Tuesday, April 19, 2005
Thank you for inviting me; it is a pleasure to be here today. My name is Michael LaFaive, and I am director of fiscal policy at the Mackinac Center for Public Policy, a research and educational institute headquartered in Midland.
Last week, the Mackinac Center released an assessment of the Michigan Economic Growth Authority program, commonly known as "MEGA." MEGA is a targeted tax credit program designed to encourage new business expansions and relocations in the state. It celebrated its 10th birthday yesterday.
During the next 20 minutes or so, I would like to briefly describe the MEGA program and the findings of our study. I will be happy to answer any questions you might have when I am done.
The MEGA Program
MEGA was first signed into law on April 18, 1995. The authority approved its first three deals just three days later. Between that day and the end of December 2004, the MEGA program has offered more than $1.8 billion in Single Business Tax relief in 230 deals.
Other incentives are often part of MEGA deals but they are not tied directly to the SBT. The state often provides job training subsidies or infrastructure grants and MEGA law mandates a contribution of local incentives as well. These non-MEGA incentives associated with the deals are valued at up to $1.2 billion.
Given the scope of the MEGA program and the increasing reliance state policy-makers seem to be placing on it, the Mackinac Center last year decided to review the MEGA program in depth. We believed that enough time had passed to do a thorough, empirical analysis of the program using advanced statistical techniques. Along with that review we intended to highlight the successes or shortcomings of the program using MEGA program data, case studies, and both anecdotal and empirical studies from outside of Michigan. The result is this 127-page study authored by myself, and my colleague, economic professor Michael Hicks.
I’d like to share its findings with you now.
MEGA’s Goals Versus MEGA’s Accomplishments
As part of our overall analysis, we went through documents obtained from the state and examined what was expected of the MEGA program by state officials and then we compared those numbers with the claims of job creation submitted by MEGA companies and tallied by state officials. These numbers are taken from a "MEGA Credits" spreadsheet that the Mackinac Center received from the Michigan Economic Development Corporation in December.
MEGA agreements are expected to create two types of jobs — "direct" and "indirect." "Direct" jobs are new jobs at the specific firm sites that are the subject of the MEGA package. "Indirect jobs" are new jobs created outside these specific firm sites in response to MEGA-related investment and direct employment.
State documents indicate that approximately 127 of MEGA’s agreements should have produced fully employed facilities through 2004 — i.e., sites hosting all of their projected direct jobs.
Of these 127, about 56, or 44 percent, have claimed credits under the MEGA program. A company can claim these tax credits, however, without meeting the initial total direct job projections. Only about 10 or 11 of these 56 cases can be shown to have created the number of direct jobs originally projected within the expected time frame. This has important implications that I will highlight in a moment. MEGA originally projected that these 127 MEGA deals would generate 35,821 direct jobs by 2005. MEGA figures obtained in December 2004 shows that these deals have actually generated about 13,541 direct jobs — roughly 38 percent of original expectations.
At this point I need to offer a brief aside: these calculations were based on very important assumptions made by the Mackinac Center. We attempted to clarify our assumptions and methodology with the Michigan Economic Development Corporation and with at least one MEGA official but were thwarted. Indeed, one MEGA official told me he had been instructed not to talk to me. This situation improved only in recent weeks, following the Commerce Committee’s meeting with Mr. Jakeway in March.
I bring this to your attention for a key reason. We made every effort to understand the figures the MEDC provided to us concerning the MEGA program, and we believe that our understanding of that data is reasonable and correct. Nevertheless, it is possible that we have misunderstood some secondary issues in areas where MEDC’s presentation of the data is incomplete or ambiguous.
If we have, it will not be for a lack of trying on our part. In fact, a partial history of our repeated efforts to clarify key points regarding the data is provided on the Mackinac Center Web site under the headline "Job Search."
Econometric Evaluation of MEGA’s Economic Impact
I have just explained how MEGA’s track record has not been very encouraging. Still, we recognized in the course of our study that the track record was not the only issue. Even if MEGA’s success rate is not very high, it could still bring important benefits if its successes were particularly effective in stimulating jobs and income growth.
