The Event Clock: Identifying Physical Information from State Prices

How much of event-price variation is spanned by physical learning rather than orthogonal valuation shocks? Under cross-system residual orthogonality, three economically distinct pricing systems identify this learning-spanned component and its variance share without a loading normalization. Bounded residual correlation yields sharp sets, while an economic anchor is required only to express the common component in physical-clock units; without one the physical scale is globally nonidentified. In 2024 U.S. presidential-election data, Polymarket, Kalshi, and Betfair place 75% of national cross-venue covariance variation by November 4 and 90% by November 5. Sampling-frequency-corrected share estimates are economically substantial but residual-correlation sensitive. A three-venue panel of seven battleground-state elections reproduces the same late timing and does not reject common normalized covariance profiles. Pre-resolution risk-premium restrictions cannot bound state-conditional valuation variation, while additional systems generate specification tests.

SSRN Scholarly Paper No. 7420838