The Geography of Need and the Proposed Foxconn Deal in Wisconsin

The Issue:

The decline of manufacturing employment has reduced the availability of higher paying jobs for less educated workers. One approach states and localities have used to address the issue is by trying to attract and retain firms with packages of subsidies and tax incentives, such as the proposed high-profile deal for a new plant by the Taiwanese electronics supplier, Foxconn, in Wisconsin. Proponents argue that Foxconn will create a "high-tech manufacturing and technology ecosystem." Debates regarding such deals tend to focus on whether the benefits of the investment subsidy will be realized. However, there are also distributional issues to be considered. While the costs tend to be broadly shared by taxpayers, the economic logic of such “ecosystems” means that benefits will be spatially concentrated. Locations that may offer the highest payoff to a place-based subsidy, for example those with excellent infrastructure or a highly skilled labor force, may not be those in the greatest need in terms of poverty or unemployment.

Discretionary place-based subsidies of the kind on offer to Foxconn hold the potential to exacerbate rather than diminish regional disparities.

The Facts:

• Counteracting stagnant earnings — and decline — for the American middle class is most often cited by states and localities seeking deals to attract manufacturing jobs. One commonly held explanation for this decline holds that job opportunities for those lacking a post-secondary education have stagnated, even as college completion rates (particularly for men) have increased only gradually. Political responses to these long-term developments place much of the blame on the decline in manufacturing employment. At 17.3 million in 2000, manufacturing employment is now down to 12.4 million, or about 8.5 percent of total U.S. employment. Most reliable projections suggest that manufacturing employment in the U.S. will remain about the same over the near term. However, since 46 percent of the manufacturing labor force includes those with only a high school education (or less) but pay is about 20 percent higher than all industries, politicians have disproportionately focused on restoring manufacturing as a pathway to middle class prosperity.
• Building on the zero-sum strategy that southern states adopted in the 1960s to attract manufacturing from the Northeast and Midwest, state and local governments nationwide have committed over $94 billion in subsidies to foster economic development since 2000; of the top 100 programs, three-quarters of subsidies have been targeted at manufacturing. The arguments for subsidizing manufacturing rely upon job multiplier studies; recent studies find that one new job in manufacturing generates an additional 0.7 to 1.7 jobs in related industries and non-export sectors. Although some place-based subsidy programs such as enterprise zones target low-income areas, others are “discretionary” place-based grants that are specific to a particular firm or project at a designated location (see here and here for comprehensive reviews of these place-based policies). Many cases of discretionary grants, such as billions in support for Boeing from Washington State, are in response to threats by large employers to relocate production elsewhere. The $7 million granted Carrier by the state of Indiana in December 2016, is a recent example of such a package. A recent study of such place-based discretionary grants found little impact on employment and only modest gains in productivity. However, discretionary grants have an added political benefit that cannot be overlooked. A local mayor, governor, or other politician can claim credit for thousands of jobs “created” or “saved.”
• The economic rather than political rationale for a discretionary place-based subsidy is that it may correct for a market failure. Benefits of agglomeration, the returns to firms stemming from locating in close proximity to one another, can accrue to all related firms. For instance, high-tech firms that cluster close together in places such as Silicon Valley can be more productive by virtue of sharing an experienced labor pool, having ready access to specialized inputs, and sharing new ideas and technologies through face-to-face interactions. While one firm's decision to locate in close proximity to an existing productive cluster can raise the productivity of other firms, a firm’s investment and employment decisions will not take account of these spillover benefits. For this reason, market signals alone may result in under-investment or under-employment. In principle, carefully targeted subsidies could correct for this failure, but designing an efficient subsidy that makes the best use of tax revenues requires information on the size of spillover benefits, the geographic scope of a cluster, and the technologies involved. Injecting a large amount of new investment can raise productivity among existing manufacturing firms in a county. While there are no hard conclusions on the magnitude of these spillover impacts (or agglomeration benefits) on employment, there is considerable agreement that they are limited in geographic scope. The most sophisticated study of the job multiplier associated with new plant openings in the automobile industry finds virtually no net gain in parts supplier employment; what new employment for suppliers that did emerge was, on average, nearby the new plant. Another study found productivity gains for the cluster as a whole when new plants located in existing high-tech clusters, but the impacts died out after a decade and were limited to the county the high-tech cluster was located in. Existing studies also only focus on the impact of increments of investment to existing clusters. The unanswered question is whether such a deal could create a new “cluster” from the ground up. A well-respected study of the emergence of the highly successful Silicon Valley cluster finds the possibility unlikely, citing the “foolishness of directive public policy efforts to jump-start” creation of such clusters. Finally, note only one of these studies of benefits evaluates whether the cost of subsidies was worth it and finds they weren’t.
