If I were to run an office pool for the after-tax reduction in equity from implementation of FAS 158 for pensions and OPEBs of S&P500 companies, rounded to the nearest billion, what number would you be looking for? Mine would have three digits beginning with a 1, and I’m seriously mulling 9 for my second digit, although I might be convinced to go with an 8 if the markets keep pushing new records.
I’ve been reading with interest Merrill Lynch’s Market Impact of Pension Accounting Reform, which features an estimate of FAS 158′s projected hit to equity due to come this yearend for most U.S. sponsors of pension and other postretirement benefits. Intriguing is that whereas most other analysts who even ventured an estimate couched their appraisals in terms as though the new accounting standard had been implemented at the end of the 2005 fiscal year, Merrill projects forward to the end of the 2006 fiscal year, for most plan sponsors the point when the equity hit will actually land. Roughly capsualized, Merrill picks $217 billion as the after-tax reduction in equity for S&P500 companies. I think it likely that guess will turn out to have been too high.
Just the S&P 500: First, let’s get through the fog that has had some pointing to that estimate as the anticipated equity hit for all pension/OPEB sponsors. As best as I know, the only guess for the entire population of U.S. companies was extrapolated by the SEC in its mid-2005 Report and Recommendations Pursuant to Section 401(c) of the Sarbanes-Oxley Act of 2002 On Arrangements with Off-Balance Sheet Implications, Special Purpose Entities, and Transparency of Filings by Issuers: $414 billion off-balance sheet liability for pensions and $121 billion for OPEBs, adding up to $535 billion. As that was a pre-tax off-balance sheet number, using Merrill’s tax rate assumption of 35% would trim even that half-trillion number down to an after-tax equity hit of $348 billion. And then as the SEC extrapolations depended on data from 2004, a year earlier than the data gathered by Merrill, and since 2005 improved the forthcoming FAS 158 outlook for most companies, it’s not difficult to imagine that were the SEC’s extrapolations to mimic Merrill’s projection to the end of 2006, we might see a number for the expected after-tax equity hit for the entire universe in the neighborhood of $300 billion. Which would not be altogether inconsistent with Merrill’s $217 billion, as long as Merrill’s figure is properly reported as representing the S&P 500 set, versus all companies.
Pension Cost Projection. So far, the part of Merrill’s report I find the least credible is its projection of net periodic costs during the 2006 fiscal year. For Boeing, which they use to illustrate their methodology, they project a net periodic pension cost of $608 million and a net periodic OPEB cost of $602 million for fiscal year 2006. That OPEB estimate might not actually be all that distant, but Boeing’s second quarterly report for the 2006 fiscal year shows net periodic pension cost for the first 6 months of 2006 ringing in at $529 million, on course to top $1 billion for the full year. That as compared with a 2005 net periodic pension cost for Boeing of $751 million (with the $552 of “Other” cost cited by the Merrill report constituting FAS 88 recognition). And although Boeing’s increase in net periodic pension cost of over 33% from 2005 to 2006 will be near the top of the list of cost increases, anticipate the median for S&P 500 pension sponsors to be in the same direction: conversely to the very much criticized early 2000s, when net periodic pension credits continued to be reported by many pension sponsors despite the onset of the perfect storm of market crashes and interest rate declines, 2006 will be a year when pension costs increase for the majority of companies, despite favorable markets and rising interest rates. (One wonders if we’ll now see the same criticism coming back in the other direction, arguing that corporate profits ought be higher for 2006 because pension costs are presumably being “overstated.”) Anyway, if Merrill’s estimate for Boeing’s pension cost serves fair indication for the entire study set, I think we can almost dismiss that part of the methodology. I’m not yet sure how their number went so far astray, and am engaged in checking with several close actuarial and accounting colleagues to get their take on the matter: even without poring through interim financial reports, these projections ought not be so significantly off target, else one starts to seriously wonder whether the remainder of the projection methodology has any similar disabling deficiencies.