We thus constructed what is known as a time-series econometric model designed specifically to measure the MEGA program’s economic impact on the state of Michigan. In building the model, we were careful to review the economic literature regarding similar models and to try to ensure that our model did not overlook key variables or repeat past mistakes. Our methodology follows a consensus view on the right way to measure the impact of targeted tax incentive programs like MEGA’s.
In running the model, we investigated the years from 1995 to 2002, since 1995 is when the MEGA program began, and since 2002 is the last year for which county-level Michigan statistics were available. We also included data on Michigan’s economy going back to 1990, so that we would have a solid picture of Michigan’s economy before MEGA was established.
Here are four general findings that emerged from the model:
MEGA did not improve Michigan’s per-capita personal income, employment or unemployment rate. In other words, the very things that state officials had hoped to improve do not appear to have been affected by MEGA.
MEGA did not improve any Michigan county’s per-capita personal income, employment or unemployment rate (our estimates of impact ranged from zero to modestly, though not significantly, negative). Similarly, Michigan counties that did not host companies receiving MEGA deals fared as well as counties that did host such companies. Unfortunately, then, MEGA did not seem to bring meaningful local improvements, either — something that would have been welcome in some of the state’s lower-income areas.
MEGA essentially did not affect aggregate income or employment in manufacturing and warehousing (the one statistically significant effect was negative, but too small to be economically significant). These are business sectors that might have been improved by MEGA, given the nature of the projects the program has pursued in the past decade.
MEGA did apparently cause a temporary shift to higher construction employment, but it did so without increasing overall employment. Our model allowed us to estimate that one temporary construction job was created for every $123,000 in MEGA credits awarded. This is a striking cost-benefit figure, particularly when you consider that targeted SBT incentives are not the only incentives involved, and when you consider that 75 percent of these jobs disappeared after one year, while the remaining 25 percent fell away after two. Indeed, the temporary increase in construction employment was accompanied by a decrease in construction wages, though the decrease was too small to be economically significant.
Broader Context: Development Agencies’ Overestimates
The findings I’ve outlined so far suggest that MEGA failed to achieve what it set out to achieve, and that MEGA has not achieved what state policy-makers had ultimately hoped to achieve — that is, a significant impact on the state’s economy. These are very important findings, but I want to take a moment to share with you several other findings we uncovered in the course of our study that provide important context in evaluating both our findings and the MEGA program itself.
First, we are not the first to question the jobs figures of economic development agencies. In Michigan, for instance, the state’s auditor general has also twice criticized (non-MEGA) Michigan development agencies for the overstated job creation numbers that the agencies had reported on audited programs. In the first instance, the Michigan Strategic Fund overstated by 39 percent the number of jobs two of its programs allegedly created. In a second instance, the MEDC was chided for relying on the claims of companies to whom it offered assistance, instead of independently verifying job creation.
We also found that in one case, MEGA’s projections for a company exceeded the company’s own expectation. In an internal memo to its employees, MEGA recipient Hemlock Semi-Conductor felt compelled to disagree with public claims of job creation made by state officials. We’ve reproduced a portion of that memo in our study.
Overly optimistic expectations are not just confined to Michigan. In Ohio last year the economic development agency, "Regional Growth Partnership" felt compelled to recant on 65 percent of the jobs whose creation they claimed to have influenced. Also in Ohio, two economists who have studied the state’s primary development incentive programs between 1993 and 1995 found that while the incentives had no positive economic impact, the businesses that received incentives were more likely to overestimate employment forecasts than businesses that had not. In other words, it is likely that the businesses inflated their job creation forecasts to help win the incentives.
Nationwide, many incentive programs come up short of expectations time and again. Economic development economists frequently review the studies made of specific programs by other researchers and summarize them. Terry Buss of the National Academy of Public Administrators reviewed roughly 300 such studies and reported that the literature is mixed, but generally negative toward tax incentive programs. Other highly respected economists have come to the same conclusion.
In light of these findings, what should be done? Our recommendation, given our review of MEGA and the economic literature on other economic development programs, is to end the MEGA program. Over 10 years it has proven itself ineffective. Time and resources spent on the program may well divert the state from more productive policies that improve the climate for all of the state’s businesses, not just a few.
