Why the Michigan Business Development Program Is More Effective than Typical Incentives

By Timothy J. Bartik

Since 2012, the state of Michigan has funded the Michigan Business Development Program (MBDP), under which grants and loans are provided to targeted firms to encourage job creation or investment.  A just-released report  evaluating MBDP, by me and other Upjohn researchers, along with the Center for Regional Economic Competitiveness, concludes that MBDP has a high ratio of benefits to costs, over 4 to 1, under most scenarios for how these incentives might be financed.Michigan Business Development Program Effectiveness Study, available both at the Upjohn Institute website and at the website of the Michigan Economic Development Corporation. One of the report's co-authors, Ellen Harpel, also wrote a summary of the report. Evaluation results are stated in terms of the ratio of the present value of gross benefits to the present value of program costs, in this case incentive costs. The benefit cost ratio is commonly used in economic benefit-cost analyses. The report states results in terms of the “Return on Investment” (ROI), which is the ratio of net benefits to costs. The benefit-cost ratio is exactly equal to the ROI plus one.

What lessons should Michigan policymakers draw from this report? What implications does this report have for incentive policy in other states?

Lessons for Michigan

Based on this report, over 80% of the gross benefits of MBDP are due to higher earnings per capita for state residents, which is similar to analyses of the benefits of other incentive programs. Therefore, the benefit-cost ratio for MBDP and other incentive programs depend on factors such as: (1) how effective incentives are in inducing job creation by firms receiving incentives that would not have occurred “but for” the incentive; (2) the multiplier effects of this direct job creation on creating other local jobs in supplier industries and retailers; and (3) what proportion of the total local jobs created go to state residents.

MBDP has a high benefit-cost ratio for several reasons:

(1)    MBDP’s average incentives per job are modest, averaging $7,500 per job. Higher incentives per job may induce more location decisions, but there are diminishing returns, so doubling incentives does not double their ability to tip location decisions.
(2)    MBDP’s incentives are heavily front-loaded. Front-loaded incentives are more effective in tipping location decisions, as firms tend to focus on short-term profits.
(3)    MBDP’s incentives are highly targeted at high job-multiplier firms, averaging around 3.7 (e.g., 3.7 total jobs created in Michigan for every 1 job directly created in an incented firm).This is the gross input-output multiplier for the project, not the net multiplier after allowing for the job creation pushing up land prices, wages, and other local costs. As discussed in my recent working paper with Nathan Sotherland, Local Job Multipliers in the United States, net job multipliers tend to be at least one-quarter below gross job multipliers. The model used in evaluating MBDP allows for these cost feedbacks to reduce the effective net job multiplier. The MBDP evaluation uses the model described in this report, but fine-tuned to Michigan and the specifics of the MBDP program  

Michigan should continue these design features as it considers its incentive programs: keep incentives modest per job, front-loaded, and targeted at high-multiplier firms.

Even with these design features, the benefit-cost ratio for MBDP is sensitive to how the program is paid for out of the state budget. The benefit-cost ratio is over 4 to 1 if the program is financed by higher business tax rates, higher personal tax rates, or cuts in public spending that does not affect the productivity of the private economy. But the report finds that if incentive programs such as MBDP are financed by cuts in public spending that is productive – the report considers cuts in K-12 spending as an example – then MBDP or other incentive programs may have negative long-term economic effects. Michigan policymakers must be careful to ensure that paying for incentives does not come at the expense of investing in workers’ skills, or other productive state government investments.

The report also finds that unemployment makes a difference. Each 1 percentage point increase in a county’s unemployment rate pushes up the benefit-cost ratio for MBDP by around 10 percent. Incentive programs such as MBDP will have higher benefit-cost ratios in areas that lack adequate jobs.

Lessons for Other States

MBDP’s high benefit-cost ratio is the exception rather than the rule for state incentive programs. Previous analyses of state incentive programs have often found benefits that are much closer to costs, with the exact benefit-cost ratio highly sensitive to the assumptions made or to details about the program or state. Why the difference?

First, MBDP’s incentive costs per job is far less than typical incentive programs. Consider MBDP’s average $7,500 per job versus: Wisconsin’s offer to Foxconn of over $230,000 per job; New York’s offer to Amazon of $75,000 per job; Virginia’s offer to Amazon of $20,000 per job. While a $75,000 offer per job may tip more decisions than an offer of $7,500, it won’t tip 10 times as many decisions.

Second, most states’ incentive programs are not particularly front-loaded. Incentives are higher in year 1 of a project, but then continue at least one-third that level through the first 10 years.   The dollar of incentive in year 10 is not as effective as the dollar in year one.Typical incentive design is presented in the Upjohn Institute’s national database on incentives and report describing the database.

Third, MBDP’s job multipliers appear to be much higher than the average for multipliers in other state, for example Foxconn and Amazon’s multipliers were reported to be no more than 2.7. Why the difference? This in part may be due to MBDP’s targeting decisions. But it also probably has to do with the Michigan context: job multipliers tend to be higher in Michigan because of the state’s extensive network of manufacturing suppliers.    

Therefore, state leaders around the country, in considering their incentive designs, might want to reform incentives to be smaller and more front-loaded, with greater targeting. Targeting could be more on high-unemployment areas as well as on higher multiplier firms. The resulting reformed incentive programs would be able to attain greater job creation at lower costs, as well as greater benefits for state residents without jobs. Lower costs would also increase the odds that paying for incentives does not undermine support for state and local public services that boost economic productivity.

Timothy J. Bartik is a senior economist at the W.E. Upjohn Institute for Employment Research. The evaluation report that is the focus of this commentary was funded by the Michigan Economic Development Corporation (MEDC), which provides administrative services for the MBDP. As with all Upjohn Institute grants and contracts, the Institute reserves the right to publish the findings. The views expressed in this commentary are those of the author, and should not be interpreted as reflecting the views of the Upjohn Institute, the MEDC, or CREC.