Nicolas Crouzet, Apoorv Gupta, and Filippo Mezzanotti
When Indian prime minister Narendra Modi announced in November 2016 that all 500- and 1000- rupee notes would no longer be valid currency, and that they would have to be redeemed by the end of the year, he set off a mad scramble to deposit those notes and profoundly shook the Indian economy. But did he also unwittingly hasten the adoption of mobile payment apps among businesses and individuals? A new study suggests that he did.
The voided bills represented approximately 86% of the total cash in circulation at the time. Although people were supposed to be able to exchange old bills for new bills, there were not enough new bills available to meet the need, so cash in circulation fell while bank deposits spiked. This contraction in available cash was particularly acute in November and December. Eventually, more cash became available, and the cash level returned to pre-shock levels that spring. The surge in deposits overwhelmed the banking sector, as millions of customers sought to deposit and withdraw cash, leading to lines of up to four hours to use an ATM in some cases, and causing the average number of visits per person to banks or ATMs to more than double in the month after the announcement. However, the shock may have also led to a wave of adoption of various electronic payment technologies.
The paper measures the adoption of electronic payments using data from one of India’s leading “digital wallet” companies. To use the digital wallet application, a consumer need only have a smartphone and a bank account. Merchants receive a unique code that enables them to accept payments from a customer, with the money transferring directly from the customer’s bank account to the merchant’s, without having to purchase any additional equipment, such as a credit card reader. High penetration of smartphones in India makes it easier for adoption to spread widely; the benefits of adoption also increase as more consumers and firms adopt the technology.
The growth of the digital wallet in the weeks following the demonetization announcement was dramatic. Prior to that, growth had been positive but relatively modest. However, in the week after the announcement, the number of transactions grew by 150%, and amounts transacted increased by nearly 200%. For the next month, weekly growth rates were consistently around 100%, in terms of both the number and total value of transactions. The growth rate eventually slowed, but growth did continue even after the amount of cash in circulation regained its former level. This suggests that the temporary cash shortage led to a permanent increase in the use of digital wallet technology by both consumers and small businesses. Interestingly, the authors show that not only did aggregate adoption increase, but that Indian districts that were more exposed to the cash contraction also experienced a larger increase in adoption.
The authors argue that the permanent increase in adoption can be explained by the importance of network externalities for electronic payments. The intuition is simple: the value of adoption a new payment technology depends on the extent to which this technology is used by others. In these contexts, the adoption of a new technology could be hampered by coordination failure — a firm would want to accept electronic payment if more firms and consumers also use electronic payment methods but has little incentive to adopt if it is the only firm doing so. In this case, even a temporary shock may help to attract a sufficient number of new adopters to create a critical mass of users that allow the technology to flourish in the long run. The authors find that the data support a set of predictions generated by this type of model, therefore highlighting the importance of externality in adoption in explaining the shift.
The paper also highlights two other interesting results. First, it shows how switching to the new technology is key to understanding the adoption rates. In fact, firms also increased their capacity to accept credit card and debit card transactions in the months following the demonetization, but neither as quickly nor as widely. This is consistent with the additional switching costs in acquiring the hardware to process card transactions compared to the smoother transition to digital wallets. Overall, credit card transactions experienced almost no change in activity around the time of the shock, while debit card use did increase. However, most of the additional debit card use is at the intensive margin, since overall debit card issuance did not increase until a couple of months after the demonetization.
Second, the paper also shows that the cash contraction caused by the demonetization had a negative, causal effect on consumption by households. This result suggests that the rise in electronic payment technologies was not sufficient to compensate for the contraction in cash. However, this negative effect was purely temporary.
Although the demonetization had negative effects for some consumers, this study highlights how the shift also helped expand the usage of electronic payment systems in India. It may be hard to evaluate the precise effects of this technological switch, but the widespread adoption of electronic payment may have substantial benefits for India. Furthermore, while the paper focuses on this specific policy event, the main inference on how a temporary shock may have a substantial impact in presence of externalities may be applied to other technologies and contexts.