Let me add briefly here, however, one thing we are not saying. We are not somehow blaming those companies receive MEGA grants for taking the tax credits when they are offered, just as we wouldn’t blame people for taking credits or exemptions on their own personal income taxes. We do believe that the evidence shows, however, that states would be better off — or at the very least, no worse off — if they refrained from trying to pick winners and losers in the marketplace with these specially targeted breaks, and instead concentrated on more broad, fundamental reform. This is what we call the "fair field and no favors" approach to economic development, where tax, regulatory or labor reform is broad-based and improves the business climate for all of the state’s businesses.
Let me also add a quick rejoinder to an argument that is frequently made in MEGA’s favor — the idea that most states have programs like MEGA, so that if we abandon MEGA, we will be engaging in "unilateral disarmament." This view gets it precisely backward. All available indications tell us that MEGA is not armament; it’s not doing the job of improving the state’s economy, even if it manages to convince a firm here or there not to locate new jobs or facilities outside of Michigan.
Remember, MEGA has managed to close 230 deals through December, yet there are 105,000 businesses in Michigan with SBT liability. The direct jobs MEGA has created amount to about 0.3 percent of Michigan’s job total. It’s hard to locate any bang for the buck — and as I mentioned before, the economic literature suggests this is true in other states, as well. Furthermore, five states operate no targeted tax incentive programs and they’re still doing better than Michigan economically.
Ending MEGA would be more like throwing aside the wrong tool for a problem. It probably frees your hand to use the right tool instead. This would be genuine armament, not disarmament, as some critics contend.
While simple reforms of the MEGA program are unlikely to increase its economic value, policy-makers can take several steps that might improve MEGA if they choose to continue it: Let me list five suggestions.
Audit MEGA. The state could consider asking the Office of the State Auditor General to conduct regular, expanded audits of MEGA’s direct job counts. Such oversight could help improve the authority’s accounting procedures. We would recommend that the auditor general’s office be encouraged to specifically review applications by MEGA candidates that were rejected. The results of such a review could help clarify the extent to which the MEGA credits have truly been necessary. The Mackinac Center for Public Policy has tried for years to obtain these applications and received a variety of different responses. Recently, the MEDC provided five applications for companies that applied for MEGA grants, but were denied. We suspect that there are more.
Take the Auditors’ Advice. Please don’t let any review of the auditor generals just collect dust. Use it. The state of Washington did such a review and basically ignored the analyses. It went on to approve a $3.2 billion incentive deal of dubious value.
Commission an Independent Econometric Review. An independent researcher could be engaged to maintain a peer-reviewed and publicly transparent econometric model that annually reassessed MEGA’s impact. The model employed in this study was crafted to detect past impact, rather than predict future performance. Regular updates of the findings would therefore be appropriate if the MEGA program continues. At the very least, hire a respected institution to review our own econometric findings.
Count Direct Jobs Only. It would probably facilitate public review and understanding of the MEGA program if indirect job benefits were no longer reported and cited by MEGA officials. "Spin-off" considerations could still be part of the evaluation process for a particular project, but MEGA would no longer make a formal or informal practice of tallying the indirect jobs its past and future projects could claim. Removing MEGA officials’ focus on indirect job counts might free the authority to more carefully document its direct job creation. Several op-eds by state officials recently made "spin-off" claims that were little more than glorified guesswork.
Develop a Transparent Framework for Tracking Success and Failure. The status of each MEGA project could be posted and updated live on the Web each month to show such basic items as the following: the state and local incentives offered in each MEGA package; the cost of these incentives so far and in the current year; the current direct job figures; what the direct job figures were originally projected to be at present; and so on. Such reporting would facilitate effective public oversight of the program’s effectiveness.
Thank you for your time. I’ll be happy to take any questions that you might have.
Michael D. LaFaive is director of fiscal policy for the Mackinac Center for Public Policy, a research and educational institute headquartered in Midland, Mich. Permission to reprint in whole or in part is hereby granted, provided that the author and the Center are properly cited.
 It was, however, statistically significant.