Are the Costs of Government on “Autopilot”?

Taxes and the costs of complying with regulation are two of the larger and more noticeable ways that private individuals pay for government services. Yet it may surprise most people to learn that a significant portion of the federal government’s expenditures and indirect costs to the US economy occur each year on “autopilot” without any action by the current Congress. These autopilot costs are the result of past legislation, interest payments, and rules created by government agencies, all of which bypass the annual appropriations process that exists to ensure the accountability of our elected officials.

Some federal government costs are included in the yearly budget. For example, discretionary expenditures—those appropriated by annual congressional vote—are budgeted at $1.15 trillion for fiscal year (FY) 2015. According to a yearly report by Susan Dudley of George Washington University and Melinda Warren of Washington University in St. Louis, MO, $62 billion of that $1.15 trillion will flow to regulatory agencies. While $62 billion is a substantial figure, it is relatively small in terms of overall government spending. The sum of regulatory costs to the economy, however, extends far beyond the salaries and spending at the agencies themselves.

The remaining costs of regulations are much more difficult to calculate than those plainly stated in the federal budget. These are the costs of complying with regulations that are borne by consumers and producers, and those costs are, in turn, paid for through higher prices, lower wages, and reduced innovation. An oft-cited report by Clyde Wayne Crews Jr. of the Competitive Enterprise Institute estimates that these costs will be approximately $1.88 trillion in 2015.

Other organizations such as the National Association of Manufacturers (NAM) and the Office of Management and Budget (OMB) offer estimates that help highlight the wide range and uncertainty in this calculation. NAM estimates these costs were $2.03 trillion in 2012, while OMB calculates that the range of annual costs from 2001 to 2013 was from $74.3 to $110.5 billion (all amounts adjusted to 2015 dollars by the authors). However, the OMB estimation only includes the costs of 116 of the 569 economically significant regulations (those with an impact of $100 million in at least one year) and none of the 36,453 regulations that do not qualify as economically significant during this period. The authors acknowledge that, because of this exclusion, “the total benefits and costs of all Federal rules now in effect are likely to be significantly larger than the sum of the benefits and costs reported.” Yet even the lower-range value from the OMB, when compared to Dudley and Warren’s figure, shows that the annual appropriations by Congress only reflect 45 percent, or $62 billion, of the total costs of regulations, $136.3 billion. The higher-range estimates from Crews and NAM indicate that this figure may be as low as 3 percent.

To get a full picture of the costs of government, we must also take into account the expenditures that are a part of the budget but are not appropriated each year by congressional votes. The three major entitlement programs—Social Security, Medicare, and Medicaid—are estimated at $1.75 trillion for the year. These programs, along with another $628 billion in other mandatory spending and $229 billion in net interest, account for the remainder of the autopilot costs. Combining the estimates of regulatory costs and the numbers from the federal budget, we can produce estimations of the total cost of the federal government in 2015. 

Then we can separate the costs that are appropriated by Congress as part of the annual budget process from those that are not to see how much is on autopilot. The appropriated costs are the discretionary expenditures, including budgets for regulatory agencies. Those remaining—non-budgeted costs of regulation, Social Security, Medicare, and Medicaid, other mandatory programs, and net interest—give us the total “autopilot” costs. Using the Crews estimate, the data reveal that the current Congress is voting on just 20 percent of the amount that we pay for their services, leaving 80 percent—an astounding $4.5 trillion—as autopilot costs in 2015. But even using the lower-bound estimate of $74.3 billion for the non-budgeted costs of regulation would only decrease the autopilot costs to 70 percent of the total.

This recurring process is not inescapable. Our Mercatus Center colleagues Jason Fichtner and Patrick McLaughlin have proposed a method of more accurately estimating costs and including them in the annual budget process called legislative impact accounting. This method would:

incorporate economic analyses of legislation and regulation into the budget process in two ways: First, when new legislation is proposed, an independent office—perhaps the Congressional Budget Office—would produce an estimate of the economic costs the legislation would create. Importantly, a legislative impact assessment would attempt to consider economic costs of proposed legislation, not just budgetary outlays. . . . Second, legislative impact accounting would require retrospective analyses of the economic effects of legislation, starting five years after the legislation passed. The idea is to learn what the real effects have been, and to then update the original estimates produced in the first stage. This would effectively create a much-needed feedback loop that communicates information about the economic effects of legislation back to Congress.

Such a process would help inform Congress and the public about the hidden costs that accompany legislation and regulation, and it would help incorporate this information into the annual appropriations process.

Indeed, any effort to account for these autopilot costs would help promote transparency and accountability within the federal government.