Earlier this year, the White House proposed spending $200 billion to repair America’s crumbling roads, bridges and other infrastructure, which it said would further stimulate $1.5 trillion in additional investments from state and local governments and the private sector. The Penn Wharton Budget Model, however, found that the plan’s stimulus would be far less than expected and it also will have little impact on the GDP. The White House has publicly disputed this analysis.
In a follow-up of this divergence in views, D.J. Gribbin, former special assistant to the President for infrastructure, and Kent Smetters, Wharton professor of business and public policy who led the development of the Penn Wharton Budget Model, spoke with Knowledge at Wharton to defend their respective beliefs. (Listen to the podcast at the top of the page.)
An edited transcript of the conversation follows.
Knowledge at Wharton: Before we talk about the White House plan, could you briefly outline what people need to know in order to have a sensible conversation about infrastructure policy?
D.J. Gribbin: One of the things I found while working at the White House is that most people don’t have a very good understanding of the federal government’s role in the federal infrastructure policy. They tend to think that the federal highway system is owned by the federal government. It’s not. It’s owned by state governments.
When you’re having a conversation about infrastructure policy and the federal government’s role, it’s important to understand a few things. First, the federal government owns about 8% and funds about 14% of the infrastructure in this country. The rest is split relatively evenly between state and local governments and the private sector. Second, the federal government’s role is primarily to impose conditions on owners in conjunction with the funding. If you receive funding for highways, and you’re a state or local government, there are about 99 additional requirements that the federal government puts on those funds. While the federal government owns only about 8%, it regulates virtually 100% of major projects in the U.S.
Also, most people think that federal funds are free. But the federal government doesn’t actually create wealth. It just redistributes it. The federal government does impose costs when funds come from a community to the federal government and back. Those costs range from 15% to 25%. If you provide a dollar to the federal government, you’re going to get somewhere between 75 and 85 cents back. People don’t take that into account when they’re thinking about it. If you send the federal government a dollar and you get 75 cents back, the headline reads, “Community Receives 75 Cents from Federal Government.” They don’t look at the other side of the equation, which is, “Well, you just sent them a dollar.” One needs to understand all of those elements when you start talking about infrastructure policy and the federal government’s role.
Knowledge at Wharton: Kent, what do you think?
Kent Smetters: I completely agree with that. The big issue, of course, is that states and localities qualify for federal funding, and these can take different forms — closed-ended grants, loans, open-ended grants. The literature has done lots of work looking at the effectiveness of these grants in actually stimulating state and local budgets and whether it truly increases spending more than it would have, had it not been for that federal money. There’s a big fungibility issue.
Knowledge at Wharton: D.J., could you summarize the White House’s infrastructure plan for us?
Gribbin: First of all, what were we trying to accomplish? We were trying to speed project delivery and provide sustainable funding sources for infrastructure. Those were the two major goals. In addition, we wanted to do four other things which would eliminate artificial barriers created by federal preferences. Sometimes, a community may have a [particular] infrastructure need but the federal government gives them money of a different color and they’re required to spend it on something else. So how do we empower infrastructure owners in terms of environmental permitting project selections? How do we let communities have greater decision rights around their own infrastructure? How do we encourage innovation? We have this fantastic system that built the interstates in the 1950s and 1960s. It was so fantastic that we have kept that model until today. It’s been 62 years. Maybe we need to rethink how this model works. Finally, we want to increase efficiency and public trust.
“The challenge with the Wharton study is that it looked at [this plan] as if we were still in the old days of giving money out in terms of block grants … when in essence what we had done is structured this as a competition.”–D. J. Gribbin
I agree with the approach taken by Wharton and what Kent said, that there is a substitution effect with federal funding. Historically, and we saw this in the Recovery Act (of 2009), when the federal government gives a community money for infrastructure, they tend to substitute federal funds for the funds that they themselves were going to spend on the infrastructure, and then take the freed-up funds and spend them on other, non-infrastructure priorities.
So how do you create a system that avoids that problem? With this in mind, we were trying to create a program that requires an applicant to do several things. First of all, they need to create a new revenue stream. They don’t receive incentive funds for a project, they receive incentive funds for new revenue. Then, you needed to identify the project on which that revenue stream was going to be spent on. We had objective criteria — because this is a competition between applicants — on how we would score that application. Show us how you’re going to spend on the project, how you’re going to spend on maintenance. Are you being innovative in your procurement? Are you using technology? What is your social and economic return on investment? Those were the five criteria. Your score is multiplied by the non-federal share of funding in your project. It’s almost like a reverse auction dynamic. The less federal funding you asked for, the more likely you were to win the competition and get funding for your revenue stream.
