I find myself in the peculiar position of having to defend CalPERS. But then again, it’s come out an accidental winner in a Godzilla-versus-Mothra contest with the New York Times’ Dealbook.
The Times’ remarkably fact-challenged article, A Sour Surprise for Public Pensions: Two Sets of Books, is false in its major contentions, as documents readily available on CalPERS’ website show. It looks like reporter Mary Williams Walsh was taking dictation from public pension fund opponents and didn’t bother with rudimentary fact checking. Note that the errors were ferreted out in less than 24 hours, with no input from CalPERS’ staff. What is the Grey Lady’s excuse for this shoddy performance?
We’ll get into the details shortly, but let’s debunk the two Big Lies at the heart of this article. First, CalPERS does not keep different sets of records. The pension fund has a very conservative, arguably punitive, methodology, which is enshrined in law for how to determine how much government entities that contract with CalPERS owe when they exit the pension system. As we’ll discuss, there are good reasons for discouraging exit. Second, the termination amount has been disclosed for every plan on CalPERS site since 2011, and prior to that, CalPERS would provide that information if an employer asked for it.
In other words, the failure of the six-employee Citrus Pest Control District No. 2 (CPCD), the itty bitty pension fund at the heart of this story, to access public information is being presented as some sort of nefarious plot. It had an overfunded pension, wanted to leave, but found that the method that CalPERS used to determine its “termination liability” resulted in the district owing money to CalPERS, not vice versa.1
That means these rubes did not do their homework and got stuck with a big bill.
Let us reiterate: all this means is that CalPERS uses a take-no-prisoners valuation methodology for departing employers. In general, parties exiting a service agreement early routinely face penalties. Intuitively, it makes sense that the punishment should be more severe if the agreement is long-term or the cancellation is disruptive. In the case of CalPERS, the concerns include having employers game the system or panicking in a severe down market when there isn’t much liquidity. CalPERS’ selling to cash out departing employers could result in losses on distressed sales, hurting the remaining members.
There are other reasons for discouraging employers from exiting CalPERS. No matter what you think of CalPERS, having the system manage a defined benefit plan is better for employees than any other deal on offer. How can a six employee plan possibly think it’s going to do better for its worker via a defined contribution plan, as in a 401 (k)? he fees and expenses, not just for any mutual funds in the plan, but for the 401 (k) administration, are going to be vastly higher than what CalPERS incurs. As John Bogle of Vanguard has stresses, higher annual fees and costs create an enormous drag on investment performance over time. And that’s before you get to the costs to CPCD of setting up the plan.
Now one can argue that CalPERS (and the legislature and unions, who participated with CalPERS in devising this approach) would have done itself a favor to by emulating the IRS and just charging a big explicit exit toll. My guess is the concern was that it could be whittled down over time. But the New York Times latches onto this strict valuation approach and uses it to try to depict CalPERS as admitting that it should be using a vastly lower return rate, one advocated by critics of public pension funds.
This episode demonstrates how the California pension fund has set itself up for the media and general public to be prejudiced against them, particularly on the core issue of the sustainability of their performance, by refusing to lower their return target (in anything other than an attenuated, Rube Goldberg manner), despite 30 year Treasury bond yields having dropped by roughly 200 basis points since 2007, the 10 year Treasury even further, foreign central banks constraining the Fed and economists widely discussing a “new normal” of lackluster growth. CalPERS’ credibility is weak also by virtue of its retreat into less transparency under former CEO Anne Stausboll, senior staff members repeatedly demonstrating that they don’t understand private equity, staff too often lying casually in board meetings, and bad-faith responses to FOIAs, which creates ill will among reporters. But none of that justifies a basic fail of journalism, which in this case results from an obviously biased account.
Let’s go through the misrepresentations in the article.
False claim 1: CalPERS kept its termination charges secret. From the article (emphasis ours):
When one of the tiniest pension funds imaginable — for Citrus Pest Control District No. 2, serving just six people in California — decided last year to convert itself to a 401(k) plan, it seemed like a no-brainer.
After all, the little fund held far more money than it needed, according to its official numbers from California’s renowned public pension system, Calpers.
Except it really didn’t.
In fact, it was significantly underfunded. Suddenly Calpers began demanding a payment of more than half a million dollars…
It turns out that Calpers, which managed the little pension plan, keeps two sets of books: the officially stated numbers, and another set that reflects the “market value” of the pensions that people have earned. The second number is not publicly disclosed.
