The Torpedo effect : myth or reality? by Jeff L. Payne and Wayne B. Thomas

General evidence indicates that managers manage earnings at three common earnings thresholds: analyst forecasts, prior period earnings, and zero earnings. We examine one market-based motivation suggested for this behavior. If managers perceive the market penalty for barely missing an earnings threshold to be disproportionately high (i.e., a torpedo effect), they may use discretion to manage earnings upward to meet the earnings threshold. This market-based incentive would explain the evidence in favor of earnings management at earnings thresholds. To test the existence of a torpedo effect, we employ a comprehensive model that measures the market’s reaction to reported earnings that barely miss earnings thresholds. This model controls for the level of unexpected earnings and several other firm characteristics known to affect the relation between returns and earnings. Overall, we conclude that there is little evidence of a torpedo effect. This conclusion holds for both low-growth and high-growth firms and is unaffected by the firm’s history of meeting the threshold. Our paper dispels some commonly held beliefs about the market’s response to earnings thresholds.