By Vivek Dehejia for the IDFC Institute and Mint
On 8 November 2016, Prime Minister Narendra Modi announced the "demonetisation" of high-value currency notes. When it became clear the ex ante goals of reducing corruption and black money were unlikely to succeed, attention shifted to a range of possible ex post facto rationales for the move. Numerous commentators, including the two of us, argued that one such goal would be to push the economy towards greater formalisation, and, relatedly, to push the financial sector towards greater digitisation and a reduced reliance on cash. This broad rationale was then quickly picked up by the Modi government and its supporters.
Now that sufficient time has elapsed, it is possible to bring data to bear on the specific question: Has demonetisation succeeded in making India a "less cash" society? To our regret, as we both had hoped the answer would be in the affirmative, the answer is clearly "no".
The chart above tells the tale. It plots currency in circulation (CIC) as a share of broad money (M3) — henceforth, CIC/M3 — using monthly data from January 2012 to May 2018, the last month for which we have data. All data are sourced from the Reserve Bank of India (RBI). We should note that the choice of a different denominator, whether a different metric of broad money or nominal gross domestic product (GDP), does not materially alter the results we present below.
The long left "tail" makes it evident that CIC/M3 had settled into an equilibrium in the range of 0.14 — with CIC equal to about 14% of the stock of broad money. The demonetisation shock in November 2016 saw CIC/M3 plunge to as low as about 0.08 within a month, not surprising when fully 86% of the currency stock was invalidated by the stroke of a pen.
The following months saw CIC/M3 starting to rebound, raising the important question: would it settle at a new, permanently lower, equilibrium, or return to where it began before demonetisation? This is not merely an arcane academic question. If the former, it would provide conclusive evidence of long-run behavioural change, with the Indian public, goaded by a temporary currency shortage, nudged into permanently preferring to hold less cash as a share of broad money, even when the cash crunch has disappeared. If the latter, we would be obliged to conclude that the pain of demonetisation did not trigger any long-lasting behavioural change in the preference for currency, thus striking off one possible long-run gain induced by the short-run pain of the cash crunch.
Look at the chart. It is not hard to see the trend: CIC/M3 is clearly converging back to its pre-demonetisation level — indeed, it is just about already back to that earlier level. In other words, the idea that demonetisation will lead to a less-cash society has come a cropper.
It is straightforward to supplement eyeballing the chart with a parsimonious but rigorous statistical analysis. As seems plausible, we assume that the statistical process governing CIC/M3 is what is known as a first-order autoregressive process, or AR(1) for short. In lay terms, this means that CIC/M3 is assumed to be a function of a constant, its own lagged value from a period ago, and a random shock. It is possible to use higher order AR processes or different statistical models, but this would not materially change the results that we report immediately below.
Linear regression analysis on the data underlying the chart validates our AR(1) model, with a "persistence" coefficient on last year's value of about 0.82 or 82%. More importantly, a simple calculation shows us that the model implies a long run or "steady state" value for CIC/M3 of about 0.14. This is exactly what we observed before demonetisation and this is exactly where we are back to today.
It should be added that the failure of demonetisation to make India a less-cash society does not necessarily imply a complete failure of the larger digitization drive — data suggest that some components of digital payments are up after November 2016, although it is not clear if this is merely a continuation of pre-demonetisation trends or an effect wrought by demonetisation. Nor can one conclude, on the basis of our analysis, that the drive towards greater formalization — in particular, bringing more Indian households and businesses under the tax and regulatory nets — has failed, although such data as we have on this question remain open to contestation and interpretation.
Making India a less-cash society was a clearly stated goal of the Modi government. For instance, on 30 August 2017, finance minister Arun Jaitley was explicit that India is a "high cash economy" and that changing this was a goal of demonetisation. That effort has evidently failed.
Vivek Dehejia is a Mint columnist and resident senior fellow at the IDFC Institute, Mumbai. Rupa Subramanya is a Mumbai-based economic researcher and writer.