The challenge with the Wharton study is that it looked at [this plan] as if we were still in the old days of giving money out, in terms of block grants for communities to make decisions [which] allowed that fungibility, when in essence what we had done is structured this as a competition. There is no block grant. When you show up to the party with your application, you may get zero. You don’t necessarily win anything. What we were hoping is that competition would force applicants to truly create new revenue streams and spend it on projects that they would not have done otherwise. So you have a net-positive effect on infrastructure funding overall.
Knowledge at Wharton: Kent, what were the findings at Wharton regarding this program?
Smetters: There were basically five steps that we did in this study. First, we decomposed the White House framework into eight categories. Five are spending categories, the way the grants and loans work, and so forth. Three are non-spending categories like expediting permitting, etc. Second, we looked at over 25 studies that analyzed past attempts of the federal government in trying to stimulate new investment that otherwise would have not occurred. And for each of the five spending categories, we did the mapping to figure out how much of the federal government spending actually leads to additional state and local spending. Third, and here the question really is, even if you believe that the government has some stimulating effect on state and local spending, whether it’s the marginal product of that additional spending. That is, what’s the value of public infrastructure, versus private, versus government debt and so forth?
The fourth step was to ask how the federal government will finance its $200 billion, because the framework doesn’t really say where it’s going to raise its $200 billion from. If it’s deficit financed, it’s a little more negative. If it’s using user taxes and lump sum taxes, that’s the opposite extreme, the most efficient. The fifth step was coming back to the three non-spending categories. The idea of the plan was to try to get permits to happen faster, and to get rid of some of the red tape. So we looked at past studies, time-to-build studies.
“The whole point of creating incentives is to try to realign people’s behavior to match with those incentives.”–D.J.Gribbin
The net effect was we found fairly small effects. We found that the White House was hoping to get — from the $200 billion — an entire additional spending of $1.5 trillion. The states and localities would essentially have a 750% multiplier on the federal government share. We estimate at the high end that the total spend would be about $230 billion dollars. When you think of that as over a period of 10 years, it’s not going to have a very big impact on the GDP.
Knowledge at Wharton: What are some of the reasons that you came to the conclusion that you did?
Smetters: The White House proposal is to spend $200 billion and hope that the state and localities will come up with another $1.3 trillion, for a total spend of this $1.5 trillion dollars. That’s a 750% multiplier on the government share. The best studies that we’ve found show that instead of a 750% multiplier, it’s about a 13% multiplier. That seems very counterintuitive. The idea is the federal government spends a dollar, and the net impact, including that dollar, is only 13 cents on that dollar, of total new spending. The question is, “Why does that happen?” There have been dozens of studies that have looked at it.
Here’s what the big issue is: How does the federal government distinguish, not just between new spending and not-new spending, or new revenue versus not-new revenue, but between new sources of revenue that would have been done even without the federal government’s help versus new revenue sources done only with the federal government’s help. Because states are doing new bond issuances, new revenue sources, all the time — even without the federal government help. The federal government has been aware of this issue for a while. The literature goes back over a half-century, trying to estimate this. And if anything, my reading of the literature is that even some of these low multipliers – 13%, maybe 20% — could actually be upward bias. In the past, D.J. has talked about this “coupon effect” — which I completely agree with. And that is, a lot of times states will actually hold back infrastructure spending, waiting for the government program to come along, so they can truly call this a “new project” or a “new revenue source.”
Knowledge at Wharton: D.J., the White House disagreed with the Penn Wharton Budget Model analysis. What were the main objections?
Gribbin: The White House disagreed strongly! When I worked as an investment banker, we would have analysts that would model transactions and benefits, and sometimes they’d get lost in the model. They didn’t pull back and look at the big picture. The Wharton study, in essence, concluded that incentives don’t work — which is a very odd conclusion for economists to make. What we were saying is, “Right now, if you’re a governor or a mayor, and you want to raise funds for infrastructure, that’s politically hard.” Plus you have this complicated effect where your taxpayers will tell you, where your constituents will tell you, “Listen, don’t raise my taxes. Get this free federal money and use that instead, so someone else pays for this infrastructure.”
So you’re walking into a political and economic environment that’s already distorted against raising new revenue. What we were trying to say is, “How do we break that dynamic?” Instead of saying, “Hey, we might give you money if you don’t do anything,” we were trying to flip that incentive structure and say, “Listen, the federal government will actually give you funds for raising new revenue. You can raise it however you want to. And you can spend it on whatever infrastructure you want to. We’re giving you lots of flexibility, lots of opportunities to access this new revenue.”