• As a place-based policy, the proposed Foxconn plant is not likely to provide many benefits to the areas in greatest need. It is sited in the prosperous southeast region of Wisconsin (the blue counties on the map). The two target counties are adjacent to counties with median household incomes of $79 thousand and above (over 40 percent above the state median). Those with limited education or work experience in the target area earn 8 to 20 percent more than the state median of $30,000 for less educated workers (see map). Although studies of manufacturing clusters find that their geographic extent varies widely by industry, most are limited to a radius of about 30 miles (see here and here). According to commuting patterns, about 84 percent of the labor force of the targeted area (Racine and Kenosha counties) lives in those counties; another 8 percent live in adjacent suburban counties. The remainder are expected to come from “the Midwest, the United States and around the world.” Residents of the 34 low-wage Wisconsin counties (in red or orange on the map) live for the most part from 100 to over 300 miles from the probable site of the Foxconn complex. State legislation recognizes that about two-thirds of these low-wage areas also qualify for economic development assistance as “Economically Distressed” (cross-hatched on the map) regions. The only economically distressed part of the state that will have any access is the City of Milwaukee (about 25 to 30 miles away).
• A complete assessment of public efforts to entice firm locations requires considering costs, which — as in the case of the Foxconn plant in Wisconsin — can be considerable. The Wisconsin State Assembly approved $3 billion in incentives to the Foxconn Technology group in exchange for construction of a $10 billion plant in the state in late August. The deal now goes to the State Senate. With an initial staff of 3,000, Foxconn projects that the plant will employ 13,000 workers by 2022. If enacted, the Foxconn incentive package would be the fourth largest ever offered a company by a state, and by far the largest offered for a new greenfield industrial development. One-half of the incentives ($1.5 billion) are “job creation tax credits,” which are paid to the firm per employee and year. Such credits now make up 40 percent of all subsidies to manufacturing firms in the United States. Foxconn job credits would be the largest such direct wage subsidy ever paid by an American state. The credits will cover 17 percent of the salary paid to any employee earning between $30,000 and $100,000 for the first fifteen years of the plant’s life; for the entire package, the total subsidy per job per year is estimated to be around  $17,000. The wage subsidy alone ($9,000) is over 3.5 times the national average of $2,500 per job per year. The new tax revenues anticipated for the project will only fully recoup the cost of the subsidies in 2043. Up through 2026, or the first eight years of the project, the net annual cost of the subsidies will grow to $200 million.
• Who benefits? Economic benefits will accrue to Foxconn, residents of the extreme southeastern part of the state, and those with high-tech degrees. The greenfield location will be chosen for its proximity to Milwaukee (to the north), Chicago (to the south), and access to the watershed of Lake Michigan (essentially the Interstate-94 corridor and east). These geographic constraints dictate that the new plant will lie in the far eastern portion of Wisconsin’s First Congressional District, home to the Speaker of the House, Paul Ryan. Although three-quarters of the projected Foxconn labor force will be production workers, more than one-sixth (2,400) will hold technical positions that require a degree in engineering or computer science, such as Process Equipment, Computer Integrated Manufacturing, and Integration Engineers. (The remainder will be business support.) Starting salaries for recipients of computer science and engineering bachelor's degrees who found work in manufacturing in 2016-2017 were about $58 to $62 thousand. For those with some experience, competitive salaries average $70 thousand and above. At a conservative estimate of an average annual salary of $75,000 for the 2,400 technical positions, at least one-quarter ($30 million) of the annual wage subsidies to Foxconn will go to these employees (or $450 million over the life of the project).
• Who will pay? Until the project breaks even in 2043, state taxpayers. Currently, about 2.85 million are employed in Wisconsin. For each employed person, the annual net cost of the Foxconn subsidies will rise to about $69 through 2026. Since state and local taxes absorb the largest share of income for those earning less than $136,000 (from 9 to 10 percent) it is likely that the burden of paying for the subsidies will fall most heavily on low and middle income taxpayers. About 250 currently employed residents will be required to cover the subsidy for the job of one Foxconn employee by 2026.