How Governments Hide Their Liabilities

In my testimony to the Senate Budget Committee the other day, I recommended that Congress set specific fiscal targets for bringing our out-of-control deficits and debt under control. My particular suggestion? Get the publicly-held debt down to 60% of GDP in 2020.

By budgeting  standards, that makes for a great bumper sticker: “60 in 20“.

But as the New York Times points out in two articles today, a measurable target isn’t enough. You also need to make sure that the government doesn’t game the accounting to hide its liabilities.

Exhibit A is Greece. The story was originally broken by Der Spiegel earlier in the week, and is described in the NYT by Louise Story, Landon Thomas Jr., and Nelson D. Schwartz in “Wall St. Helped Greece to Mask Debt Fueling Europe’s Crisis“:

As worries over Greece rattle world markets, records and interviews show that with Wall Street’s help, the nation engaged in a decade-long effort to skirt European debt limits. One deal created by Goldman Sachs helped obscure billions in debt from the budget overseers in Brussels. …

Instruments developed by Goldman Sachs, JPMorgan Chase and a wide range of other banks enabled politicians to mask additional borrowing in Greece, Italy and possibly elsewhere.

In dozens of deals across the Continent, banks provided cash upfront in return for government payments in the future, with those liabilities then left off the books. Greece, for example, traded away the rights to airport fees and lottery proceeds in years to come.

Critics say that such deals, because they are not recorded as loans, mislead investors and regulators about the depth of a country’s liabilities.

The winning quote:

“Politicians want to pass the ball forward, and if a banker can show them a way to pass a problem to the future, they will fall for it,” said Gikas A. Hardouvelis, an economist and former government official who helped write a recent report on Greece’s accounting policies.

Exhibit B are all the contingent liabilities of the United States government, of which Fannie Mae and Freddie Mac have been the most prominent (and expensive). In “Future Bailouts of America,” Gretchen Morgenson interviews budget expert Marvin Phaup (now at George Washington University and previously a colleague of mine at the Congressional Budget Office). She writes:

“If we are extending the safety net, extending the implied guarantee to the debts of a lot of other financial institutions, and we know those guarantees are valuable and costly, then we ought to start budgeting for it,” Mr. Phaup sad in an interview. “We can’t reduce the costs of these subsidies if we can’t recognize them.” …

As the number of firms with implicit government backing has risen because of the crisis, so too have the expected costs of those commitments, Mr. Phaup said. And yet, under current budget policy, those costs will be ignored until the recipient of the guarantee collapses — the precise moment when the guarantee is likely to cost taxpayers the most.

If we are going to set an explicit target for the publicly-held debt–60 in 20!–, we need to think carefully about what politicians may strategically omit from the calculation of the 60.


10 thoughts on “How Governments Hide Their Liabilities”

  1. Great post. There are lots more examples of masked borrowing. One example if the provision allowing conversion of traditional IRAs into Roth IRAs. Taxpayers get enormously valuable future tax breaks in exchange for paying a little tax now. I think I estimated that the internal rate of return on Roth conversions was something like 12 percent for a hypothetical 40 year old, implying a staggering borrowing cost to the government. But instead of debt, the rollovers are scored as revenue and reduce measured deficits for a couple of years.

    State and local governments have turned hidden borrowing into an art form, since most are required to balance their budgets. Selling future tobacco tax settlement revenues at a steep discount, borrowing from state employee pension plans (and/or underfunding them), and myriad other creative schemes exist.

    Correctly measuring debt if policymakers have an incentive to hide it will be an enormous challenge.

  2. Donald,

    Great point re: hiding liabilities (or viewed differently in some cases, overstating projected revenues — from the government-run lottery, etc.). Isn’t the opportunity to engage in such deception largely dependent on a limited time horizon — e.g., a debt/GDP goal set for 2020 without accompanying goals for 2030, etc.? Eventually such hidden elements must come to the surface (reducing projected revenues and/or increasing projected spending*), and more so with time, right?

    * Unless there is explicit deception (rather than implicit), with the government dishonestly claiming access to revenues (from the lottery, for example) that it simply doesn’t have.

  3. Thanks Len, Marvin, and Brooks.

    Brooks is right about the time horizon. The specific items that Len mentions would likely be captured if you looked at projections that stretch far enough (but I am sure Len has other ones in mind as well that might not). The contingent liabilities that Marvin wrote about would not, however, since the current convention is to leave them out of the budget until they actually happen. On this point, by the way, kudos to the Administration for including in the budget a line item for future disaster costs, which are another form of contingent spending that often gets overlooked.

    In terms of a fiscal target, we will definitely need more than a single target in a single year, but it may not fit on the bumper sticker :). Although, 60 in 20 and declining thereafter might work.

    1. Thanks Donald. In addition to preventing/reducing such accounting gimmicks, another advantage of setting goals per a longer time horizon is, of course, better planning and more rational prioritization. For example, if we want debt/GDP to be 60% in 2030, we could choose between two approaches, (1) reaching 60% by 2020 and maintaining it through to 2030, which I assume would require further reduction in seniors’ entitlement spending per beneficiary and/or reductions in eligibility (given the continued increase in the number of seniors), and at least greater reduction than per Option #2, or (2) getting debt/GDP below 60% by 2020 so we can avoid/mitigate that reduction in spending per beneficiary and allow more such spending than in Option #1 and allow increases in debt/GDP, eventually reaching 60% in 2030.

      Of course, I’m just picking 2030 for illustration. The same would apply for a target in 2040 or whenever, and a target for 2030 may not suffice for the same reason (i.e., the drivers of the fiscal imbalance not yet leveling out). The point is just the equivalent of saving for our own retirement or saving for college or for a rainy day that is inevitable. If we want our balance sheet to end up a particular way in 20 or 30 years, we may prefer to sacrifice more over the initial 10 years (and exceed that ultimate balance sheet target at the end of the first 10 years) so we don’t have to sacrifice even more than we otherwise would in the subsequent 10 or 20 years.

    2. Donald,

      Regarding contingent liabilities (in particular, loan guarantees), wouldn’t an ideal solution be to require the government to, in effect, purchase insurance via some sort of credit default swap, so we wouldn’t just be including some “expected value” type cost in budgets, we’d actually be making the cost explicit and known (since we wouldn’t have to pay more if the contingent liability became an actual liability)?

      Of course, that would require a sufficient market and require that we could reasonably assume that the seller of that “insurance” will be able to pay up if and when necessary (and without government sending them money so they can send it back to government), as opposed to our doing the rough equivalent of buying insurance on the Titanic from a passenger on the Titanic.

      Do you agree this approach would be ideal, and if so, what are the prospects, given the requirements I mention and/or others?

  4. I can’t help but think of Polk’s and dispute over Oregon “54-40 or Fight!”

    From a more current perspective – are we not over due for a discussion of the effects of accrual accounting would have on the 60 in 20 campaign and other fiscal debates?

Comments are closed.