over 4% to just over 4.5%.
For the record, PERA’s inflation assumption was dropped from 4.5% to 3.75% in 2003, where it has stayed. Both the investment return and inflation numbers are higher than the national average and national median, though.
I don’t really think that the inflation numbers here are unreasonable. And my problem with PERA’s 8% return assumption goes beyond the average itself – 8% has been the historic return on stocks, and doesn’t take into account the additional volatility and risk that come with higher return. But it’s clear that PERA and other plans have been dining out on their flexibility on returns, while the increase in real expected returns goes unremarked-on.
The disconnect also highlights the price we’re going to pay – in accuracy, and eventually in dollars – for using the rate of return as the discount rate. Interest rates are closely tied to expected inflation, and here the funds themselves are admitting that the gap between the rate of return and the proper discount rate has been growing.
This chart is more or less self-explanatory. Over the last 20 years, the number of workers per beneficiary within PERA has dropped from about 3.6 to just over 2.0:
PERA’s CAFR includes the following disclaimer:
By itself, a declining ratio of actives to retirees and beneficiaries does not pose a problem to a Division Trust Fund’s actuarial condition. However, to the extent that a plan is underfunded, a low or declining ratio of actives to retirees and beneficiaries, coupled with increasing life expectancy, can complicate the Division Trust Fund’s ability to move toward full funding, as fewer active, contributing workers, relatively, are available to amortize the unfunded liability.
This is about right, although even a fully-funded system won’t stay fully-funded for very long under these conditions. Indeed, PERA was fully-funded as late as 2001. In the 90s, PERA’s long-term problems were masked by a tech bubble, and when that burst in 2000, the fund started to fall into an under-funded state that it’s never recovered from. Since under an underfunded defined-benefit plan, current expenses have to be paid for out of current contributions, and fewer workers are pulling the cart for each retiree, the deficient horsepower will have to be supplied by the taxpayers.
Rep. Mike Coffman took a lot of heat for suggesting that perhaps it was time to re-evaluate these priorities, with his words being distorted into an attack on the liberal arts in particular, and higher education in general. Turns out he was onto something.
- Fine Arts
- Religious Studies/Theology
Just for grins, I looked up how many of CU’s undergraduate degress over the last quarter-century (well, since 1989), have been awarded in these majors:
For fun, I added in almost anything with the word “Studies” in it, and that’s what the percentage line (right-hand axis) shows.
Peaking at just over one-third of all bachelor’s in 2004, over 30% of all undergraduate majors are still in these low-return majors. The dropoff occurred between 2004 and 2008, but has since – astonishingly – stabilized since the popping of the housing bubble, when job prospects for graduates have almost never been worse. Students are taking on crushing burdens of debt to graduate with these degrees.
You and I will be subjected to sob stories about how Colorado is under-funding its higher education. Instead, perhaps we ought to be taking a closer look at what we’re funding.
One of the favorite tropes of PERA apologists runs like this: PERA was fully-funded in 2001 or so, at which point the state legislature began failing to make 100% of its Annual Required Contribution (ARC). It was then that PERA’s funded level began to drop off. Therefore, if the state legislature had fully-funded the ARCs, today, PERA would not face a massive unfunded liability.
It’s a rhetorical masterstroke, redirecting blame for the current situation on a stingy state legislature that put PERA last in its priorities. And while grounded in a grain of truth, it severely understates and misattributes the nature of PERA’s financial crisis-in-the-making.
The grain of truth is this: the state legislature, beginning in 2003, began to under-pay its Annual Required Contribution. Of course, this affects not only the immediate year, but all future years going forward. Not only is the money from the shortfall not there, but the accumulated return on those dollars aren’t there, either. For this post, I’m just going to focus on the two largest divisions, the School Division and the State Division. They were combined in 1997, and separated again in 2006, so I’ll consider them as a unit.
Here are the yearly shortfalls, along with their future values to the end of 2011 (the latest year for which we have data). The first year for each underpayment is dollar-cost-averaged, so we give half the year’s return, and full yearly returns thereafter:
For 2003, the legislature underpaid by about $177 million, costing about $142 million in future returns, for a total effect in 2011 of $319 million. If you add up the total effect, year-by-year, you get the following result:
So by the end of 2011, the cumulative effect of 9 years’ worth of funding shortfalls is a little over $4 billion. The argument by PERA hinges on the fact that it’s at 2003 that PERA began to be underfunded:
As you can see, though, PERA was already suffering from poor 2001 and 2002 returns, even though there was no shortfall from the state those years. What did increase substantially was the size of the liability; the size of the assets actually held steady. From 2004 to 2007, solid returns managed to keep the dollar amount of the gap from growing. But then 2008 hit, and the size of the obligations continued to increase even as the fund got clobbered in the market. The liability dropped as a result of certain stop-gap changes that were made in 2009, but has since resumed its upward march, even as the actuarial value of the divisions’ assets has continued to fall.
