The primary sources of financial data concerning MEGA and its related local incentives are documents produced by the MEDC and its predecessor agency, the Michigan Jobs Commission, obtained by the Mackinac Center under Michigan’s Freedom of Information Act. Graphics 2 and 3 (see links below) provide two-page samples of the "All MEGA Projects" and "MEGA Credits" spreadsheets that are referenced throughout this study.
Since its inception, MEGA has offered targeted business tax relief exceeding $1.8 billion in a total of 230 deals to more than 200 companies or related subsidiaries. (As mentioned earlier, the precise count of the companies involved in these deals can be debated, depending on how one treats acquisitions, subsidiaries or suppliers.) Graphic 1 shows a year-by-year breakdown of total MEGA deals and the approximate, inflation-adjusted value of their respective MEGA credits.
The number of MEGA packages and the total size of the SBT credits offered each year has generally been rising, despite dips in 1997, 2001 and 2003. In 1996, MEGA’s first full year, MEGA offered just 15 deals, totaling $89.9 million in SBT credits; only 25 MEGA deals were allowed annually at the time. In 2004, however, MEGA produced 41 packages valued at $253.3 million (for up to 20 years) in SBT relief alone.
The amount of MEGA state tax credits for individual firms has varied over the years, from a tiny offer of $160,000 to Integrity Design Inc. in April 2001 to an $88 million MEGA "retention" credit to the Ford Motor Company in November 2000 (these credits could be earned over five and 20 years, respectively).
Each MEGA deal has included additional state or local tax incentives for the recipient firms. To date more than $987 million has been offered to MEGA companies by local units of government or local economic development agencies. As mentioned earlier, local incentives offered in MEGA deals usually involve tax abatements on real or personal property, permit waivers, road improvements or municipal perquisites for company employees.
The incentives offered through MEGA packages total more than $3 billion since the beginning of the program. These incentives include state tax credits, local abatements and other state and local inducements, such as job training and road improvements. Graphic 4 shows the approximate value of all known incentives offered to recipients of MEGA deals by category.
According to MEGA Briefing Memos produced by the MEDC, the smallest local incentive in MEGA history appears to be a $28,000 break to Universal Forest Products in 2002 (though it is possible that a smaller incentive was offered that was not discernible in the MEDC documents obtained by the authors). The largest local incentive was $165 million over 25 years to General Motors Corp. in June 2000.
General Motors Corp. has been the direct beneficiary of six MEGA deals, by far the most of any corporation. Additional MEGA deals have been concluded with GM suppliers as part of an overall package benefiting GM. For instance, CMI‑Schneible Co., Machining Enterprises Inc. and Michigan Production Machining Inc. have all received MEGA deals. The MEGA credits and other incentives offered to these companies were all designed to support the "General Motors Nodular Iron Redevelopment Project" in Saginaw.
While calculations about jobs creation are, in theory, straightforward, they depend on assumptions that may differ. MEDC has not provided all of the clarifications we had originally sought about the basis of some of the projected MEGA jobs figures, so we adopted two key assumptions: first, that all predictions of direct job creation by MEGA corporations by a certain year (1999, for example) meant the jobs would exist at midnight on Dec. 31 of the previous year (1998); second, that this previous year roughly corresponded to the tax year listed on the "MEGA Credits" spreadsheet. Changing the methodology would change the precise totals, though probably with no significant impact on the findings.
Another issue that must be decided in reviewing MEGA’s performance involves distinguishing between "direct" and "indirect" jobs purportedly created by the program. The MEDC defines "direct jobs" attributable to a MEGA package as the increase in the number of jobs at the specific firm sites that are the subject of the MEGA package. It defines "indirect jobs" as those jobs that are created outside the specific MEGA business sites following the MEGA-related investment and direct employment.
