Prior studies have confirmed that the greater the investment in information technology (IT), the better the performance of the firm. In the present study, we use bootstrap data envelopment analysis (DEA), which has previously been used in the accounting literature, and which has an edge over traditional accounting-based measures, to compute the performance efficiency of Indian banks. Specifically, we measure the impact of automated teller machine (ATM) investment intensity on the production efficiency of Indian banks. We also study the impact of bank ownership, soundness, size and risk on efficiency. The study contributes to the literature on the productivity paradox and also draws on structure–conduct–performance theory. We find that ATM intensity has a significant negative association with bootstrap DEA technical efficiency. These results differ from prior research in developed countries. The results could be ascribed to heavy investment in IT such as ATMs by banks, and their inability to reduce labour costs given that many processes still continue to be manual. Accordingly, the study suggests that investment decisions with regard to IT need to be taken with great caution. If the related processes are not simultaneously automated, such investments are unlikely to yield the results that management may have envisaged.