Blogging Employee Benefits

June 1, 2006

Still of the Opinion: Retire the Number $450 Billion

Filed under: Legislation, Pensions — Fuguerre @ 7:31 pm



No, I still don’t believe that $450 billion number. And the more I see, the more elusive I find that fiction. Ultimately, absolute truth on it might give us a whole new meaning to the term “actuarial error.” But from what I can tell, $450 billion may be several multiples above even the high end of any reasonable range for the true level of current underfunding among PGBC-covered defined benefit pension plans.

Two weeks ago, coming up on Memorial Day empty-handed of anything but ephemeral rumors of progress on compromise pension reform legislation and impatient with waiting for the appellate court’s crucial ruling on the IBM cash balance plan age discrimination controversy, I found opportunity for a brief rant against that $450 billion number, which is almost universally accepted on faith as the premise for caricature of a pension system in crisis, on the verge of collapse. Comment posted in response to my rant takes issue with my doubts: we won’t question the lack of substantiation for the number itself as long as it comes from an official Administration source, but we need to document proof if we take issue with it. Actually, much of my own frustration comes from the sheer void of hard evidence: the Administration demands from pension plans what they characterize as “transparency,” yet plays a poker-faced bluff when it comes to their own hand. But OK, fair enough: my earlier rant didn’t merely seek reason to believe, rather went on to express very strong disbelief, so it’s not altogether inappropriate to expect me to proffer somewhat more solid a basis for my opinion than the Administration cares to give for its number.

Except I should first be quick to point out that I myself am no numbers person. I’m prepared to leave the Schedule B to the actuary (at least, more so than the lack of confidence expressed in actuaries by the departing PBGC Executive Director), prepared to leave financial reporting to the accountants (although I share with many others some serious misgivings about where FASB is taking us), prepared to leave the markets to investment advisers (that is, as long as my own 401(k) account gives me the retirement income that would have been promised by a traditional pension plan). Even so, no benefits practitioner can avoid constantly making concrete decisions based on dollars and percentages and formulas – ours is an industry defined as much by the numbers as by the words. So in that context, and constantly reminded by this weblog’s disclaimer, here’s where I come from on the issue of the $450 billion myth.

At Which Pension Plans Are We Pointing? When we march the $450 billion around the block, are we referring solely to single-employer defined benefit pension plans? Or to an aggregate of single-employer and multiemployer pension plans? Have I missed where the Administrative might have already made that most basic starting point clear?

Let’s look to a recent London roundtable discussion on U.S. pension reform, where Assistant Treasury Secretary Mark Warshawsky cited our target number, stating, “In 2000, underfunded DB pension plans reported total underfunding of $7 billion. As of 2005, that figure had risen to $450 billion, a more than 60-fold increase.” Does that give us any clue to whether or not the $450 billion might include multiemployer plans? Perhaps. That $7 billion number appears to represent an underfunding number for single employer pension plans reported for the year 2000 on ERISA 4010 filings, as reported on Table S-47 of the Pension Insurance Data Book 2004 published by the PBGC.

Which would suggest that if we do ultimately accept the $450 billion number as underfunding in single employer pension plans, then the situation is even more dire than is already being portrayed, since Table M-10 of the same report suggests multiemployer pension plan underfunding for the year 2000 in the that would add over $21 billion to the $7 billion starting point cited by Warshawsky. That is, believe the $450 billion for single employer plans, and unless multiemployer pension plans have done significantly better than single employer plans over the past 5 years, the total hole for plans covered by the PBGC could easily be deeper than $1 trillion. But I’ll scream if I ever catch any press quoting me as forwarding so outrageous an estimate: already I feel the press has far too easily simplified discussion of the $450 billion to the point of severely misleading meaninglessness.

But the point here being that if a $7 billion figure for 2000 defines our universe, then we’re looking at qualified U.S. single employer pension plans. And to reiterate a recurring theme in my rant: all of this would be more easily addressed if the Administration itself better documented and communicated its own numbers and analysis.

Single Employer Pension Assets – About $1.5 Trillion? In the same roundtable discussion, Warshawsky also stated, “At present private DB [pension plan] assets are a bit less than $1.8 trillion . . . .” Except that shortly thereafter Warshawsky clarifies that, “Single-employer private DB pension assets are around 85 percent of the DB total . . . .” In other words, although his earlier discussion appears to focus solely on single employer plans, we apparently now switch to waving our hands more ambiguously to the PBGC’s entire universe of plans. OK, but then my simple math would suggest we’re looking at current single-employer defined benefit pension plan assets in the neighborhood of $1.53 trillion. I have some serious doubts even about that number, but let’s accept it for now.