This is utter horseshit and the reporter, Mary Williams Walsh, who has been reporting on pensions for decades, ought to be ashamed. You can easily find a report on CalPERS’ website for the “MISCELLANEOUS PLAN OF THE CITRUS PEST CONTROL DISTRICT #2 OF RIVERSIDE COUNTY” for fiscal year 2013. I’ve also embedded it at the end of this post for your convenience.
The report is dated October 2014. That means this information was public and the Times is flat out wrong to claim otherwise. Moreover, it was clearly available to the Citrus Pest Control District before it made its decision to terminate its plan with CalPERS “last year,” which is clearly sometime in 2015.
CalPERS has a very large staff that fields questions of beneficiaries and employers. The text of the report is consistent with it having been mailed to the CPCD (“Section 2 can be found on the CalPERS website…”). So the CPCD apparently couldn’t be bothered to read reports from CalPERS in full.
The report also encourages the CPCD to call the actuary who signed the report if they had any questions. CalPERS can’t be faulted for the CPCD not reading the report and calling to clear up anything they didn’t understand before pulling the exit trigger. This oversight failure is even less defensible given that the board for this plan has a fiduciary duty to the plan beneficiaries.
False claim 2: The plan was underfunded. No, it wasn’t. As the report shows, the bill to the CPCD was solely the result of the decision to leave the CalPERS system. That triggered a termination liability. As you can see from the CPCD’s October 2014 report, it has a section that is clearly flagged in its Table of Contents: “HYPOTHETICAL TERMINATION LIABILITY”. If you read the earlier sections on the plan’s valuation, you’ll see caveats2 before you get to the meat, which are the estimates of funding status. As you can see on pages 8-19, the plan was over 40% overfunded using a 7.53% return assumption, had no projected funding requirements over the forecast horizon (through fiscal year 2020-21) and even under the most pessimistic return forecast of -3.8% returns for 2014-2017, would still not require any extra employer contribution over that period.
CalPERS then on page 20 has its Termination Liability section which shows how CalPERS arrives at this result.
False claim 3: CalPERS has “two sets of books”. This isn’t true. This is not “two sets of books”. This is “Here is what happens when you terminate your contract early because we have the right to make it painful to you because it’s painful for us.” And the punitive logic is to protect the remaining plan beneficiaries.
CalPERS makes termination liability estimates only for government agencies that join CalPERS. Those employers have effectively contracted out their pension fund management to CalPERS. The workers in those plan account for only 1/3 of CalPERS’ total beneficiaries. Roughly an additional third is for community college and non-teacher school employees. The final third is state employees. State employees and schools must participate, save those charter schools that are charter schools (they are treated as government agencies and thus can opt in). Since the schools and the state employees cannot exit CalPERS, there is no termination computation for them. You can see the most recent actuarial reports for the state employees and the “schools pool” for confirmation. So how can one possibly assert that there are “two sets of books” when the item is question is a computation listed as a single item in a lengthy report prepared for particular employers, and not done for 2/3 of the CalPERS beneficiaries?
False claim 4: How CalPERS makes its Termination Liability computation is a big secret. As the Times breathlessly writes:
“One of the first things I think you should do is publish that number for every city,” said William F. Sharpe, an economist who won the Nobel in economic science in 1990 for his work on how the markets price financial instruments. He is also a California resident who voluntarily helped his city, Carmel-by-the-Sea, crack the secret pension code — figuring out the market value of its debt to its retirees in 2011 before Calpers resolved to start divulging the information later that year.
This is just bizarre. Current and former board members tell us that CalPERS would provide the information before 2011 if asked. And the article even more peculiarly fails to mention that Sharpe used to be a consultant to CalPERS. Has the fact that he’s not happy with his community being exposed to a large bill if it were to leave CalPERS led to him over-egging the pudding and the New York Times uncritically picking it up?3
A footnote to “Termination Liability Discount Rate” describing how it is constructed.4 At a high concept level, CalPERS is taking all the future plan liabilities, meaning payments that are estimated to be due to beneficiaries upon retirement, discounting them at the comparable Treasury bond rate (or if there is no comparable Treasury bond because the liability is longer than the maturity of any Treasuries, coming up with a kludge). This is an absolute worst case scenario because it assumes zero in the way of future contributions from employees and assumes the lowest-risk investment approach, one that is not used in the wild because it’s now widely seen as far too conservative.
