Thread by @MichaelDFarren

I’m happy to announce research released today with @MattMitchell80, @jmhorp, and @OliviaJGonzalez.

This is a thread summarizing our findings.
mercatus.org/publications/g…

The study broadly analyzes targeted economic development subsides, using Foxconn in Wisconsin as a case study. It uses @TimBartik’s prev. research to build a cost-benefit analysis tool to evaluate the long-run effect.

Spoiler: It’s not good.

We illustrate how arguments in favor of econ. dev. subsidies (multiplier effects, positive externalities, & industry clusters) don’t hold much water, and how the analysis done by industry consultants is almost always a “benefits-only” approach (I.E.: ignoring the economic costs)

Most prev. research on the Foxconn deal was excessively optimistic, never accounting for the cost.

We find that WI could lose $20 billion in GDP due to the higher taxes needed to fund the $3.6 billion subsidy. Any econ. benefit from the subsidy has to counteract this loss.

But the econ. benefit from the subsidy also has to incorporate the likelihood that it didn’t actually do anything. Subsidies typically only sway such decisions in 2-25% of situations.

Most often businesses would make the decision to relocate or expand *without* the subsidy.

Even assuming a 50% prob. that the subsidy swayed Foxconn’s decision—2x more than the highest estimate of the typical subsidy’s efficacy (which seems unlikely given Foxconn’s utility needs and WI’s 0% corp. inc. tax for manuf. firms)—the net effect of the subsidy is *negative*

Over the next 15 years, we estimate that the Wisconsin’s subsidies will actually *reduce* state GDP by at least $370 million and potentially as much as $19.2 billion.

That’s for the Gen 10.5 LCD plant that Foxconn initially promised to build. It now claims to be building a Gen 6 plant that would be about 75% smaller. The expected economic losses don’t get much better, though, ranging from $1.2-6.0 billion.

Note, these estimates don’t include the $1.4 billion in other subsidies (local gov., utility, & infrastructure) which, due to methodology, couldn’t be part of the analysis.

So the actual econ. harm will be much larger—our findings should be considered an underestimate

This shouldn’t be shocking news—empirical research has long-found that subsidies have negl. or even neg. effects on the local economy.

A forthcoming literature review by @MattMitchell80, @jmhorp, & @OliviaJGonzales will illustrate the broad findings of academic research.

But the political understanding is quite different:

84% of surveyed mayors think that subsidies are good policy & @NateMJensen’s research shows that politicians believe that voters will reward them for “doing something” to help the local economy.

citylab.com/life/2019/02/b…

But our research shows that “doing something” is probably the WRONG thing.

There’s two ways that local leaders can go about encouraging local development:
They can take a “targeted approach” by directly promoting the development of particular firms or industries via subsidies.

Or they can take the “general approach” by creating an environment that is conducive to economic development for all businesses by generally-applicable tax, spending, regulatory, and legal policies.

In his Nobel prize lecture, F. A. Hayek described the targeted approach as an attempt “to shape the results as the craftsman shapes his handiwork” contrasting it with what he called the “environmental approach to growth.”

In fact, because the targeted approach to economic development often comes at the expense of other taxpayers and local residents, you could argue that the gardeners (politicians) are fertilizing some plants by composting others.

And in reality, governments’ attempts to artificially create economic growth can derail real growth from occurring! @MichaelEPorter

But while some subsidy costs can be estimated (with difficulty!), many more defy quantification.
These include deadweight loss,

anticompetitive effects, like X-inefficiency and dynamic inefficiency,

and rent seeking (pursuing politically-granted privileges), which creates a host of problems.

The distorting and corrupting effects of subsidies don’t just stay in the economy, though. They are a pathology that spreads to infect political systems as well.

We illustrate how mixed signals, perverse incentives, & politicians’ and voters’ lack of knowledge, combined with a lack of feedback mechanisms, means that accurately determining the “right” company to subsidize is practically impossible, assuming that such a unicorn even existed

Consider this fanciful but feasible scenario:
The $3.6 billion that WI promised Foxconn could instead have built 7 sq. mi. of greenhouses to attract orange growers from FL.

With benefits-only analysis, this might sound like a good idea!
@ggerlock npr.org/sections/thesa…

Even worse, the privileges that politicians grant can petrify the political landscape because the subsidized firms don’t want to give up their benefits—meaning that it’s hard to escape the resulting policy trap. In essence, we can get stuck in dysfunctional doldrums.

But that’s what our forthcoming research on political reforms to avoid the subsidy arms race is all about. More on that soon, but see my testimony in Utah for an initial glimpse:
le.utah.gov/av/committeeAr…

mercatus.org/system/files/f…

Lastly, thank you to a number of scholars
@TimBartik @NateMJensen @BellmanEquation @Richard_Florida
& reporters
@RickRomell @CorriHess @BruceMurphyMKE
@arthur8823 @reckless @JillRozell @AustinCarr

We extensively cited your work, without which our research wouldn't be possible

@TimBartik @NateMJensen @Bellmanequation @Richard_Florida @RickRomell @CorriHess @BruceMurphyMKE @arthur8823 @reckless @JillRozell @AustinCarr Please unroll, @ThreadReaderApp

@TimBartik @NateMJensen @Bellmanequation @Richard_Florida @RickRomell @CorriHess @BruceMurphyMKE @arthur8823 @reckless @JillRozell @AustinCarr @threadreaderapp Unroll @threadreaderapp