Pension blogger Leo Kolivakis recently wrote about the importance of good pension governance. He rightfully points out that the governance model at US public funds is one, if not the main culprit of our pension woes. I agree with Leo Kolivakis on that.
I believe, however, that for as long as we are unable to measure the effects of poor plan governance, not much will change. Up until now, the talk about good plan governance has been a lot of prose without hard numbers attached to it. And I worry and fear – based on my own experiences – that these hard numbers will be difficult to come by. Not for lack of an algorithm to determine such “Returns on Governance”, but for lack of detailed input data.
“PenPIRR goes beyond the investment sphere. PenPIRR translates the funding status effects of typical plan sponsor actions into return space. All factors that influence overall pension plan success (e.g., investment returns, benefit and funding decisions) become comparable to each other. PenPIRR measures overall pension plan success and is thus the only return measure that can meaningfully be compared to one’s expected rate of return.”
I did not mention, however, my PenPIRR-based performance attribution system. In short, PenPIRR is the top measure of ultimate pension plan success. It is a combination of investment returns (measured in money-weighted return space) and a return component called the “Returns on Governance”. It is my hypothesis that the “Returns on Governance” have been negative for most pension plans, especially for public plans.
To calculate PenPIRR and the Returns on Governance, however, I need detailed liability cash flow projections for each of the years within the performance evaluation period, typically all Projected Benefit Obligations (PBO) data. If we want to calculate a 25-years PenPIRR and the “Returns on Governance”, I would need 26 such PBO structures. I need to quantify the changes in funding status that are not caused by the investment sphere, but by mere plan sponsor actions. For instance, by a decision to grant higher benefits. Or by a decision to not fully fund new benefit accruals.
Let me describe my recent journey to quantify the “Returns on Governance” in an intended pilot study for the three statewide public pension plans in my home state Minnesota. PERA and TRA responded that the data requested does not exist. TRA furthermore claims that “the actuaries never consistently prepared PBO schedules for valuations.”MSRS simply responded that “the data is unavailable” and that its actuary said that “there are no easy (or inexpensive) ways to recreate projections for prior years.”
I find this perplexing as the PBO data are the basis for all subsequent pension calculations. The present value of liabilities and its actuarially distorted relatives such as the “Accrued Actuarial Liabilities” (AAL) and “Unfunded AAL” (UAAL) are all based on PBO data. I even believe that the asset management process is severely impaired without knowledge of the liability structures: What are the upcoming liquidity needs? What is the duration of pension liabilities? Additionally, the lack of past PBO data will now make it all but impossible for the pension plans to calculate a number that they can actually compare to their expected rate of return.
One “huge” success in public pension reform here in Minnesota includes a stipulation that going forward, these PBO structures will have to be published by the pension plans. Yet PenPIRR and the “Returns on Governance” are most useful when viewed over a longer investment horizon. Here is my hypothetical long-term ranking of return measures:
Standard Money-Weighted Return > Time-weighted return > Expected Return > PenPIRR (still positive) >> Returns on Governance (negative by a large degree)
Well, of course, I cannot prove this hypothesis for as long as pension plan executives refuse to publish their PBO data. But below is the rationale behind this ranking.
The National Association of State Retirement Administrators (NASRA) usually goes back 25 years or more to report impressive time-weighted returns that exceed the expected returns as indication or “proof” that taxpayer’s dollars have been saved over this time period. The particular return sequence of good asset returns early on followed by below average returns later on, however, lifts the corresponding money-weighted return above the reported time-weighted return. Pension plans experienced a positive “Path Dependency Effect” since financial markets were benevolent to plan sponsors. Given that return sequence, pension plans should be well overfunded by now – if it weren’t for plan sponsor actions such as contribution holidays or additional benefit increases. While the effects of these actions show up in a plan’s funding status, they are not captured by time- or standard money-weighted returns. But because of our reliance on expected rates of return when deriving funding cost expectations, we need a return measure that is comparable to an expected return. This return measure is PenPIRR.
Aforementioned plan sponsor actions typically lead to negative “Returns on Governance”. Since
PenPIRR = Investment Returns + Returns on Governance,
it may very well be that PenPIRR, the relevantplan sponsor return, is well below the expected rate of return, even if the investment manager succeeded in earning a higher time-weighted return.