We have chosen to work primarily with direct job counts in this study. Estimating indirect job counts is a subjective exercise, and econometricians and accountants with the best of intentions can produce widely varying figures, depending on their assumptions and estimation techniques. In addition, the indirect job totals implicitly claimed by state officials in recent press commentaries are, at best, rule-of-thumb estimates, and they appear to be unreliable, as we will detail below. Thus, when we do discuss indirect job counts (also known as "spin-off" jobs), we will say so specifically.
Another key issue in evaluating MEGA’s track record on jobs is captured by the remarks of David Hollister, Director of the former Michigan Department of Consumer and Industry Services, in discussing MEGA’s job program:
There is usually a lag time of several years between MEGA board approval of a project and reporting of jobs created. Many of these projects involve the construction of new plants and facilities, which can take several years. Therefore the number of jobs created by MEGA-aided projects … is only a fraction of the actual number of jobs that MEGA will eventually help to stimulate. … MEGA has been in existence for less than eight years, (so) we are just beginning to see the benefits of credits granted early on.
It is true that economic and financial investments can involve lag time. Thus, it is necessary to factor in the delays projected by MEGA officials in each specific tax credit package, and we have done so in our calculations in the following section. We do not assume, however, that time lags can effectively be infinite.
Finally, we would note here that we have kept track of MEDC incentives differently than MEDC officials do. MEGA officials often remove data from their "All MEGA Projects" spreadsheet for deals that fail early on, and MEGA then subtracts from the spreadsheet the amount of the incentives originally offered to those firms. We have maintained this data, so we often report different and higher MEGA incentive totals.
MEGA Job and Project Performance
Based on state documents compiled since 1995, we estimate that 127 of MEGA’s agreements should have produced fully employed facilities through 2004. (For a discussion of the assumptions we employed in calculating this number, see "Appendix C: Determining MEGA’s Job Counts.") Of these 127, about 56, or 44 percent, have claimed credits under the program.
The ability to claim credits does not mean, however, that a firm has achieved the job goals that it projected in its MEGA agreement. For instance, for most MEGA deals, a company must create an initial 75 "qualified new jobs," usually within a year of commencing operations a particular site. If it does reach this minimum job threshold, it can begin claiming its tax credits, even if it never achieves the job count it originally projected.
Many of the 56 cases in which tax credits were claimed did not fully meet original projections. In fact, only 10 of the 56 can be shown to have created the number of direct jobs originally projected within the expected time frame — although three of the 10 (Kmart, for instance) have had setbacks following their initial success in meeting the targets (see "Appendix C: Determining MEGA’s Job Counts").
According to MEDC figures, MEGA originally projected that 35,821 direct jobs would be created at MEGA companies by the 127 MEGA deals that were supposed to be fully operational by 2005. Based on a MEGA document obtained in December 2004, about 13,541 jobs exist at those companies, or roughly 38 percent of what was originally hoped for. Given this figure, MEGA’s direct job total represented about 0.3 percent of Michigan’s 2004 workforce (see Graphic 5a by clicking on the hyperlink "More Images...," which appears beneath Graphic 4 above).
This finding, we note, is similar to an independent estimate made by The Detroit News in 2003. The News used state documents similar to those employed in this study and calculated that 10,787 direct jobs had been created at authority-aided projects."
Our direct job figure of 13,541 is admittedly lower than the job estimate MEGA officials have sometimes cited. For example, in a November 2004 commentary in Business Direct Weekly, MEDC Chief Executive Officer Donald Jakeway claimed, "Of the MEGA projects that have collected SBT credits for their projects to date, 28,812 total jobs have been created at an actual SBT credit cost of $75.1 million."
According to the MEDC’s "MEGA Credits" spreadsheet, however, the $75.1 million in MEGA SBT credits referred to in the commentary actually created 13,541 direct jobs. Our exchanges with MEGA over the discrepancy eventually indicated that the additional 15,271 jobs are MEGA’s estimate of the indirect jobs created by these MEGA projects. As a result of additional inquiries in the ensuing three months, we received the following explanation of how the indirect jobs figure was calculated:
The MEDC staff member [who produced the total jobs number] figured the proportion of indirect to direct jobs project[ed] and applied that proportion (approximately 1.1 to 1) to the number of actual jobs (13,541) that had been reported to us up to that point in time. The number of actual new indirect jobs was thus estimated to be approximately 14,709, for total actual new direct and indirect jobs of a little more than 28,000.
