Bankruptcy risk and productive efficiency in manufacturing firms

The paper investigates the determinants of bankruptcy on three representative unbalanced samples of Italian firms for the periods 1989-1991, 1992-94 and 1995-97 with two different estimating techniques. The first (logit) approach shows that: i) the impact on the probability of bankruptcy of both level and trend balance sheet variables is sample - and business cycle specific since financial factors are predominant over real factors before periods of financial distress but not otherwise - with the exception of the interest charge to value added ratio which is always significant; ii) qualitative regressors such as customers' concentration and strength and proximity of competitors have significant predictive power and suggest that banks should not restrict their monitoring activity to balance sheet variables. The paper also shows, through a stochastic frontier approach, that productive efficiency and the probability of bankruptcy are significantly correlated and that the significance of the relationship between the distance from the efficient frontier and the probability of failure is robust across the last two sample periods. The distance from the frontier is in fact shown to have marginal explanatory power, net of balance sheet and qualitative variables considered in the previously mentioned logit model.