This study investigates the causal effects of proactive and reactive enforcement strategies employed by the Norwegian Tax Administration to improve the accuracy of wage income reporting. Using a large-scale field experiment targeting wage earners and pensioners who reduced their reported income—despite high-quality third-party reporting from employers—we examine the impact of behavioral prompts, audits, and penalties (added tax). A key strength of the design is the random assignment of all enforcement tools: Taxpayers were randomly assigned to receive a behavioral prompt or not during the tax return process. Those who continued to moderately underreport were randomized to immediate or delayed audits, while substantial underreporters were all audited but randomly assigned to immediate or delayed penalty assessments. Importantly, the experimental variation in audit and penalty timing was maintained into the subsequent tax year, allowing us to study the persistence of effects. While we estimate compliance effects using administrative data, an important contribution of the project lies in its use of survey data to study broader attitudinal outcomes. The survey, conducted before the next filing deadline, measures trust in the administration, tax morale, fairness concerns, and perceived detection probability. These softer outcomes offer rare insights into how enforcement strategies are perceived by taxpayers and how they may shape long-term compliance behavior.