Timothy J. Bartik
Senior Economist
W.E. Upjohn Institute for Employment Research
Kalamazoo, Michigan

This paper summarizes what we know about eight issues for public policy toward economic development incentives. The evidence supporting many statements made here can be found in the references listed at the paper's end.

1. Benefits vs. costs of incentives.

The issue isn't whether economic development incentives can work; empirical evidence suggests they can. The issues are whether benefits of incentives outweigh costs, and how benefits and costs are affected by local conditions and incentive design.
Empirical research on taxes and business location suggests that state and local taxes have a statistically significant effect on business location decisions. Local job growth has significant effects on the earnings of local residents.
But do the earnings benefits justify the costs of incentives? Incentives are costly per job created. The effect of taxes on business location is modest. Many location and expansion decisions are unchanged by incentives. The cost of incentives to businesses whose location decisions are unchanged exceeds the taxes from businesses whose location decisions are changed.
When one considers all the evidence, the costs to local residents (in higher taxes or lower services to pay for incentives) from an "average" economic development incentive program are of similar size to the earnings benefits for local residents. This implies that the average incentive program does not make sense in a low-unemployment area. If unemployment is low, local residents can easily find jobs, and the earnings benefits from greater employment rates will overstate the social benefits of new jobs.
Economic development incentive programs are more likely to pass a benefit cost test if (1) local unemployment is high, so the new jobs are needed by local residents; (2) the jobs pay higher wages; (3) more of the jobs go to local residents.

2. Targeting.

Targeting some firms for greater incentives than others may be rational. Such targeting should not, however, be based on political pressure or media attention. Targeting should be based on which firms are likely to provide greater social benefits at lower incentive costs. Reasonable targets include firms that provide greater social benefits because they pay higher wages, or are more likely to employ local residents.
Because large new branch plants attract media attention, there is political pressure to target incentives at these plants, and not smaller or existing firms. A more rational basis for targeting incentives is toward firms that provide greater social benefits at lower incentive costs. Firms that provide greater social benefits, and therefore might be targeted for greater incentives, include firms paying higher wages and firms that hire the local unemployed or disadvantaged.
Incentives could also be more cost-effective if targeted to firms that are more responsive to incentives. This type of targeting is more questionable. Research has not provided clear evidence of the types of firms or industries that are most responsive to incentives. Economic development officials lack sufficient information on particular firms to tell whether an incentive will prove decisive.

3. Rules vs. discretion.

Economic development incentives should be distributed with more reliance on rules and less on discretion. The balance between rules and discretion should vary for different incentives. For example, job training and infrastructure incentives should be customized to individual firm needs and therefore must be partly discretionary. On the other hand, most financial incentives could be folded into the state and local tax code.
It is more efficient and fairer to provide similar incentives to all firms providing similar social benefits per dollar of incentive. Formulas or rules for distributing incentives can help do this. In theory, a benefit-cost analysis of each firm could do a better job of evaluating the social benefits from the firm's increase in employment. In practice, some simple rules may do reasonably well at matching the larger incentives to the firms offering greater social benefits. Formulas and rules help a state or local government maintain its bargaining position against political pressure to increase the bid for a particular company.
Financial incentives--tax abatements or credits, subsidized loans--could be governed by rules by being incorporated into the regular tax code. Firms could be entitled to a "New Jobs Tax Credit" that based the amount of the credit on factors such as the firm's increase in employment, the wage rate paid by the new jobs and who gets the jobs.
Job training incentives or infrastructure assistance will be more useful if customized to the needs of the particular firm. Therefore, these programs should not be designed as entitlements, although there should be guidelines on which types of firms are to be helped and how.

4. Incentives vs. small business assistance.

Customized government assistance to particular businesses does not always represent an "incentive" designed to attract businesses and increase job growth. Small business assistance programs provide customized services to small businesses to improve their productivity. Such services to small business include help with job training, technology, exporting, financing and business management. These programs are efficient if they increase business productivity by more than their costs.
There are many possible private "market failures" that adversely affect small businesses, including: inefficiently low job training; poor information on management, technology and exporting; and problems in obtaining financing. Many economic development programs seek to address these problems, either through government directly providing services, or encouraging a private organization to provide services. Evidence from surveys, and comparisons of assisted and unassisted firms, suggests such assistance can be effective.
Small business assistance programs should not be confused with incentive programs that are solely justified by their effects on local job growth. Small business assistance programs can be efficient even if they do not increase local job growth. For example, suppose a small business assistance program provides training to help disadvantaged individuals start new restaurants. Suppose these new restaurants compete so successfully with other local restaurants that they drive them out of business. Even if total local employment is unchanged, the local restaurant sector will be more productive, because presumably the restaurants that survived had better food or lower prices. If the greater productivity of the new restaurants exceeds the costs of training new restauranteurs, then local consumers are better off.

