Why Tangier made cars while Wisconsin waited for LCDs

From Wisconsin to Ethiopia, large incentive packages and ready-built factories have repeatedly fallen short of creating durable industrial ecosystems. Morocco and India show the more durable formula: build the operating system first, then pay for verified production, exports, jobs and supplier depth.

In 2017, Wisconsin was promised a $10 billion display-manufacturing complex from Foxconn, the Taiwanese electronics manufacturer. The project was meant to create 13,000 jobs and seed an LCD supply chain in Mount Pleasant.

It was the kind of announcement politicians like because the number appears to settle the argument before the hard work begins.

By December 2024, Foxconn had invested nearly $717 million and created 1,242 verified jobs. It had earned $62.9 million in credits. The LCD ecosystem did not arrive as advertised.

Now put that against a factory 4,000 miles away. Renault’s Tangier plant in Morocco began production in 2012 and made 312,669 vehicles in 2024. That same year, Tanger Med, the port complex serving Morocco’s automotive industry, exported 368,843 vehicles from Renault’s Tangier and Casablanca plants.

The contrast arrives early because it changes the whole story. Foxconn had an enormous announcement attached to a particular corporate plan. Renault had a working route from factory floor to 24 countries.

The Wisconsin site endured, but the original LCD manufacturing promise did not. Photo: Chad Davis from Minneapolis, United States / Wikimedia Commons, CC BY 2.0.

Wisconsin reset its agreement with Foxconn in 2021. Credits were tied to actual jobs and investment, reducing taxpayer exposure by $2.77 billion. In November 2025, the state approved a smaller expansion: $569 million and 1,374 additional jobs, with up to $16 million in new credits, focused on data infrastructure rather than the promised LCD complex.

The serviced site and infrastructure could still be useful. The original story could not be forced back into existence with a bigger cheque.

So what makes a factory stay once the incentive has done its work?

The belief that refuses to die

The conventional view is simple. Offer the right tax holiday, discounted land, grant or industrial loan and a region can outbid its rivals for a strategic factory.

There is some truth in that. If two sites already work, an incentive can tip a close decision. It can fund training, machinery or the early months of production.

But governments often treat incentives as the product rather than the closing argument.

The OECD’s 2026 practical guide to investment tax incentives reaches a blunt conclusion: tax relief is unlikely to overcome instability, poor infrastructure or skills shortages. Its investment-promotion survey found that infrastructure, workforce, markets and the legal and administrative environment ranked above incentives.

Imagine you run a factory. A tax reduction is welcome. But it does not keep a production line running during an outage. It does not move parts through a port. It does not train a technician or make a missing component appear on Tuesday morning.

Those are not side issues. They are the business.

This is where most people stop looking. They compare one subsidy headline with another instead of asking whether the factory can make, move and sell its product at a cost that survives outside the press release.

Morocco built the machine around the factory

Tanger Med’s vehicle terminals can process one million vehicles a year and connect directly to 24 countries. Renault’s Tier-1 supplier count, meaning companies supplying major systems directly to the carmaker, rose from 26 in 2015 to 76 in 2020. Local integration exceeded 60%. Morocco also established a dedicated automotive training institute beside the Tangier plant.

Tanger Med gave Renault something tax breaks cannot provide: repeatable access to export markets. Photo: Adam Cli / Wikimedia Commons, CC BY-SA 4.0.

At first glance, this can look like another story about export-zone benefits. Morocco did offer them. But firms needed authorisation and at least 70% of turnover from exports to qualify.

That condition matters. The benefit sat on top of a logistics machine: factory, rail, port, training and suppliers. Renault’s continuing model allocations, including hybrid production from 2024, suggest an operation embedded in a system that is difficult to replace.

Renault’s Tangier operation grew alongside suppliers, training and port capacity. Photo: Mert Dinçer / Pexels, Pexels licence (free commercial use).

Put Morocco beside Wisconsin and a pattern appears that neither announcement says outright. Foxconn was tied to a promised product and scale. Renault was woven into physical and human capabilities that made more production practical.

One bought attention. The other reduced operating risk.

For a manufacturer searching for every incentive available, that should reverse the order of work. Test power, logistics, labour, permits, suppliers and customer access first. Then calculate whether an incentive closes the remaining gap. Starting with the tax credit is like choosing the colour of a lifeboat before checking whether the ship floats.

