The idea isn't as crazy as it sounds — it's actually a debate that exists in mainstream economics. The core concept is the fiscal multiplier.

Here's the logic his claim rests on:

1. Welfare puts money into the hands of people with a high marginal propensity to consume — meaning they spend almost all of it immediately on rent, food, transport, etc.
2. That spending becomes income for businesses and workers, who then spend it again.
3. This creates a chain of economic activity that generates income tax, VAT, payroll taxes, corporate taxes, and reduced welfare spending (as people find work).
4. Some Keynesian and post-Keynesian economists argue that certain types of government spending — particularly targeted transfers to low-income households — can have multipliers above 1, meaning the total economic activity generated exceeds the initial outlay.

So, could the total tax revenue generated by that multiplier chain exceed the original welfare payout? In theory, yes — if the multiplier is high enough and the tax take is sufficient.

The problems with his claim

Verdict: Luminary or mad?

Neither extreme. He's repeating a real (but contested) economic argument, often associated with MMT (Modern Monetary Theory) and left-Keynesian circles. It's not mainstream consensus, but it's not delusional either.

If he presents it as obvious and settled truth, he's overstating. If he presents it as a debated hypothesis with some empirical support, that's fair.

The more interesting question is: under what conditions does this hold? And what's the evidence for those conditions in the specific case he's discussing? That's where the nuance lives.