We find that nonlinear estimation of profitability decomposition frameworks yields more accurate out-of-sample profitability forecasts than forecasts from both a random walk and linear estimation. The improvements derive from nonlinear estimation and synergies between nonlinear estimation and profitability decomposition frameworks. We analyze three essential financial statement analysis design choices to provide insights for the practice of fundamental analysis and find robust evidence that higher levels of profitability decomposition, focusing on core items, and using up to three years of historical information improve forecast accuracy. We find that our forecasts predict returns and profitability changes before and after controlling for analyst forecasts and common asset pricing factors.

This study examines the relation between exposure to social tax discontent—public expressions of frustration about others’ tax avoidance on social media—and individual tax avoidance. Economic and psychological theories predict bidirectional effects: discontent may reduce avoidance by inducing internal sanctions (e.g., shame and guilt) or increase avoidance by normalizing noncompliance through perceived prevalence. Using a novel, large-scale measure of tax discontent derived from geolocated Twitter data and a language classification model, we find that increases in social tax discontent are associated with reductions in estimated tax avoidance. The effect is stronger among high-income groups, in response to tweets targeting wealthy individuals, and in areas with higher political engagement and social media activity. Our results reinforce the importance of taxpayers’ social and behavioral motivations and suggest that jurisdictions’ efforts to manage taxpayer beliefs and leverage social norms can act as a complementary tool to traditional enforcement. Data Availability: The data used in this study are derived from public and third-party sources described in the text: the IRS Statistics of Income, the U.S. Bureau of Labor Statistics (BLS) Quarterly Census of Employment and Wages, the Bureau of Economic Analysis, Google Trends, and the National Neighborhood Data Archive. Twitter/X data were collected via the Twitter Application Programming Interface (API) under its academic-access terms and are subject to Twitter’s/X’s terms of use; they cannot be redistributed by the authors, but the collection and classification procedures are described in Section IV.