Cheap labour was not cheap enough

Hawassa Industrial Park in Ethiopia opened in 2016 with projections of 60,000 jobs and roughly $1 billion in annual exports. By early 2019, it employed about 25,000 people.

It had modern factory sheds, low wages and export-oriented manufacturing conditions. On paper, it had much of what an apparel company might want.

PVH, the US apparel company behind Calvin Klein and Tommy Hilfiger, operated a plant there. A World Bank policy brief recorded that PVH’s operation was running at only 25% to 30% efficiency in late 2018.

The problem was not one missing tax break. The International Labour Organization reported annual labour attrition around Hawassa of approximately 100%.

A 2025 World Bank review adds an important complication. Hawassa had not disappeared: 15 active companies and 18,000 workers remained. Yet firms still relied on diesel generators during outages.

Modern factory space could not by itself solve Hawassa’s productivity, retention and power constraints. Photo: Amein Eskinder / Pexels, Pexels licence (free commercial use).

So this is not a clean tale of collapse. It is more useful than that. Ready-built factories created activity, but did not automatically create productivity, worker retention, dependable production or lasting export economics.

Low wages can mislead because wages are only one part of unit cost, the cost of making each item. If workers leave rapidly, output is weak and diesel generators cover outages, the apparent labour advantage can evaporate. A cheaper worker is not cheaper if training has to start again every few months.

For a closer look at why industrial parks need evidence before they seek development capital, see GI Network’s analysis of East African industrial parks.

An assembly plant is not yet a car industry

Saudi Arabia’s AMP-2 plant, operated by Lucid, the US electric-vehicle maker, opened in September 2023. The Saudi package offered Lucid access to up to SAR5.19 billion, approximately $1.4 billion, through industrial loans, milestone-based capital support and a government agreement covering at least 50,000 and potentially 100,000 vehicles over ten years.

That sounds like a fully formed automotive bet. It is not, at least not yet.

Lucid’s 2025 SEC filing described operating capacity of 5,000 vehicles a year at AMP-2, where kits manufactured in Arizona were being reassembled. The planned 150,000-unit fully integrated capacity was still under construction.

That is not a criticism of early assembly. It may be sensible sequencing while Saudi Arabia builds technicians and suppliers around Lucid; Ceer, the Saudi electric-vehicle brand; Hyundai, the South Korean vehicle manufacturer; and Pirelli, the Italian tyre maker.

Here is the twist: the contract design is more realistic than the headline. Support can be suspended if operational milestones are missed. Abandonment or chronically low use can trigger transfer of the plant or repayment of amortised support.

The policy is quietly admitting what headlines conceal. Capability takes time. Public support should not pretend that capability already exists.

India proves incentives can work, then asks a harder question

India offers the strongest case against the lazy claim that incentives never matter. Its smartphone production-linked incentive programme paid companies after verified incremental output, rather than primarily before construction.

By March 2026, large-scale electronics beneficiaries reported ₹20,587 crore of investment, ₹11.62 trillion of production and ₹6.43 trillion of exports, above programme targets. Smartphones became India’s largest individual export commodity in FY2025-26, according to the Government of India’s Press Information Bureau.

The programme accelerated Apple-linked and other mobile production in a country that already had large labour pools, domestic demand, established electronics locations and export infrastructure. The incentive landed on ground that could carry it.

But the story has a second act. Official estimates placed domestic value addition at 18% to 20%. An external evaluation for FY2023-24 put it at 23%.

That distinction is the whole point. India has proved it can scale production. It has not yet proved that every crucial component, engineering capability and slice of supply-chain value will follow automatically.

The test of a manufacturing cluster is not whether the first line opens. It is whether the second, third and tenth supplier have a reason to locate nearby.

The price of paying for proof

Investors and taxpayers often hear that milestone-based support is slower and more demanding than an upfront subsidy. It is. That is precisely why it is useful.

The evidence does not provide a clean like-for-like comparison of every incentive package, and anyone claiming one is selling false precision. Yet the public downside in these cases is strikingly visible. Wisconsin’s contract reset reduced taxpayer exposure by $2.77 billion because credits were tied to realised employment and investment. Germany offered €9.9 billion towards Intel’s planned Magdeburg expansion, but paid €0 after Intel delayed the project in September 2024 and discontinued the German expansion in 2025.

Lucid’s SAR5.19 billion package is available, not automatic. India’s electronics payments followed verified incremental output. Four different structures, one basic idea: do not transfer the full risk to the public before the factory has shown it can perform.

