How Cities and States Can Stop the Incentive Madness

Recent years have seen a startling surge in the use of economic-development incentives by local and state governments. Amazon unleashed a fierce competition for its HQ2, with a number of contenders offering incentives worth billions of dollars (one package was estimated to be in the range of $8.5 billion). Before that, in 2017, Wisconsin handed more than $4 billion to the electronics manufacturer Foxconn. In 2016, Nevada gave Tesla more than $1 billion to build a battery factory, and two years earlier, Oregon gave Intel $2 billion for a new semiconductor chip plant.

Today, nearly all U.S. cities and states use financial incentives to attract companies, even though the bulk of research on the subject shows they are an ineffective waste of taxpayer money. In the first part of a two-part conversation with Timothy Bartik, we discuss the problem of incentives. Later this week, in the second part, we’ll talk about how effective place-based policies can combat regional inequality and help revive distressed places. Bartik is an economist at the W.E. Upjohn Institute for Employment Research and the author of a new book, Making Sense of Incentives: Taming Business Incentives to Promote Prosperity. Our conversation has been edited for length and clarity.

What exactly are business incentives, and how much do states and cities spend on incentives?

I consider incentives to be providing businesses with assistance that is customized to individual businesses, or at least particular types of businesses—a tax break that wouldn’t ordinarily be provided to all businesses, for example. The amount that I estimate is devoted to incentives is around $50 billion per year. Almost all of that is in tax incentives. Since 1990, I estimate, incentives have about tripled in the U.S. A lot of that occurred in the 1990s to the early 2000s, and [by] 2001 to 2015, things had roughly stabilized. Some states cut back on incentives; some states expanded incentives.

But the recent Amazon and Foxconn location competitions led to huge incentive offers, well beyond what most states typically have done. Foxconn was about 10 times as expensive per job as ordinary incentives. And some of the Amazon offers were also very high. So, a concern is that we’re going to see another rapid escalation of incentives in the next few years.

Do you think that Amazon HQ2 has changed the way that incentives are viewed by local and state-level actors?

There were some huge offers for Amazon. But then in fact Amazon ended up choosing Virginia and New York, which did not even make the biggest offers within [their respective] metro areas. Newark offered a lot more than New York City. Virginia’s offer was far less what Maryland offered, and it was certainly in line with Virginia’s prior policies. Virginia tends to be a low-incentive state.

It’s also hard to tell how this politically will play out because of the backlash to the New York offer. I think some of the backlash occurred because New York offered at least four times the amount per job that Virginia did, and people looked at that and said, “Wait a second.”

In Virginia’s case, a lot of the “incentives” were public-service enhancements: a new campus of Virginia Tech in Northern Virginia, improvements in public transit, enhancements in job skills. This kind of package makes more sense than just throwing a lot of cash at the company.

So, will states in the future imitate some of the huge offers to Amazon, or will they imitate what Virginia did? That’s the political choice faced by governors and state legislators.

What kinds of firms typically get incentives?

The evidence strongly suggests that larger firms are much more likely to get incentives, and get larger incentives. First, many state incentive programs have minimum size requirements, so that if you’re only creating 10 jobs, you get nothing, [but] if you create 500 jobs, you get a lot per job. So that type of incentive program inevitably favors large firms.

Second, large firms can get special legislation passed. The Foxconn deal was not something that was part of Wisconsin’s normal incentives. They passed special legislation to do it. They’re not going do that for one firm that is promising to create 50 jobs. This raises a key question: Does providing the largest incentives to the very largest firms increase the market power of these firms in a way that is damaging to our economy and society?

What are the main factors that attract firms to certain locations, and of those factors, where roughly do incentives rank among them?

The cost of labor and its productivity is clearly a more important factor than incentives or even overall state and local business taxes. That doesn’t mean incentives have zero effect, but obviously other cost factors are more important. The typical incentive package offsets about 30 percent of the state and local business taxes that the firm would otherwise pay.

In your book, you argue that business incentives are based on bad ideas and misleading claims.

Politicians and economic developers often claim that every firm they touch with an incentive dollar would never locate in their city and state but for the incentive. And they often claim that incentives are self-financing. But the empirical evidence strongly suggests that at least 75 percent of the time, the same jobs would’ve been located in the state and local economy anyway, so the incentive is then all cost and no benefit.

You devote a chapter to the concept of multiplier effects. Tell us more about this—why is it important in understanding business incentives?

People talk about multipliers as though they are magic. One of the rationales for trying to give incentives to a firm is that potentially there are multipliers that occur, because the firm will buy supplies from local suppliers, or because the firm’s workers will buy more things locally.

But when you attract firms and you add these multiplier effects, the resulting growth tends to drive up local wages, local land prices, and other local prices. So the area becomes a higher-cost area, and that’s going to discourage some job growth in other firms. Estimates I’ve done suggest that these kind of negative cost feedbacks reduce the multiplier by perhaps one-third.