Overall Pension System Funded Status Versus Underfunding in Underfunded Plans. We now come to the next serious fork in the road to any sort of truth, and a very serious fork it is: Are we looking at the pension system as a whole, aggregating both overfunded and underfunded plans together to give ourselves a single funded status number? Or are we only focusing on pension plans that are underfunded, without offsetting underfunding in those plans with any overfunding in plans with sufficient assets?

Let’s start out with the wrong answer here, both because it will explain one of the major distinctions of the Administration’s estimate as contrasted with the general conclusions we’ll point to from broad industry studies, plus this false premise – in maybe being at least closer to a basis large enough to carry the weight of a $450 billion number – suggests at least to me the overwhelming unlikelihood that the final answer is anywhere close to that $450 billion. OK then, the wrong answer here would be that we’re looking at all single employer defined benefit plans, both overfunded and underfunded. Again with my simple math, that would mean that against the $1.53 trillion of assets, we’re staring down $1.98 trillion of liabilities. Which would mean that even when we include plans that have sufficient assets, the average funded ratio for the entire single employer industry is only about 77%. And even my most simple math can’t make that work: against all of the excess assets out there for the overfunded plans, our typical underfunded plan – not the rare distress termination that gets the worst press, but common enough to balance out the underfunding in the system – would need to be somewhat less than 50% funded. God help us all, we’d have far more pension plans already forced to make deficit reduction contributions even under the current rules than is the case.

The press all too frequently blindly passes along the $450 billion number as though it represents the funded status for the entire single employer pension universe, but in all likelihood that’s not what the Administration means to be fingering. Indeed, Table S-42 of the PBGC data book points to an aggregate funded ratio of 105% for the overall single-employer system at the worst point of the “perfect storm” of 2000-2002. Favorable investment markets since then, combined with FAS 87-driven advance contributions, make it highly likely that 105% was our low water mark, that overall funded ratios have increased since 2002 instead of continuing the 2000-2002 decline. But to reiterate a recurring theme in my rant: all of this would be more easily addressed if the Administration itself better documented and communicated its own numbers and analysis.

Only the Underfunded Plans? Then Let’s Get Some Balanced Press. I’ll continue my simple math in a minute, but first a bit of a swipe at the alarmist one-sided reporting we too frequently see on this issue. If in fact we are only looking at underfunded plans, then even if the $450 billion had any credibility, it sure would be nice whenever the press or any policymaker repeated that figure, if in the same breath it were quickly pointed out that overfunding in that same pension system exceeds $450 billion. Sure, pension math can get terribly complicated rather quickly, but please don’t strip it down to the extreme of a grossly misleading oversimplification without playing the flip side of the record.

Can the Average Funded Status of Underfunded Plans Be As Low As 63%? Before I return to my 4-function hand calculator, let’s first nail down that we’re probably focusing on only the underfunded plans when we try to reach $450 billion, ignoring overfunding in overfunded plans. In my rant earlier here, I find it impossible to reconcile the numbers if the balance of both overfunded and underfunded plans were to net out to a $450 billion deficit, but we probably have some strong evidence in the PBGC data book of what the actual basis might be. Returning to Warshawsky’s reference to the $450 billion as having grown from 2000’s $7 billion number, Table S-47 of the PBGC data book correlates that $7 billion number with a $6.57 billion number from Form 5500 filings, which on Table S-43 of the PBGC data book is identified as the underfunding for underfunded plans, as distinguished from the overfunding amounts presented on Table S-44.

Now on Table S-43 we’re left in the lurch back over 3 years ago at the worst moment of the “perfect storm,” where the PBGC tells us we had some $130 billion of underfunding in the underfunded single employer plans. If you see things colored by suspicions voiced by at least one congressman, you might infer that we don’t have a more recent version of Table S-43 because that may in fact have been the lowest point, with total underfunding in underfunded plans less than one third of our $450 billion number. And under the conspiracy theory view of things, since it would not serve the Administration’s purpose in perpetuating the image of a pension system in crisis if we were to see any signs of pension funded status improvement since that lowest point, the suggestion has been made that updated data is being withheld until after action on the pension reform legislation has been completed. Whatever. For now, let’s take what we can get and try to work from there.

