The Costs of Government
By Robert P. Murphy
“Trying to calculate the waste of government by analyzing total expenditures is a poor method. A large fraction of government spending consists of transfers.”
One method for generating a ballpark estimate is to calculate the fraction of the economy absorbed by government at all levels. In the 4th quarter of 2009, federal, state, and local government spending accounted for 35 percent of U.S. GDP.1 Now, suppose that an extreme critic of the government thought that, on net, it provides no services—meaning that whatever good things the government does (locking up murderers, teaching most kids to read, etc.) is counterbalanced by the bad things it does (locking up speculators who promote market efficiency, teaching most kids that the New Deal ended the Great Depression, etc). If that critic is right, then getting rid of the government altogether would return that 35 percent of GDP to the taxpayers, immediately increasing the real income of the average American by about 54 percent.2
However, trying to calculate the waste of government by analyzing total expenditures is a poor method. A large fraction of government spending consists of transfers, which do not destroy the actual resources being redistributed. For example, retired individuals may have railed against “Big Government” during their peak earning years, but now many of them would lose out if Social Security and Medicare spending were cut, even if the rollback went hand-in-hand with massive tax cuts. Similarly, many poor residents in urban neighborhoods might not gain from the “services” of local vice squads and the CIA, but they might gain from free (to them) public schooling for their children and income-assistance programs.
Because transfer payments create obvious winners and losers, this article does not examine overall government budgets. Rather, it looks at specific government functions and comes up with quick estimates for how much money is wasted, relative to the private sector providing the same services. We also consider some of the major categories of “deadweight loss” associated with government policies.
Consumption and Investment Spending at Various Government Levels
To avoid the problem of simple transfers, we can focus on government purchases of goods and services. But because government typically has higher expenses to produce an equivalent output, we can’t treat government purchases as equivalent to, say, a private corporation’s investment spending.
To provide some empirical mooring, we can look at input versus output in K-12 schooling. Andrew Coulson recently estimated that the Washington, D.C. public schools spend an average of $24,606 per pupil annually. In contrast, D.C.’s private schools—which include some of the most elite institutions in the country—spend only $14,534.3
Despite their lower funding, private-school students consistently outperform their public-school peers. However, much of the disparity disappears once differences in student background are taken into account.4 Assume for the sake of argument that public and private schools do a comparable job educating equivalent students.
Superficially, these results suggest that even for a local task, subject to immediate feedback from voters (and parents), the government consumes almost 70 percent more resources to produce the same output as the private sector. Yet, because Coulson’s figures include some federal funds and because the effect we are measuring might be exaggerated near D.C., let us conservatively say that for standard services that could be handled just as well by the private sector, local governments spend 50 percent more than the private sector to do an equivalent job.
We have reason to believe that this estimate should be much higher for state and, especially, federal activities. For one thing, Charles Tiebout suggests that as political jurisdictions grow larger, the political class becomes less responsive to the voters and caters more to special interests.5 Surely a compilation of wasteful federal uses of tax dollars would dwarf a compilation of state or local abuses.
Besides weaker accountability to the voters, higher levels of government tend to waste more resources per dollar of expenditure because of the mix of programs offered by the federal versus state and local levels. When local officials waste money, they might, say, subsidize a convention center or sports stadium and use downtown real estate in a suboptimal way. But when federal officials waste money, they might spend millions on antitrust prosecutions of large firms that have kept prices low and, thus, made competition too intense for their competitors. There is no analog in such cases for private provision because it’s not a “service” at all.6 It’s true that local governments provide dubious “services” just as federal politicians do, but we assume they constitute a smaller proportion of local activities, leading us to choose a lower estimate of the overall waste per dollar of purchases.
With these considerations in mind, let’s assume that local governments use 50 percent more resources than necessary to fund their legitimate services, that state governments use 75 percent more, and that the federal government spends twice as much (100 percent) on its legitimate services, compared to a private counterpart. What this implies, using simple algebra, is that 33 percent of local government payroll, 43 percent of state government payroll, and fully 50 percent of federal civilian payroll is wasteful.
According to the Bureau of Economic Analysis, annualized federal nondefense purchases were $377 billion in the 4th quarter of 2009. State and local governments spent $1,786 billion purchasing goods and services during 2009.7 If we apply the 50- percent discount to the federal figure, and a (weighted) 35-percent figure to the state and local expenditures, we come up with an estimated ($377bn × 50% + $1786bn × 35%) = $813 billion in wasted resources.