Thus, MEGA officials have used a ratio to estimate the number of indirect jobs created by existing MEGA projects. An MEDC spokesman notes that this calculation is based on projected job estimates produced by an economic modeling software program known as REMI (Regional Economic Modeling Inc.), and that MEDC officials "trust the statistical model used in the widely used REMI analyses to provide a good estimate that helps us determine the economic value of a proposed new facility or expansion."
While we respect the power of the REMI model, we would note that the method described above was developed by an MEDC official, not the professional REMI modelers on contract with the state. Further, we would add that this method of estimating existing indirect job totals is unreliable for at least four reasons.
First, the MEDC starts with 13,541 existing direct jobs and applies to all of them a constant ratio (of indirect jobs to direct jobs). But this ratio would not be constant; rather, it would vary for each of the 56 projects that produced the 13,541 jobs. Indeed, one key reason for employing the REMI model is to determine what the indirect job creation rate will be for each project, since spin-off job creation can vary in complex ways.
Second, the constant ratio that the MEDC employs is based on projected direct and indirect job creation for all of MEGA’s projects. In many of these cases, however, the MEGA-related investments and job creation will happen many years from now, well past 2005. Many of the assumptions made about the creation of these future jobs are not valid for job creation during the past 10 years — the very interval for which the projection of 28,812 jobs was made.
Third, even if the ratio of indirect jobs to direct jobs were constant and based on relevant figures, the 14,709 indirect jobs calculated by the MEDC would necessarily overstate the number of indirect jobs that have been created. In REMI models, indirect job projections do not stop once the direct jobs have been created. Instead, the model assumes that indirect jobs will continue to be generated for years into the future. Thus, the indirect jobs estimate calculated by the MEDC method will inevitably include some jobs that have not yet come into existence. Unfortunately, the public claim of 28,812 jobs created implies that all of these jobs already exist.
Fourth, the jobs claim output of the REMI model is based on a series of assumptions (such as estimates of direct job creation) that we know in retrospect to be false. This observation is no criticism of the REMI model or of those who use it; assumptions of this kind are always necessary in economic modeling, and they are often seen to be inaccurate in hindsight. Nevertheless, any current determination of MEGA’s indirect job creation should acknowledge that these projections and any ratio based on them are no longer sound. The MEDC’s method does not do this.
As a result of these and other concerns about estimating MEGA’s indirect job figures, we have not used the 28,812 job total in our analysis. Nevertheless, we show in Graphic 5b what MEGA’s job contributions to state employment would represent if the 28,812 figure were correct (view Graphic 5b by clicking on the hyperlink "More Images...," which appears beneath Graphic 4 above).
We would note that the performance figures we have calculated for MEGA’s projects are consistent with reports that the MEDC sends to the state Treasury each year. These reports include tallies of the size of the SBT credits that are projected to be claimed under MEGA agreements that year, and the Michigan Treasury in turn reports and revises these numbers periodically.
Based on state Treasury figures, between 1996 and 2004, MEGA originally estimated that more than $220 million in SBT credits would be redeemed as a result of the MEGA program. (Graphic 6 shows the year-by-year tax credit claims projected for the MEGA program; the graphic can be viewed by clicking on the hyperlink "More Images...," which appears beneath Graphic 4 above.) As mentioned earlier, however, credits of just $75 million have been claimed, or about 34 percent of the amount originally expected.
According to the State of Michigan’s "Executive Budget Appendix on Tax Credits, Deductions, and Exemptions Fiscal Year" documents, MEGA officials expect foregone SBT revenue to total only $9.7 million in 2005, down from $37 million in 2004 and from an all-time high of $57 million in 2002.