I can understand the motivation by pension plan executives to not publish their PBO data – do they really want to know the extent to which their actions negatively impacted investment returns? But giving them the benefit of a doubt that they really don’t know, I am interested if readers could let me know their experiences with regard to PBO or ABO data (either publicly in the comment section below or by writing to Ehrentreich@ldi-research.com). Are you a corporate or a public plan sponsor? Do you know your PBO and/or ABO numbers? Are you an actuary? Do you provide these numbers to your clients? Do they request them? How long do you have to retain those numbers?
It is my hypothesis that most corporate plans would know their PBO data by now, but my experience with the Minnesota case makes me quite pessimistic for public plans. I believe that public pension plan administrators and their investment managers are flying blind – they do not have the necessary data to effectively manage their plans. And without knowing what was going on the past, they will continue to make the same mistakes that brought the funding crisis upon us.
The financial crisis did not cause the funding crisis of pension plans. It just exposed their structural flaws. If we won’t address them by the time the next crisis – or a simple recession – comes around, I believe that we will see many more DB to DC conversions in the public sphere. A couple years ago I wrote in my blog ("Quo Vadis, Public Pensions?") that if
“we do not act soon with meaningful pension reform, I believe that the few states that are currently considering a DB to DC switch are just the beginning of a tsunami wave that will eventually lead to an almost complete abandonment of DB plans in the public sector – similar to the one we are currently witnessing in the corporate sector. It will play out over several decades, but economic reality is going to catch up eventually with public plan sponsors.”
The pension reforms that we have seen so far are steps in the right direction – but those steps usually pale in comparison to the magnitude of the funding problem. It is thus not surprising that Pension & Investments is now going to host a “Public Funds Defined Contribution Summit.” We are well on our way to the scenario that I depicted two years ago.
P.S. After receiving a few remarks and questions, I thought I should clarify what the Returns on Governance concept covers and what not.
One reader asked about administrative costs such as investment fees, due diligence, legal fees, advice and other services. Paying excessive administrative costs – either in relation to the invested funds or in comparison to a peer group – might point to some serious governance problems. However, those are not covered in the concept I introduced.
Another reader opined that “governance is … substantially based and required in law and by regulation: non- compliance is a real potential risk with a significant cost. There are real examples of the cost of bad governance and it can be explained.” He thinks that calculating an ROI on that kind of governance is a “waste of time.” That kind of governance typically deals with board composition, independence, due diligence, political influences, conflicts of interests, etc. etc. For as long as those governance aspects do not affect the benefit and contribution policies, they won’t show in my Return on Governance concept either.
My Return on Governance concept is thus quite narrow – but it is quantifiable if the PBO data were available. It only answers the question whether the benefit structure is adequate or too generous for the amount of funding the pension fund receives from the plan sponsor. Or, framed differently, it analyzes whether plan sponsors sufficiently supported their benefit policies through adequate contributions.
“In Baltimore, Md.; San Jose, Calif.; and a number of other municipalities, a similarly idiotic "COLA" surrogate or supplement was installed, giving retirees a pension increase or "13th paycheck" in years that markets outperformed the average. That policy guarantees that the pension fund will never achieve its expected return. The good years are skimmed off while the taxpayers have to cough up for market losses. Even worse, these schemes siphon off more from the plan when markets become more volatile as they have since 1997.”
I find this comment remarkable because it hints at something that I have been working on over the last few years and that I have injected into the pension reform debate before. Miller suggests that it isn’t investment returns alone that determine whether a pension plan achieves its expected rate of return. Obviously, plan sponsor actions such as “idiotic” COLA increases or a “13th paycheck” when markets are doing well will affect the relevant plan sponsor return.
It would have been good if Miller would have elaborated on what that relevant plan sponsor return really is. I believe that Miller is onto something interesting here, yet I also believe that up until now, no return metric exists that can be meaningfully compared to a pension plan’s expected rate of return. However, if no such measure exists, there is nothing that prevents us from creating one.
