Bayesian Panel Data Model With Application To Debt Financing Decision Analysis Of Private Listed Companies

Debt financing has always been one significant problem in the field of corporatefinance. A reasonable debt ratio can provide appropriate financial leverage effect, andhave a positive role in enterprises’ expansion business. From the point view of currentempirical research the influencing factors of debt financing decisions can bedetermined. As the private listed companies play a more important role in the nationaleconomy, how to solve the problem of debt financing and provide basis for d ecisionmaking for the private listed companies’ debt financing has become a urgent problem.Bayesian panel data model is introduced in this paper to research the debtfinancing decisions. Comparing with the existing linear regression methods, paneldata model combine the time-series data and cross-sectional data, so it contains moreinformative which can increase the observation data and freedom, therefore, paneldata model can get more information and provide more efficient and reliableparameter estimator. Moreover, Bayesian analysis methods apply a priori informationin statistical inference, which can further improve the quality of statistical inferenceand solve the problem of parameter uncertainty. Simulation results show thatBayesian panel data model based on the MCMC sampling algorithm can effectivelyimprove the accuracy of parameter estimation.The empirical study is conducted with the financial data of fifty private listedcompanies. A Bayesian static panel data model and Bayesian dynamic panel d atamodel describing the debt financing decisions of private listed companies. The resultsshow that the MC error of the Bayesian static panel data model and Bayesian dynamicpanel data model is much smaller than the standard deviation, which proves thereasonableness and accuracy of the model; the debt financing level of private listedcompanies have a positive correlation with previous debt level, the size of thecompany, the growth stage of company and business-government relations, butnegative with external macro-financial environment, the profitability of company andthe short-term debt paying ability; Bayesian dynamic panel data model cancharacterize the inertia of debt level, and comparing with Bayesian static panel datamodel, the parameter convergence of Bayesian dynamic panel data model is muchbetter, and the standard deviation and MC error is much smaller, which indicates that Bayesian dynamic panel data model can get more accurate results in the study of debtfinancing decisions of private listed companies.