These anticipated drops in annual SBT revenue effectively represent expected drops in the employment totals of MEGA businesses. If MEGA were meeting its original job creation projections, the size of the claims of annual SBT tax credits would be rising, not dropping, as more MEGA companies brought their new or expanded facilities online.
Three Major News Events
One of the purported benefits of an economic development program like MEGA is the positive publicity about the state’s business climate that can follow the success of a MEGA project. We therefore reviewed in greater depth several MEGA packages that provided the potential for significant job creation under well-publicized agreements with well-known firms.
The MEGA agreements described immediately below were concluded in 2000, 1996 and 1999 respectively. State government press releases described them as follows:
In further promotion of the MEGA agreements, former MEDC President Doug Rothwell told Site Selection magazine that Webvan, the subject of the third news release, was "one of the best-financed retailers on the market for the next wave of e-retailing."
Seven companies are included in the MEGA packages described in these press releases: Altair Engineering, Case Systems, Delphi Automotive Systems, LDM Technologies, National TechTeam, Shape Corp. and Webvan. Of the 5,249 projected jobs described in the news releases above, 3,455 were direct jobs. All of these direct jobs were to have been created by 2005 at the seven companies involved.
According to the official "MEGA Credits" spreadsheet received by the authors in December 2004 (and on which the figures in this study are generally based), four of the seven companies have not claimed the tax credits they would have been entitled to if they created the jobs stipulated in their MEGA agreement. One of these four, Webvan, went bankrupt, losing most of its value in the 12 months after its MEGA deal had been approved.
Of the three companies that have received credits — Shape Corp., National Tech Team and Case System — none directly created the number of jobs that were forecast in the time frame expected. National Tech Team qualified for about $180,000 worth of MEGA credit through 1998 before stumbling and temporarily losing access to credits.
Thus, as of December 2004, these seven companies appeared to have created a total of 514 direct jobs through tax year 2003 — or about 15 percent of what state officials had said would be created.
In fairness to MEGA and the companies involved, MEGA’s February data, received shortly before this study’s publication date, indicate a somewhat better result. Two of the seven companies have improved their job performance. National Tech Team has received MEGA credits for tax years 1999 through 2003, and in 2003, the company’s employment stood at 115 at the facility for which it received the MEGA package. Also, Delphi Automotive was awarded tax credits for 600 jobs created at its facilities in 2002 and 2003.
Including this new information about National Tech Team and Delphi raises the jobs total for all seven companies mentioned in the three news releases to 1,229. This represents about 36 percent of the total jobs that were projected originally.
The February data shows that MEGA’s job figures can improve with time. Still, it is worth noting that these figures can decline again, as well. For instance, according to published reports, Delphi Corp. is expecting losses of $350 million in 2005, and the company’s stock has tumbled from $17 in February 1999 to a low of $4.15 in late March 2005.
Kmart was twice awarded MEGA deals by the state of Michigan. The total value of these two deals, including MEGA’s state tax credits and local incentives, exceeded $34 million.
The first MEGA deal for Kmart occurred in May 1998 and included an offer of MEGA state tax credits valued at as much as $14.3 million in total over 20 years; job training subsidies worth as much as $297,500; and a local commitment from Troy worth $450,000 in "public improvements" and other favors. The second deal was approved in August 2000, less than 17 months before the corporation declared bankruptcy. Its MEGA credit was worth as much as $15.9 million in total over 14 years; job training subsidies of up to $450,000; and local infrastructure assistance of $2.6 million to "add a third lane to Kmart’s entrance and … improve traffic flow to and from the project site."
In exchange for these incentives, Kmart promised not to let its base employment levels drop below 3,637 in the 1998 deal and 4,084 in the 2000 deal. By February 2003, however, employment levels had dropped to about 3,500, which was below the conditions of both MEGA agreements. A November 2004 Detroit News article suggested that the job count is now as low as 2,000; another Detroit News report indicated that Kmart Corp. was tentatively offered a third package of incentives estimated to be worth at least $40 million if it promised to maintain just 1,500 jobs in the state.