That changes the financing psychology. An upfront promise can make a project look cheaper before it has earned anything. A milestone condition leaves more early risk with the company, which makes the initial proposal tougher to finance. But it also makes management prove that demand, operations and execution are real.

Even €9.9 billion has limits

Germany offered €9.9 billion towards Intel’s planned investment of more than €30 billion in Magdeburg, with 3,000 direct jobs projected. Intel delayed the project in September 2024 and discontinued the German expansion in 2025.

Its annual report cited reassessed demand, financial constraints and a requirement for clearer returns on capital. Germany confirmed that no Magdeburg subsidy had been paid.

This is where the argument becomes more honest. Reliable infrastructure is necessary, but it cannot make an uneconomic corporate investment economic. A factory also needs customers, a financially capable parent company and a return that management can defend against other uses of its money.

Germany’s conditional approval did something quietly valuable: it protected the state from paying for an announcement that no longer made commercial sense.

What operators should prove before asking

A company seeking incentives should not pitch only jobs and capital expenditure. Those are promises. A credible application shows why production can continue after public support ends.

Start before a location is publicly committed, when construction choices and supply routes can still change. In Morocco, eligibility rested on export performance. In India, payment followed verified incremental output. In Wisconsin, the reset tied awards to realised jobs and investment.

The programmes are different, but the discipline is the same: prove performance before collecting the full benefit.

An operator should make four things plain. What will be verified: production, exports, investment, jobs or a combination. What local depth means: suppliers, trained workers or locally integrated processes. What happens if targets are missed. And whether the site can function through ordinary risks, especially power, logistics and labour turnover.

That is why founders should not treat advisers as a route to a larger subsidy. Their more valuable role is turning vague claims into evidence that can survive scrutiny.

GI Network’s view: The strongest incentive package is not the biggest one. It is the one that pays only after the factory has demonstrated that it can operate, employ and export under real conditions.

What investors see in the small print

For investors, subsidy conditions are not legal decoration. They change downside protection, expected returns and where capital should be allocated.

A payment linked to verified output reduces the chance that public money fills a hole created by a project that never reaches production. Wisconsin’s reset reduced taxpayer exposure by $2.77 billion. Germany paid no Intel subsidy after the project was discontinued. Saudi Arabia can suspend support if Lucid misses operating milestones and has remedies for abandonment or persistently low use.

Those protections matter to private capital too. If incentives are paid only after production, jobs or exports are verified, the company must fund more of the early risk itself. That can make a proposal harder to finance at the start. But it also distinguishes a business with an operating plan from one built around collecting support.

Experienced investors therefore ask a less glamorous question than, “How large is the subsidy?” They ask, “What has to happen before it is paid, and what happens if it does not?” They then compare that answer with the project’s power exposure, export route, supplier dependence, demand assumptions and workforce plan.

GI Network would test the operating case behind a proposed industrial investment: the power and logistics exposure, labour and supplier assumptions, export route, approvals and evidence supporting each. We would then align the capital plan and incentive application with measurable milestones, map the objections an investment committee will raise, and structure the case around what can actually be verified rather than what can merely be announced. That is the same discipline behind our guide to funding the evidence a factory needs.

Use the STICK test

Before valuing any manufacturing incentive, use the STICK test. Score each category from 0 to 2 before assigning any value to public support.

0 means unproven. The application relies on an intention, an announcement or an untested assumption.

1 means partly evidenced. There is a plan or early capability, but a material dependency remains unresolved.

2 means operating evidence. The capability exists, is contracted or has been demonstrated under real conditions.

S: Supply chain. Are critical suppliers present, or is there a believable route to bring them in?

T: Throughput. Can power, transport, customs and the port keep materials and finished goods moving every day?

I: Investment logic. Does the company still have demand, financial capacity and a return case without assuming the subsidy solves everything?

C: Capability. Are workers, managers and technical training available, and can the operation retain them?

K: Keep-it-honest conditions. Are payments linked to verified output, jobs, investment or exports, with clear consequences if the project falls short?

A score of 8 to 10, with no zero in Supply chain, Throughput, Investment logic or Capability, means an incentive may be solving a narrow and credible gap. A score of 5 to 7 means fix the weak operating link before negotiating the package. Below 5, do not price the incentive into the investment case. And a zero in any of the first four categories is a stop sign, however attractive the subsidy looks.

If a location fails the first four, the fifth cannot rescue it.

Incentives should make a good factory decision better. They should not be asked to impersonate one.