Note: This post was published about 12 hours prior to publication of Justin Hyde's article on the same topic in the Detroit Free Press. Justin was the one who brought the topic to my attention, and I made the decision to write this post after a conversation with him. I thank him for allowing me to go ahead with publication, even though his own article would be appearing later.
In an earlier post, I expressed my strong suspicion that top managers at General Motors were utilizing big bath accounting. By 'big bath', I mean a violation of GAAP that permits delayed recognition of relatively small losses over time, so as to recognize the whole enchilada in some later period. For some reason that others may wish to ponder, managers prefer the big bath to the accounting equivalent of death by a thousand cuts. In GM's case, they appear to have improperly delayed as much as $11 billion in writedowns of their deferred tax assets.
Now comes another enormous red flag out of GM's public disclosures. In fact, the numbers -- in the neighborhood of $50 billion -- make the big bath look like a glass of water. This new one is of the 'cookie jar' variety: the improper deferral of a gain so as to spread its sweet goodness to the benefit of many subsequent accounting periods. But, sad to say, this tale has another annoying twist: if GM doesn't get SEC approval for the accounting they are aiming for, they can -- for no other good reason -- opt out of their recent milestone agreement with the United Auto Workers.
How this Opportunity for Accounting Shenanigans Came to Be
Before I get into the details of the current situation, some background information may help. GM has an 'OPEB' ('Other Post-Employment Benefit') liability on its balance sheet that is somewhat north of $50B. It represents the present value of estimated future payments to employees as reimbursement of health care costs during their retirement years. In all, the plans cover about 500,000 current and retired employees. I have read that the expected future payments add about $1,600 to GM's per-vehicle cost, which is about eight times the cost incurred by foreign competitors (who benefit from more generous state-sponsored health care programs). Note 15 to the financial statements in GM's 2007 10-K indicate that they spent in the neighborhood of $6 billion on retiree health care costs in that year.
Yuck. How did GM let itself get eaten alive by an OPEB in the first place? The story starts with accounting standards -- or more accurately, the appalling lack thereof. FAS 106, though significantly flawed, filled a gap in GAAP, but it was birthed only in 1990 -- long after the horses galloped through the open barn door. My recollection from reading the financial press in the years just preceding is that corporate America was already buried under approximately $1trillion in off-balance sheet liabilities relating to retiree health care costs. Why did management keep them off-balance sheet? Because they could. Why did managers let the liabilities get to be so humongous? Because they were off-balance sheet.
Let me explain. When negotiating with unions, companies could either grant wage rate increases that would affect the bottom line starting at Day 1, or provide deferred compensation that would not hit the income statement for decades. Such was the case with retiree health care benefits prior to FAS 106. The "generally accepted" accounting prior to then was "pay as you go." In other words, you expensed only that portion paid out to employees and their health care providers. Actual payments (and thus, expenses) at the outset were low because so few of the employees to whom benefits were promised were old enough to begin receiving them. By the time FAS 106 came to require accrual of benefits as the employees earned them, the unionized rust belt was already awash in unfunded, gold-plated retiree health care plans. To make matters worse, health care costs looked like they might increase faster than inflation forever.
Back to Now
Late last year, GM and the UAW entered into a compromise ('Settlement Agreement') whereby GM gave its commitment (albeit with an escape clause I shall address anon) to pre-fund, in 2010, its $50 billion accumulated retiree health care obligation. In exchange, GM would be relieved of any future obligation to make payments, except for funding annual plan shortfalls up to a paltry $165 million per year for the next 20 years. (The UAW thinks that GM's money should last them 80 years, but that's another story.)
$165 million? What's up with that? The numbers I gave you earlier make it abundantly clear that it's but a drop in the bucket compared to the expected plan costs and the number of employees in the plan. If we assume that expenditures are the current amounts paid by GM and ignore inflation, $165 million amounts to about 10 days worth of coverage. If we further assume that there are about 1 million beneficiaries (retirees plus spouses), that's a safety net of only $165 per beneficiary. That would be like a safety net made of thin-sliced swiss cheese.
As to the real purpose of the $165 million, it's much akin to a fly on a cow's hindquarter: maybe just enough to get the cow to swish her tail -- or to get the 'right' accounting. The 'right' accounting for GM is "negative plan amendment" treatment under FAS 106, or else they're gonna pick up their marbles and go home.
And just what is negative plan amendment accounting? It's a cookie jar reserve. Basically, the accounting treatment of transactions of this ilk boil down to three possibilities:
- Settlement: The liability would be taken off the books, and a gain (around $50 billion) would be recorded in 2010 when the settlement occurs. The GM-UAW agreement looks like a settlement and quacks like a settlement, but FAS 106 (para. 90) defines a settlement as "...a transaction that (a) is an irrevocable action, (b) relieves the employer ... of primary responsibility ... and (c) eliminates significant [emphasis supplied] risks related to the obligation and the assets used to effect the settlement." Thus, the result of settlement accounting would be no cookie jar: just a blob of earnings that can't be used to juice any earnings-based compensation of top management.
