Loss of Control: Look the Other Way

Don’t measure it so that you can ignore it

Just saying “show me the money” doesn't work

A previous posting pointed out that the federal government does not have a balance sheet. It operates on a cash basis. The point was made in connection with the discussion of the Medicare and Social Security trust funds. Those trust funds illustrate one problem created by the federal government accounting. As ridiculous as it seems, the only asset that the federal government treats the same way other institutions treat assets is its own debt. Treating one's own debt as an asset makes no more sense for the federal government than it would for you or me. Further, the only obligation it treats as a liability is the national debt. So it's treating its debt as a liability in some instances, but if it holds that debt, it treats it as an asset.

The problem with the approach is that owing oneself money is hardly an asset. Plus, debt is not the only form of a liability. So, the first step is to identify what is not being taken into account. In some respects, liabilities are easier to identify. Easier because a major liability is recognized as the national debt, but even there a more realistic approach would be to delete that portion of the national debt the federal government owes to itself.

The national debt, however, is not the only liability of the federal government. Liabilities are contractual obligations to future payments. In the case of the federal government, contractual obligation is a nebulous term. Federal obligations originate from the promises of politicians when they are written in the law. But even then, laws can be changed. So the true extent of federal liabilities depends upon one's belief that the federal government will honor previous commitments.

A common approach and one that is highly useful is to assume that current laws will not be changed and project the obligations based on that assumption. The advantage of the approach is that it recognizes unfunded mandates as liabilities. Some of those mandates have been discussed in previous postings. Even recognizing them as liabilities only goes halfway to establishing an accurate accounting for the liability side of the federal government.

One doesn't need to make a direct promise of payment in order to create a liability. Liabilities can also be created by guaranteeing or cosigning on a liability of someone else. In the federal government case, such liabilities almost always take the form of a guarantee. Essentially, the federal government is subsidizing the borrower by lending them its good credit rating. The subsidy has value. But the implied liability has a cost that the federal government does not recognize. That sets up a terribly perverse incentive. The federal government treats loan guarantees as if they had no cost.

There are multiple ways one could go about estimating the liability implied by the loan guarantee. One appropriate approach would be to estimate the probability of default on the loans that the federal government has guaranteed. In the private sector regulated financial institutions would be required to create a loss reserve to cover that cost. That reserve would then become an asset that would balance the liability.

Absent a balance sheet no similar procedure exists for the federal government. However, the federal government could estimate the loan losses and treat that as if it were the cash outlay in the year the loans are made. It's unlikely such a procedure would be acceptable since it would necessitate recognizing that the federal government is making bad loans.

Perhaps a more appropriate approach given the federal government's practice of doing cash accounting would be to impose the same charge off rules that are applied to private sector loans. Then as the loans become delinquent the federal government would have to make it a cash charge against its budget. That would be an improvement on the current practice of just allowing the delinquent loan to exist and ignoring the fact that the federal government (i.e., taxpayers) are going to have to pay it off.

What is being ignored is not trivial. Multiple previous postings have noted the magnitude of the loans the federal government is guaranteeing. Most recently there were postings on student loans where the risks are already in the hundreds of billions of dollars range. However, there have also been postings on the government's inclination to assume the role as the leading subprime mortgage lender. The recent financial crisis makes it impossible to overlook that risk, and it automatically puts the magnitude of the risk in perspective.

The ability to create liabilities without having to account for them until they come due encourages financial mismanagement. It provides an incentive for the government to incur liabilities beyond what would be considered prudent if they were accounted for properly. But the total reliance on cash accounting also creates perverse incentives when it comes to assets.

Cash accounting means that any expenditure whether for immediate consumption or for a long-lasting asset is accounted for in the same way. Consequently, if the benefit of investment accrues over 20 years, the current period return on that investment is only about 1/20 of its lifetime value. By contrast, any expenditure for current period consumption results in 100% of the value of the expenditure in the current period. Is it any wonder then that they get expanded food stamps, relaxed qualifications for disability payments, and subsidies to buy health insurance while infrastructure is allowed to degenerate. If that seems partisan, I could've used the Medicare prescription drug plan or a tax cut as examples. The point is not partisan; the point is there is a structural problem with the incentives created by cash flow accounting.

At the state level where they have balanced budget requirements, they also use cash flow accounting. Their solution to the problem of capital investments is often to borrow specifically for the construction of a particular capital investment. School bonds are a familiar example. The effect is to create more congruence between the cash outlay requirements and the flow of benefits from the capital investment. At the federal level, no similar procedure is common. Another way to address the issue is to have separate operating and capital budgets. That creates its own problems, but at least it makes explicit investment decisions.

The deficiencies of cash flow accounting when assets are involved doesn't just pertain to investment decisions. It also creates perverse incentives with regard to managing existing assets. One would have to be incredibly naïve not to realize that the federal government has tremendous assets. They include Federal lands, offshore drilling rights, microwave spectrum, buildings, and in some cases, mineral rights just to name a few.

At any given point those assets can be viewed either in terms of their current cash value or the stream of benefits they can provide to Americans over centuries. Just as there is an incentive to focus on the immediate cash flow when investing, cash flow accounting also encourages the view of assets as a means of managing current cash flow. Thus, if the budget is out of balance, an asset can be sold to rectify the cash flow situation. Since there's no balance sheet, it has no other consequence in terms of the financial position of the federal government.

The comprehensive balance sheet for the federal government is an unreasonable request. After all, the greatest asset of the government is the ability to tax. Determining that would involve speculating on the attitude of future populations toward taxes. Nevertheless, the net effect of cash flow accounting is to encourage the natural propensity of politicians to have a short-term focus coinciding with their reelection cycles.

Cash flow management is an important aspect of financial management, but it isn't a blueprint for a complete view of appropriate financial management. At a minimum, the federal government should try to estimate those aspects of the balance sheet that can be subjected to reasonable accounting rules. 

However, far more important than trying to apply private sector accounting rules to the federal government is getting elected officials to think in terms of a balance sheet rather than election cycles. Elected officials are only part of the problem; executive employees, the bureaucrats, need to think in terms of the balance sheet rather than processes and throughput. After all, various federal agencies have tried applying private sector accounting practices with almost no improvement in performance.