Different Assumptions about Foxconn Job Creation Yield Much Higher Estimates of the Cost of Jobs

August 3, 2017

pdficonPDF version     pdficonPress release

The cost of the proposed new tax credits for the tentative deal with Foxconn could be far larger per job created than some people have suggested.  Those costs will vary greatly depending on the ratio of spending for payroll versus the capital expenditures.  The more that Foxconn invests in its facilities and state-of-the-art automation, rather than employees, the more the proposed deal will cost state taxpayers per job created.

The following report describes straightforward calculations of the tax credit costs for four different scenarios – each of which is consistent with the memorandum of understanding between the state and Foxconn. In brief, we found that:

  • The cost to the state of the new tax credits (excluding other subsidies) could range from about $219,000 per job for 13,000 Foxconn jobs to $587,000 each if the new enterprise employs 3,000 people.
  • On an annual basis, the proposed credits would cost state taxpayers $17,273 per job per year for 15 years under the scenario of 13,000 Foxconn jobs described by state officials, but the cost could exceed $54,000 per job per year under a scenario of 3,000 jobs that last for 10 years rather than a minimum of 15 years.
  • On top of the very substantial cost of the new tax credits, there’s the cost of existing tax breaks for manufacturers (which would already essentially zero out Foxconn’s corporate tax liability), plus the costs of huge local infrastructure investments, the cost of prioritizing work on the I-94 corridor, potential costs to state taxpayers from making a guarantee to cover up to 40 percent of local losses from spending for the project, and the unknown costs from unprecedented exemptions from environmental regulations.

Calculating the new costs

Before we get into the details of our analysis, it’s important to clarify a few general points because many people have misimpressions about the proposed subsidies for Foxconn.

The new tax credits are simply state payments to Foxconn, not tax abatements

Some people hearing the initial descriptions of the proposed deal probably got the impression that the new state “tax credits” that have been proposed – which will pay up to 17 percent of payroll costs and up to 15 percent of capital expenditures – will be offsets against the corporate income taxes that will be owed by Foxconn. That is not the case; Wisconsin currently has two very generous tax breaks that make it highly unlikely that manufacturers like Foxconn will owe any corporate income tax on their net profits.  (See the note about those tax breaks.)

As a result of the already very generous tax breaks for corporations like Foxconn, the new tax credits would have very little to do with the state tax code. Instead, they would simply be large checks written each year by the Department of Revenue to Foxconn – totaling up to $2.85 billion over the next 10 or 15 years.

In addition to that $2.85 billion, the proposed new tax breaks for Foxconn include $150 million for a sales tax exemption for material and certain services relating to the construction of the new manufacturing facility.  In contrast to the new corporate income tax breaks, the sales tax exemption is an abatement of potential new revenue. Our analysis in this report focuses just on the two new tax credits that will cause a net drain on the state treasury.

Over the past few years, Wisconsin has phased in something called the Manufacturing and Agriculture tax credit, which provides a nonrefundable tax credit equal to 7.5 percent of net profits. That credit completely eliminates the business income tax for corporations that are structured in a way that passes the profits through to the owners. For other corporations it nearly eliminates their income tax liability, reducing it from 7.9 percent of net profits down to just 0.4 percent – regardless of whether they are increasing or decreasing their Wisconsin workforce. This tax credit would not only mean that Foxconn has little or no income tax liability in Wisconsin, it would do the same for Wisconsin manufacturers selling components to Foxconn.

In addition to that relatively new tax credit, a change to Wisconsin’s corporate tax law initiated in 2008 means that a manufacturer like Foxconn will not owe any state income tax if all its sales are to out-of-state entities. Because of that tax change, Wisconsin no longer determines how much of a business’ profits are taxable in our state by taking into account what portion of its employees and property are located here. The new formula for allocating profits is based entirely on the percentage of its sales that are in Wisconsin. If Foxconn sells all of what it makes in Wisconsin to a company like Apple located outside our state, then Foxconn would have no profits taxable here.

The proposed deal allows Foxconn to emphasize investments in robots over investments in employees

For Foxconn to receive the maximum amount of new tax credits, it would need to spend more than $8.8 billion for payroll and at least $9 billion for capital investments in its new facilities. However, as the proposal is currently structured, it does not preclude Foxconn from using the new capital expenditure tax credits to execute a business plan that stresses state-of-the-art robotics, with a much smaller investment in payroll. Under such a plan, Foxconn might receive a smaller amount of total credits, but the state would pick up a much larger portion of the cost for each new worker. We illustrate that in the scenarios that we describe below.

