The best plants aren't in another country. They're in another decade.

Jim Farley said Ford's best factories aren't in America. He named plants in Mexico and China. Around the same time, China had overtaken Japan as the world's largest auto exporter. Commentary was immediate — wages, unions, trade policy, currency. All the comfortable arguments. Nobody asked the one question that actually matters: how old is the quality system running inside those walls?

I have run quality in plants on both sides of this equation. Greenfield sites where I drafted the first control plan before concrete was poured. Legacy sites where the process flow hasn't matched the PFMEA since someone's now-retired predecessor last touched it. The output gap between these two categories isn't geography. It is the volume of unvalidated drift between what the certificate says and what the floor actually does.

The variable nobody is measuring

The Mexican and Chinese plants Farley admires are newer. Built under current IATF 16949 expectations, with APQP discipline embedded from the first shovel. That is not a national advantage. It is an architectural one.

Two decades crossing borders in this industry — European plants, Airbus North America, greenfield launches, inherited messes. The pattern is consistent. Plants built in the last ten years start with a quality architecture designed before the first part. Plants built in the last forty start with one designed before anyone currently working there was hired. Both carry the same certificate. Only one is still running the system it was certified for.

What greenfield actually buys you

When I built the QA/QC department for SNOP's greenfield plant — 900+ employees, launched from bare ground — I had something most quality leaders never get. A blank control plan. Not a revised one. Not an inherited one with seven layers of redline and a footnote from 2016. Blank.

We designed PFMEA from scratch, mapped every failure mode before the first part ran, and wrote QRQC protocols into shift handover before the first shift started. Mistake-proofing logic went into the tooling specification — not bolted on after the first 300 PPM escape reached a customer.

That is what greenfield buys you. Not cheaper hands. Zero quality debt. No operator who was told "just mark it and we'll sort later" during a launch crisis and turned that workaround into a permanent habit. No gauge relocated during a capacity crunch that never moved back. No supplier deviation that became a standing exception because nobody remembered to close it.

We hit 70% defect-cost reduction in that plant. Not because the location had a secret manufacturing advantage. Because we designed the quality system before the first part, not after the first customer complaint.

The quality debt no balance sheet tracks

Every legacy plant carries what I call quality debt. It functions exactly like technical debt in software — every workaround, every informal adjustment, every "we've always done it this way" is a small loan against future compliance. It compounds.

IATF 16949 and AS9100 certificates tell you none of this. A plant can be fully certified and still carry years of process drift behind the paperwork. I have audited enough of them across Europe and North America to recognise the pattern. The control plan says one thing. The floor does another. The gap is bridged by tribal knowledge, nicknamed shortcuts, and operators who quietly compensate for engineering failures the engineers don't know about.

You walk the gemba and find a station where the documented cycle time hasn't matched reality for three years. A poka-yoke bypassed during a launch and never re-engaged. A supplier PPAP approved conditionally in 2019 with the condition never verified. None of it shows up in a document audit. All of it shows up in your warranty numbers, your escalations, your cost of poor quality — the metric everyone tracks and nobody volunteers to present.

When I reduced EASA audit findings by 50% in one cycle, it was not by adding documentation. It was by closing the gap between documented process and actual process. Sometimes that meant rewriting the document. Sometimes it meant fixing the process. Rarely did it mean adding a new form.

A certificate proves you built a quality system once. Your scrap rate proves whether you still have one.

Don't build a new plant. Re-launch the one you have.

The instinct when you hear "our best plants are in another country" is to consider moving. That is the expensive answer to the wrong question.

The right question: what would happen if you treated your existing plant as a new product launch? Run full APQP on your own processes. Revalidate every control plan against what actually happens on the floor. Pull the PFMEA out of the QMS and walk it station by station with the operators whose names aren't on the document but whose hands are on the process. Re-examine every standing deviation and every legacy grandfather clause with the scrutiny you'd apply to a new supplier's first PPAP submission.

This is what I did leading quality across a 2,000+ multi-site workforce. Not as a project with a start and end date. As an operating model. The plants that sustained results were the ones that adopted launch discipline for running processes — treating every discovered gap like a nonconformance, every workaround like an open 8D, every undocumented change like an audit finding waiting to happen.

Farley named the wrong variable. Wages don't explain why a five-year-old plant outperforms a forty-year-old plant carrying the same certificate. System age does. And system age is the one competitive variable you can change without moving a single machine — if you're willing to admit that your compliant plant and your capable plant are not the same thing.