On July 19, 2023, an EF3 tornado tore through a Pfizer plant in Rocky Mount, North Carolina. It was not an overseas factory in a country with a shaky regulator. It was an American plant, in the American heartland, making sterile injectables for American hospitals. Pfizer and news reports at the time said the site produced nearly a quarter of Pfizer’s own sterile injectables sold in the United States, and roughly 8 percent of all sterile injectables used in US hospitals. Heparin, lidocaine, midazolam and naloxone were among its products.
The drug was local. The risk was still concentrated.
That episode is a useful corrective to the current enthusiasm for bringing pharmaceutical manufacturing home. Governments from Washington to New Delhi to Brussels now treat local production as the answer to supply fragility. The instinct is understandable after the pandemic. But the thesis of this piece is simple: location is not the same as resilience. A domestic single point of failure is still a single point of failure, and a local plant that nobody is contractually obliged to buy from will not survive the first price war.
What localization actually fixes
Local manufacturing does solve one real problem: political exposure. In a crisis, governments can restrict exports, and they did during COVID-19. A country that makes its own essential medicines cannot be cut off by a neighbour’s export ban.
That is worth something. It is not worth everything, because most shortages are not caused by export bans. They are caused by quality failures, plant accidents, a single supplier exiting a low-margin market, and demand spikes. None of those respect borders.
Sterile injectables show why concentration happens even without any foreign dependence. Making them requires aseptic filling lines, cleanrooms and validation work that take years to build and qualify. Once a line exists, the cheapest way to add volume is to run it harder rather than build a second one somewhere else. Economies of scale reward a few large sites. So when buyers push prices down, the industry responds rationally by concentrating output in fewer, bigger plants, and a single tornado, fire or failed inspection can then take out a meaningful slice of national supply. Rocky Mount was not an anomaly. It was the logical product of a market that pays for low unit cost and nothing else. Moving those same lines onshore, under the same purchasing rules, would reproduce the same concentration with a different postcode.
India’s lesson in the depth of the chain
India is often called the pharmacy of the world, and on finished generic formulations the label is earned. Further upstream the picture changes. Indian government statements and analysts, including the Observer Research Foundation, have for years put China’s share of India’s imports of bulk drugs and active pharmaceutical ingredients at around 70 percent, with even heavier reliance for some antibiotic inputs.
India responded in 2020 with a production-linked incentive scheme for bulk drugs, paying manufacturers to produce specific critical ingredients domestically. Some capacity has come online. But the dependence has proven stubborn, and the reasons are instructive. The Chinese advantage was never only cheap labour. It was clusters of chemical suppliers, cheaper utilities, effluent treatment capacity and years of process engineering. A subsidy can build a plant. It cannot quickly build an ecosystem.
The deeper point is that “local” has layers. A tablet pressed at home from imported ingredients, made from imported starting materials, is local in the way a car assembled from imported parts is local. Resilience depends on the most concentrated layer, not the most visible one.
The uncomfortable economics
Here is the trade-off that policy speeches tend to skip. Many of the medicines most prone to shortage are old, off-patent generics sold at very low prices. Buyers, whether hospital purchasing groups or national tender agencies, have spent decades driving those prices down, and they succeeded. Margins became so thin that manufacturers consolidated, cut redundant lines and stopped investing in quality systems.
A new domestic plant enters that market with higher costs than the incumbents it is meant to replace. If purchasers keep buying on lowest price, the new plant loses. It gets subsidised for a few years, then quietly closes or pivots to more profitable products. The country ends up where it started, minus the subsidy.
So the real question for any local manufacturing push is not “can we build it?” It is “who will commit to buy from it at a price that covers its costs, and for how long?” Without multi-year purchase commitments, local production is a photo opportunity.
A back-of-envelope on the price of reliability
Here is a stylised calculation, with invented numbers, to show why purchasing decides the outcome. Imagine a generic injectable that a national buyer purchases at 50 cents a vial, ten million vials a year. The total bill is $5 million. A domestic entrant with higher labour, energy and compliance costs needs 70 cents to break even.
Awarding the whole volume to the domestic plant at that price would cost an extra $2 million a year. Splitting the award, say 30 percent to the domestic supplier at 70 cents and 70 percent to the cheapest bidder at 50 cents, costs an extra $600,000. That is a rounding error next to the cost of a shortage, when hospitals substitute more expensive alternatives, delay procedures and pay pharmacists and nurses to manage the scramble.
The catch is that the $600,000 is visible on a budget line every year, while the avoided shortage is invisible because it never happens. Purchasers are measured on the first number. That is the whole problem in one sentence, and it is why resilience rarely survives the annual budget round unless someone writes it into the rules.