And incentives don’t pay for themselves, as a rule. When you bring in jobs, you also bring in people, and you have to provide this added population with public services. Usually, the added public-service costs eat up at least 90 percent of any increased revenue.  

So if incentives don’t pay for themselves, where do you get the money to pay for the incentives? You have to increase taxes, or you have to cut public spending. Increased taxes reduce residents’ incomes. They don’t have as much to spend. And that’s going to decrease jobs in the area. If you cut public spending, that may directly lead to laying off public workers—which directly reduces jobs, and also reduces the spending power of those workers who were laid off.

Also, tax increases or spending cuts can have what economists call supply-side effects on the productivity of the local economy. If you cut spending on public schools, that’s going to have some negative effects on the quality of the local labor supply. The same could be true for cuts in university spending, cuts in public-health spending, cuts in infrastructure spending.

The point is that the money to pay for incentives is not coming out of thin air. State and local governments have to somehow pay for incentives, and that has some economic consequences. So you can’t simply look at the positive side of the multiplier effects of incentives and job creation, and ignore some of these negative effects. You’re going to drive up costs, which is going to drive away some jobs, and you’re also going to have to pay for the incentives, and that’s going to drive away some jobs.

How can policymakers and journalists better evaluate incentives and incentive packages?

I would suggest they ask some tough questions. They should ask: what happens if, instead of assuming 100 percent of the jobs are due to incentives, we only assume 10 or 20 percent are due to the incentives? How does the benefit-to-cost ratio change? They should ask whether the claimed multipliers account for the negative cost feedbacks that will tend to reduce multipliers.

They should be asking: to what extent are these jobs likely to go to local residents? If the local economy is already booming, if you don’t have a shortage of jobs, it seems strange to pursue a policy of just creating jobs for the hell of it. Especially in a booming economy, maybe you need to be more targeted in how you create jobs—what types they are, who gets them.

And you also need to look at how the jobs match up with the local workforce. If all the jobs seem unlikely to be the type that local workers—particularly local workers who are unemployed or underemployed—can access, you should raise some issues about who actually benefits from providing incentives to encourage job growth.

What would a model incentive package look like?

What I would do is target incentives [to] distressed places. Right now, that’s not the case. There’s no rhyme or reason in which state or local areas are most aggressive in offering incentives. If we’re trying to create jobs, distressed areas need the jobs the most. In fact, there’s empirical evidence that job creation has greater benefits in increasing employment-to-population ratios in distressed areas.

We also need to cut back on long-term incentives. Many incentive packages provide assistance for at least a decade, and some provide assistance for 15 years or 20 years. Yet firms are pretty myopic in making investment and location decisions. Long-term incentives are bad economically because they have less bang for the buck, in that they tip fewer location decisions per dollar. They’re bad politically because they’re so tempting for state and local political leaders. The current governor or mayor can get political credit now, but the fiscal cost falls on a future governor or mayor. I would restrict the term of incentives to just a few years.

By cutting back long-term incentives, you also cut the cost of incentives by quite a bit. I would limit incentives to be around $10,000 to $20,000 per job. If you restrict incentives to be only short-term, and if you target distressed areas, it would likely cut the overall cost by at least two-thirds, maybe more.

The resulting savings could be used to invest in services to small and medium-sized businesses, and to invest in local job skills, land development, and infrastructure, all of which has a higher bang for the buck for local economies. If you put together a package like that, you’re going to have more job creation per dollar, it’ll be more targeted at distressed areas, and more of the jobs will go to local residents. In contrast, what we’re doing with incentives now is just redistributing jobs among geographic areas according to random political factors.

Why can’t state and local policymakers do a better job of policing themselves and limiting wasteful incentives?

It’s in the interest of residents for state and local governments to restrain themselves. A state that was wise, a local government that was wise, would say: “Look, we’re spending a lot of resources on these incentive programs. They have a relatively low bang for the buck, they have high opportunity costs, and we’re not getting the results we want. We’re going to restrict these incentives more to distressed areas. And we’re going to focus more on providing various programs that are more effective, such as services to smaller businesses, and investments in skills development and land development.”

These investments would be far more cost effective in promoting local prosperity. That would be a much more sensible strategy than just handing out lots of tax breaks and other cash to firms. Unfortunately, handing out tax breaks is easy and provides immediate political benefits, while postponing costs to the next governor or mayor. So politics outweighs what would be in the best interests of a state’s residents.

What should the federal government do to rein in incentives?

The federal government could prohibit incentives, or at least prohibit the most excessive incentives or the most poorly targeted incentives, or put an extra tax on those. It could limit incentives that go to the very largest firms, with more than 10,000 employees. I have some doubts about whether the federal government is likely to have the political will to ever take such actions.