Given Kmart’s declining job numbers, the firm was unable to claim all of the credits it was originally eligible for under its MEGA agreements. Nevertheless, before Kmart’s new jobs disappeared, it did receive $6.1 million in state tax relief, as well as job training and road improvements.
Covisint was born in December 2000, the child of several automotive companies, including Ford Motor Co., DaimlerChrysler AG, General Motors Corp., Renault SA and Nissan Motor Company Ltd. The firm was operating out of Southfield, but apparently had not settled on a permanent headquarters. Covisint was going to be an online automotive supply electronic auction site.
Industry experts and others had high hopes for Covisint, and the company was expected by some to broker more than $300 billion in annual sales, while producing $5 billion in annual revenue for the company. Officials from other states, such as Georgia, were wooing the company in hope of landing Covisint’s new headquarters.
Official "Meeting Minutes" of the MEGA board indicate that MEGA officials thought that by 2021, the Covisint project would bring 1,000 direct new jobs and 966 indirect new jobs statewide. As part of its deal the firm needed only to maintain a base employment staff of 169. In April 2001, Covisint announced that it had chosen Michigan for its permanent headquarters.
Unfortunately, by the end of 2002 Covisint was struggling due to unforeseen challenges, such as competition from other business-to-business auction sites, and the fear of suppliers that fake auctions were posted on the site in order to get a glimpse at what tier-two suppliers might offer. Less than three years after it received its MEGA deal, Covisint’s prospects had plummeted, and its approximate value in 2003 was $25 million. By October 2004, parts of the company were reportedly sold to FreeMarkets, an auction site, and Compuware Corp. (the latter portion for about $7 million).
As of Jan. 2005, total employment at Covisint (under Compuware) stands at 122. We have found no evidence that the firm met its job creation goals and or claimed its MEGA credits.
These are not the only cases where widely publicized MEGA projections did not come to pass. With Aspen Bay, for instance, MEGA officials offered nearly $22 million in MEGA credits and other incentives for construction of a new pulp plant, certifying that the company was "financially sound and that its plans for the expansion or location are economically sound." But the company was unable to find private financing for construction of the plant after it had received approval of its MEGA package.
Of course, MEGA has picked some winners, as well. Shape Corp., one of the seven companies mentioned above under "Three Major News Events," has apparently prospered. While it’s true that the company failed to achieve the original projection of 400 new direct jobs by 1999, by 2003 it had created 462, exceeding their expected job creation total by 62.
Similarly, Lacks and Quicken Loans have managed to meet and exceed their expected job output (by 56 and 193 jobs, respectively). Magnesium Products, Meridian and Robert Bosch could also be seen as MEGA investments that have done well so far (though Robert Bosch has stumbled recently).
Still, MEGA’s success rate, as we found earlier in "MEGA Job and Project Performance," is not very high. The cases above indicate that MEGA’s miscalls have included some of the better known projects that might have sent positive signals about Michigan’s business climate if they had proved successful.
Michigan’s Economic Climate Under MEGA
MEGA’s broader goal in stimulating job development is to improve Michigan’s economy. Michigan’s broader economic record since 1995, however, has not provided clear evidence of the program’s success.
From December 1995 through December 2004, Michigan finished 50th out of 50 states in percentage employment growth. Even focusing only on private-sector employment growth from December 1995 to December 2004 (thereby excluding public sector employment which MEGA does not directly affect), Michigan placed 50th in the nation. (Michigan ranked 50th in the same category from 2000 to 2003.)
Other economic measures seemed little better. From 1993 to 1997, Michigan’s percentage increase in per-capita gross state product was 18th in the nation, but from 1998 to 2003, it had fallen to 44th (a discontinuity in the methodology of this federal metric between 1997 and 1998 necessitates this temporal division of the data). From 1995 to 2003 (the most recent data available), Michigan’s per-capita personal income growth was 43rd in the United States.
