Specification of Models that Estimate Response Coefficients in the Market–Accounting Relation
Published online on November 07, 2025
Abstract
["Abacus, EarlyView. ", "\nWe argue that log‐linear models based on a power‐law relation between market and accounting data, which use elasticities to measure response coefficients in regression models of the fundamental market–accounting relation, are well specified and provide precise, readily interpreted and valid estimates of the long‐run relation between market and accounting values. Using this approach, we show that financial statement data is sufficient, with little or no extra data, to explain fundamental firm market value. We illustrate the use of the measures of elasticity response coefficients by discussing the evidence for dividend irrelevancy, the relationship of the market‐to‐book ratio with growth and its uncertainty, and the existence of abandonment options. Our method of estimating parameters in the market–accounting relation facilitates replication. We use all active Compustat firms between 1971–2020, without deletion or special treatment of outliers, demonstrating the virtues of log‐linear models in capital market research.\n"]