- Negative plan amendment: Even though a plan amendment immediately affects the calculation of the liability recorded on the balance sheet, FAS 106 requires that it be deferred and recognized over the time that current employees become eligible for retirement (para. 55). If that amortization period is, say, 20 years, then negative plan amendment accounting creates an earnings cookie jar to be drawn on at the rate of $2.5 billion per year.
- Partial Settlement: GM is insisting that the recognized liability be written down to about $1.5 billion, the present value of a 19-year annuity of $165 million per year. It is conceivable that one could find that GM is exposed to more risk than that amount, and that, therefore, the liability should be higher.
Section 21 of the Settlement Agreement (Exhibit 10(m) of the 10-K), is where stated that GM can hold up the agreement if they can't get the liability on their balance sheet down to $1.5 billion. Both settlement and negative plan amendment accounting will do that, and there is some chance that the Settlement Agreement may qualify for neither. That's the scenario under which everybody has to sit down and renegotiate. However, a presentation that GM gave to analysts reveals that the brass ring is negative plan amendment accounting. That's where the measly $165 million comes in; it's supposed to be just enough to be considered "significant." They want the SEC to say that because of it, settlement accounting is not appropriate, and that accounting as a negative plan amendment is the result. It's a ridiculous charade, well-hidden by the following 10-K disclosure appearing under the caption "Risk Factors":
"We are relying on the implementation of the Settlement Agreement to make a significant reduction in our OPEB liability. Under certain circumstances, however, it may not be possible to implement the Settlement Agreement. The implementation of the Settlement Agreement is contingent on our securing satisfactory accounting treatment for our obligations to the covered group for retiree medical benefits, which we plan to discuss with the staff of the SEC. If, based on those discussions, we believe that the accounting may be some treatment other than settlement or a substantive negative plan amendment that would be reasonably satisfactory to us, we will attempt to restructure the Settlement Agreement with the UAW to obtain such accounting treatment, but if we cannot accomplish such a restructuring the Settlement Agreement will terminate...."
I have a couple of things to say about this disclosure:
- First, the possibility of not getting the accounting treatment one wants is not a risk factor. Risk factors have to do with the possibility of real losses; paper losses are just that -- unless, perhaps, recognizing a paper loss has an indirect real effect like tripping a loan covenant. In fact, the SEC has said as much quite recently, and I wrote about it here. I admit to not having read the 10-K completely (I do have a life), but I can't see that the accounting treatment has any such indirect effects. If there were any, that surely is a substantive risk factor, and should have been disclosed.
- Second, what does Section 21 of the Settlement Agreement and the risk factor disclosure say to providers of capital about the focus of GM's management on the real business of running a car company? Exactly why is a particular accounting result so darn important that GM is willing to go back to the table with the UAW in order to get it? Everything else equal, you gotta expect that in a renegotiation GM will end up giving more to the UAW; they will get nothing more in return than a new "economic substance" to run up the SEC's flagpole.
When the ball is in the SEC's court, what will they do with it? It doesn't appear that anyone at the SEC has lifted a finger to follow up on GM's $11 billion big bath deferred tax asset charge, and I don't expect they will. My money says the fix is in for this one, too. The only question is how Chief Accountant Conrad Hewitt is going to fall over himself to give GM the negative plan amendment accounting they crave, resulting in what may be the largest legitimized accounting cookie jar in history.
I've been blogging about financial reporting for a little over six months now, and so far I haven't had to overly tax my brain to find something to write about once or twice a week. For whatever reason(s), there are many tales of wealth destruction that begin with a bad accounting rule. Vast destruction of shareholder wealth ensues by the deliberate actions of managers who realize they can paper over their self-serving behavior with rosy short-term earnings reports. The cases of retiree health care costs at company's like GM are particularly notable because it takes multiple manager and employee turnovers spanning decades to merely begin the process of exterminating the termites eating away at shareholder wealth and employee job security.
The GM case is particularly emblematic of corporate governance run amok because the older generations of managers skimmed accounting cream going into questionable deals with unions when more discipline was called for; now, the latest generation is trying to do the same on the back end. As they go about their business of re-arranging the deck chairs, current management seems to be doing quite well for themselves. It is even more certain that their scheming progenitors have retired and shielded themselves with ironclad contracts, signed and sealed by board members who effectively serve at the pleasure of the CEO. Those managers became rich while at the same time bequeathing their legacy of unsustainable labor costs.
I've been wondering about this too.
Posted by: Independent Accountant | March 15, 2008 at 09:13 PM
I think settlement accounting is correct in this case, and for the SEC to require a negative plan amendment would be an error.
I think your emphasis on the word significant is misplaced; it doesn't matter whether the remaining liability is significant, but rather the reduction in risk & assets. The problem is that settlement accounting has been nixed in cases where the employer is left with a residual risk or opportunity in connection with the liabilities & assets.
Posted by: Dan Moore | May 13, 2008 at 10:52 AM