Introducing the Pension Plan Internal Rate of Return (PenPIRR)
A better measure of total pension success is a money-weighted type of return that also considers changes in funding status due to non-investment related influences – typically plan sponsor decisions to change the benefit structure or to not fully fund new benefit accruals.
The only tricky thing is to properly separate the changes in funding status due to the peculiar nature of investment returns (which average return and which particular return sequence) from those that are solely attributable to plan sponsor decisions and are thus completely outside the investment manager’s realm.
In order to avoid a terminological confusion with traditional money-weighted returns, I suggested to GASB calling this new return measure the Pension Plan Internal Rate of Return (PenPIRR). It uses a money-weighted return methodology, but considers one additional cash flow that traditional money-weighted returns ignore. A funding deficit that is caused by an increase in pension benefits, for instance, would require one additional hypothetical cash flow from the plan sponsor into the fund.
PenPIRR is a return measure that goes beyond the investment sphere. PenPIRR translates the funding status effects of typical plan sponsor actions into return space. All factors that influence overall pension plan success (e.g., investment returns, benefit and funding decisions) become comparable to each other. PenPIRR measures overall pension plan success and is thus the only return measure that can meaningfully be compared to one’s expected rate of return.
Different Scenarios – Same Money-Weighted Returns, Different PenPIRRs
Let us briefly investigate the shortcomings of standard money-weighted returns as a measure of total plan sponsor success in more detail. Since they do not take funding status information into account, they can be the result of very different funding cost scenarios. Assume that we have two pension plans A and B with initially identical liability structures and identical cash flow patterns over the entire reporting period.
Scenario I – Fully Funded at the Beginning and at the End of the Investment Period
When both pension plans start and end the performance evaluation period fully funded, no further modification to the cash flow structure is necessary. Traditional money-weighted returns and the Pension Plan Internal Rate of Return coincide.
Scenario II – Benefit Increase at the End of the Investment Horizon
For scenario II let us assume that everything unfolds identically to scenario one – except for a benefit increase for plan B at the end of the investment period. All cash flows during the performance evaluation period remained identical; therefore, the standard money-weighted return measure would not be affected by the final increase in the present value of liabilities. However, everybody would agree that the financial health of the now underfunded pension plan B is worse than that of a pension plan A.
PenPIRR accounts for that change in financial condition by immediately recognizing in full the funding status effect of the benefit increase. Traditional money-weighted returns would gradually recognize this benefit increase as it will affect only future cash flows.
Scenario III – Insufficient Contributions to Fund New Benefit Accruals
Net cash flows are generally the result of contribution and benefit payments. However, the standard money-weighted return methodology does not relate the actual contributions to the required contributions (normal cost + shortfall amortization). Assume that contributions are sufficient for plan A, but insufficient for B to fully fund their normal cost. Plan B will accumulate a funding deficit over time.
All these scenarios would result in the same standard money-weighted return measure as they have not (yet) affected cash flows. Each of these scenarios, however, depicts a different pension success story. The new return measure PenPIRR is able to differentiate between these different scenarios and ranks them accordingly.
I commend Girard Miller for pointing out that plan sponsor actions do affect the relevant plan sponsor return. Investment returns alone aren’t responsible for total plan sponsor success. Once we can agree on this, then we can better measure our past performance – the prerequisite of learning the right lessons from our past.
On May 15th, 18 minutes before a deadline, Minnesota’s governor Tim Pawlenty signed an omnibus retirement bill into law that he had opposed earlier as it lacked “significant reform, continues to provide for future increases, mandates additional costs to local government and makes the structural budget worse by tens of millions of dollars.”
What governor Pawlenty may have been missing in this bill was an amendment that would have placed new public employees hired after July 1st into a defined contribution plan. This amendment was defeated by the Senate before it moved on to the House. However, the final version now contains a section that orders a study of deferred contribution and defined contribution plans, to be finished by June 1st, 2011. The study “must include an analysis of the feasibility, sustainability, financial impacts, and other design considerations” of these retirement plans. It is entirely possible that the next time a pension reform bill is considered in Minnesota, it will contain the stipulation to move towards a defined contribution system.
