[hire_charts]

Can't you essentially fabricate this statistic by controlling for the amount of time that counts as "before", assuming the inverted curves happen more often than recessions?

[mannykannot]

I am guessing that "before" means that the inversion persisted until the economy was technically in recession.

[tfehring]

Not quite - see https://fred.stlouisfed.org/series/T10Y3M for reference. The grey areas are recessions, and values below 0 are inversions.

To address the GP’s point, the relationship could be fabricated, but the data don’t indicate that that’s happening.