What resilience actually looks like
If localization is a partial answer, what is the fuller one? A few principles hold up better than flags on factories.
- Diversity over proximity. Two qualified suppliers in two different countries beat one supplier next door. Tenders that award the whole volume to the cheapest bidder destroy that diversity. Splitting awards across suppliers costs a little more and buys a lot of safety.
- Pay for reliability. Purchasers can score suppliers on their quality record, redundancy and inventory commitments, not just unit price. That creates a market reason to invest in the things that prevent shortages.
- Know your map. Many health systems cannot say where the active ingredient in their critical medicines is made. Mapping the chain down to the starting materials is tedious, and it is the precondition for every other decision.
- Regional over national. For smaller countries, a plant sized to national demand is often uneconomic. Regional manufacturing, with pooled procurement across neighbours, can reach viable scale. The African Union’s push for continental vaccine manufacturing rests on this logic, and its success will depend on whether African purchasers, and the agencies that procure on their behalf, actually buy the output.
The strongest case for building at home
Advocates of localisation have a better argument than the one usually made for them. It is not that domestic plants are more reliable. It is that in a true crisis, contracts stop meaning much. During a pandemic or a war, a government can requisition output inside its borders and cannot requisition output inside someone else’s. Diversity across countries only works if those countries keep exporting when it matters most.
That point holds, and it is why the strategic reserve comparison is apt. For a short list of products where a sudden cut-off would be catastrophic, paying a premium for capacity you control is rational, even if that capacity is uneconomic in normal years. The error is extending that logic to thousands of generic molecules, where the more common failure is a quiet market exit rather than a border closing. For those, a domestic plant solves the rarer problem and does little about the common one.
Second-order effects
Local manufacturing drives, if poorly designed, can make supply less resilient. Subsidised domestic entrants can undercut and push out existing foreign suppliers, reducing the number of qualified sources. If the new plant then struggles, the market ends up more concentrated than before.
There are regulatory side effects too. Every new site needs inspection capacity, and regulators with stretched inspectorates may find that a wave of new plants dilutes oversight just when quality is the point. A domestic plant with a weak quality culture is not safer than a foreign one with a strong one.
And trade policy can backfire. Tariffs meant to encourage domestic production raise the cost of imported ingredients for the manufacturers who are already local, squeezing the thin margins that caused shortages in the first place.
What each actor can do
For health ministries, the priority is to separate the two goals. Name the short list of products that justify strategic domestic capacity, fund it explicitly, and treat everything else through procurement reform rather than industrial policy.
For national tender agencies and group purchasers, the lever is contract design: multi-year awards, split volumes, and scoring that rewards verified redundancy and quality history alongside price.
For hospital supply chiefs, the work is mapping. Knowing which critical products depend on a single plant or a single upstream source lets a hospital hold more buffer stock where it matters instead of spreading inventory thinly across everything.
For funders and development banks backing regional manufacturing, the condition for money should be an offtake commitment. Financing a plant without a buyer is financing a future closure.
For manufacturers, the message is to make reliability legible. Buyers can only pay for redundancy they can verify, so disclosing sourcing depth and backup capacity becomes a commercial asset rather than a trade secret. A supplier that can show two qualified sources for its active ingredient, and spare capacity on a second line, has something to sell beyond price, but only if buyers ask for that evidence and reward it when they get it.
Where to push back
None of this means local manufacturing is a mistake. For some products, especially vaccines and a handful of critical injectables, domestic or regional capacity is a legitimate security investment, like a strategic oil reserve. The mistake is treating it as a general answer and funding the building while ignoring the buying.
Rocky Mount recovered. Pfizer said most lines had restarted by late September 2023. But hospitals spent months rationing, substituting and scrambling, and the plant’s American address did nothing to prevent it.
The takeaway for health ministers and hospital supply chiefs is this: before you pay for a factory, decide how you will pay for its output, and how many other factories you will keep alive alongside it. Resilience is bought through purchasing policy far more than through construction.
What to watch
In the next 12 to 24 months, watch whether governments that announced domestic manufacturing incentives pair them with purchase commitments, or leave new plants to compete on price alone. Watch India’s progress on the bulk drug incentive scheme, where the real test is whether domestic ingredient production holds up without continued subsidy. Watch whether African regional manufacturing initiatives secure binding offtake from the agencies that buy vaccines for the continent. And watch the tender documents themselves. The clearest signal of change will not be a ribbon cutting. It will be a procurement rulebook that prices reliability.
Has your health system changed how it buys, not just where it buys, since the pandemic? We’d like to hear what has worked.