Would it have made a difference if the state had made good on its entire ARC for 2003-2011. The answer is yes, but not very much. Adding in the cumulative shortfall each year, here’s the effect on the assets and the funded ratio for the State and School Divisions:
Despite some increased, they’re still seriously underfunded. Since it’s difficult to see the difference between the two charts, I’ve made the comparisons here. First, the difference in assets:
An increase of total assets from $31 billion to $35 billion is not nothing, as they say down at the station, but it’s also not nearly enough to start to close the gap with liabilities. So little that the difference in funding ratio barely moves the needle:
For those of you who want it all on one chart, possibly for optical exams, here it is:
In reality, it’s worse than this. Prior to 2006, PERA didn’t report a sensitivity analysis on its assumed rate of return, so we have only the values for 8%. If we assume a more realistic 6.5% return going forward the unfunded liability grows from $25 billion to about $40 billion, and the extra $4 billion makes even less of a dent.
PERA isn’t suffering from a legislature that isn’t keeping its promises, it’s suffering from having made promises it can’t keep. And it’s the very PERA members who are going to get hurt the most, the ones who’ve been sold a bill of goods about what’s waiting for them when they retire.
Ultimately, large changes will be needed to keep PERA solvent. But little changes can have an effect, too. Today, the state House Finance Committee will be hearing HB13-1040 from Republican Rep. Kevin Priola:
Current law averages the 3 highest annual salaries of a member of the public employees’ retirement association (PERA) when calculating that member’s retirement benefit amount. The bill increases the number of highest annual salaries used from 3 to 7 for anyone who was not a member, inactive member, or retiree of PERA as of December 31, 2013.
In 2010, SB1 changed some rules to make it more difficult for employees to suddenly spike their salaries and other compensation at the ends of their careers, in order to game the system and maximize their PERA benefits. For instance, for benefit calculation purposes, raises and other increases in compensation – like saved vacation being cashed in – were limited to 8% in any given year. Employees could receive larger raises, but benefits could only be calculated on the first 8% of the increase.
This bill would make it even harder to game the system by averaging the highest seven years’ compensation instead of the highest three. It’s a reasonable measure, and it would only apply to employees who join PERA after the end of this year. The Democrats enjoy a large majority on the Finance Committee, so they may well kill the bill. But the fact that it got assigned to the Finance Committee at all, rather than relegated like SB13-055 to the State, Veterans, and Military Affairs Committee, makes the outcome less certain.
The state Democrats, led by State Senate President John Morse (D-Colorado Springs), are continuing their deeply unserious approach to Colorado’s massive unfunded PERA liability this session. Incoming Speaker of the House Mark Ferrandino (D-Denver) had already indicated as much, first by characterizing those who would seek to deal with the problem as wanting to use the economic situation as an excuse to take away public employee’s pensions, and then by appointing Lois Court (D-Denver) as Chairman of the House Finance Committee.
The latest sign comes with respect to a bill proposed by Republicans Sen. Kent Lambert of Colorado Springs and Rep. Lori Saine, a freshman Republican from Dacono. The bill, SB13-055, would require PERA to use the state’s long-term borrowing interest rate as the discount rate for its liability, and would require that the CAFR be released by May 31 of each year. It would also require that contribution and/or benefit levels for any individual fund be adjusted when the amortization period for that fund climbs beyond 30 years.
All of these are eminently reasonable proposals. For reasons discussed before, the discount rate should be the required rate of return of PERA’s investors, the members. There is also no real reason why PERA can’t produce a CAFR within five months of the end of the calendar year; those dates have been slipping in recent years, but the fact is, the bulk of the CAFR is boilerplate or easily-written text, and the same financial statements and charts each year. And since the stated goal of PERA is to be within a 30-year amortization window, requiring them to be so would simply put teeth into an existing target.
Yet Sen. Morse has chosen to assign the bill not to the Senate Finance Committee, which has jurisdiction over PERA oversight, but to the State, Veterans, and Military Affairs Committee. That committee, often referred to as the “kill committee,” is traditionally staffed with the most partisan members of both parties, and used to kill inconvenient bills. Often that’s because “no” votes on the bills might be embarrassing to the majority, perhaps sufficiently embarrassing that some members wouldn’t be able to resist the temptation to vote for them.
That Senate President Morse has chosen to put this bill in front of the kill committee is practically an admission of its common sense, and his lack of it, but it does make clear the two halves of the Democrats’ full-court defense of their public employee clients’ pension plans.
First, they’ll protect them from any votes they might lose, and also protect the House and Senate Finance Committee members from having to cast “No” votes they’ll later have to explain.
Second, both Ferrandino and Morse have claimed that 2010′s SB1 “solved” the PERA problem for good, though tremendous liabilities remain, even given the plan’s own overly-optimistic assumptions and accounting practices. That will be their explanation for dodging these votes, but even if true, it’s a non sequitur. If PERA is truly fixed, surely members of the committee charged with its oversight would not only be the best-informed of that fact, but also best able to explain it to constituents. Putting the bill in front of a committee whose actual charter has nothing whatsoever to do with PERA, and whose reputation is one of chief enforcer, can’t inspire much confidence that the Democrats will be dealing with the state’s financial problems in good faith in the next two years.