While extensive research has examined the external consequences of fraud, its impact on managerial decision-making within firms remains underexplored. Drawing on the Behavioral Theory of the Firm (BTOF), this study explores how fraud influences firms' feedback-react processes through two distinct mechanisms. Before detection, fraud strengthens problem-hiding motivations, with firms overstating their performance, attributing underperformance to external factors (e.g., industry uncertainty), and avoiding disruption of the status quo. Once detected, however, fraud triggers problem-solving motivations, as firms seek to restore their reputation and reduce the likelihood of re-detection. Detected fraudulent firms are more likely to attribute underperformance to internal factors, such as technological or innovation weaknesses, and engage in riskier problem-solving behaviors, such as increasing R&D investments. Using a dataset of Chinese firms and focusing on fraudulent financial reporting, the study provides robust empirical evidence supporting these claims. Fraud risk is predicted using a machine learning model, with firms divided into two stages—before and after their first fraud detection—based on official enforcement actions. The findings show that, prior to detection, fraudulent firms reduce their R&D intensity compared to honest firms, as they overstate their performance and attribute performance issues externally. After detection, these firms shift their attributions internally and increase R&D intensity to address performance deficiencies. This study contributes to the fraud literature by clarifying its impact on firms' decision-making processes and to the BTOF literature by highlighting fraud as a common, yet neglected, form of intentional motivational bias.