What seemed unthinkable a few years ago – public pension plans rethinking their commitments to defined benefit plans – is happening in more and more states across the country. So I decided to take a closer look across the country and find out what the status of the public DB versus DC debate currently is.
States that were or are considering a DC option
The first state to actually implement a DC alongside a DB option was Nebraska in the mid-sixties. Anna Sullivan, Director of the Nebraska Public Employees Retirement System until 2006, summed up Nebraska’s experience by stating, “Our experience with the defined contribution plans has been mixed. We have had over 35 years to ‘test’ this experiment and find generally that our defined contribution plan members retire with lower benefits than their defined benefit plan counterparts.” This statement is the result of a series of studies conducted by Buck Consultants in 1994, 1998, and 2000.The 2000 study showed that the DC returns of state and county workers were substantially below – up to 4 percentage points – of those who stayed in the traditional DB pension plan. Consequently, Nebraska lawmakers passed a law in 2002 that required new employees to participate in a cash balance plan while existing employees had a choice to either stay in the DC plan or participate in the CB plan.
In September 2009, the National Conference of State Legislatures published an overview report of state DC retirement plans. According to this report, just two states (Alaska and Michigan) and the District of Columbia had defined contribution plans as primary retirement systems for new employees.
In May 2009, the Center for Retirement Research at BostonCollege published a set of factsheets on states with DC pension plans:
Alaska: Mandatory DC plan (since 2006): For all general state government employees and teachers hired after July 1, 2006. Since 2008, a switch back to a DB system is being considered
Indiana: Mandatory Combined Plan (since 1997): All general employees and teachers hired after 1955 are covered by a defined benefit plan and are required to contribute to an annuity savings account (ASA).
Michigan: Mandatory DC Plan (since 1997): For all general state government employees hired after March 31, 1997.
Oregon: Mandatory DB and DC Plan (since 2003):All state employees hired after August 29, 2003 are in both a defined benefit plan and a defined contribution Individual Account Program (IAP). Active members hired prior to that date kept their previous defined benefit plan, but paid their future contributions into an IAP account.
West Virginia: Type of Plan: A) Mandatory Defined Contribution Plan; B) Mandatory Defined Benefit Plan Year Effective: A) 1991; B) 2005; Workers Covered: A) All teachers hired after June 30, 1991 and before July 1, 2005 excluding those who switched back to the defined benefit plan. B) All teachers hired before July 1, 1991 and after June 30, 2005.
A 2007 fundfire article (“Public DB vs DC Debate Flares Up Again”) mentioned legislative attempts by Georgia, Pennsylvania, and Rhode Island to shift the states’ retirement systems from defined benefit to defined contribution, in hopes of fixing hefty funding shortfalls. Yet Keith Brainard, research director at the National Association of State Retirement Administrators (NASRA), said in the same article that “[v]irtually every state and a large percentage of cities and counties in this country that have evaluated changing from a defined benefit to a defined contribution system have decided they would be better off sticking with their original pension arrangement. That ought to say something.” Given NASRA’s reluctance to even admit a funding problem and a deep structural reform deficit on the public side – see my previous blogs – I am afraid this time may be different.
That this concern is not too far-fetched shows the web page “Assessment of Threats to Public Defined Benefit Pensions” by the Communications Workers of America (CWA) union. It lists, state by state, past and pending pension regulations that may threaten the existence of the public DB system as we know it. The Utah Senate recently voted to close the state’s DB pension plans to new entrants. Senator Bill Brady, the Republican nominee for governor in Illinois, wants a shift to a DC system. Similarly, for California’s parting governor Arnold Schwarzenegger, all options are on the table and he made it clear that he wants to tackle pension reform before he leaves office. I am sure that there are other states in which at least some legislators are warming to the DC option.
The widely publicized funding shortfalls and budget crises in many states contribute to increased public pressure to abandon public DB plans, and predictions that state and local governments will shift to DC retirement plans are easy to find on the web, for instance here. So the threat is real and we should not discard it as easily as Keith Brainard did a few years ago.
Are we learning the wrong lessons?
