16.2 Problems in Measurement


The government budget deficit equals government spending minus government revenue, which in turn equals the amount of new debt the government needs to issue to finance its operations. This definition may sound simple enough, but in fact debates over fiscal policy sometimes arise over how the budget deficit should be measured. Some economists believe that the deficit as currently measured is not a good indicator of the stance of fiscal policy. That is, they believe that the budget deficit does not accurately gauge either the impact of fiscal policy on today’s economy or the burden being placed on future generations of taxpayers. In this section we discuss four problems with the usual measure of the budget deficit.

Measurement Problem 1: Inflation

The least controversial of the measurement issues is the correction for inflation. Almost all economists agree that the government’s indebtedness should be measured in real terms, not in nominal terms. The measured deficit should equal the change in the government’s real debt, not the change in its nominal debt.

The budget deficit as commonly measured, however, does not correct for inflation. To see how large an error this induces, consider the following example. Suppose that the real government debt is not changing; in other words, in real terms, the budget is balanced. In this case, the nominal debt must be rising at the rate of inflation. That is,


where π is the inflation rate and B is the stock of government bonds. This implies

ΔB = πB.

The government would look at the change in the nominal debt ΔB and would report a budget deficit of πB. Hence, most economists believe that the reported budget deficit is overstated by the amount πB.

We can make the same argument in another way. The deficit is government expenditure minus government revenue. Part of expenditure is the interest paid on the government debt. Expenditure should include only the real interest paid on the debt rB, not the nominal interest paid iB. Because the difference between the nominal interest rate i and the real interest rate r is the inflation rate π, the budget deficit is overstated by πB.

This correction for inflation can be large, especially when inflation is high, and it can often change our evaluation of fiscal policy. For example, in 1981, the federal government reported a budget deficit of over $7 billion. But inflation was over 12 percent, and after correction for inflation, the deficit turned into a small surplus.

Measurement Problem 2: Capital Assets


Many economists believe that an accurate assessment of the government’s budget deficit requires taking into account the government’s assets as well as its liabilities. In particular, when measuring the government’s overall indebtedness, we should subtract government assets from government debt. Therefore, the budget deficit should be measured as the change in debt minus the change in assets.

Certainly, individuals and firms treat assets and liabilities symmetrically. When a person borrows to buy a house, we do not say that he is running a budget deficit. Instead, we offset the increase in assets (the house) against the increase in debt (the mortgage) and record no change in net wealth. Perhaps we should treat the government’s finances the same way.

A budget procedure that accounts for assets as well as liabilities is called capital budgeting, because it takes into account changes in capital. For example, suppose that the government sells one of its office buildings or some of its land and uses the proceeds to reduce the government debt. Under current budget procedures, the reported deficit would be lower. Under capital budgeting, the revenue received from the sale would not lower the deficit, because the reduction in debt would be offset by a reduction in assets. Similarly, under capital budgeting, government borrowing to finance the purchase of a capital good would not raise the deficit.

The major difficulty with capital budgeting is that it is hard to decide which government expenditures should count as capital expenditures. For example, should the highway system be counted as an asset of the government? If so, what is its value? Should spending on education be treated as expenditure on human capital? These difficult questions must be answered if the government is to adopt a capital budget.

Economists and policymakers disagree about whether the federal government should use capital budgeting. Opponents of capital budgeting argue that, although the system is superior in principle to the current system, it is too difficult to implement in practice. Proponents of capital budgeting argue that even an imperfect treatment of capital assets would be better than ignoring them altogether.

Measurement Problem 3: Uncounted Liabilities

Some economists argue that the measured budget deficit is misleading because it excludes some important government liabilities. For example, consider the Canada and Quebec Pension Plans. People pay some of their income into the system when young and expect to receive benefits when old. Perhaps accumulated future public pension benefits should be included in the government’s liabilities.

One might argue that pension liabilities are different from government debt because the government can change the laws determining pension benefits. Yet, in principle, the government could always choose not to repay all of its debt: the government honours its debt only because it chooses to do so. Promises to pay the holders of government debt may not be fundamentally different from promises to pay the future recipients the public pension system. In the mid-1990s, the unfunded debt of the Canada Pension Plan was just about the same size as the entire federal government debt that is usually reported, so this measurement issue is important. The government responded to this public pension problem by scheduling a series of increases in the associated payroll taxes (employer and employee contributions to the public pension) to take effect over a 10-year period. As a result, there is no actuarial underfunding problem remaining for our public pension. Some other countries have not taken similar bold action. As a result, you will continue to read about ongoing debates in both the United States and the United Kingdom about whether they should address their underfunded pensions in the same way (by increasing contributions) or in another way (by decreasing benefits or delaying the age that citizens can begin to collect their pensions).


A particularly difficult form of government liability to measure is the contingent liability—the liability that is due only if a specified event occurs. For example, the government guarantees many forms of private credit, such as student loans, mortgages for low- and moderate-income families, and deposits in banks and trust companies. If the borrower repays the loan, the government pays nothing; if the borrower defaults, the government makes the repayment. When the government provides this guarantee, it undertakes a liability contingent on the borrower’s default. Yet this contingent liability is not reflected in the budget deficit, in part because it is not clear what dollar value to attach to it.