First, let’s pretend that the plans that were underfunded in 2002 remained underfunded and the plans that were overfunded remained overfunded. Even my non-math head knows that would not have been the case, but let’s at least start with that. At the very least, I think it completely safe to say that we cannot extend the progression of plan asset amounts from Table S-43 for the years 2000-2002 to arrive at any reasonable estimate of where the total assets of underfunded plans might have stood at the end of 2005: even with my simplest additions, extending the perfect-storm-driven growth of that number through 2005 would completely swallow up the $1.53 trillion number suggested as currently comprising all single employer pension plans, and even the Administration is not going to pretend that to be the case. So for starters, let’s just take 2002’s underfunded plan assets of $768 billion as a very rough draft marker for the assets of underfunded pension plans at the end of 2005 – quite coincidentally, that would be about half of the $1.53 trillion total, which would be saying that about half of pension assets are in overfunded plans against half in underfunded plans. And then working off that rough draft marker, together with other data suggesting that the system overall has excess assets in overfunded plans that very likely equal or exceed the deficits in underfunded plans (more on that score later in this rant), I again quickly arrive at numbers that simply don’t make sense. Proceeding from that rough draft marker: if the end of 2005 were to see assets in underfunded plans of $768 billion, then $450 billion of underfunding on top of that would give us liabilities of over $1.2 trillion, for an average funded ratio of the underfunded plans in the neighborhood of 63%. That would be different from the earlier 77%, this time supposedly representing average funding solely for the underfunded plans, but even there I can’t see it. Interestingly, if we were to revalue the liabilities of the largest ulcers in the PBGC’s worst scenario future – a revaluation I’ll come back around to if I rant long enough here – then we can actually come close to reaching down to 63%. But even the elephantine weight of those worst case threats is not sufficiently representative to get us to an average funded ratio of 63% for underfunded plans. OK, if that rough draft marker of $768 billion is too low, despite overall improvement in pension funded levels since 2002, then we would have an asset base for underfunded plans that might give us a more reasonable funded ratio to stretch toward; except then we have a smaller asset base for the overfunded pension plans in the system, which with the overall net positive balance systemwide would mean that there are some extraordinarily rich plans out there. Which is not the case, not sufficiently so as to make these numbers balance out. And as seems very strongly suggested by evidence beyond the 3-year-old PBGC data, if in fact the proportion of the system’s assets in underfunded plans is less than 2002’s figure of $768 billion, then a current deficit of $450 billion becomes absolutely impossible to accept, since then we’d be talking of average funded levels for the typical underfunded plan that are lower than 63%. Even on alternative valuation assumptions, I see no way to inflate Table S-43’s $130 billion of underfunding to $450 billion by the end of 2005 and have everything balance out to what we know from more current evidence. And again to reiterate a recurring theme in my rant: all of this would be more easily addressed if the Administration itself better documented and communicated its own numbers and analysis.

Other Evidence – Corporate Financial Statements. I’ve probably already ranted long enough here to have killed off any but the weirdest of readers; so reminding myself that I continue to blog more for my own purposes than for anyone reading over my shoulder, for now I’ll try to keep this part of my rant mercifully brief. Except that although as discussed to this point I can’t see anything in the Administration’s numbers that makes any sense, I’m not putting the most weight on that leg of the argument; rather, I see strong evidence elsewhere that directly disputes the $450 billion myth. Namely, the pension numbers reported by companies in their annual financial statements. Which, as studied by Credit Suisse and various researchers, indicated average funded levels near 100% as of the close of 2005 for the combination of overfunded and underfunded pension plans.

Although I myself have not dissected that financial statement data directly and to a great extent am here relying on discussions I’ve had with an analyst who has done so, it bears repeating a major issue raised in my weblog post two weeks ago: studies of pension numbers based on companies’ financial statements invariably are taking the numbers directly as reported, including foreign plans (which by and large tend to be less well funded than U.S. pension plans) and unfunded non-qualified pension plans. Taking ExxonMobil as but one of numerous illustrations, the numbers going into the typical study of pension funded status sees a worldwide number less than 72%, while isolating only the U.S. plans would have the higher figure of 77% (and even a higher funded status for U.S. qualified plans, were we to eliminate nonqualified plans from the analysis). So too for almost every single pension sponsor included in the typical study. The long and short of it being that if we look solely at U.S. qualified single employer plans, then employers’ financial statement data indicates an overall average funded status, again for now aggregating both overfunded and underfunded plans, that actually exceeded 100% as of the close of 2005. Again, as developed earlier here, our $450 billion has already too often been miscommunicated as the deficit for the system as a whole, referring instead to the underfunding of underfunded plans, and I don’t mean to repeat that mistake here: I’m not saying that the $450 billion should be eliminated altogether, even if the overall funded status for the system does exceed 100%. Even so, in the pooled overfunded+underfunded data from employers’ financial statements we have an independent starting basis that gives an overall context within which that $450 billion, if it in fact has any validity, must ultimately fit.

