Liquidity, Investment and Risky Debt: Some Evidence

Abstract

Utilizing bond data for U.S. non-financial corporations, this study finds a relationship between liquid assets and default premia. The paper argues that this can be explained by endogenous adjustments in liquid asset holdings by firms concerned with the possibility of a liquidity shortage, which, in the presence of external financing constraints, may trigger default and the costs of financial distress. As a result, riskier firms have not only higher default premia, but also larger holdings of liquid assets. Thus, the endogeneity of liquidity may be central to the interaction between default risk and the condition of the firm’s balance sheet. Simple regressions that treat liquidity as an exogenous parameter may therefore yield unreliable results. In our model, a positive relationship arises when future firm cash flows cannot be fully pledged as collateral for borrowing and the benefits of firm investment can only be realized if the firm does not default. At the same time, exogenous changes in the firm’s liquidity are expected to be negatively related to default premia. The prior expectation that firms with high liquidity should be safer explains the direct effect of liquidity on default premia, but does not capture the indirect effect arising from the endogeneity of liquidity, which appears to dominate in practice. Previous studies have focused on leverage and volatility as determinants of default premia, and ignore the potential effect of liquidity. Others assume that default is triggered by a shortage of cash, but they do not allow for an endogenous determination of liquidity. The results obtained here suggest that firm characteristics that determine default premia, such as leverage, volatility, profitability and financing constraints, may also affect the firm’s liquidity, which, in turn, affects default spreads. Our results also suggest that common empirical specifications in studies of default premia, which use default risk variables suggested by extant models, may be overlooking important variables. Int Adv Econ Res (2009) 15:492–493 DOI 10.1007/s11294-009-9221-z