The California Government Code (sections 02571-02572) sets forth the treatment of employers who terminate their plans. There’s also a lengthy discussion of CalPERS approach in board documents, such as this report for the September 2016 board meeting.
The analytic approach that CalPERS mimics the computations that corporations and municipalities use when discharging obligations via defeasance. Defeasance is often cheaper than buying back outstanding bonds to retire them. The entity can remove the obligation from its accounts by escrowing a portfolio of of low-risk securities, typically Treasuries, of a similar amount and maturity to the outstanding debt.
False claim 5: The row with CPCD shows CalPERS and other public pension funds are pulling a fast one on their members and taxpayers. From the article:
The two competing ways of valuing a pension fund are often called the actuarial approach (which is geared toward helping employers plan stable annual budgets, as opposed to measuring assets and liabilities), and the market approach, which reflects more hard-nosed math.
The market value of a pension reflects the full cost today of providing a steady, guaranteed income for life — and it’s large. Alarmingly large, in fact. This is one reason most states and cities don’t let the market numbers see the light of day.
If Sharpe and the New York Times want to object to CalPERS’ approach, they could contend that the giant pension system is using unduly risk-averse approaches to screw departing employers to the benefit remaining CalPERS beneficiaries. But that isn’t the case being made here.
The insinuations are that CalPERS (and other pensions) are trying to pull a fast one on all their members. The Times has no evidence for its explicit claim that public pension funds routinely calculate “market value numbers” but hide them.. Indeed, as we pointed out in “False claim 3” above, CalPERS does not make this computation for 2/3 of the members in its system. You can’t hide what you haven’t done.
And this section makes clear that it regards the “hard nosed math” of providing for the “full cost” today, when as we explained above, that’s clearly conservative by virtue of assuming no future contributions by workers who still are likely to continue to work and pay into the system, as well as assuming what any investment professional would regard as a very risk-averse approach.
Tellingly, even actuarial publications don’t agree with how the New York Times presents the “market based method”.5 For instance, consider a 2013 article from the American Association of Actuaries, Measuring Pension Obligations Discount Rates Serve Various Purposes. The Times seems to think it should get a pass for providing a link to the article in the online version and then ignoring what it says.
The approach CalPERS used was a subset of the market-based approaches called a solvency method. And you can see CalPERS’ implicit justification for using it with exiting employers: it assumes no more funds will be added to the pension plan:
The solvency value is the amount needed to fulfill all benefit obligations when invested in a portfolio of securities free of default risk whose cash flows match the future benefit payments.
An important characteristic of the solvency value is that it is intended to fulfill the benefit obligation without additional funds. This requires that the portfolio be free of default risk or else additional funds may be needed.
The method that CalPERS and public pension funds use is called the budget approach. It reflects the way pension funds operate, which is that they invest in higher-return instruments than Treasuries. That in turn creates over and underfunding risks. It’s all too easy to forget that CalPERS was overfunded in the 1990s; indeed, the decision to cut contributions based on the belief that dot-com valuations would be durable is part of the reason the system is underfunded now.
This cycles back to our ongoing beef with CalPERS: there’s nothing wrong with a budget-based approach provided you don’t overestimate future returns (or at least not by much). That is the risk CalPERS is running now by refusing to lower its return assumption in the light of central banks around the world continuing to push interest rates lower and lower.
False claim 6: CalPERS is deviating from sound practice in not using the Treasury discount rate that it used in its Termination Liability computation for its liabilities generally. See the Times play fast and loose with financial economics concepts:
With everybody either retired, or about to be (Mr. Houser will retire later this year), there is no guesswork in determining everybody’s pensions. The actuaries at Calpers project each of the future monthly payments due to Mr. Houser and the other five retirees, assuming they will live to age 90.
This is already false, as anyone who knows bupkis about actuarial methods or just plain common sense will tell you. Do you know how long you will live? No. Therefore even if you have a defined benefit pension, you have no idea what the total payout will be. You could be flattened by a bus tomorrow or live to be 115. Actuaries use the rule of large numbers. With only 6 lives the estimate is far less confident.6
An economist would say the right rate for Calpers is the one for a risk-free bond, like a Treasury bond, because public pensions in California are guaranteed by the state and therefore risk-free. And that’s what Calpers does when it calculates market values.
