Two competing principal–agent models explain why firms pay dividends. The substitute model proposes that corporate insiders pay dividends to signal and build trust with outside shareholders who lack legal protection. The outcome model, in contrast, surmises that when shareholders have legal protection, they demand dividends from insiders to prevent them from expropriating corporate funds. Either way, dividends represent an agency cost paid to align the interests of shareholders and insiders. Expropriations by insiders and reduced investment by shareholders are also agency costs, but they are difficult to identify with archival data. Using a laboratory experiment, we identify the impact of strengthened shareholder protection on all three typ...