Pick A Number: Why All GDP Estimates Post-Demonetisation May Go Wrong

One of the interesting things about India’s demonetisation of Rs 500 and Rs 1,000 notes is that it is completely unprecedented in scale and impact. So, no matter how it plays out for us, it will be one of the world’s biggest live experiments in severe short-term financial compression. We will see books and case studies being written on it for decades.

But, on a more practical level, the first thing to do is to take all predictions of doom (or, the rarer forecasts of quick revival, including mine) as best guesses. We will know who is right, or closer to the truth, only some time in the second half of 2017. We will be able to come to a tentative conclusion on the full impact of demonetisation only over a period of 12-18 months, when gross domestic product (GDP) figures are recalculated, sectoral impact figures show up in corporate earnings, bank non-performing assets (NPAs) surface, and we learn how agriculture and rural wages have been impacted.

So when former prime minister Dr Manmohan Singh says demonetisation may cut GDP growth by two per cent, we should note it with an “oh, ah”. The same should be done with all recent estimates made by rating agencies and brokerage houses. All deserve the proverbial pinch of salt, and some more than a pinch.

What everybody agreed on is that GDP will be impacted in 2016-17. While Goldman Sachs thinks GDP growth will fall 1.1 per cent, Care Ratings sees a drop of 0.3-0.5 per cent, Icra and ICICI Securities 0.4 per cent, Emkay Global 0.9 per cent. The Centre for Monitoring Indian Economy (CMIE) even has a precise figure for economic loss: Rs 128,000 crore – Rs 61,500 crore for industry due to supply chain disruptions, Rs 15,000 crore losses to workers in terms of wages, Rs 16,800 crore cost to government for printing new notes, and Rs 35,100 crore costs to banks, calculated in terms of the time lost due to handling demonetisation logistics.

The problem with calculating costs this way is that we assume that what is lost won’t be recovered. Sales lost in November-December may well be regained in the next quarter as suppressed demand resurfaces. Of course, wages lost now may never be recovered, but can we be sure of this? Isn’t it a normal human tendency to work harder if there has been a loss? How many one-day strikes have we seen which ultimately made no impact on total sales or revenues?

The really outlandish number on GDP impact is that of Ambit Capital, which believes the growth rate will be down by a massive 3.3 per cent. It means negative growth in the current and next quarters, for the previous two quarters and the first half of the third quarter (October-December 2016) are already over.

The only thing one can say is that when there are so many estimates about the impact of the same event, almost all of them are likely to be wrong. Anyone who got it right would have done so by fluke.

A lot of the modelling that economists do relies on which scenario they are trying to compare the current monetary compression with. So while they can predict what could happen in broad terms (ie, that growth will slowdown, etc), in a live economy, where millions of players are responding to the compression in different ways, no one can accurately predict how all the individual decisions taken by those affected will add up in terms of growth.

In 1991, for example, when Dr Manmohan Singh presented his first reformist budget after an external crisis, the sharp fiscal compression he imposed in Year One (1991-92) led to a crash in GDP growth rates from 5.3 per cent in the year before to 1.4 per cent in 1991-92. But growth rebounded sharply in Year Two, to 5.4 per cent.

The Dr Manmohan Singh and Ambit estimates are probably wrong for one simple reason: 1991 was a huge crisis; 2016 is a limited crisis manufactured by a specific man-made incident (demonetisation). So the impact is bound to be less severe than in 1991 when all things were changing simultaneously – licensing, exchange rates, interest rates, et al.

This time the only change is in cash availability. And even this can revert to normal in the next four months. Also, we have several alternatives to cash, and thus the rebound could be sharper this time. The only uncertain factor is how much consumer and investment sentiment has been hit, and how long it will take to return to normal.

My guess, made last week, is that the sense of crisis will abate by the end of the next quarter, and the first quarter of 2017-18 will see a sharp rebound compared to the previous quarter.

And with a budget coming in-between on 1 February, and several mid-season stimulants in terms of monetary easing also in the works, India will put this setback behind faster than it did in 1991.

If you want numbers, let’s take 1 per cent GDP growth loss and CMIE’s cost numbers. A 1 per cent loss of growth means a sacrifice of around Rs 150,000 crore of potential output, and if we add CMIE’s loss figure, we will get a demonetisation cost of Rs 278,000 crore.

Demonetisation would have been worth it if the gains are larger than that.

That seems possible if more than Rs 2 lakh crore of black money fails to come back to the system, and tax revenues surge by Rs 1 lakh crore or more in the coming four quarters. The Rs 2 lakh crore of liabilities written off by the Reserve Bank of India on notes that don’t come back can resurface as public investment, stimulating the economy from the first quarter of 2017-18.

And if the shift from cash to formal systems of banking becomes more permanent in urban India and parts of rural India, the gains will be incalculable.