I am convinced that maintaining DB retirement plans in the public and in the corporate sector is something worth fighting for. While I am not opposed to offering a DC option alongside a traditional DB retirement plan, I fear that when plan sponsors look at current funding deficits and the projected costs to close these funding gaps, they will draw the wrong conclusions and rather replace existing DB plans. But DB plans are inherently more efficient than DC plans. They offer a longevity insurance that DC plans do not offer (or only at higher costs). When employees are switched from a DB to a DC plan, they should increase their saving to make up for the loss of longevity insurance – depending on where in their life cycle they are, it could be an increase of up to 40%. Various studies show a reluctance of people to opt for annuities and rather enter their retirement with a lump sum distribution. If they would rationally increase their personal saving rates to make up for the loss of longevity insurance, we would end up with inefficient over saving in the economy. Given where our personal saving rates are, I believe that this is not the case and people will enter their retirement period with insufficient assets.
But a determination to save DB pension plans needs to be supported by realistic projections what level of benefits is sustainable. For far too long we have favored Casino style investment strategies over prudent investment strategies. I have pointed out in my previous blogs that investment strategies with heavy return seeking components such as equities and hedge funds are regularly more expensive than expected. And because we look at the wrong success metric – time-weighted average returns – we do not even realize on an ex post basis that they have cost us more than lower returning, but less volatile LDI strategies. The moment we implement an LDI strategy we know what it costs to fund future benefit payments. If that seems too expensive, we need to adjust those benefits down.
Moving to an LDI strategy in the public sector would mean to abandon the expected rate of return as a discounting rate. In a previous blog I have shown that this is a bad choice for a discount rate as it neglects intra-period volatility and systematically underestimates funding costs. Using realistic market rates of return could potentially more than double the present value of liabilities. If lawmakers consider that liability as absolutely impossible to fund, then we have already crossed the threshold and it is only a question of time before they will ask for a pension bail out.
If we do not act soon with meaningful pension reform, I believe that the few states that are currently considering a DB to DC switch are just the beginning of a tsunami wave that will eventually lead to an almost complete abandonment of DB plans in the public sector – similar to the one we are currently witnessing in the corporate sector. It will play out over several decades, but economic reality is going to catch up eventually with public plan sponsors.
I really dislike geometric average returns as a proxy for long-term funding costs of DB pension plans. However, the recent report by the National Association of State Retirement Administrators (NASRA) that the long term investment returns of public plans beat their return assumptions shows that among public pension officials, the connection between long term average returns and funding costs is undisputed.
There has been a recent wave of critical studies that all question the funding levels reported by public plans by pointing out that lower discount rates should be used.
Not only is there a pattern emerging from these independent studies, there is also one emerging from the responses from CalPERS, CalSTRS, or the National Council on Teacher Retirement (NCTR). They refer to the studies as “funny” or “fuzzy” math, use qualifier such as “so-called study”, call them “fundamentally flawed” or “grossly misleading” and that they should not be trusted. The main argument for their unanimous rejection is the NASRA claim that long term average returns beat long term return expectations and thus billions of tax payer dollars were saved. In an emailed statement to Global Pensions, Ricardo Duran from CalSTRS claimed that “long-term investment returns were the most appropriate measure of the fiscal health of the pension system and the soundness of contribution levels.”
Yet I continue to object to geometric average returns as a valid proxy for funding costs. As I have claimed before, they are misleading as they can differ greatly from the relevant dollar weighted rates of returns. Now I wanted to know how misleading they can be and here are some new results “straight from the lab”. Be aware, NASRA, NCTR, CalPERS and CalSTRS would surely label this as funny math.
I have argued before that dollar weighted rates of return are path dependent on the specific return sequence with which a certain geometric average returns was achieved. I argued that the discrepancy between the geometric average return rate and the dollar weighted return rate can be substantial, especially when asset returns are highly volatile and plan durations are low. Now I wanted to test this hypothesis and see how big the differences can be.
In a first step, I simulated 10,000 return sequences that all share the common property of having a geometric average rate of return of 8%, the most common return assumption among public pension plans, over a ten year investment horizon. Return volatility is calibrated to resemble that of a typical 70/30 investment portfolio. With that setup, I assume that there is absolutely no uncertainty about the realized 10 year geometric average return rate. All funding cost risk stems solely from intra-period return volatility, which NASRA and other public pension officials completely ignores, both in their ex-post analyses and their ex-ante risk assessment.
