With money cascading into the Pension Benefit Guaranty Corp. in recent years, it's easy to be perplexed by the fact that the PBGC's deficit keeps getting bigger just the same.
Let's look at the numbers. Between September 1990 and September 1993, the PBGC's assets tripled to $8.4 billion. During the same time, its assets per dollar of annual payout of benefits and expenses rose 57% to $10.18.
Despite these glowing results, the deficit reported by the PBGC still rose $1 billion in the same period to $2.9 billion. How could this be?
By a stroke of good fortune, while attending an actuarial convention on Maui last July Fourth, I bumped into an old friend, Akihn Akchuarian, sunning himself happily on the beach. Akihn was brought here from Moscow as a child, became an actuary, and was recruited by the Central Intelligence Agency to be a Kremlinologist and interpret the statements of the Russian command. With the collapse of the Communist party, he was laid off and came to Maui to relax and check out the actuarial world.
With the bright sun and the foaming waves as background, I told him about the PBGC's puzzling numbers and he snapped, "Tell me what you know about the PBGC."
I replied, "I will, but it's mostly what I read in its press releases, because the annual reports are pretty tough to read."
He immediately became interested. "Maybe," he said, "my combination of actuarial and CIA training will be useful to you, and I have nothing better to do right now."
Akihn called me on Labor Day to invite me to his New York pad. As he poured vodka martinis, he said: "You were right about the PBGC's annual reports. To understand the actuarial values, I had to apply CIA approaches, such as looking for big numbers that suddenly showed up with no explanation or one that was hard to understand or find. I will show you what I mean."
I pulled up a chair to his desk.
"The PBGC," he began, "has widely publicized that its deficit grew to $2.9 billion in 1993 from $2.7 billion in 1992, while the underfunding of still active defined benefit plans, insured by the PBGC, mushroomed to $53 billion from $38 billion. Given the work you have done on the PBGC's financial condition and the underfunding of defined benefit plans, which you view with much less alarm than the PBGC does, I only need tell you about something you haven't delved into yet, namely, how the PBGC's financial statements stack up against generally accepted accounting principles, or GAAP."
I told Akihn I didn't know if my heart could take the excitement, and he said, "Do your best because you'll soon appreciate how important GAAP is."
He continued. "Let's look closely at the $2.9 billion deficit from the point of view of GAAP. The biggest part is $1.6 billion of 'net claims for probable terminations.' This represents the PBGC's estimate of claims it probably has incurred by fiscal year end. Both the PBGC and its auditor, the General Accounting Office, claim the $1.6 billion is computed in accordance with GAAP as codified by the Financial Accounting Standards Board's SFAS* 5, 'Accounting and Reporting of Contingencies.' This requires that for an estimated loss to be accrued by a charge to income, it must be probable that a liability had been incurred as of the financial statement date and that the amount of the loss can be reasonably estimated.
"But the PBGC's probable terminations do not meet these criteria in large enough amounts to satisfy them. As you pointed out, in 1993, for instance, only 17% of the probable terminations became actual liabilities, and it is not known when or if the 83% balance will do so.
"Effectively, then, the PBGC, with the GAO's blessing, carries what appear to be largely speculative 'probable terminations' as real liabilities for up to several years. SFAS 5 clearly warns against this practice, cautioning that its requirements for financial statements are 'intended to prevent accrual .*.*. of amounts so uncertain as to impair the integrity of those statements.'
"The statement emphasizes that if it is not probable but only 'reasonably possible' that the amounts were incurred, then they should not be accrued as a charge to income, but disclosed instead in the statements. It would seem the PBGC would then wish, for balance, to also disclose the approximately $2 billion of future annual receipts it now can anticipate, to make the public aware of these substantial amounts of income."
"Does that mean," I ventured, "that you would throw out the $1.6 billion of what the PBGC now designates as probable terminations?"
Akihn promptly replied: "I would. A deficit is commonly defined as liabilities less assets. You can include incurred but not yet reported claims, but certainly not what may be speculative claims as if they were real liabilities.
"This position is supported, incidentally, by SFAS 60, 'Accounting and Reporting by Insurance Enterprises,' which has no provision for the addition of any liability for probable terminations. But SFAS isn't relied on by the PBGC, although the PBGC's primary characteristics are those of an insurance enterprise."
