Since Modigliani and Miller (1958,1963), the relationship between debt and value still remains an open question. The numerous contributions of US research seem to show that the relationship between debt, performance and value is complex. The influence of debt on performance is affected by many factors such as economic cycle (Platt and Platt,1994), corporate ownership, the reputation of management and/or the particular industrial sector.
Although it is difficult to take into account all of these factors; it would be interesting to understand the influence of economic cycle on the debt-performance relationship. Through the use of SOM (Self-organising maps), three main results appear.
Firstly, the coexistence of both negative and positive effects of debt on performance highlights the fact that the relationship between debt and performance during an economic downturn is non-linear. From a sample of 200 firms in France, our results suggest that debt can be a source of either "good stress" or "bad stress".
Secondly, our investigations lead us to suggest that debt accelerates a decrease in performance for industrial firms with an extensive operating cycle. For these firms, debt is a factor of financial distress in an economic downturn.
Thirdly, for firms in financial distress the debt-performance relationship seems to be dynamic. If highly leveraged firms have the poorest performance, this decrease in performance has a negative (i.e. increasing) influence on the level of debt (p value < 5 percent). We can note the relationship of Altman (1968,1984) which shows that decrease in performance leads to an increase in debt.
Keywords: Neural Networks, Financial Distress, Performance, Economic Cycle