Beyond the explicit expenditures, government involvement in schooling has another large cost: By keeping millions of students cooped up in classrooms who would otherwise be working for pay, mandatory attendance laws and subsidies for post-secondary schooling suppress economic activity.
Of course, there is a potential tradeoff in which attendance laws and student aid reduce potential output from young workers as they stay longer in school, but lead to higher output after they have graduated with better skills. However, it is not obvious that the massive government support for schooling actually yields a more-educated population, particularly as the education pertains to on-the-job skills. Indeed, a leading economic model interprets college and advanced degrees as signals to the employer of the aptitude of a particular student.8 Many businesses will pay a high starting salary to a 4.0 history major from Harvard, not because they need someone with a thorough knowledge of medieval Europe, but because the type of student who achieves this will probably succeed in many job environments.
To the extent that this “signaling game” characterizes formal education, schooling is like an arms race between students. If they could collectively agree to spend less time in school—with far fewer people going on to grad school and with fewer people taking it for granted that “everybody should go to college”—then employers’ expectations would adjust. Students, on average, wouldn’t be hurt in their career prospects and would gain by having the option to start their careers earlier.
Things would be better for the students who remained in school, too. As a former college professor, I would have much preferred preparing to teach, say, the top half of my students in each class, to having to cater to the ones who always asked, “Will this be on the exam?”9
Let us arbitrarily assume that without government encouragement, enrollment in college and graduate programs would fall by half, and that this shift would yield a citizenry (on average) just as educated. With these assumptions, the increase in output from half of the current enrolled students entering the workforce would be a pure gain.
In 1999, there were 12.6 million Americans enrolled in undergraduate degree-granting institutions. If we assume that half of these young people would get jobs earning $20,000 on average, were it not for government support, we can add another $126 billion to our tally. If we assume that half of the 1.8 million enrolled in graduate programs could get jobs earning $35,000 on average, we can add another $32 billion.10
One of the biggest components of federal spending is the military budget. Including supplemental funding for the operations in Iraq and Afghanistan, total military spending for Fiscal Year 2010 will probably be some $720 billion.11
It is difficult to evaluate this aspect of government activity because there is no obvious private-sector analog to the operations of the Department of Defense. However, consider this: Despite the hundreds of billions spent over the years to influence foreign events and establish a worldwide intelligence network, 19 men with boxcutters were somehow able to bring down the World Trade Center and attack the Pentagon on September 11, 2001.
For the sake of argument, let’s suppose that a “Fortress America” defense, in which the United States did not try to police the world but, instead, concentrated a smaller force at home to deter explicit attacks, could provide the same general level of security. And we’ll arbitrarily pick $150 billion as the annual expense of maintaining this deterrent. Note that even by spending this smaller amount, the U.S. would still lead the world in military spending, and many of the other contenders are explicit U.S. allies. With the U.S. already starting with its stockpile of advanced weaponry and trained forces, and still spending more than any other country in the world on its military, it is difficult to see what power would cross the oceans and credibly threaten an invasion, especially considering how many private citizens would rush to join militias and the National Guard in the months before such a large-scale attack.12
If, indeed, a “Fortress America” strategy costing $150 billion annually could provide American citizens with comparable protection from foreign attack, this means that the military currently uses some $570 billion per year that could be returned to the private sector.
Tax Compliance and Disincentives
Beyond the actual income extracted from the private sector, the tax codes make the country poorer because of explicit compliance costs (hiring accountants, keeping separate books, etc.) and because of the disincentives to produce additional income. In a December 2005 study, the Tax Foundation projected that 2009 tax compliance costs—focusing just on the burden of households and businesses filling out tax forms—would be $347 billion.13 In light of the massive income tax cuts that we discuss below, we’ll chalk $300 billion up to the tax compliance costs of Big Government.
In addition to the resources wasted on tax compliance, we also must consider output not produced in the first place, because of incentives. Mankiw and Weinzierl (2004) concluded that 17 percent of a tax cut on labor income and 50 percent of a cut on capital taxes are eventually recouped through higher economic growth.14 Because most taxes fall on labor rather than capital, we will assume that any tax cut will yield new income (through greater economic activity) equal to 25 percent (a factor weighed toward 17 percent rather than 50 percent) of the size of the tax cut.
