I’ve noticed many economists from Wall Street and elsewhere on cable news shows and other media claiming that more federal infrastructure spending would lift the economy.

The Wall Street Journal reported, “Goldman Sachs economists estimate that the substantial stimulus that would come with a Democratic sweep, along with longer-term spending increases on areas such as infrastructure and health care, would provide the economy with a boost.” Meanwhile, Penn-Wharton economists assume that added federal infrastructure spending under a Biden administration would be highly productive and boost GDP.

I think these economists are missing something fundamental. The federal government owns very little infrastructure. State and local governments and the private sector own 97 percent of America’s nondefense infrastructure. Just looking at public nondefense infrastructure, state and local governments own 87 percent and the federal government owns 13 percent. State and local governments, for example, own the entire interstate highway system and virtually all the nation’s commercial airports.

If additional spending on these public assets was high value, wouldn’t state and local governments increase spending themselves? If the benefits of expanding highways, airports, and urban transit systems outweighed the costs, wouldn’t the owners of these assets be in the best position to spot the opportunities and act on them?

Economists and pundits who claim federal action is needed seem to assume that federal policymakers have superior knowledge or better foresight than state and local officials and infrastructure managers. They seem to assume that the D.C. legislative process produces more efficient outcomes than state and local processes in allocating investment dollars and balancing the benefits of spending with the costs.

Perhaps the view is that the states can’t raise enough money for needed infrastructure. But that is not correct. The states have huge fiscal power to fund high-value projects by taxing, borrowing, or reprioritizing existing spending. If a $1 billion highway project in Texas was expected to generate $1.5 billion in benefits, then Texas highway planners would be the ones to recognize it and the Texas legislature could fund it or attract private financing to build it.

The Penn-Wharton modelers find that Biden’s plan to spend $1.6 trillion more over a decade on infrastructure would boost GDP substantially. Based on a prior study, they assume, “… more public capital investment raises the productivity of private capital and labor. Based on past experience, our analysis assumes that an additional dollar of public infrastructure generates over 10 cents of output per year, everything else equal. In contrast, we estimate that an additional dollar of private capital generates around six cents of output per year.”

Thus, spending another $1 billion on a government high-speed rail line in California will add more to the economy than, say, Intel Corporation building a $1 billion chip plant in the state. That seems unlikely. But if it were true, then the government of California or entrepreneurs in the state would be in the best place to recognize it and fund such high-value opportunities.

For every type of infrastructure, there is an optimal amount of spending. If governments spend too much, the marginal benefits will be less than the marginal costs, which would undermine the economy. Infrastructure advocates and economists such as the Penn-Wharton modelers seem to assume that we are always below the optimum, that more spending is always better. Supposedly, state and local governments across the nation routinely fail to fund marginal investments that have positive net benefits, and the federal government can swoop in and efficiently fill the void.

Trump and Biden want the federal government to spend hundreds of billions, or trillions, more dollars on infrastructure, and many economists in the media cheerlead for it. Economic models assume it will be greatly beneficial. But at the margin, why would it be?

State governments show little hesitation to increase funding of their infrastructure when they see opportunities. Just since 2013, 31 states have raised their gas taxes to fund transportation. Even if there were gaps in state funding, experience shows that federal aid would not fill them efficiently because of red tape, high costs, and pork barrel problems. This study describes why federal spending on state activities is less efficient that state spending on state activities. Greater federal intervention would undermine the efficiency of America’s infrastructure investment, not improve it.

Whichever candidate wins the presidential election is likely to push for higher federal infrastructure spending next year. I’d like to see reporters ask federal policymakers why they think state and local governments can’t fund their own assets. And I’d like to see economists re-think their assumption that Washington is better able to determine optimal infrastructure investment than officials in hundreds of state and city governments that own and manage the assets.

For more on America’s infrastructure, see here. To understand why the federal government fails at much of what it does, see here. To appreciate the disadvantages of federal aid for state activities such as infrastructure, see here. For studies on transportation infrastructure, see here.