The total state and local costs for Wisconsin taxpayers could be far greater than $3 billion

The first round of media coverage of the proposal created the impression that the subsidies for Foxconn would be capped at $3 billion. It’s true that the new tax breaks are capped at that amount, but they are just part of the total costs that will fall upon state and local taxpayers.  Other costs or potential costs include the following:

  • Local infrastructure spending – Huge costs for infrastructure improvements, such as streets, sewers and possibly new school facilities, will be incurred by local government bodies. The cost of some of those improvements can be financed using bonding – by establishing tax incremental finance (TIF) districts that channel the increase in property tax revenue to gradually pay for the principal and interest costs. However, this approach has considerable risks for local taxpayers, especially under the new bill’s proposed changes to the TIF law, which envision larger public infrastructure investments that could take up to 30 years to pay off. In addition, TIFs often penalize school districts because during the life of a TIF district they are precluded from getting any increase in property tax revenue from that area (not even from the increase in property values likely to have occurred in the absence of the new development).
  • Interstate highway improvements – The proposed legislation authorizes the Department of Transportation to borrow $252 million for the I-94 North-South corridor in southeast Wisconsin.  The principal and interest for the new bonds would be paid from the state’s General Fund, at an estimated cost of more than $20 million per year once the bonds were fully issued.  The bill makes the new bonding contingent upon the receipt of federal matching funds that can be used to help finance the I-94 North-South corridor project.
  • $10 million for grants to local governments – The bill includes an appropriation of $10 million during the current fiscal year for grants to local governments for costs related to the Foxconn project.  Any unused funding could be carried forward by the state for use in future years.
  • A new state guarantee to bear a portion of local losses – An unusual provision in the Foxconn legislation expresses the state’s intent or “moral obligation” to pay 40% of the cost of losses incurred by local governments who finance costs related to Foxconn, if the local government financing had been approved by the Department of Administration Secretary.  Those losses could be quite substantial if Foxconn ended the project earlier than anticipated.
  • Training costs – There are likely to be significant costs for training workers who lack the skills needed for the new jobs; however the proposed bill does not currently set aside any new funding for that purpose.

In addition to all of those direct costs to taxpayers, there are other potential costs – particularly from the unprecedented exemptions the bill would authorize from environmental regulations. Those could have significant costs in adverse health effects and future environmental remediation expenses.

The proposed deal does little to protect taxpayers if Foxconn pulls out of Wisconsin

Proponents of the deal with Foxconn have asserted that it contains protections for Wisconsin taxpayers, and those assurances may have been a significant factor in winning over some of the lawmakers and commentators who had doubts about making such a large state commitment to Foxconn’s “Project Flying Eagle.” But if “Flying Eagle” takes flight from Wisconsin, the current legislative proposal and “memorandum of understanding” (MOU) between Foxconn and the Governor would do little to keep the state’s expenditures prior to that date from having been a huge expense for state and local taxpayers with little return.

The proposed bill allows The Wisconsin Economic Development Corporation (WEDC), which is not directly accountable to Wisconsin voters, to do two things to protect the public’s investment in Project Flying Eagle:

  • First, WEDC is directed to revoke Foxconn’s certification to receive additional tax credits if it moves its operations outside the area of the new project, ceases operations for a year, or submits false or misleading information to receive tax benefits.
  • Second, WEDC “may require a business to repay any tax benefits the business claims for a year in which the business failed to maintain employment levels or a significant capital investment in property required by an agreement between the business and the corporation.”

The second item, which is often referred to a claw-back provision, is the most important of the potential protections for the public’s substantial investments in Project Flying Eagle. As things stand now, that provides no safeguard for the public investments because the current MOU merely says that Foxconn will invest “up to $10 billion” and will create “up to 13,000” jobs.  However, the claw-back measure could become more useful if the state enters into a contract with Foxconn to create a minimum number of jobs, and that may be what WEDC intends.

A positive aspect of the proposal is that one of the two new tax credits is tied directly to job creation because it will increase in direct proportion to payroll (up to $100,000 per employee).  That’s a welcome contrast to the state’s Manufacturing and Agriculture Credit, which offsets corporate tax liability regardless of whether a business is increasing or slashing its jobs and payroll.

At least there appears to be some degree of accountability for a corporation receiving the proposed new tax credit. Nevertheless, if Foxconn exits Wisconsin much sooner than anticipated, it seems unlikely that the claw-back language would allow the state or local governments to recover much of their massive investments in infrastructure for the Flying Eagle.