I then take the liability structures of six actual DB pension plans with durations ranging between 6.8 and 21 (this plan used PBO numbers instead of ABO numbers).
Figure 1: Liability structures of six different defined benefit pension plans. All monthly benefit payments are normalized such that the present value of liabilities equals $100 million using an 8% discount rate. Numbers in parentheses are plan durations.
I then conducted a Monte Carlo simulation with the following simulation structure:
For all plans and all return paths
Initialize plans with xx% of the PV(Liabilities)
For i = 1 to 120 months
Pay out benefit
Determine dollar weighted rate of returns and analyze terminal funding levels
With this structure it is implicitly assumed that these pension plans are hard frozen and that occurring funding deficits are not closed by additional funding contributions.
Now let’s look at the histograms and descriptive stats of dollar weighted rates of return.
Figure 2: Histograms of dollar weighted rates of return for six DB pension plans, given a geometric average rate of return of 8% over a ten year investment horizon (10,000 simulation runs).
Table 1: Descriptive statistics for the dollar weighted rates of returns given a geometric average rate of return of 8% over a ten year investment horizon (10,000 simulation runs).
Now let’s see how these dollar weighted rates of return translate into final funding levels.
Figure 3: Histograms of final funding levels for six DB pension plans, given a geometric average rate of return of 8% over a ten year investment horizon (10,000 simulation runs). 1 represents a final funding level after 10 years of 100%.
Keep in mind when looking at figure 3, that we started fully funded given an 8% discount rate and every single return sequence with which these final funding levels were achieved represents a geometric average rate of return of 8%. We see that low duration plans exhibit a larger variation in dollar weighted rates of return and final funding levels. This is the result of reverse dollar cost averaging. We can also see from these graphs that return volatility seems to result in more risk than reward opportunities as the dollar weighted return distribution is heavily skewed to the left. It will be interesting to find out why.
I may be impressed with the public plan’s claim of having earned 9.25% over the last 25 years given the difficult market environment, but I am not convinced that this resulted in lower funding costs to the tax payer.
The critics of public pension accounting reject expected geometric average returns as an appropriate discounting rate for public pensions since they do not match the riskiness of the underlying pension promises as standard financial theory requires. I fully agree with them. Here I presented an additional argument why expected asset returns are a bad choice for a discount rate: They are not an appropriate measure of long‐term funding costs.
My simple analysis is only meant to illustrate the discrepancy between geometric and dollar weighted rates of return. It is not meant to be an accurate quantification. Yet I believe that by incorporating more and more realistic assumptions, the distribution of dollar weighted rates of return would be shifted to smaller values. Typical pension behavior – e.g., contributions holidays, the inability to close funding gaps as soon as possible, average funding levels below 100%, asset and liability smoothing, etc. etc. – ensures that the relevant dollar weighted rates of return are substantially lower than the reported geometric average returns. Until public pension officials start publishing the correct return metrics, I will remain convinced that their investment strategies and funding policies lead to higher funding costs to the tax payer.
On March 15th, the National Association of State Retirement Administrators (NASRA) published a report that showed that the median investment returns for the 20‐ and 25‐year periods ended 12/31/09 exceed the most‐used investment return assumption of 8.0 percent. For example, for the 25‐year period ended 12/31/09, the median investment return was 9.25 percent.
NASRA immediately used these findings to rebut the recent criticism about the long-term return assumptions and the risky investment strategies of public pension plans. Pointing to the long-term focus of governmental plans, Keith Brainard, NASRA’s director of research,exclaimed in the accompanying press release that a “short term look at investment returns can be misleading, as returns can fluctuate dramatically over brief periods of time. For example, public plans have had a one-year investment return of nearly 20% and a 3-year return below negative one percent.”
