The textbook trade — El Niño droughts Australia, wheat rallies — tested across six grain instruments and four horizons. Nothing survives, and the significant results carry the wrong sign.
No effect at any horizon tested, across 24 tests.
El Niño suppresses rainfall across eastern Australia and northern India during the southern growing season; wheat is the most exposed major because Australia is a swing exporter into Asia.
Corn at six months returns a Newey–West t-statistic of 3.30 and a p-value of 0.001. Under the persistence-robust null it returns 0.98, and its coefficient carries the opposite sign to the hypothesis.
The three-month SOI has an autocorrelation of 0.93, and overlapping forward returns share five of every six months. Two persistent series correlate by accident far more often than the t-distribution admits. The circular block-shift test preserves both series exactly — autocorrelation, seasonality, fat tails — and slides one against the other, so persistence appears identically under the null and cannot manufacture significance.
| Rev | Stage | Status | Change |
|---|---|---|---|
| 01 | Draft | — | Direction and criteria fixed before the experiment ran |
| 02 | Tested | Null | 24 tests; 0 below p=0.05, none surviving FDR |