The risk-free rate is the Treasury bond rate. California is not risk free. It is rated AA- now, which is the best bond rating it has had in 14 years. In 2009, the state resorted to issuing interest-paying IOUs to work around a budget crisis, including meeting payroll, and some analysts forecast the state would default.
On top of that, Sharpe and the Times conveniently ignore that these obligations are not iron-clad. In California, the sections of the California Government Code regarding public pension promises are backed up by a commitment to them written into the Constitution. But that change was made in 1992 and the Constitution could well be amended in the future. In most other states ex Illinois, which also has very strong state protections for pensions, they are similarly vulnerable to the deals being retraded, for instance, the way private pension plans have often been cashed out into defined contribution plans. So the notion that there is a great deal of certainty, much the less Treasury level certainty, in the odds of actually getting a public pension payment, is considerably overstated. Hence using Treasuries as the basis for discounting their liabilities is wrong-headed.
Sharpe chooses to ignore pension realpolitik when he claims:
“Every economist who has looked at this has said, ‘It’s crazy to use what you expect to earn on assets to discount a guaranteed promise you have made. That’s nuts!’” Professor Sharpe said.
No, it isn’t, because these are not guaranteed promises in the overwhelming majority of cases. For most government bodies, the ugly truth is that any large underfunding will be defaulted in part on rather than honored.
In addition, and more subtly, as the article makes clear, the employers who subcontract to CalPERS can and do withdraw. Thus those pensions are the obligations of those particular governmental bodies, not the state of California, since CalPERS is acting as an agent of these entities. One way of thinking about it is that they’ve upped the caliber of their credit-worthiness considerably by opting into the CalPERS system. But natively? What is the discount rate that would apply to the CPCD? I doubt they can borrow for five years at anything like a retail rate (over 10%). And 10 to 30 years? Fuggedaboutit.
You can see that issue made explicit in the discussion at CalPERS’ current board meeting about some defaulting employers in its system. If CalPERS cannot collect the amounts owed, it will reduce the benefits to the workers in the plan.7 Again, these pension obligations are not “a guaranteed promise,” as Sharpe is trying to claim. The promise is only as good as the party that is making it.
So if you discount their liabilities by the discount rate appropriate to them, the net present value would be small because they are crappy credits and the value of their pension promises ex participating in the CalPERS system would not be worth much, according to this methodology. In other words, this approach (and the Times article) does not address where the actual legal liabilities sit, as in when the CPCD pulled the trigger and left CalPERS, it was reminded of a central fact it chose to overlook: the pension plan responsibility sat with them.
If you extend this logic to other public pension plans, you’ll grasp that this approach implies that the actuarial value of a plan is a function of the creditworthiness of a plan sponsor, in much the way the value of an annuity is a function of the credit rating of the insurer from which you buy it. California is about as good as it gets for public pension funds.
Yet the Times article is taking sides in a row that dates back to 2003, when the Bush Administration was trying to privatize Social Security. The article cites a paper that argues that public pension funds which are not even remotely US Treasury level credits should use Treasury-type discount rates to value their liabilities (curiously, the Times did not link to it but we’ve seen summaries elsewhere).8 That flies in the face of corporate finance logic. The only justification for using a risk free rate to discount a liability with some risk is an investment based method of the budget sort that this cohort says it rejects. In other words, to argue for Treasury-based discount rates says you are using an investment-return-based method, not a liability discount method based on the riskiness of the pension fund sponsor. You are just arguing over how conservative an investment return assumption to use.
I hope Naked Capitalism readers will write the New York Times’ Public Editor, Liz Spayd, at firstname.lastname@example.org, with a link to this post, and call for a retraction. This article was a piece of shoddy advocacy masquerading as journalism. But it should be no surprise to see the Times putting its finger on the scales by attacking defined benefit plans so as to make it easier for Wall Street to press for them, as in the case of the clueless CPCD, to be converted to much higher fee 401 (k)s.
1 CalPERS acts as administrator for many smaller government entities in the State of California at their choosing. They are not required to join the CalPERS system; some entities, like the Los Angeles City Employees Retirement System (LACERS) are independent of CalPERS.
2 Such as: “The estimate for 2016-17 also assumes that there are no future contract amendments and no liability gains or losses (such as larger than expected pay increases, more retirements than expected, etc.) This is a very important assumption because these gains and losses do occur and can have a significant effect on your contributions. Even for the largest plans or pools, such gains and losses can impact the employer’s contribution rate by one or two percent of payroll or even more in some less common circumstances. These gains and losses cannot be predicted in advance so the projected employer contributions are estimates.”
