Also I am not sure I get why this signal is supposed to be nearly foolproof. Doesn't it just rely on the wisdom of the crowd. If most traders think there will be a recession then the yield curves inverts, but the majority could still be wrong? Or is there another mechanism which inverts the yield curve?
Inflation and deflation are tied to recessions because less economic activity, meaning lower demand for goods and services, leaves companies with surplus goods. To make up for the excess in supply and stimulate demand, they'll deflate the prices. At least that is conventional wisdom.
There are a lot of variables that go into bond rates (much more so the back end of the yeild curve); current central bank rates, expected central bank rates, inflation, credit, etc. In the investor pool you have a lot of different parties with different motivations; pensions, governments, corporates, short term traders etc. Add to these factors the insane amount of liquidity and changes to the yield curve do have some meaning.
A good answer to this question could easily be a finance or economics dissertation and I’m pretty sure a bunch of them are out there. Economics and finance are complex non linear systems with feedback. When a rule is uncovered, it changes the system in ways that can invalidate the rule. The answer, thus, is it depends.
Examples:
1. In the late 70s inflation in the US was raging largely due to the cost of fighting the Vietnam war while pursuing LBJs Great Society and not raising taxes to pay for it. To fight it, the fed raised short term rates to 20% which had the effect of creating deflation in the since that it brought inflation down rapidly.
2. The US had an inverted yield curve for several years in the early 90s due to Greenspan trying to keep a lid on the punch bowl. It never caused deflation, but instead assured investors that inflation was under control.