The cost per job will be far higher if Foxconn invests heavily in automation

The amount that Wisconsin spends on the new tax credits for each job created will depend on the ratio of spending on payroll versus Foxconn’s capital spending for its new manufacturing facilities in Wisconsin.   To get a handle on what the spending per job might be, we performed some straightforward calculations for four hypothetical scenarios:

Scenario 1 – Foxconn spends $9 billion over a 6-year period for capital investments in its new enterprise, and employs 13,000 full-time workers (earning an average of $53,875). The number of workers begins at 3,000 in 2020 and grows steadily to 13,000 in 2025, and then remains at that level until at least 2034.

Scenario 2 – Assumes the same level of capital investments, but those investments are much more focused on automation.  The number of jobs quickly grows to 3,000 in 2020 and remains at that level until at least 2034.

Scenario 3 – Also assumes 3,000 new full-time jobs and $9 billion of capital spending, but Foxconn ends Project Flying Eagle in 2030, so there is no additional spending for the tax credit after tax year 2029.

Scenario 4 – Same as scenario #3, except the total capital spending is $5 billion rather than $9 billion.

In each of these scenarios we used some common assumptions. We assumed, for example, that the average salary would be $53,875 – as the Walker administration has asserted – and for the sake of convenience we assumed that all of the salaries would be in the range that qualifies for the new tax credit ($22,000 to $100,000 annually).   Because we did not build in any adjustments for inflation, we are essentially keeping all the amounts in 2017 dollars.

Table 1: New Tax Expenditures for Different Levels of Foxconn Jobs and Capital Spending

Scenario 1 Scenario 2 Scenario 3 Scenario 4
New Foxconn jobs 13,000 3,000 3,000 3,000
Duration of jobs 15+ years 15+ years 10 yrs (2020-29) 10 yrs (2020-29)
Total payroll through 2034 $8.9 billion $2.4 billion $1.6 billion $1.6 billion
Total capital investment $9 billion $9 billion $9 billion $5 billion
Total jobs tax credits $1.5 billion $412 million $275 million $275 million
Capital expenditure credits $1.35 billion $1.35 billion $1.35 billion $750 million
Total of the new tax credits $2.85 billion $1.76 billion $1.62 billion $1.03 billion
New tax credits per job $219,231 $587,381 $541,588 $341,588
New credits per job per year $17,273 $39,159 $54,159 $34,159


What our analysis illustrates is that even under the relatively optimistic assumptions about job creation that have frequently been suggested by the Walker administration, the cost of just the new tax credits would be roughly $219,000 for each of the 13,000 Foxconn jobs.  That amounts to an average of more than $17,000 per job per year, without factoring in the costs of other state and local subsidies.

If Foxconn creates a factory that uses state-of-the-art robots and only creates 3,000 new jobs, while still spending at least $9 billion for capital costs at the new facilities, the cost per job could exceed an average of $587,000 per year, but the total cost of the new credits would be significantly lower – about $1.76 billion over 15 years, rather than $2.85 billion.

Scenarios 3 and 4 illustrate the effects on spending per job per year if Foxconn were to end Project Flying Eagle at the beginning of 2030, after the new jobs had been in place for 10 years.  That change doesn’t significantly affect the total spending for credits for each new job created, but it would substantially increase the average annual spending for the credits per job.  Under scenario 3, the cost of the new tax credits would be more than $54,000 per job per year.  On the more positive side, the total cost of the new tax credits would drop to $1.62 billion.

If Foxconn were to terminate the project much sooner than expected – and we only raise that as a theoretical possibility, not as our prognosis – the biggest loser would probably be the municipality that takes on the very substantial infrastructure costs, under the assumption that the new debts will be paid off over the 30-year lifetime of a new TIF district.  However, the proposed legislation could also result in state taxpayers bearing 40% of the cost of local losses.


State and local lawmakers and Wisconsin citizens throughout the state need to take a very careful look at the potential costs and benefits of the proposed Foxconn deal, and we need to remove out rose-colored glasses before we analyze the risks and rewards.  Because this analysis only covers a small portion of the important factors to consider, we urge policymakers to proceed slowly and to provide ample opportunities for public input.

One positive aspect of the proposal is that new tax subsidies are tied to job creation and capital investments, and the subsidies will increase as the number of Foxconn jobs grows.  Nevertheless, the structure of the proposed deal would allow Foxconn to do very well by investing in automation, rather than trying to maximize the new jobs tax credits.

The proposed subsidies are far greater per job created than anything Wisconsin has offered in the past.  We need to think carefully about whether a corporation induced to locate here by subsidies of that magnitude is going to stay when those subsidies end, or when another state offers an even better deal. In addition, policymakers should think about how these subsidies would affect the demands of other companies who play the game of threating to locate elsewhere.

Finally and most importantly, we have to think very carefully about the effects of the proposed legislation on the state budget and whether there are more effective ways to invest billions of dollars of state and local tax revenue.


Jon Peacock