Innovation and business survival: A long-term approach

This paper explores the influence of innovation on the probability of survival of two hundred top British firms founded throughout the nineteenth and twentieth centuries. To this end, we have collected the firms' significant innovations and classified them by Schumpeterian types, patented and non-patented and domestic and imported. The number of patents registered by the firms throughout their lifetime -a rough measure of their incremental innovation activity- has also been recorded. In addition, twelve control variables - five characteristics of the firms and seven of their business leaders have been included. Both log-normal and gamma duration models have been used in the analysis. They have been estimated, firstly for the whole set of firms and, secondly, for the manufacturing and the service firms separately to control for industry differences. The results of the log-normal and gamma estimations are highly coincident, with some nuances. The significant innovations - particularly new processes, non-patented and domestic ones- have been found to positively influence the probability of business survival. The number of patent applications seems to increase the survival probability of the manufacturing firms, but not of the service ones. Among the control variables, the firm's size, its international dimension, and the age of the business leader at entry seem to be the most influential ones on business survival, although there are some differences between manufacturing and services. The main results are robust to the division of the sample by entry period.