Revealing tax evasion: Experimental evidence from a representative survey of Indonesian firms

Abstract

Measuring tax evasion by firms is notoriously difficult: direct questions yield biased responses while administrative data captures only detected violations. This paper examines the prevalence of reported tax non-compliance in Indonesia through a randomized, double-list experiment embedded in a nationally representative survey of 2,955 registered firms that preserves anonymity. The results indicate that around one-quarter of firms indirectly report not paying all taxes they are required to pay. An indicative back-of-the-envelope calculation suggests that the associated revenue losses could be economically meaningful, in the order of two percent of GDP. Heterogeneity analyses suggest higher reported non-compliance among firms that do not export, face informal-sector competition, and view tax administration as a major obstacle. These findings help inform tax authorities, which face substantial challenges in estimating levels of tax evasion and identifying non-compliant taxpayers, by providing the first representative estimates of firm tax evasion in a major middle-income country.