Corporate Financial Statements Again, This Time Just on Underfunded Plans. Thing is, companies’ annual reports not only give us pooled numbers, but separate out numbers for underfunded plans. Again, I rely on input from an analyst who has amassed extensive data on this, not yet having had the time myself to do much more than scan through the numbers he has sent me, but at least that data is far more accessible than what we’ve seen supporting the $450 billion claim. And solely isolating underfunded plans without offsetting by overfunded plans, again segregating out foreign and non-qualified plans, the amount of underfunding for underfunded plans reported by S&P 500 pension plan sponsors as of the close of 2005 does not even reach $100 billion. Having already stretched this rant miles farther than I intended this morning, I won’t here indulge in the exercise of extending S&P 500 conclusions to the pension universe at large, save to suggest that my own back-of-the envelope scratch calculations suggest that if that $100 billion for the S&P 500 is at all reliable, then the underfunding of all underfunded plans could not be too much more than $140 billion unless we hypothesize that non-S&P 500 companies’ pension plans are far more underfunded than those of S&P 500 companies. And by the way, since the data I’ve seen for the S&P 500’s underfunded pension plans actually suggests that the total amount of underfunding in those plans has remained virtually constant over the past several years (to oversimplify here, with asset gains for 2003-2005 having by and large been cancelled out by liability remeasurement attributable to declining interest rates during that period), those scratch calculations might actually be echoing the PBGC’s S-43 number for 2002 underfunding: about $130 billion for all underfunded plans in the entire system.

Are We Using Reasonable Assumptions? Ultimately, I expect that the issue is going to have to come down to reconciling the actuarial assumptions used to value pension plan liabilities. Which is where my 4-function calculator even in its most valiant adventures dares not tread: I would be more than happy to leave that debate to the actuaries (although I personally am still of the opinion, as previously blogged here, that pension actuaries have not yet been as present and active in this debate as so important a national policy has desperately needs them to be). And that said, I can once again reiterate a recurring theme in my rant: all of this would be more easily addressed if the Administration itself better documented and communicated its own numbers and analysis.

This particular non-actuary wonders that if the settlement basis snapshot that is required under the accounting standards for companies’ financial statements is so drastically off as to require us to triple the level of underfunding that has been reported for underfunded plans to take things from the FAS 87 basis to the Administration’s basis, then don’t we have a story for the SEC and investors that is as gravely serious if not more so than is being told to congressional conferees? And notice here that this is not the same press as the significantly higher number being touted as the potential hit to shareholder equity that would arise from FASB’s proposed new pension accounting rules. That particular number is largely where it is not so much because of underfunding, but rather because of significant existing prepaid pension costs, recently made even higher as companies scrambled to avoid drains on shareholder equity arising under from existing accounting rules (alas, such scrambling to have been in vain if, as proposed, the new accounting standard would ignore all that advance funding). No, if the $450 billion is supposed to represent the more reasonable estimate of current funded levels for underfunded plans, then figure there to be some $300 billion of liability that has not yet been even acknowledged for financial reporting purposes, and a comparable further hit representing similar remeasurement of the obligations for overfunded plans. Without any apologies whatsoever for the limits of my simplistic math, that size of underreporting simply does not exist. In other words, if financial reports being prepared on the settlement basis under FAS 87 are anywhere near being representationally faithful, then $450 billion is overstated to the tune of 3 times the true level of underfunding among underfunded plans; conversely, if the assumptions being used to value those financial statement numbers are so severely liberal as to lend credence to the $450 billion number, then we should be given solid evidence of that allegation instead of the unsubstantiated nebulous explanations we’re currently being fed, and then in that instance the policy implications should be far more reaching than even the most aggressive pension reform, reaching through to FASB and the SEC and the investors and creditors of pension plan sponsors. By and large, although “actuarial assumptions” can too easily carry some water in this debate, I find no evidence to suggest that even the most liberal assumptions being used for financial reporting (and truth be told, there are some rather optimistic assumptions being used by a very small minority of pension plan sponsors) are so outrageous as to bridge the gap from at most $150 billion of underfunding as reported under FAS 87 to the Administration’s alleged $450 billion.