"Is there anything else you found interesting?" I asked.
Akihn then revealed the following: "Within the $1.6 billion, there are puzzling individual amounts. There is the $429 million 'change in benefit liabilities' in 1993 compared to minus $43 million in 1992; it is also nine times the average for the prior three years. This isn't further explained, and we can only guess at what it could be.
"Seven pages later, we find a footnote in tiny print stating that $175 million of the net claims for probable terminations is for 'not-yet-identified probable terminations,' an item that never appeared before 1993."
He turned to me and asked, "Do you have any idea how the PBGC could justify carrying $175 million as a real liability under SFAS 5 when no one has even identified it?"
I didn't know, but I noted the coincidence that the deficit rose in 1993 by almost the identical amount of $0.2 billion.
"Well, we certainly don't know if there is any connection, do we?" Akihn asked reprovingly, and I nodded agreement.
He went on. "There are two quite large items, other than probable terminations, that I didn't feel were adequately dealt with in the annual report - the treatment of interest and mortality changes. Because of a drop in interest rates, the report shows liabilities rose $1.1 billion. A PBGC official has yet to say, publicly anyway, that the large rise in rates since 1993 has probably wiped out most of the $1.1 billion. This would, of course, be offset by a decrease in the market value of bond holdings, so we're talking about a net drop in liabilities of $0.5 billion or so.
"Further piquing my curiosity," Akihn added, "were two major changes in the mortality tables the PBGC uses to value its liabilities, one in 1992 and one in 1993."
"And what were those?" I asked.
"First," he said, "PBGC adopted in 1992 the 83GAM Table, a standard used by life insurers. It's a bit odd, but the PBGC did not develop its own mortality experience table, which it could easily do, given that it pays monthly benefits to more than 150,000 annuitants. It could then compare its actual mortality rates to those in its then current UP84 Table to see if, in fact, a new table was needed. A Society of Actuaries' study indicates the 83GAM Table rates may, in fact, be too low for the PBGC's annuitants and, if so, would increase its liabilities more than necessary.
"Second," he went on, "it is unusual without good cause - and none has been stated - to make major changes in the mortality table two years running. The additional change in 1993, you might note, projected the 83GAM rates for 10 years, assuming a decrease in mortality in each of those 10 years.
"In asking around," he continued, "I learned that your Washington colleague, Paul Jackson, also an independent pension actuary, estimated that these two changes increased the PBGC's liabilities by about $1 billion, some of which he said may be justified to strengthen annuity reserves. But Jackson believes the existing actuarial basis was already throwing off gains, an argument for not making any changes in the mortality table at all."
"Well," I said, making a quick calculation, "this means that the deficit has practically vanished. If we eliminate from the $2.9 billion deficit the $1.6 billion in net claims for probable terminations, $0.5 billion due to the upsurge in interest rates, and $0.5 billion up to $1 billion because of the questionable mortality table changes, then there may be nothing left of the $2.9 billion."
"Yes," Akihn said, "that is distinctly possible. And given that the PBGC is now flush with money - it holds $10.18 of assets for each $1.00 it pays out - the agency can now seriously begin to think of cutting its premium rates. As you noted, it probably shouldn't be holding more than $5 or so."
It was starting to get late. "I may be getting preachy because of my martini," Akihn said, his voice becoming husky, "but, the accumulated effect of all the negatives presented by the PBGC is to present a picture that I personally believe is in violation of both the letter and spirit of GAAP, meaning those who need accurate data concerning the PBGC will probably be misled.
"The Congress, for instance, is now deliberating the PBGC reform bill and relies on the PBGC's numbers. You can't blame Congress for accepting what the PBGC puts forth as factual - there is no other organization examining it critically besides the GAO, and it appears the GAO is not staffed for it. The GAO also may have compromised its independence as an auditor by its strong support of the PBGC's position that its financial condition is vulnerable and needs to be remedied."
I got up to leave. "Akihn, I am sure glad I bumped into you at Maui. Your CIA training was superb. Unfortunately, that is what it seems to take to make sense of the PBGC. I now know to whom I can turn whenever I hit a snag."
*Statement of Financial Accounting Standards, published by the Financial Accounting Standards Board
David Langer is president and consulting actuary at David Langer Co., New York.