“Can the federal government really tease out what’s truly going to be not just new, but it’s going to be new and additional to what the states and localities would have otherwise done?”–Kent Smetters
The challenge is that the literature is focused on past efforts to increase infrastructure spending, which were matching grants or other forms of direct revenue from the federal government. And there is a strong substitution effect there. I don’t think there’s any literature on this idea, because this is a very new approach. We haven’t really done anything like this before. Will there be some gamesmanship? Yes. In any incentive program, you’re going to have a subset of people who would have engaged in that behavior, anyway. But the whole point of creating incentives is to try to realign people’s behavior to match with those incentives. So we were surprised that the conclusion of the Wharton paper, in essence, was that “Economic incentives for people to change their behavior won’t work in this field.”
Knowledge at Wharton: Kent, what do you think? Do incentives work?
Smetters: Absolutely. But I think people are very clever, and they can figure out how to get a dollar. What the past literature has shown is that open-ended grants really do create marginal effects, because if I spend more, I’m going to get additional money. … This plan does not have this open-ended approach to it, because it’s much more expensive to do it that way. It’s [called] “inframarginal.” And so if I’m going to give you a buck for you to do what you’re going to do anyway, it’s very challenging for the federal government to figure out if you’re going to do that or not. Because I can always say I wasn’t going to do that.
States have pretty much figured it out. It’s such a small spend relative to total infrastructure spending that it becomes easy to absorb it. Don’t get me wrong. I think the White House’s plan was very laudable in recognizing the problem. It’s not obvious to me that they were able to solve it, unless they go to this much more expensive, open-ended approach. It’s also true that the focus on the average doesn’t mean that all infrastructure spending is not going to be productive. Averages are good to focus on because it in part reflects the political process that we’ve seen in the past. But there are clearly projects — especially repairs projects on bridges, and so forth — that might have a fairly high ROI relative to other projects.
A lot of the White House criticism was focusing on technicalities, too. They said that we’re doing all this in what’s a closed economy, and that wasn’t correct. Or that we’re assuming a very low marginal productive capital, and that was not correct, either. But nonetheless, it really comes back to this issue: Can the federal government really tease out what’s truly going to be not just new, but it’s going to be new and additional to what the states and localities would have otherwise done?
Knowledge at Wharton: D.J., since you are our guest, we’ll give you the last word on this particular debate, before we move on to other things.
Gribbin: I would just say that I don’t think I disagree with anything in the Penn Wharton study. I just think it was applied to something the administration didn’t propose.
Knowledge at Wharton: Let’s turn to some broader questions about infrastructure. Kent, states seem to be strategic in how they approach federal cost share by withholding projects until a federal grant program comes along. And you referred to the “coupon effects” that D.J. has spoken about in the past. Do you agree with that? And what can be done about it?
Smetters: D.J. is absolutely right. He’s the first person I’ve heard use that term – “coupon effect.” I think it’s exactly right, where states will often withhold doing projects because they’re anticipating the federal government coming up with some type of matching money later on. It’s somewhat analogous to corporations holding so much money overseas hoping that the federal government would eventually change the tax code so that they could repatriate that money at a much lower tax rate. It’s a completely rational thing to do, from a company or from a state perspective. From a social perspective, it just creates a bigger substitution effect. It now becomes harder to distinguish what’s truly new and additional.
Knowledge at Wharton: What do you think, D.J.?
Gribbin: Yes, I agree. If we can get additional information or different benefits in the future, then it’s highly economic-rational for us to say, “Let me delay that decision.” Right now, we talked about the program having limited funding. The federal government opposes costs on that funding, so you net fewer dollars as a result of federal restrictions. But I think you need to add onto it this very real political dynamic, which is this coupon effect of, “I am incentivized as a leader — a political leader — to delay any increase in revenue.” Because state and local governments provide three to four times more funding for infrastructure than the federal government does, it can have this very distorting impact, where a federal dollar could, in essence, lock up three to four state local dollars — just the prospect of that hanging out there. So again, part of what we’re trying to do in our proposal is to flip that around and say, “Instead of engaging in strategic delay, act now and raise revenue, and we’ll incentivize you for that behavior.”
Knowledge at Wharton: Kent, what areas of infrastructure seem to produce the biggest ROI?
Smetters: It’s not the cool, new projects. Typically, the ROI is highest on the boring stuff, like repairing stuff. In Pennsylvania, for example, we have more than 400 bridges that will probably fail inspection in a couple of years. Could you imagine a critical bridge going out of service and all that transportation having to be re-routed? That’s not to say that there’s not going to be a decent ROI on new projects.