Total federal tax receipts in 2009 were $2.1 trillion.15 If the federal government restricted itself to providing defense (“Fortress America”), making sure the president had a sharp wardrobe, and a few other bare-bones functions, it would have spent only, say, $200 billion in 2009, allowing for an immediate tax cut of $1.9 trillion. State and local governments collected $1.9 trillion in receipts in 2008.16 If we assume half of current state and local spending could be eliminated by a return to the traditional role of government, that would allow an additional tax cut of almost $1 trillion.
Scaling back Big Government would therefore allow an immediate tax cut of $2.9 trillion. With our weighted 25 percent factor, that means the economy would be an additional ($2.9tn × 25%) = $725 billion richer, if businesses and individuals didn’t have to forfeit such a large share of their output to the government.
|Civilian Purchases Waste, All Levels||$813 billion|
|Undergraduate Over-Enrollment||$126 billion|
|Graduate Over-Enrollment||$32 billion|
|Excess Military Spending||$570 billion|
|Federal Income Tax Compliance||$300 billion|
|Deadweight Tax Loss, All Levels||$725 billion|
|Total Wasted Resources:||$2.6 trillion|
Government artificially restricts economic activity in hundreds of ways. These include the restrictions the FDA imposes on new drugs and “alternative” medicine; the hundreds of thousands of drug offenders (who would otherwise be productive citizens) living off the taxpayer; the tremendous macroeconomic uncertainty caused by poor Federal Reserve policies over the decades; and, finally, the moral hazard engendered by the trillions of dollars in direct and indirect bailouts of the financial sector since 2008. This article has considered none of these additional sources of loss. Doing so would only add to the already high estimate of the gain from reducing the size of government.
Despite this article’s narrow focus, it has, nonetheless, documented a $2.6 trillion increase in annual economic output if the government were to restrict itself to its classic role of protecting the country’s residents from criminals and invaders. Such an increase would have meant an 18 percent increase in GDP in 2009.17 It also works out to an average of about $8,500 per man, woman, and child in America, per year, starting almost immediately after a hypothetical rollback of Big Government. The only long-term component in our calculations was the disincentive effect from high tax rates. The other categories would yield their savings almost immediately, just as soon as businesses could adjust to the new influx of workers (previous students) and the influx of cash in consumers’ pockets.
Total government spending was $5.1 trillion, according to Table B-83 (p. 428) of the 2010 Economic Report of the President [.pdf]. Total GDP for the quarter was $14.5 trillion (Table B-26, p. 360).
If the government consumes 35 percent of GDP, then the private sector retains 65 percent. Returning that 35 percent to the citizens would thus raise their real income by (35 / 65) = 54 percent.
See the U.S. Department of Education’s 2006 report [.pdf], “Comparing Private Schools and Public Schools Using Hierarchical Linear Modeling.”
For the BEA figures, see the FRED chart on federal nondefense consumption and investment expenditures, and the FRED chart on state and local expenditures. Federal Reserve Bank of St. Louis.
A good discussion on this topic is The Chronicle of Higher Education, “Are Too Many Students Going to College?”, November 8, 2009.
For the estimates of undergraduate and graduate enrollment, see the Tables 188 and 189 of the Department of Education’s “Digest of Education Statistics 2001” [.pdf].
The Wikipedia entry on the Military budget of the United States explains the various components, though we do not include their category of “defense-related expenditures outside of the [Department of Defense]” because it includes spending on Homeland Security and interest on the debt due to previous military expenditures.
For a ranking of countries by military spending (as of 2008), see the Wikipedia entry, List of countries by military expenditures.
See Scott A. Hodge, J. Scott Moody, and Wendy P. Warcholik, “The Rising Cost of Complying With the Federal Income Tax,”Tax Foundation Special Report No. 138 (December 2005).
See N. Gregory Mankiw and Matthew Weinzierl, “Dynamic Scoring: A Back-of-the-Envelope Guide,” National Bureau of Economic Research, NBER Working Paper No. 11000, December 2004.
Federal receipts broken down by category at available at Wikipedia: United States federal budget: Major receipt categories.
See the U.S. Census Table 418, “State and Local Government Current Receipts and Expenditures” [.pdf].
This calculation actually is conservative, because it takes the official U.S. GDP figures at face value. But if it’s true that a sizable fraction of government consumption and investment expenditures are wasteful (as argued earlier in the article), then the $2.6 trillion in calculated savings should be divided by a smaller GDP figure, yielding an increase in GDP greater than the reported 18 percent figure in the text.
For more articles by Robert P. Murphy, see the Archive.