During the two-year period 2016-18, currency with the public went up by 10.2%. That was less than half of the growth in nominal gross domestic product (GDP), which went up by 21.7% over the period.

Why are we taking a two-year instead of a one-year period? Because year-on-year comparisons are skewed due to the demonetisation episode.

Next, let’s consider the preceding two-year periods. During 2014-16, currency with the public moved up by 28.3%, well above the growth of nominal GDP, which was 22.5% (see chart). In other words, we have moved from a situation in which growth in currency with the public was much more than growth in nominal GDP to one where it’s less than half the growth in nominal GDP.

One reason for demonetisation was to correct the sharp rise in currency growth, compared to the level of activity in the economy. But perhaps the correction has been overdone.

For instance, growth in currency with the public was lower than nominal GDP growth in the two-year period 2012-14, but it wasn’t as low as in 2016-18. The nudge to digital modes of payment seems to have become a push, perhaps even a shove.

It doesn’t really mean much when people say currency with the public is back to levels seen before demonetisation. Of course it will be—the economy has grown handsomely in the intervening period. The point is that currency with the public at the end of March 2018 is up 3.4% compared to where it as at end-October 2016, before demonetisation. Compare that with the 17.8% year-on-year growth in currency with the public at end-October 2016.

If the growth in cash is already at such low levels compared to the growth in the economy, it will require only a small rise in demand for currency to upset the apple cart. It’s like running a business with the absolute minimum number of personnel—if a few of them fall ill, or demand suddenly rises, you can’t meet the expectations of your customers.

Could it be that demand for currency has gone up due to the impending elections in Karnataka? That begs the question why there wasn’t a cash crunch during the Gujarat elections in the December quarter. It is interesting that cash on hand with banks fell sharply between September and December 2017—it’s possible that while the Reserve Bank of India (RBI) may have printed the same amount of money, the demand for extra cash was met by drawing down cash with banks. At present, cash levels with banks are already low—the level at end-March 2018 was 23.6% lower than at end-September 2017—so it could be that further drawdowns are not possible. Or it may just be that supply is not forthcoming.

What is obvious is the demand for cash is now more than the supply. As we have seen, supply has been constrained. Demand, on the other hand, may have gone up due to the usual suspects—elections, a pick-up in the economy, hoarding cash, or simply a return to old habits of cash preference.