What standards make sense for economic development tax incentives?

Woman points at building in a development model while others look on

By Timothy J. Bartik

Economic development tax incentives are often controversial. By “economic development tax incentives,” I mean providing some sort of cash tax break, compared to normal taxation, for a business that is locating or expanding in a state. The controversy arises in part because the immediate beneficiary of the tax break is the firm. But the ultimate beneficiaries may include state residents if the tax incentive successfully creates new jobs, and if those new jobs help state residents by boosting their employment rates and real wages.

As I have reviewed in much prior research, economic development tax incentives can pay off if properly designed and in the right local circumstances. The devil is in the details. For example, incentives in distressed local labor markets, financed without undermining productive public services such as education, for export-base businesses with high multiplier spillovers for other jobs, can have high benefit-cost ratios, even if the tax incentives per incented job are large. But if these conditions are not met—incentives are in a booming local labor market, incentives lead to downward pressure on state education spending, multipliers are low—then incentive benefits may be much lower than costs.

One incentive design issue is whether incentives can have their benefit-cost ratios improved by various standards. Frequently discussed standards include higher wage standards, more local hiring standards, and industry targets. Less frequently discussed standards include targeting distressed local areas and making sure incentives are paid for by tax increases. This memo provides a very brief discussion of these design issues. Fuller discussion can be found in my book on incentives, in the various papers discussing my benefit-cost model of incentives, or in my recent paper on the Ford Soar project in Marshall, Michigan.

Higher wage standards

Wage standards are sometimes set for a firm’s eligibility for incentives or the magnitude received of incentives. How should such standards be set?

Ideally, such standards should be set based on the wages paid by the firm for particular jobs relative to the educational credentials and other credentials that these jobs required. For example, Amazon HQ2 supposedly offered jobs paying $150,000 per year—a very high wage. But since these jobs also had high credential requirements, it is not clear that they offered a “wage premium,” which is the key. A firm that paid far less than Amazon per job but was willing to hire workers with only a high school diploma might offer a higher wage premium.

To my knowledge, no state’s wage standards adjust for credentials required by the firm. This could be readily done. A state could calculate what the median wage rate per hour is in various local labor markets in the state for workers with different educational credentials, from a variety of federal statistical sources. And then the firm’s wage rates for jobs with different educational credentials could be compared with those standards.

Simply setting wage standards by some wage relative to the area median wage runs into the problem that it may target incentives to high-paying firms with very high credential requirements, and penalize firms with lower wages but with a greater likelihood of hiring workers with lower educational credentials.

Sometimes proposals have been made to tie wage standards to median household incomes. This has some additional problems. In particular, since many households have more than one worker, this standard may penalize jobs that may in fact offer good wages relative to credentials required for the many households with more than one worker.

Hiring standards

Sometimes proposals are made to tie incentives to local hiring, so that state residents benefit more from the local jobs. Some proposals go further to tie incentives to hiring the local non-employed. In other cases, particularly in states that are suffering from slow population growth or population decline, the idea has been floated of tying incentives to bringing in new residents.

To discuss hiring standards, we need to understand how job creation due to incentives may affect local employment rates and population. First, if the incented jobs are induced, in some “export-base industry” (a firm that mainly sells its goods and services outside the state), the new jobs in the incented firm will have some multiplier effects on other state jobs. Some suppliers to the incented firm will add jobs. Furthermore, the added workers at the incented firm and its suppliers will buy more goods and services in the local economy, creating local retail jobs.

How do these induced jobs, both the incented induced jobs and the resulting multiplier jobs, lead to higher employment rates and population? The new hiring in the induced incented jobs and multiplier jobs comes from three sources: 1) already-employed state residents, 2) non-employed state residents, 3) in-migrants to the state. But the first source, already-employed state residents, creates a job vacancy, which is filled in the same three ways. The resulting job vacancy chains are only terminated when the initial new jobs—both incented and multiplier jobs—ultimately result in additional jobs for either the state’s non-employed or additional in-migrants. There is no other alternative.

