Ebook Leverage, Profitability Shocks and the Dynamic Responses of Agency Cost Measures
Many models have suggested the use of debt in a firm’s capital structure can reduce the managerial agency costs, such as insufficient effort and excessive perquisite consumption, for outsider managed firms. Some theories hold that debt can increase agency costs, however, especially at an excessive level.
Thus it is possible that the disciplinary role of debt is related to its level in a non-linear way. Furthermore, this function should reveal itself through the dynamic responses of agency costs to the unexpected shocks in the firm’s profitability in the data since the conditions in reality are less stringent than theoretically assumed. This paper analyzes the dynamic responses in agency cost measures to profitability shocks of firms with different leverage levels using the panel data vector autoregression (Panel Data VAR) technique.
Understanding the agency behavior at different leverage levels can shed light on the puzzling results for leverage in the recent literature on ownership structure and agency costs. This literature has successfully provided direct evidence on the relationship between different owner ship factors and agency cost levels. Leverage has always been included as a control in various cross sectional regressions due to the consensus on its disciplinary function.
The coefficients on leverage, however, can be either statistically insignificant or sometimes even have the wrong sign. Besides the non-linear relationship between leverage and agency costs as a plausible explanation for this result, we also would like to stress the fact that leverage level alone cannot guarantee efficient agency behavior since the agent may not always be under enough financial pressure. During the life of long-term debt, firms usually face a random income steam and if the manager is fortunate for some periods, the cash at his or her disposal can become more than enough to guarantee the repayment. Having a good track record on hand, the manager can sit back, create more pet projects and consume more perquisites under insufficient corporate governance. On the other hand, following lackluster realizations of operating outcomes, the manager may have to cut perks, halt empire-building activities and work harder to improve performance in order to avoid bankruptcy. The effect of debt should thus be reflected in the way agency cost measures respond to unexpected profitability shocks and it should be different for firms with different levels of leverage.
In this paper, we test the above hypothesis using a large U.S. panel data set of more than 2000 publicly traded industrial companies from 1989 to 2006. We first apply the Arellano and Bond (1991) dynamic panel data estimator to estimate the system of equations for a VAR system upon subsamples grouped by industry-adjusted leverage quartiles. The VAR system includes the agency cost measures, a proxy for profitability and other control variables proxies such as growth opportunities and short-term asset balances.
The Arellano-Bond estimator eliminates both the firm fixed effect and the cyclical factors in the equations of the VAR system. To illustrate the responses of agency cost measures, the orthogonalized impulse response functions of the agency cost measures to profitability shocks then are plotted following standard VAR techniques with a set of conservative and economically sensible assumptions. The benefit of using the orthogonalized impulse response analysis is that it only counts the influences on agency cost measures due to a one-period pure innovation in profitability. Thus the identified responses are neither contaminated by contemporaneous innovations of the other variables nor by future innovations of all the variables in the VAR system.
The confidence intervals for these responses in the Panel Data VAR framework, however, are different from those in the standard time series analysis. They are calculated according to the asymptotic formula derived by Cao and Sun (2006) for the Arrelano-Bond estimator to reflect the additional estimation uncertainty when dealing with dynamic panel data.
Download
PDF Ebook Leverage, Profitability Shocks and the Dynamic Responses of Agency Cost Measures
Posted in :
