Beyond Tail Co-Movement: How Temperature Extremes Shift Financial Return Distributions

This paper extends the Conditional Probability of Exceedance (CPE) framework into the domain of physical
climate variables. Using daily temperature data from 15 cities across Europe, North America, and Asia
(2000–2026), we construct regional temperature exceedance indices and test four co-movement directions
between temperature extremes and 26 financial instruments. The central finding is structural: temperature
extremes produce financial distribution shifts, not tail co-movement. When regional temperatures exceed
their 90th percentile, related instruments exceed their median forward return with CPE 0.60–0.76 (lift
1.20–1.84×) at 63–126 day horizons across 89 surviving signals — but only 1 signal survives at the
75th–90th financial percentile tail. This contrasts with financial®financial CPE pairs from Papers 1–5
achieving 97.0% hit rates at the 93rd percentile. The strongest channels are sugar (CANE, lift 1.84×), base
metals (DBB, 1.51×), gold (1.37×), and natural gas (1.33×). A critical limitation is that city-based temperature
indices — appropriate for energy demand signals — are poor proxies for agricultural zone temperatures
relevant to crop stress. Paper 7 addresses this directly using ERA5 gridded crop-zone temperature data.