September 4, 2018

The Congressional Budget Office (CBO) released a working paper last week summarizing all the available economic studies of whether or not federal highway funding supplements, or crowds out, state and local highway funding. Entitled “Fiscal Substitution of Investment for Highway Infrastructure,” the paper can be read here.

The consensus of economic studies appears to be that yes, additional federal highway funding will, to some extent, replace money that state and local governments would otherwise have spent on their own, but the estimates vary widely on just how large the displacement is. The CBO paper states that “Researchers have found evidence that state and local governments reduce spending on highways from their own funds as federal grants increase. For an increase of $1 in federal grants, most estimates suggest, state and local governments would reduce spending on highways from their own funds by between $0.20 and $0.80.”

The CBO paper lists the following studies supporting “fiscal substitution,” some of which date back quite a ways:

The fact that the two most recent studies reach opposite conclusions (Dupor vs Leduc and Wilson, both from 2017) is interesting. Both studies focus on the $27.5 billion in supplemental highway funding provided by the 2009 ARRA stimulus law. But Dupor only looks at the effects of the federal grants directed to capital funding and their effect on highway capital spending by state DOTs directly, while Leduc and Wilson look at capital but also at grants for highway operation and maintenance funding – and their effect not just on spending directly by state DOTs, but also on road grants by states to local governments. The lesson that there is more fiscal substitution on the capital side than on the O&M side seems clear and intuitive, but the effects that federal grants to state DOTs have on grants by state DOTs to local governments probably needs more study (and varies from state to state based on their differing governmental structures).

It has already been conclusively proven that ARRA’s highway funding to some extent displaced the regular federal-aid highway funding – outlays to state DOTs for the regular highway program dropped significantly during fiscal year 2010, the year in which most ARRA funding was outlaid. After all, the ARRA money had a 100 percent federal share (no state matching funds needed) and had a quick “use it or lose it” deadline, whereas most regular federal-aid highway money required a state matching share of up to 20 percent of project cost, and could be postponed until after the ARRA money had lapsed. (In its initial modeling, while Congress was still debating stimulus proposals, CBO got this right and OMB got it wrong.)

Section 1201 of the ARRA law had a “maintenance of effort” requirement for the DOT funding to state governments, requiring each state governor to sign a certification for how much the state planned to spend in its regular highway, transit, airport, etc. budgets through the end of FY 2010, and threatening a financial penalty if the state’s actual spending over the next 20 months failed to match that level. However, (a) state governors were free to fib on their certification, and (b) the only penalty for failing to maintain effort was to be locked out of the fiscal year 2011 August redistribution of highway obligation limitation. Six states (Alaska, Delaware, Idaho, Kansas, Maryland, and New York) were eventually penalized for failure to maintain their pre-ARRA spending levels as certified by their state governors in March 2009, but their only penalty was being blocked out of their share of the $1.2 billion August redistribution – New York, for example, may have lost $80-90 million in the August redistribution, but they were free to cut their state spending by far more than that.

Dupor’s research (in the 2017 study and others) posits a political theory that rings true with what we remember at the time ARRA was being drafted – that the main purpose of ARRA was not to build a lot of new infrastructure. The main purpose of ARRA was to prevent massive state and local government job layoffs. To the extent that preventing state and local employee (and downstream contractor) layoffs was the point, then the fiscal substitution of ARRA funds for regular federal funds, or state and local funds, was not the worst possible outcome.

The CBO paper cautions that the highway studies cited don’t necessarily translate to other modes of transportation. Their lessons would be most applicable to mass transit spending, where all of the projects are carried out by local transit agencies or state governments with federal aid that requires a standard non-federal matching share. They would be less applicable to aviation and water infrastructure spending, where the federal government spends a lot of money directly (instead of passing it through state and local governments), and the whole rail sector has so much non-governmental investment that it can’t be compared with government-dominated programs at all.