The diffusion of process innovation in the UK financial sector: an empirical analysis of automated teller machine (ATM) diffusion

Recent policy initiatives have identified that the diffusion of innovation constitutes an important component in technical change and progress and is the impetus behind changes in firm productivity. To date, however, the main emphasis of economists has been on the diffusion of process innovations in the industrial sector with diffusion in the financial sector either ignored or, at best, summarised by a number of stylised facts relating to the spread of information. The objective of this thesis is to explore the inter-firm determinants of ATM adoption and diffusion in the UK financial sector and identify firm-specific and market factors in the diffusion process. The empirical analysis draws on duration analysis which represents the current state-of-art modelling approach to inter-firm diffusion. This approach conceptualises inter-firm diffusion as a cross-section of durations of nonadoption from which, most importantly, hypothesised factors (or `covariates') can be examined by their significance or otherwise on the conditional probability of adoption. The main findings of this thesis support the stylised fact often made in the diffusion literature that the inter-firm diffusion curve is sigmoid and characterised by a nonmonotonic hazard function. Furthermore the empirical analysis supports the hypothesis that firm-specific characteristics and expectations have played a crucial role in the interfirm diffusion of ATMs. In addition, the results indicate that the diffusion of ATMs in the UK has been characterised by the existence of positive network externalities. The results are also shown to be robust across a number of model specifications and assumptions concerning the time-path of covariates.