However, the NASRA brief and Brainard’s comments evoke the opposite of the intended reaction in me. I now have even less faith in the long term viability of the public pension system. It is disheartening to see that pension officials and their advisors (a.k.a. research director) seem unable to ask the right questions. Here is a small sample of questions they should ask themselves:
If we earned, on average, 1.25% over our long term return assumption of 8% over a 25 year horizon, how is possible that our public plans are severely underfunded. Shouldn’t our plans be healthily overfunded by now if actual investment returns exceeded our assumptions?
Is it possible that our better than expected long term returns did not automatically translate into lower funding costs?
What’s the likelihood of higher average returns leading to higher funding costs?
Are the expected long-term return assumptions really the best predictor of long-term funding costs?
It is common that the moment an average return exceeds its expectation or that of a competing investment strategy, plan sponsors and their advisors pat themselves on the back in their belief of having chosen the right investment strategy. Unfortunately, looking at geometric average rates of return can even be more deceiving than focusing on short term investment results. The mental short cut of assuming that higher average returns automatically mean lower funding costs is too often only wishful thinking on the plan sponsor’s part.
Turning higher average returns into lower funding costs is not an automatism. For that, certain requirements need to be fulfilled, and most public and corporate plan sponsors in the U. S. (and in most other countries) consistently violate them.
Almost more important than geometric average returns are the asset balances on which those returns are earned. Instead of focusing on geometric average returns, plan sponsors need to look at dollar weighted average returns. We need to do more serious math than just calculating geometric average returns to see the funding cost implications of various investment strategies. A simple back-of-the-envelope approximation of dollar weighted rates of return is the funding status weighted rates of return. By this estimate, I believe that most public pension plans racked up higher long term funding costs than liability driven investing strategies which are designed to keep a plan close to fully funded.
If public plan sponsors turn to NASRA for guidance in these turbulent times and receive a “stay the course” message, then NASRA is doing a disservice to its constituents. I believe that the public pension system is only salvageable if we start looking at the correct success metrics. If plan sponsors, their advisers and investment staff do not even realize on an ex-post basis that their high returning, but risky investment strategies lead to higher funding costs, how will they ever change to more reasonable investment strategies?
If plan sponsors believe that LDI strategies are too expensive for them to fund their future pension promises, then they need to seriously rethink their benefit formulas. With any decent LDI strategy that deserves its name one pretty much knows at the time of implementation what it costs to pre-fund future liabilities. If plan sponsors consider that as too expensive, then they need to acknowledge that they are overpromising. In the long-run, their equity strategies will, most likely, be more expensive than the LDI strategies they currently shun. The lure of equity heavy investment strategies is that one can cloak the true funding costs behind unrealistic investment assumption far too long. And one can cling onto the belief that above average returns will help to close deep funding gaps.
As long as pension officials keep focusing on the wrong return metrics, as long as they don’t understand the impact of reverse dollar cost averaging on their long term funding costs, and as long as they want to address funding deficits with risky investment strategies and not with contributions, until then I will remain pessimistic about the future of our public and corporate pensions.
Granting funding relief for DB pension plans seems to have become a new tradition up on Capitol Hill. In December 2008, then President Bush signed the Worker, Retiree and Employer Recovery Act (WRERA) into law. Given the magnitude of the financial crisis, that granted relief was nowhere sufficient for cash strapped plan sponsors.
Consequently, we are now treated to another round of funding relief bills, of testimonial hearings, and open letters from employee and employer organizations. The Preserve Jobs and Benefits Act by Earl Pomeroy (D-ND) and Pat Tiberi (R-OH) would provide funding relief of up to 15 years for single-employer plans and up to 30 years for multi-employer plans. To take advantage of these longer amortization periods, employers would have to continue offering retirement benefits.
Nowadays, it is surprising to see any piece of legislation garner universal support from both the Republican and Democratic Party as well from employee and employer organizations alike. And I support it, too – although half heartedly and largely because of the proposed benefit guarantee, which is exactly the part that remains contentious. The benefit guarantee section of the bill would require plan sponsors to keep their plans open for their employees. Delaying funding contributions, however, does nothing to guarantee that the already accrued pension promises will be honored. On the contrary, absent the PBGC, it would increase the probability of breaking that pension promise.
While I agree that plan sponsors desperately need funding relief, I think it is also worth looking at the efficacy and associated costs of granting more and more funding relief.
