Infrastructure Spending Must Justify Itself
Infrastructure spending is worthwhile only if we need the infrastructure.
The debate over President Trump’s plan to bolster infrastructure spending is simmering on the backburner as we await specifics. But this debate is merely the latest chapter in a larger, economic one, which has persisted since the days of John Maynard Keynes, over how much government spending benefits the economy as a whole.
Our new research suggests that economic stimulus and jobs should never be the primary justification for government spending.
The economic argument for government spending on infrastructure projects, such as roads, bridges, and schools, should be about whether they provide us with more in benefits than in costs. So, for instance, if an old bridge has become unsafe and can no longer be repaired, then a new bridge should be built. But if it is more cost effective simply to repair the old bridge, then a new bridge should not be built. Similarly, a new school is only needed when refurbishing the existing one won’t satisfy kids’ educational needs. Every infrastructure project should be evaluated by this type of cost-benefit analysis.
Politicians and activists often attempt to sway the argument in favor of public spending by claiming an additional source of value: the jobs involved in constructing the project. But the truth of the matter is that, under normal circumstances, government spending creates essentially no net jobs.
At first glance, it might seem obvious that government spending does create jobs. After all, we can see the workers constructing the project. But a complete economic analysis must consider not only what we can easily see, but also what we cannot, for instance, a corresponding loss of jobs in the private sector. In an economy at or near full employment, the analysis is fairly straightforward. To pay for infrastructure projects, government must pull money out of the private sector through some combination of taxing and borrowing. With less money available, private spending falls, and, as a consequence, private employment declines. This effect mostly or entirely offsets the jobs supported by the government spending.
By contrast, if the economy is in a depressed state with high unemployment, as it was from 2009–2010, for instance, the economic analysis is more complicated and more controversial. In a depressed economy, most economists believegovernment spending can stimulate the economy and create jobs. This stimulus effect strengthens the argument for government spending because the public benefits from both the new infrastructure and the boost to the economy.
Many prominent economists have taken the idea a step further, however, arguingthat in a depressed economy, government spending is beneficial for the stimulus effect alone. In this view, getting the economy moving again takes priority, so any spending by the government — even if it’s entirely wasteful or results in no useful intrastructure improvements — is desirable.
During the Great Depression, Keynes argued that even digging and refilling holes would be beneficial. Accordingly, the federal government’s Works Progress Administration sometimes employed people in low-value activities such as picking up roadside litter. During the most recent economic crisis, Nobel Laureate Paul Krugman advocated spending by the government for any reason, including even an imaginary one such as a space alien hoax requiring “a massive buildup to counter the space alien threat.”
Our research comes to a different conclusion: Even in a depressed economy, the benefits of fiscal stimulus are not sufficient to justify indiscriminate spending.
Specifically, we account for the fact that government spending requires taxation, which lowers national income by reducing the incentive for people to work. Furthermore, even new jobs for previously unemployed workers require foregoing the value of alternative activities, such as working at home or in the underground economy.
After subtracting these costs, we find that stimulus spending is economically justifiable only if what it produces, such as infrastructure improvements, provides the public with at least 70 cents on the dollar in value. While this result allows for a relatively small amount of waste (no more than 30 cents per dollar, and only in a depressed economy), it clearly rules out make-work schemes such as those described by Keynes and Krugman.
Moreover, since a government-spending project still needs to deliver enough value to cover most of its cost, it must be justified primarily on its merits. A new bridge or school is either needed or it is not. Hence even during times of economic distress, defenses of government spending that appeal to “jobs” or “stimulus” are essentially a distraction; informed citizens would be wise to discount them.
The bottom line? Infrastructure spending is worthwhile only if we need the infrastructure. The administration and Congress would do well to keep that dictum in mind as they consider infrastructure spending.