3 One of the reasons that Sharpe might have found it hard to figure out what was up with Carmel- by-the-Sea is that that small community has eight pension plans, as you can see here by inputting the city’s ID (7011478504) or “Carmel” then clicking on the city’s name here to see all the recent actuarial reports.
By all accounts, Sharpe is a very personable man. However, as we point out in ECONNED, intellectual integrity is not his long suit. Sharpe, whose Nobel Prize rests on being the father of the Capital Assets Pricing Model, admitted that his model was hopelessly flawed. Once you abandoned the assumption that investors could borrow unlimited amounts at the risk free rate, as he put it:
The consequences of including such aspects of reality are likely to be disastrous in terms of the usefulness of the resulting theory…. The capital markets line no longer exists…. Moreover, there is no single optimal combination of risky securities; the preferred combination depends upon the investors’ preferences…. The demise of the capital markets line is followed immediately by that of the security market line. The theory is in a shambles.
As we noted:
At that time (1970) Sharpe nevertheless argued for continuing to use CAPM, invoking [Milton] Friedman’s endorsement of unrealistic assumptions and the belief that even a flawed theory was better than no theory.
4 “The discount rate assumption used for termination valuations is a weighted average of the 10 and 30-year US Treasury yields in effect on the valuation date that equal the duration of the pension liabilities. For purposes of this hypothetical termination liability estimate, the discount rate used, is the yield on the 30- year US Treasury Separate Trading of Registered Interest and Principal of Securities (STRIPS). Note that as of June 30, 2014 the 30-year STRIPS rate is 3.55 percent.”
5 Note that the article implies that there is only one “market based” computation when it is a general technique that can be implemented in various ways: “Using a market-based method, a discount rate is selected by looking at observable data in the nancial markets at the measurement date. Market-based methods use xed-income yield data because xed-income securities are similar to the pension obligations – both make xed payments in future years. Market-based methods vary in the amount of default risk recognized. For example, nancial statement disclosures for private-sector employers use AA corporate bond rates, plan-termination measurements use insurance company premium quotes, and solvency measures (discussed further below) often use U.S. Treasury bond rates.”
6 In another sign of the dubious integrity of this article, the “they will live to be 90” claim looks to be made up. If you take a life expectancy calculator, put in the age of 65 (the CPCD members are reported to be retired, presumably just retired, or about to retire), you get a life expectancy of 17.6 years for men and 20.2 for women. If you charitably increase the age to 70, you get 14.1 years for men and 16.3 for women. Where does this 90 years come from?
8 “For all three agencies, failure to pay the termination liabilities required by the contracts will result in a reduction of retirement benefits under Government Code § 20577.”
8 The author again is visibly less than even-handed in how it handles this dispute. The article links to an August 9 letter, in which the long letter from the editor states that it disputes the claims made by the authors of a paper that had been in the review process as to why a joint Academy/Society of Actuaries Pension Finance Task Force had been dissolved. Here was the gist of the dispute. You’ve never infer this was the Academy’s version of events from reading the Times:
This was not a paper that was submitted to the Academy and Society for publication, as you or I might submit a paper to the North American Actuarial Journal or Contingencies. Rather, it was developed by a joint Academy/Society task force—including work by others on the task force, and by staff of both organizations. We cannot afford to encourage a precedent where individuals who are frustrated by the review/edit cycle can short-circuit it by simply taking a draft Academy document and publishing it elsewhere under their own names. We certainly recognize that members of the task force could open a new Word document and start writing, with the result being a paper that’s very similar to the one developed by the task force—and doing so would be entirely appropriate (and I would encourage them to do it). But we don’t believe it’s appropriate for anyone to take work done by an Academy task force or committee and publish it under his or her own name—especially not as a way to get out of the hard work of going through the review process.
The Times then cherry-picks the August 22 update: “The academy’s president..said the paper ‘could not meet the academy’s publication standards.’” This comes close to being a misrepresentation. If you read the entire piece, the Academy continued to review the draft, even though some, and maybe all of the authors had walked out of the review process. The president said (emphasis ours), “The PFTF draft could not meet the Academy’s publication standards.” The Times version gives readers the impression the Academy rejected a completed paper, when all that can be concluded is that it made a good faith effort to see if it could publish a draft even though the authors had abandoned the edit/review process and couldn’t find a way to do that.