Tax Audits and Firm Value: The Moderating Role of Tax Risk in Publicly IDX-Listed Companies

Abstract

Tax compliance oversight has long been treated by capital market participants as a proxy for governance quality, yet whether a tax audit itself moves the needle on how a firm is valued remains an open empirical question, particularly outside the handful of jurisdictions where such studies have concentrated. This paper investigates whether tax audits affect firm value among IDX-listed companies and whether tax risk, measured through the volatility of the effective tax rate, conditions that relationship. Drawing on a purposively selected panel of 117 firms administered under the Tax Office for Listed Companies (KPP PMB) and observed over five years following their audited tax year (2015–2019), the study estimates a random effects model with firm-clustered standard errors, following selection through the Chow, Hausman, and Lagrange Multiplier tests. Firm value is proxied by Tobin’s Q; tax audit is captured through a dummy indicator, capturing audit occurrence rather than audit outcome as a distinction from prior studies that have typically linked firm valuation to assessment results, penalties, or settlement amounts; tax risk is proxied by the standard deviation of the effective tax rate. Neither the direct effect of tax audit on firm value nor the interaction between tax audit and tax risk reaches conventional significance thresholds, and the model’s explanatory power is minimal (R² = 0.0014). These results suggest that, at least within this institutional setting, a tax audit’s mere occurrence provides too little economic information to move investor valuation, and tax risk, as operationalized here, has not yet become a lens through which the market filters audit-related signals. By isolating the valuation effect of audit occurrence itself, independent of audit findings, this study extends Signaling and Legitimacy Theory to a dimension of tax enforcement that prior audit-outcome-focused research has left largely untested. The paper closes by situating these null findings within Signaling and Legitimacy Theory and by outlining measurement refinements that future work might pursue.