5. The timing of incentives.

Both economic efficiency and theories of public decision-making suggest that better decisions will be made if incentives are required to be provided upfront. "Clawback" provisions should be attached to these upfront incentives, allowing some incentive funds to be recovered if the promised jobs do not arrive or later disappear.
Because empirical evidence suggests that business executives heavily discount future cash flows, incentive dollars provided 10 years or more in the future have little effect on business location decisions. Furthermore, allowing politicians to give away the future tax base of their jurisdiction is likely to encourage profligate use of incentives.

6. The zero-sum game argument.

The theoretical case for economic development incentives being a zero-sum or negative sum game--that is, for rational competition among state and local governments inevitably leading to economic inefficiency--is weak. The national concerns about economic development incentives should be twofold: (1) For political reasons, governors and mayors often do not pursue the general interest of their jurisdiction, resulting in inefficient use of incentives; (2) economic development competition among states and local governments is making it harder for these jurisdictions to redistribute money or services to low- and moderate-income groups.
If one assumes that each state or local government serves the interests of all its residents, and that these governments can use many tax instruments, economic development competition should be efficient. Competition should bid business taxes down to equal the marginal cost of providing businesses with public services, plus the marginal environmental and congestion costs created by the business, minus the marginal social benefit the business provides by creating jobs. Such a business tax system is economically efficient.
This competition might lead to some expansion of national output, employment or wage rates. Subsidies for employment and capital in many jurisdictions might increase total national employment and capital stock. (This is more likely if greater incentives are provided in high-unemployment areas, which have the greatest social benefits from job growth.) Support for higher wage industries and firms might increase wages nationally. The provision of information and training to small businesses could increase national productivity.
One problem with this argument for the efficiency of economic development competition is that governors or mayors may not pursue the interests of all local residents. Governors or mayors in low-unemployment areas may offer large incentives, though the benefits for local residents are small, to help land developers. Governors or mayors may use long-term tax abatements to pass the costs of incentives on to their political successors. Governors or mayors may favor the large branch plants that get media attention over helping smaller businesses.
A second problem is that economic development competition is making it harder for state and local governments to redistribute public services or money to low-income households. Traditionally, state and local fiscal systems have been designed so that businesses and upper-income households pay a bit more in taxes than they receive in public services, and low-income households receive more in public services than they pay in taxes. The mobility of businesses and upper-income households has limited but not eliminated this redistribution. The increased mobility of businesses and economic development competition is making redistribution more difficult for state and local governments.

7. The feasible federal role.

The federal government cannot eliminate economic development competition among states and local governments. The federal government can encourage a more rational competition. Federal policy could discourage discretionary financial incentives and encourage economic development policies that would have greater national benefits, for example, policies to encourage growth in high-unemployment areas or increase labor productivity. But greater federal intervention in economic development could also discourage state and local experiments, and hurt distressed areas.
As long as state and local governments can design their own business tax systems and mix of public services, they can use such power to attract new businesses. Federal policy might more feasibly be used to penalize, through withholding federal grants, discretionary financial assistance given to particular branch plants, but not to similar businesses. Federal policy could encourage and support economic development incentives in high-unemployment areas, where the social benefits are greater. Federal policy could encourage experiments with small business assistance, which may increase productivity. Finally, federal policy could support more consistent information on and evaluation of state and local economic development policies. Better information on the size and effects of incentives, and other economic development policies, should put pressure on state and local governments to adopt more rational policies.
The problem is that federal intervention could also eliminate some important experiments in economic development policy. A federal policy to discourage discretionary incentives might discourage them everywhere, including high-unemployment states and cities. Furthermore, federal intervention could discourage job training, infrastructure and small business assistance programs that could help increase productivity.

8. Incentives, federalism and redistribution.

The problems of economic development competition reinforce the argument that the federal government should have primary responsibility for redistributional policies. Current proposals to turn over responsibility for welfare and Medicaid to the states are going in the wrong direction.
The traditional wisdom in public finance is that income redistribution should be a federal responsibility, because mobility of households and businesses makes this task difficult for state and local governments. This traditional wisdom seems more sensible as business has become more footloose and economic development competition intensifies. The problem with turning over welfare and Medicaid to the states is that states have strong incentives, for economic development reasons, to cut funds for welfare and Medicaid. Taxing businesses to finance welfare and Medicaid does not make sense for a state's economic development.


References

Bartik, Timothy J. "The Market Failure Approach to Regional Economic Development Policy," Economic Development Quarterly 4, (November 1990): 361-370.
Bartik, Timothy J. Who Benefits from State and Local Economic Development Policies? Kalamazoo, Mich.: W.E. Upjohn Institute for Employment Research, 1991.
Bartik, Timothy J. "Jobs, Productivity, and Local Economic Development: What Implications Does Economic Research Have for the Role of Government?" National Tax Journal 47, No. 4 (December 1994): 847-61.
Bartik, Timothy J. "Strategies for Economic Development," Management Policies in Local Government Finance, 4th Edition, J.R. Aronson and E. Schwartz, eds., International City/County Management Association Press, forthcoming.

Top - Introduction - The Economic War Among the States