Has trade protectionism ever actually worked?

Sometimes — but far less often than its advocates claim, and almost never in the blunt "slap on tariffs to save jobs" form. The honest answer is: it has worked in specific, disciplined cases and failed badly in most others. The difference is how it was done.

The cases usually cited as successes

• Postwar East Asia (South Korea, Taiwan, Japan). The famous "infant industry" wins. They protected nascent industries (steel, shipbuilding, electronics, autos) — but crucially paired protection with brutal export discipline: firms had to compete in world markets to keep their support. Subsidies were conditional on performance, and withdrawn from losers. That's the key feature people forget.
• 19th-century US and Germany. Both industrialised behind tariff walls. Real — but they were also huge internal markets with strong institutions, and it's genuinely debated how much the tariffs caused the growth versus coincided with it.
• Targeted strategic/temporary cases. Some defence-critical or genuinely-infant sectors have been nurtured to viability and then exposed to competition.

The common thread in the wins

Protection worked when it was: temporary, conditional, and export-oriented. It bought time for an industry to reach competitive scale, with a hard requirement to actually become competitive — and a sunset. It was industrial policy, not just a tariff.

Where it fails (most of the time)

• No sunset. Protection meant to be temporary becomes permanent. The industry never has to get good because it never has to compete. (Much of Latin America's mid-20th-century import-substitution went this way — initial growth, then stagnation and inefficiency.)
• Protecting the mature, not the infant. Shielding a declining industry from competition preserves jobs briefly but raises costs for everyone downstream and delays adjustment. Steel and textiles in rich countries are the classic examples.
• Retaliation. The other side responds, and you lose export markets. Smoot-Hawley in the 1930s is the cautionary tale — it deepened the collapse in world trade.
• Concentrated benefits, diffuse costs. A few protected producers gain a lot; millions of consumers and steel-using manufacturers each lose a little. The losses usually exceed the gains in aggregate; they're just spread thin enough to be invisible.

So, the verdict

Protectionism can work as a temporary scaffolding for an industry that has a real path to global competitiveness, if the state is disciplined enough to demand results and remove the support on schedule. That's a high bar most governments fail to clear, because the politics push the other way — toward permanence and toward protecting the politically powerful (declining incumbents) rather than the economically promising (genuine infants).

As a permanent shield for mature or uncompetitive industries, the record is poor: it mostly makes a country's own downstream firms and consumers poorer while delaying the adjustment it was meant to prevent. The successful version looks much more like conditional industrial policy with a stopwatch than like tariffs for their own sake.