But note that the extent to which jobs go to the state’s non-employed or in-migrants depends on far more than who is hired by the incented firm. The proportion of jobs for the non-employed versus in-migrants also depends on hiring patterns for the multiplier jobs and hiring patterns for every employer along the various job vacancy chains.

Having said that, it is still the case that who the incented firm hires matters to the ultimate impact. The incented jobs will be a sizable share of the initial jobs created, although multiplier jobs also matter. If more of the incented jobs are filled by hiring the local non-employed, this will tend to raise the ultimate effects of the project in helping the state’s non-employed. On the other hand, if more of the incented jobs are filled by bringing in workers from other states, this will tend to raise the ultimate effects of the project in increasing state population.

Which is more valuable, more jobs for the non-employed or inducing in-migration? In most circumstances, increasing jobs for the non-employed is much more likely to increase benefits relative to costs. More jobs for the non-employed reduces the social costs of non-employment (e.g., substance abuse, crime, family break-ups, etc.) and reduces the need for welfare payments. More jobs for the non-employed also increases tax revenues without much increase in public service requirements.

On the other hand, inducing in-migration does not reduce social problems for state residents. Furthermore, although the new workers and residents will increase state and local tax revenues, these new residents will also require more public spending to maintain the quality of public services. Teachers will need to be hired to keep class sizes from exploding; police and fire personnel will need to be hired to keep response times reasonable; roads and other infrastructure will require costly investments to accommodate population growth. The resulting public service costs in many cases will outweigh the revenue gains, resulting in a net fiscal loss.

In deciding on hiring standards, one must further consider that requiring firms to hire the local non-employed may in many cases be perceived by the firm as a costly unfunded mandate. Sometimes a state or local area with many new job prospects can succeed in insisting on such local hiring, but this is a difficult position to maintain for any state or local economy that perceives itself to be short of jobs.

One alternative to a hiring standard mandate is to encourage firms to engage in cooperative arrangements with local job training agencies. This cooperation will often lead to firms hiring workers who otherwise would be unemployed or underemployed. Such cooperation could be encouraged by giving firms extra “points” on incentive applications if they are willing to cooperate with local job training agencies. Alternatively, states can provide funds to firms for customized job training programs that are managed by local job training agencies.

Furthermore, if a state incentivizes a new firm that is “large” relative to the local labor market, or with potentially large multiplier effects relative to the local labor market, local hiring of the non-employed might be encouraged by providing supplemental funding to local job training agencies. Policymakers should want to provide training and job placement help to encourage more diverse hires along all the job vacancy chains in the local economy, not just at the incented firm.

Industry standards

Which industries should be targeted? First, targeted firms should be in export-base industries. Second, targeted firms should have high multipliers. Third, targeted firms should have some reasonable prospect of being sustainable into the future.

On export-base firms: it makes no sense to target non-export-base firms that sell goods and services mainly within the state. Even if the incentive induces a non-export-base firm to expand, its expansion will simply take sales away from some other non-export-base firms that are competing for the same customers. The net gain in state jobs is likely negligible.

On multipliers: the higher the multiplier, the higher the job-creation impact on boosting employment rates and real wage rates, and hence the higher benefits in raising state residents’ per capita earnings.

Multipliers are to some degree higher for industries whose production involves more local suppliers. Higher wages also tend to increase multipliers.[1] Multipliers also may be higher in some high-tech industries if the area already has some local jobs in that high-tech industry, which leads to some synergy benefits for the existing high-tech sector.

On the third point: while everyone probably agrees that there is little point in incentivizing an industry that will be phased out in five years, the issue is that there may be no consensus on future industrial structure. For example, although U.S. manufacturing declined in jobs quite a bit since 2000, it seems likely that U.S. manufacturing in the future will do somewhat better due to concerns about the reliability of global supply chains and foreign policy tensions. Therefore, we should be cautious about declaring that some industrial sector has no future, and that some other industrial sector will clearly be booming for the next 10 years. We should be prepared to be wrong in our forecasts.

Targeting Distressed Areas

As I have argued for many years, benefits of incentives are far higher in local labor markets that are distressed. Why? Because as a result, local hiring of the non-employed will be far higher along all the various job vacancy chains.