Knowledge at Wharton: Closely related to the big ROI is a project that has been called “the most important project in the nation” — the Gateway Project. D.J., could you help us understand why it is so significant? You’ve also mentioned that it would be a great case study on the challenges of federal funding.
Gribbin: Yes, I’d be happy to. Before that, I don’t want to miss the chance to violently agree with Kent and his assessment of ROI. The administration’s plan, in fact, takes this into account and says, “Your new project can be maintenance. You don’t need to be building something brand new.” But then the question becomes, “Who makes that ROI decision?” There are lots of views on this. Our general approach is the community in which the infrastructure exists is probably best positioned to make that decision.
The Gateway is a nice segue to take this theoretical conversation and bring it down to a live, practical project. It is enormously important. It’s a $30 billion series of projects in the Northeast Corridor — to expand trackage and train service, to double the Hudson Tunnel, to improve some bridges and other areas in the New York-New Jersey region. If you follow the debate, right now it is highly centered on the fact that the region, the community, wants the federal government to pay for half of that. And the federal government is saying, “We shouldn’t have to pay for half of that because while there is a federal component, it’s not 50% of the value.”
“The further removed you are from the actual taxpayers, the more likely you are to end up with a project that doesn’t make any sense to the community.”–D.J. Gribbin
The Gateway has been labeled the “most important project in America,” primarily by the New York press and the D.C. press. Interestingly, when I travel around the country, [I find that] the most important projects for people tend to be more local. But what makes the Gateway Project interesting is that it is massively important. Also, it builds a conversation around who should pay for which component when you have a project that has local benefits, state benefits, regional benefits, and national benefits.
We think the best answer is to figure out who is benefiting from it, and then if it’s affordable, let the beneficiaries pay for the share that they’re benefiting from. So if it’s primarily a local project, then it should be primarily locally funded. If it’s primarily federal, it should be primarily federally funded. But that is not the level of debate right now. As a result, this extremely important project is just stuck in a process that doesn’t have much hope of concluding any time soon.
Knowledge at Wharton: Kent, what do you think? What are the main lessons to be learned from the Gateway Project?
Smetters: I completely agree that Gateway is one of those projects that likely has a pretty high ROI, because there’s a big congestion in the Northeast Corridor, and relieving that congestion is a good thing. It’s also one of those projects that the federal government has a lot more control on, and things that the federal government directly owns, like Amtrak. There’s obviously a state and local component to tracks and things like that, but, you don’t have to worry as much about the substitution effect if they actually own those assets.
It’s also instructive, as D.J. mentioned. Seeing the congestion in the Northeast Corridor, you think about the wait time, and about the value of that economically, how much that wait time costs, and so forth. But then politically, you notice that Amtrak doesn’t just run in the very profitable Northeast Corridor. It runs throughout the country, because everybody wants their Amtrak, even if it’s not very profitable. And that’s been the real tough thing. That even if you can identify these high ROI projects, how do you get it through the political process?
Gribbin: I want to come back on the Amtrak issue, because this is a very important dynamic. When the federal government spends funds, it’s coming from another community. Right now, the estimate from the department of transportation — I haven’t studied these numbers independently — is that about 90% of the benefits are local. This means that 90% of the benefit of that system is getting people from New Jersey into New York for work and back. So when we say we want the federal government to help fix that, what we’re saying to the people in Columbus and Austin and Seattle is that we want to take money from you and send it to these New York commuters.
Knowledge at Wharton: Recently, I was watching this film, Fiddler on the Roof , where Tevye, the milkman, is singing a song about what he would do if he were a rich man. One of the lines in that song is that he would build a house with one staircase that goes just up, and another staircase that just comes down and another leading nowhere, just for show. The infrastructure equivalent of that is the Bridge to Nowhere. Kent how do we avoid that problem?
Smetters: People have talked about this a lot in the past — how do you stimulate new local and state spending but also avoid the bridge over the swampland? It’s challenging, because ultimately, Congress has to appropriate these funds. People have talked about having these commissions — these blue ribbon groups — to try to figure out what’s the high ROI. I’m all in favor of that, but we know what happens with those types of reports. They become part of a political process.
One criticism, I think, is valid. As D.J. has pointed out, so much of the benefit in the Northeast Corridor is for commuters trying to get into work every day. Tremendous value, but one reason why people are living in New Jersey is because zoning is so restrictive in New York. Why are we rewarding these cities for having such regressive zoning that pushes people out of the city? New York should be solving this problem with less regressive zoning.
Knowledge at Wharton: D.J., what do you think?