Accounting for TARP

In 2008, many U.S. banks found themselves in serious trouble, and the federal government put substantial taxpayer funds into rescuing the financial system. A Case Study in Chapter 11 discusses the causes of this financial crisis, the ramifications, and the policy responses. But here we note one particular small side effect: it made measuring the federal government’s budget deficit more difficult.

As part of the financial rescue package, called the Troubled Assets Relief Program (TARP), the U.S. Treasury bought preferred stock in many banks. In essence, the plan worked as follows. The Treasury borrowed money, gave the money to the banks, and in exchange became a part owner of those banks. In the future, the banks were expected to pay the Treasury a preferred dividend (similar to interest) and eventually to repay the initial investment as well. When that repayment occurred, the Treasury would relinquish its ownership share in the banks.

The question then arose: how should the government’s accounting statements reflect these transactions? The U.S. Treasury under the Bush administration adopted the conventional view that these TARP expenditures should be counted as current expenses, like any other form of spending. Likewise, when the banks repaid the Treasury, these funds would be counted as revenue. Accounted for in this way, TARP caused a surge in the budget deficit when the funds were distributed to the banks, but it would lead to a smaller deficit in the future when repayments were received from the banks.


The Congressional Budget Office (CBO), however, took a different view. Because most of the TARP expenditures were expected to be repaid, the CBO thought it was wrong to record this expenditure like other forms of spending. Instead, the CBO believed “the equity investments for TARP should be recorded on a net present value basis adjusted for market risk, rather than on a cash basis as recorded thus far by the Treasury.” That is, for this particular program, the CBO adopted a form of capital budgeting. But it took into account the possibility that these investments would not pay off. In its estimation, every dollar spent on the TARP program cost the taxpayer only about 25 cents. If the actual cost turned out to be larger than the estimated 25 cents, the CBO would record those additional costs later; if the actual cost turned out to be less than projected, the CBO would later record a gain for the government. Because of these differences in accounting, while the TARP funds were being distributed, the budget deficit as estimated by the CBO was much smaller than the budget deficit as recorded by the U.S. Treasury.

When the Obama administration came into office, it adopted an accounting treatment more similar to the one used by the CBO, but with a larger estimate of the cost of TARP funds. The president’s first budget proposal stated, “Estimates of the value of the financial assets acquired by the Federal Government to date suggest that the Government will get back approximately two-thirds of the money spent purchasing such assets—so the net cost to the Government is roughly 33 cents on the dollar. These transactions are typically reflected in the budget at this net cost, since that budgetary approach best reflects their impact on the Government’s underlying fiscal position.” image

Measurement Problem 4: The Business Cycle

Many changes in the government’s budget deficit occur automatically in response to a fluctuating economy. For example, when the economy goes into a recession, incomes fall, so people pay less in personal income taxes. Profits fall, so corporations pay less in corporate profit taxes. More people become eligible for government assistance, such as welfare and employment insurance, so government spending rises. Even without any change in the laws governing taxation and spending, the budget deficit increases.

These automatic changes in the deficit are not errors in measurement, for the government truly borrows more when a recession depresses tax revenue and boosts government spending. But these changes do make it more difficult to use the deficit to monitor changes in fiscal policy. That is, the deficit can rise or fall either because the government has changed policy or because the economy has changed direction. For some purposes, it would be good to know which is occurring.

Many economists believe that government spending and tax rates should be set so that the budget would be balanced if real GDP was at the natural level. If this were accomplished, we would observe deficits during recessions (when unemployment is high and the government is making more transfer payments and collecting fewer tax dollars), and we would observe surpluses during booms (when employment and tax revenue are high and employment insurance and welfare payments are low). To assess fiscal policy, then, we need to know what the deficit would be if we were not undergoing a business cycle.


To solve this problem, the Department of Finance calculates what it calls the cyclically adjusted budget deficit—what the excess of spending over revenue would be if Canadian GDP were at it natural-rate value. We can now clarify how these data are used. For example, in 1994, when the federal government’s annual deficit was at its pre-2008–2009 biggest, the deficit was about $40 billion, while the cyclically adjusted deficit was estimated to be approximately $25 billion. According to these calculations, about $15 billion of the $40 billion total resulted from the fact that unemployment was so high in 1994. According to this approach, any attempt to push the deficit below $15 billion is regarded as inappropriate, since that part of the deficit was simply due to the state of the economy. It would vanish automatically when the economy returned to the natural rate. Efforts to eliminate it any earlier just prolong and deepen the recession. It is true that the national debt increases while we wait for this automatic elimination of that fraction of the deficit to take place. However, since there should be budget surpluses in the boom years, there should be no tendency for debt to grow over the longer run.

As the 1994 example makes clear, the cyclically adjusted deficit is a useful measure because it reflects policy changes but not the current stage of the business cycle.

Summing Up

Economists differ in the importance they place on these measurement problems. Some believe that the problems are so severe that the budget deficit as normally measured is almost meaningless. Most take these measurement problems seriously but still view the measured budget deficit as a useful indicator of fiscal policy.

The undisputed lesson is that to evaluate fully what fiscal policy is doing, economists and policymakers must look at more than just the measured budget deficit. And, in fact, they do. No economic statistic is perfect. Whenever we see a number reported in the media, we need to know what it is measuring and what it is leaving out. This is especially true for data on government debt and budget deficits.