More recent research I have done suggests that the same thing is true when considering firm locations among counties within local labor markets. Jobs in a more distressed county—say, a central city county whose baseline employment rates are lower—will have greater benefits than jobs in a suburban county with higher baseline employment rates.

How does this affect incentive design? Incentives should either give a preference to distressed local labor markets or counties or be greater in such places.

Incentive Financing

Incentives are almost never a fiscal free lunch. That is, although incentives may have some “fiscal benefits”—induced state and local tax revenue that exceeds the required added public service costs due to more population—these fiscal benefits are typically significantly less than the gross costs of incentives (their “sticker price”).

The lack of a free lunch is true even when incentives take the form of simply forgoing tax revenue associated with the new jobs in the incented firm. Economic developers frequently argue that there is no fiscal cost here, as without the new jobs, this tax revenue would not have been created. Therefore, there are no net costs—as long as the incentive forgoes some percentage less than 100 percent of the normal revenue collected, there is no fiscal cost.

But this argument overlooks two realities. The first reality: not all firms awarded incentives actually lead to new jobs being created. This is in part because some incented firms would have created these new jobs in the state even without the incentives. In other cases, even if the incented firm would not have chosen the state “but for” the incentive, some other firm might have located at the same site and created some jobs without receiving incentives.

As a result of this reality, some of the forgone tax revenue really is forgone—some or all of the jobs have some positive probability of being created anyway, and therefore the state would have had higher tax revenue if the incentive had not been provided.

This first reality is not a small issue. For average-sized incentives—not “mega-incentives”—I have in the past estimated that at least 75 percent of the incented jobs would have been created anyway in the state, either by the incented firm or by some other firm.

The second reality: even if 100 percent of the incented jobs are net new induced jobs for the state, these net jobs bring along accompanying public service costs due to the expanded population. Part of how local public services have traditionally been financed in this country is that households and businesses share the costs of paying for population services. If we forgo the normal tax revenue associated with new job creation, we will be short of needed tax revenue to pay for added public services due to the added state residents, and either we will have to skimp on public service quality or increase taxes on households or other businesses, or maybe both.

This second reality is also a sizable issue: on average, about 80 percent of new jobs result in increased population. This population growth can have sizable public service costs. If one does realistic calculations, the revenue gain from added business activity, minus these increased public service costs, rarely offsets the gross incentive costs. Typical analyses find that the fiscal benefits from incentives might offset 20 percent of the gross incentive costs.

If incentives are not a free lunch, then incentive analysis has to account for the “opportunity costs” of paying for the net costs of incentives: state and local taxes must be increased, or public service quality must be reduced. Either alternative has some negative economic impact. However, the most negative impact is from reducing spending on productive public services. For example, estimates are that cutting back on public school expenditures has very negative effects on a state’s economy by reducing the future earnings prospects of a state’s children, most of whom will end up being the state’s future workers.

Therefore, one desirable standard for any proposed expansion of incentives is to explain explicitly how all or at least 80 percent of the costs will be paid for without reducing productive public services. What taxes will be raised to pay? Or if tax increases are infeasible, what outmoded or inefficient current public spending program will be cut?

In other words, a good design standard for incentives is to have an incentive budget. As part of that budget, we must explain how that incentive will be paid for without harming the state’s economy.


Incentive policy must aim to achieve the main goal of economic development policy: broadly shared higher earnings per capita for state residents. The design of incentives should go beyond wage standards to consider how to encourage more inclusive local hiring. The selection of incented projects should target incentives at the distressed places that are most short of jobs. Incentive budgets should make sure that incentives are not financed in a way that undermines the state’s long-term economic development.

Timothy J. Bartik is a senior economist with the Upjohn Institute for Employment Research


[1] This may seem to contradict the prior argument that higher wage standards should be judged relative to credentials. However, in general, the benefits from increasing the wage premium relative to required credentials by 10 percent far exceed the benefits from increasing the multiplier due to increasing wages by 10 percent with no wage premium boost.

Date: April 4, 2024
Categories: Commentary