Corroboration From Schedule Bs. Yet again I must look to the actuaries to flesh out this side of the debate. Suffice it for me to comment that based on information I’ve been walked through for several key Schedule Bs, the FAS 87 settlement assumptions used for companies’ financial statements may not be so drastically understating liabilities as the $450 billion number might suggest. To be sure, Schedule B information is almost as outdated as the PBGC data book, at this point publicly available only through 2003. But line up those Schedule Bs with the concurrent FAS 87 numbers, and they’re not so widely divergent as to suggest drastic understatement of the FAS 87 results. So to my recommendation that the SEC be brought into the debate if the FAS 87 numbers are as far off as the Administration would have us to believe, add the Actuarial Standards Board and the IRS and even the PBGC itself, since we’d pretty much be accepting that all actuarial valuations – not just for FAS 87 purposes, but for Schedule B reporting as well – are badly understating liabilities. But I simply don’t see evidence of so severe an understatement. And again to reiterate that recurring theme in my rant: all of this would be more easily addressed if the Administration itself better documented and communicated its own numbers and analysis.

But That’s Not What I Came To Talk About. That’s right, after all this “extension of remarks,” so to speak, that’s not even what actually set me off two weeks ago. What I found so unbelievable then was the Administration’s response to American Academy of Actuaries research that points to significant improvements in pension funded status since the close of 2005, based on favorable investment experience and rising interest rates during the first half of 2006: $450 billion remains the “official government estimate.” While I’m not going to be the one to steer anywhere close to seeing the long-term commitment represented by a company-sponsored defined benefit pension plan as legitimately treated like some day trader’s options contract, I find it terribly difficult to reconcile that ostrich stance with the Administration’s continual calls for greater transparency in pension reporting. The American Academy of Actuaries stands as the credible authority on this point: although serious underfunding still must be addressed in the minority of plans that continue to pose a threat to the PBGC, pension funded status has improved quite dramatically from the darkest days of the perfect storm, and through the beginning of 2006 has continued to show clear signs of improvement. I seriously don’t find reason to trust the validity of the $450 billion figure at any point during the history of the U.S. pension system; but if it was even distantly representative of some strained worst-case scenario at any point during the on-going legislative process, then most certainly the number is already materially lower now. For the Administration to suggest that interest rates can increase significantly while investments are outperforming expectations, yet that those changes have left pension funded status untouched, asks me to believe a financial reality that I must reject outright, without any apology whatsoever for personally lacking the actuarial credentials to back up my opinion.

So What Number Would Feel Right? On the basis of PBGC data through 2002, Schedule B data through 2003, and companies’ financial statements through 2005, for underfunding as of the end of 2005 for underfunded U.S. single employer qualified plans, I’d feel $200 billion to be a bit on the high side, but I’d not rant nearly so much. Even that number would allow for a degree of remeasurement sufficient enough to warrant closer attention from financial accounting auditors and the SEC, but at least does not so drastically depart from the best settlement basis estimates we have at our disposal. Even then I would insist on pairing that number with a comparable number for the overfunded plans on the opposite side of the fence, so as not to be so atrociously misleading about the overall state of the pension system. And even if I did not recalculate the number more regularly than annually, I would at least acknowledge the plain financial truth that rising interest rates paired with favorable investment results would not leave those figures untouched.

And I’d dig a deep grave for that $450 billion number.

1 Comment »

  1. The Pension Benefit Guaranty Corp. – the federal agency that insures private pension plans much like the Federal Deposit Insurance Corp. insures bank accounts – estimates that $100 billion of that [pension underfunding] is with companies with serious funding problems.

    AP 6/30/06

    Oh, indeed! That matches the information I’ve seen, as close as one might ever expect to see when it comes to measuring pension obligations.

    Now if only the Administration, congressional policy-makers, the press, and everyone else would carry through the rest of the math: you can’t tally up only $100 billion of underfunding in the companies with the serious funding problems (as in, auto, airlines, steel, and other mature industries), then expect to find $350 billion additional overfunding sitting around in other companies that do not have serious funding problems.

    Once again, kill that $450 billion fable.

    Comment by Fuguerre — July 1, 2006 @ 10:03 am

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at

%d bloggers like this: