Carbon Market Signal Quality, Greenwashing Proxies, and the Cost of Debt in China

Carbon allowance prices may inform green lending, but regional price variability is not a direct measure of information noise or borrower exposure. We formulate a two-signal Bayesian screening benchmark and examine what the available public data can identify. Published summary moments for seven Chinese pilot markets motivate an explicitly illustrative classification sensitivity exercise. The available merged archive permits independent re-estimation of debt-cost models for 4205 firm-years from 952 firms in 2012–2022; a separate 2022 overlap contains 110 firms across six pilot jurisdictions. In firm and year fixed-effects regressions, the standardized greenwashing proxy has a coefficient of 0.0037 for standardized accounting debt cost (95% CI −0.0385 to 0.0459). Alternative debt-cost results are similarly imprecise. The six-market descriptive comparisons change substantially with proxy choice and market exclusion. These estimates cannot distinguish small effects from measurement and design limitations, nor test the proposed lender screening channel directly. We report the observable mapping, calibration assumptions, sample construction, and remaining data requirements to delimit the claims supported by this evidence.