Gribbin: I agree that there’s a challenge with allocated spending. It tends to be politically allocated and not, say, market-allocated. So how do you overcome that obstacle? I think the answer to avoiding bridges to nowhere is to not have Congress pick projects. The challenge is, the further removed you are from the actual taxpayers, the more likely you are to end up with a project that doesn’t make any sense to the community. Another benefit to incentivizing state and local revenue is that the more removed the revenue is from the people who are using it and from the project itself, the more likely you are to have bridges to nowhere and a misallocation of resources.
Smetters: I agree with that. In fact, that’s what this idea of having commissions or having an agency pick the projects is all about. But ultimately, Congress has to do the appropriation. They cannot outsource that. What they could do is do some type of appropriation that’s very broad, then hand it over to some group or some agency. But in the past, at least, Congress has resisted because people want their stuff in their backyard.
Knowledge at Wharton: Or Congress could create an incentive program which encourages state and local governments to generate revenue that then is capped at the local and state level.
Gribbin: I think part of this is that we need to rethink the federal government’s role. I think the best way for us to have infrastructure is to send money to Washington and have it come back. And I would argue, actually, that’s still an effective way to do certain things. But for the vast bulk of our current infrastructure needs, which are primarily local, that’s an incredibly inefficient way.
Knowledge at Wharton: One of the most important infrastructure needs seems to concern air traffic systems in particular, which is a somewhat outdated relative to Europe. Kent, do you agree with that? And what can be done?
Smetters: I do. It’s the ability to move people effectively and fast. That is where you get tremendous ROI. [Former U.K. Prime Minister] Margaret Thatcher took the controversial position of privatizing lots of the infrastructure in the U.K. It was very controversial at the time. Not controversial anymore. For the most part, people say that worked. The U.K. system is not perfect. I still sometimes pull my hair out trying to get through the security line at Heathrow [airport], but nonetheless, it is much more effective in terms of moving people than what we see in the United States. It’s challenge having these conversations. Whenever you put the word “privatization” in front of something, you always get people just focusing on that word. It looks like a right-wing, conservative type of word.
“How do you stimulate new local and state spending but also avoid the bridge over the swampland? It’s challenging.”–Kent Smetters
But there are strategic uses of privatization. You can now get the localities to take much more ownership. They certainly own the airports already. But in terms of the air traffic systems, they can take much more control of that. On top of that, it’s not obvious as to why localities even own the airports, versus having a for-profit company that actually gets more money, the more people that they move. If you want to deal with market power, you can still put price caps and things like that on it that have a reasonable return, like we do with utilities. You can still create much more people movement in these systems — where it’s air traffic control, which is controlled by the federal government. In the local airports, which are controlled more locally, all that stuff could be probably incentivized much better to move people through more private effort.
Gribbin: Yes, again, I totally agree with what Kent said. One of the challenges we have around air traffic control is we’re asking the FAA (Federal Aviation Administration) to do something that they’re not well designed to accomplish, which is in essence, “We’d like you to run a technology business inside a government bureaucracy with all of the very important rules around procurement and everything else that comes with government operations.”
Smetters: D.J. is absolutely right [regarding] the federal government’s procurement and use of technology. Look no further than the Affordable Care Act. A billion dollars [has been] spent for a website that literally would cost less than a million dollars to build.
Knowledge at Wharton: When you look around the world, in which countries do you see the most innovation on infrastructure? And by innovation, I’m referring not just to the use of technology, but also to the way in which these projects are funded. What are some of the main lessons for the U.S. from these experiences?
Gribbin: Canada has done a remarkable job of providing a quasi-government service in a very efficient way. Australia has done a really good job of developing world-class infrastructure. I think the core of this is what Kent was mentioning earlier. It is governmental infrastructure with a heavy private sector component to it. I think part of the answer is that there are some things only the government can do, and some things only the federal government can do. And the government should be doing those things. But to the extent that we can get the private sector involved in truly non-governmental activities, that would be helpful.
When the federal government, or any government, builds infrastructure, its mission isn’t just to build infrastructure. Its mission is to build infrastructure, train employees, help disadvantaged businesses and purchase domestic products. That laundry list has lots of social benefits, but it also dramatically increases the cost and slows the timetable of the delivery of infrastructure. When we can strike a balance and give the private sector more responsibility for doing the commercial activities, you end up with a much more robust, lower cost, quicker delivery system than we have in the U.S. today.
And some of those things are not complicated. For example, when a private operator took over the Chicago Skyway project, a highway that comes from the [Indiana state line] into the city, which every Sunday had massive backups, they had a person in the tollbooth collecting tolls and another person in traffic collecting tolls — this is before automatic tolling. So they collected from two cars at a time. This reduced the backlog dramatically. There was no great technology being used. The difference was that now you had an operator that was incentivized to do that.