This paper tests traditional capital structure models against the pecking order model. The pecking order model, which predicts external debt financing driven by internal financial deficits, has greater explanatory power than the static trade-off model, which predicts firms adjust toward an optimal debt ratio. The paper questions whether empirical evidence supports the notion of an optimal debt ratio. It shows that the power of tests for the trade-off model is low, and that the pecking order model better explains debt ratio variance. The paper also finds that the pecking order model can be rejected if financing follows the trade-off model, but the trade-off model appears to work when financing follows the pecking order. The paper concludes that the pecking order is a better first-cut explanation of debt-equity choices and questions the evidence for an optimal debt ratio. It also finds that the static trade-off model is not rejected even when false, while the pecking order can be easily rejected when false. The paper suggests that more sophisticated models are needed for empirical research on capital structures.This paper tests traditional capital structure models against the pecking order model. The pecking order model, which predicts external debt financing driven by internal financial deficits, has greater explanatory power than the static trade-off model, which predicts firms adjust toward an optimal debt ratio. The paper questions whether empirical evidence supports the notion of an optimal debt ratio. It shows that the power of tests for the trade-off model is low, and that the pecking order model better explains debt ratio variance. The paper also finds that the pecking order model can be rejected if financing follows the trade-off model, but the trade-off model appears to work when financing follows the pecking order. The paper concludes that the pecking order is a better first-cut explanation of debt-equity choices and questions the evidence for an optimal debt ratio. It also finds that the static trade-off model is not rejected even when false, while the pecking order can be easily rejected when false. The paper suggests that more sophisticated models are needed for empirical research on capital structures.