One often-cited reason for funding relief is that required pension contributions would drain much needed money from companies at a time when they and the economy need it most. Companies could use that money to preserve existing jobs, to create new jobs, or to invest in new capital. It is the cyclicality of required funding contributions that would, according to this reasoning, exacerbate an ongoing economic downturn and would dampen an economic recovery.
But the “Preserve Jobs” part of the bill title could turn out to be a misnomer. Jeremy Gold and Daniel Cassidy pointed out in a 2008 P&I article that what is true for an individual company does not hold for the economy as a whole. They rightfully argue that the jobs argument ignores what happens to those cash contributions: “The argument implies that the money goes down a rat hole, never to be seen again. In fact, cash contributed to pension plans is immediately reinvested in capital market securities, so there is no reduction in total supply of capital in the economy. The money contributed by one company will find its way, through the capital markets, to those companies with the best opportunities for productive investment and hiring. The very purpose of the capital markets is to ensure that such opportunities are funded.” Of course, that argument only holds true to the extent that the contributions are invested in local capital markets. For foreign investments, we’d still have spill over effects, though.
Another concern is that spreading funding contributions over such long term horizons with no requirement to substantially reduce risk exposure is a recipe for disaster. When the next recession, stock market crash, or financial crisis hits, pension plans will still be underfunded. Without taking risk off the table, plan sponsors will then find themselves in a similar, if not worse, funding situation. A stipulation in the relief bill that funding contributions for domestic liabilities are to be invested in local corporate bonds would reduce funding status risk. It would also ensure that funding contributions are invested productively in the domestic economy.
I also feel that discussing funding relief without addressing the underlying reasons for our current funding crisis is woefully inadequate. Yet I see no serious attempt by lawmakers to address the contradicting pension regulations that have plagued our pension system since the mid 1980s when Congress passed the reversion tax on excess pension assets. In combination with the non-deductibility of excess contributions, plan sponsors suddenly had no incentive to overfund their plans. Yet substantial overfunding in good economic times is absolutely necessary if plan sponsors want to pursue risky investment strategies with higher expected returns. Unfortunately, no one noticed when Congress killed the viability of equity strategies for DB pension plans and changed their asset allocations accordingly. Instead, plan sponsors adopted minimum funding policies and took contribution holidays whenever possible. This directly increased long- term funding costs. Ever since plan sponsors have faced a typical time inconsistency problem – short term tax optimization contradicted their long term interests of sufficiently funding their plans to weather bad times.
Establishing disincentives for plan sponsor to adequately fund their plans while punishing them for being underfunded is not the only example of sending mixed signals to plan sponsors. Funding relief itself is another mixed signal. A recent P&I editorial puts it this way: “First, Congress will consider President Barack Obama’s fiscal 2011 budget proposal, unveiled Feb. 1, that would see corporate income tax revenue rise 89% to $297 billion in fiscal 2011, and includes other measures that would raise corporate tax revenue, including a curtailment of tax benefits of multinational companies regarding offshore earnings. It also contains small tax relief proposals for small and midsize companies. Second, Congress is considering a bill to provide funding relief to companies — mainly large ones — that sponsor defined benefit pension plans. In other words, on the one hand, the administration believes corporations, in aggregate, are so rich they can afford to pay more to the government, even in this weak economy. On the other hand, Reps. Earl Pomeroy, D-N.D., and Patrick Tiberi, R-Ohio, believe corporations are so poor they can’t meet their pension funding obligations, and so have sponsored legislation-to provide them relief from funding the obligations.”
And then there is the cost argument against funding relief: by delaying funding contributions, pension plans will remain underfunded for long periods of time. This will directly increase their long-term funding costs as asset returns will only be earned on physically existing assets. In my opinion, the impact of average funding levels on long-term funding costs are too often neglected or underestimated.
After contemplating all the cons that I just discussed, I still find myself supporting the Preserve Benefits and Jobs Act. Funding relief is unlikely to have an effect on the unemployment rate. Funding relief will increase long term funding costs. Funding relief increases the risk of not honoring already accrued benefits. But if it is funding relief that is needed to keep plan sponsors from freezing their plan, then I have to support it.