Hey y’all, it’s Austin. Whether it’s Wisconsin’s much-trumpeted deal to get a Foxconn manufacturing hub, Tesla’s solar agreement with Buffalo, New York, or the imperiled bid for half an Amazon HQ2 in New York City, there’s one pattern: Advocates for the deals often claim they’re low-risk for taxpayers. As it turns out, that is rarely the case.

Over the past few months, for a Bloomberg Businessweek cover story, I’ve been buried in reporting on Foxconn’s $4.5 billion subsidy from Wisconsin to create a massive production facility and 13,000 jobs in the Badger State. After publishing the story, which details how the project has so far failed to live up to expectations, Wisconsin Senator Ron Johnson said in an interview on Bloomberg TV that the Taiwanese manufacturer wouldn’t get any of the promised government subsidies until they actually hit their job and capital expenditure targets. Former Governor Scott Walker, the chief architect of the Foxconn deal, has made the same defense. And countless state officials and project boosters made similar arguments to me during the course of my research.

Are taxpayers really protected in the event these types of corporate-welfare deals fall apart? In the case of Foxconn, it was perhaps the biggest myth of this government-incentive package.

Let’s start with the upfront costs. Taxpayers are already paying for Foxconn’s project. Racine County and Mount Pleasant, where the company is constructing a campus, have spent at least $130 million on related infrastructure costs. (Moody’s has even published cautionary notes on their mounting debt loads.) One estimate puts the state’s commitment at as much as $122 million for related road improvements. And then there are the ballooning public costs of the time that has gone into putting this deal together and monitoring its progress, which, as one economist recently noted, is “incalculable” at this point.

Officials can point to examples when they withheld incentives and spared taxpayers from certain burdens. Foxconn missed its maximum first-year hiring targets by 82 percent and forfeited its right to state tax credits for 2018. Perhaps the only thing more costly than Foxconn continuing to under-deliver is if it starts living up to its promises. The Wisconsin Legislative Fiscal Bureau, a nonpartisan government agency, reported that the state would not likely see a return on its investment until at least 2042. Newly elected Governor Tony Evers said the cash subsidies that would be paid to the company will come with their own budgeting challenges for the state, which could effectively be writing checks to Foxconn as high as $312 million per year in the decade ahead. That would bring the net cost of the project to a whopping $1.04 billion in 2032.

Some even argue that analysis is too sunny. It didn’t, for example, include the kinds of increased public-services costs associated with population growth. The state’s early analysis also made the unrealistic assumptions that every single one of Foxconn’s employees at the plant would live in Wisconsin and contribute tax revenue—rather than commute from nearby Illinois—and that Foxconn would rely wholly on suppliers located in-state—rather import some parts from abroad, as the company has already done.

Such suspect assumptions could seriously impact the crux of Foxconn’s whole deal with the state: the so-called “multiplier effect.” The hope was that once Foxconn set up its so-called Wisconn Valley, the campus would create a massive economic ripple that would make up for the public spending. Wisconsin’s incentives are projected to cost at least $219,000 in tax breaks and other incentives per job. If the ripple doesn’t prove as large as anticipated (or if Foxconn scales back its commitments), that could have untold consequences for the state.

If Amazon does decide to pull back on its plans in New York City, as it’s reportedly considering, the people of Queens may find it to be an unexpected gift. —Austin Carr