Nor have Michigan’s recent job figures been encouraging. In December 2004, Michigan’s unemployment rate was 7.3 percent, tied with Alaska for the worst in the nation. Michigan and Ohio were the only two states to lose jobs in 2004, and unlike Ohio, Michigan lost a significant number — 46,500, as opposed to Ohio’s 200, according to the U.S. Department of Labor.
A January 2005 United Van Lines study further suggests that Michigan is experiencing a net emigration. The company’s annual survey of moving figures found that in 2004, Michigan was one of only 11 states in the continental United States that qualified as "high outbound" — i.e., a state in which more than 55 percent of the moves handled by United represented an exit, rather than an entrance. Michigan’s outbound traffic was 61 percent of its total, the highest percentage in the Great Lakes State since 1982.
Summary of Findings
The evidence suggests that the MEGA program has fallen well short of its stated goals on several levels.
As noted earlier, about 44 percent of its eligible projects have claimed tax credits since the authority’s inception. In only 10 cases — about 8 percent of the total — did MEGA projects achieve their estimated job totals on schedule.
At the same time, direct job creation in the MEGA program appears to have lagged, reaching only 38 percent of MEGA’s original projections. This 38 percent represents 13,541 jobs, or about 0.3 percent of Michigan’s overall job count. MEGA’s underperformance in job creation is mirrored by the finding that only one-third of the dollar value of the originally expected SBT tax credits granted has actually been claimed, and by the fact that the projected rate at which these credits are expected to be claimed has dropped in recent years.
Direct job figures do not account for all of the jobs that can be attributed to MEGA packages; indirect jobs have probably been created, as well. Still, the low success rate in MEGA’s direct job creation suggests a similarly low success rate for indirect job growth, since spin-off job growth is driven in part by the activities of employees in the jobs created directly.
The case studies we discussed also suggest that there are two sides to the issue of the publicity generated by MEGA deals. While successes in the case studies described above possibly could have encouraged other businesses to consider new investment in Michigan (at least with the aid of a MEGA package), the failure of MEGA’s deals could likewise send the signal that Michigan is not a good place to do business. If firms cannot create jobs even when they are offered tax credits and other state and local incentives, the implicit public message about the state’s business climate is probably negative.
Another interesting observation arises from the case studies. State officials often justify the granting of tax incentives by noting that a MEGA recipient must achieve its job goals in order to actually receive its MEGA credits; if the company fails, the state grants no credits and forgoes no revenue. The Kmart deal, however, shows a wrinkle in that view: Kmart received $6.1 million in state tax relief for temporarily creating jobs that now no longer exist. It was also offered subsidies to train workers that may no longer be employed at the corporation, and it obtained infrastructure improvements to widen roads that probably won’t have the projected traffic. There was a cost to MEGA’s investment in Kmart, and it is not obvious that this cost was worth the temporary jobs that were promoted by the plan.
Moreover, even in cases where a firm never manages to receive state tax credits, it often receives the local incentives anyway. (R.J. Tower in Delta Township is one example of this dynamic, having begun to collect a local property tax abatement, even though the firm’s poor jobs record means it has not claimed MEGA SBT credits.) This fact, together with observations in later sections of the study (including "The Problem of Ensuring MEGA Credits Are Necessary") suggests that MEGA is not as "cost-free" as it is sometimes described.
Michigan’s economy has not shown obvious signs of strength in response to 10 years of MEGA investment. In fact, all major indicators suggest slow growth, including state employment growth. It would seem that MEGA’s 127 projects should have been able to influence the state’s economic growth during this period if the assumption on which the program is based were sound.
Of course, it is conceivable that Michigan’s jobs and economic performance during these years would have been worse without the MEGA program. We thus undertook a more detailed econometric investigation in order to determine whether MEGA might have buffered the state (and its counties) during an economic downturn. This investigation appears in the next section of the study, "Econometric Evaluation of MEGA’s Effectiveness."