OK, but if Merrill’s pension cost projection is similarly understated for the entire S&P 500, what does that do to their $217 billion result? Actually, less than one might think. Material underestimation of net periodic pension cost might mean we should increase Merrill’s projection of a $397 balance sheet liability, yes. But much in the same way that, as Merrill points out, employer contributions do nothing to the hit to shareholder equity arising from FAS 158 implementation, likewise that hit is relatively immune to 2006 cost. The only cost components that matter, actually, are the amortization pieces, since they will affect the amount of unrecognized balances that are the basis for the equity hit. So back to comparing Merrill’s Boeing with Boeing’s 10-Q: Merrill projects amortization cost components of $823 million for 2006, while annualizing the second quarter report suggests the actual number will be $1.1 billion. Still a significant difference between projection methodology and reality, but less than for the full pension cost, and particularly less in the context of the understatement relative to the full balance of unrecognized amounts. And interestingly, of the opposite effect as with the overall cost understatement. So here’s another one for those who actually believe that we’ve made pension accounting more understandable: although by underestimating net periodic pension cost, Merrill may have underestimated the projected balance sheet liability, underestimating the amount of reduction in unrecognized balances means that they overestimated the expected hit to shareholder equity, at least with respect to that piece of the puzzle.
Ignoring the Freeze. Merrill’s recent reports have done as much if not more than any other research to track and analyze the pension ice age, and this report continues that commendable effort. Yet despite the intricate complexity of the rest of its projection methodology, Merrill stopped short of incorporating known pension freezes into its estimate of the FAS 158 hit, stating that, “For these companies, a figure closer to their ABO is likely more accurate; appropriate adjustments should be made for these specific situations as the charge to equity will not be as severe.” Yes, to put it rather mildly so. In practice, although the decrease from PBO to ABO can be significant enough to add up to some billions for the pension ice cubes out there, for some cases the effect on our hit to equity can actually exceed that PBO-ABO number, since there we’re looking at an immediate acceleration of recognition of significant accumulated net losses. Which under FAS 88 sends the effect of the freeze through the company’s P&L, but then accordingly removes that effect from FAS 158′s equity hit. In other words, as Merrill implicitly acknowledges but without fully completing the thought, some of that $217 billion number represents amounts that will have already been recognized in 2006 net income.
FAS Implementation Date or FYE 2006? As near as I can make out, Merrill is updating data from fiscal years ending in 2005 to the end of the 2006 fiscal year. For employers with calendar fiscal years, that does in fact bring us to the FAS 158 effective date. But for almost all others (other than the very rare employer with fiscal year ending from 9/29 through 12/15 who could delay implementation until the year ending in 2007, but instead elects early adoption), the projection leaves us a year short of actual FAS 158 adoption. And for almost every one of those employers, we’re leaving out a good year, one when the much-discussed new volatility inherent in FAS 158 goes in the employer’s favor, significantly cutting into the eventual hit to shareholder’s equity. OK, the overwhelming majority of the S&P 500 have a calendar fiscal year, but characterizing FYE 2006 numbers as projected FAS 158 numbers still overstates what the aggregate hit to equity will eventually be at actual FAS 158 implementation by at least several billions.
Is Merrill Warning of a Q4 Bear? Assets during 2006 projected using a 5% rate of return?! Not only have we already seen a median pension asset return exceeding 10% for the S&P 500 companies with non-calendar fiscal years (such as those ending 6/30/2006) that have already released their 2006 annual financial statements, but these past several months have been more than kind to pension funds. OK, sure, reports such as this one do not constitute any formal Merrill prediction of future markets; and granted, a firm in Merrill’s position probably does better to err on the conservative side rather than risk raising false hopes. Even so, we’re close if not well beyond the point at which in order for Merrill’s 5% to be anywhere near reasonable for the full year, they essentially had to be factoring in a very major bear market for the final quarter of 2006. Interestingly, their assumption on the liability side of the equation of a 25-basis-point increase in discount rates might not be so far off the mark: although 2005-to-2006 rates roller coastered for companies with mid-year financial yearends so that some increased discount rates by 100 basis points and more, we’re already back down to levels where Merrill’s view of interest rates makes sense. But that investment return assumption! Sorry, but I don’t buy that one.
All of Which Drops My Own Guess Below $200 Billion. I see enough right in the background data and in the rest of Merrill’s work that says if I can start with their $217 billion, then shave off some of the unrecognized balance for underestimated cost amortizations, drop some more unrecognized balance for 2006 freezes, put everything in terms of FAS 158 implementation in lieu of FYE 2006, and be a little more relaxed with my investment return expectations, I’m going to be pretty solidly below $200 billion if I go into my hypothetical office pool. For the S&P 500. (And completing the loop back to the SEC’s estimate, I’d see that as consistent with about a quarter trillion after-tax hit to equity for the entire universe at the point of FAS 158 implementation.) Just one blogger’s guess, though.