The chaos created by demonetization is not because of a failure in implementation. It could not have been avoided by calibrating ATMs on time. It is a function of how much of India’s black money is held in cash.

To understand this, consider a thought experiment. Suppose that there was no cash black money in India, and yet the government, for some vague reason, announced demonetization. Would there have been a similar level of chaos?

Or consider another thought experiment. Suppose that the government announced demonetization, but also announced that the 500 and 1000 rupee notes could be exchanged at the bank without limit, no questions asked. Would a similar panic have spread through the economy?

My answer to both questions is no, and it would stay no even if you told me that in both thought experiments, the government was as unprepared with the task of ATM calibration as in the real world. That is because in both those situations, shops would continue to take the demonetized notes even though it would be illegal to take them.

In the real world, shops are refusing to take those notes, not because it would be against the law (we aren’t a very law-abiding society), but because of the difficulty of converting them to legal tender. This difficulty is because of the existence and quantum of cash black money (thought experiment 1) and the difficulty of laundering it (thought experiment 2).

Because of these two factors, the demonetized notes are trading at a discount. Because of this discount, there is a huge impact on the economy. Remember that the black economy is also a part of the economy, and a hit on the black economy is necessarily an attack on the economy, and the people affected aren’t necessarily, or even primarily the holders of black money.

Another way to say the same thing is to note that the discount on the demonetized notes affects everyone, not just those who hold unaccounted money. It affects daily wage labourers, it affects the supply chain for food, everything.

My initial reaction to demonetization was that it was a mildly positive move, but ultimately more of a PR stunt rather than one that would have a large impact one way or the other. That is because my thinking ran on the lines of ” Who keeps their black money in cash?” I thought was that people mostly maintained only their working capital in cash, and as a great believer in the efficacy of markets, I thought that money laundering channels would quickly take care of the black money that was in cash. The net result would be a PR victory for Modi, a one-time shock to the cash economy and a net cash transfer to the poor, as it would be the poor whose Jan dhan accounts would be used to launder all the black money.

From the chaos that is occurring, it appears that I was wrong about how much black money is stored in cash. Of course, I don’t have a baseline level of how much chaos was expected to compare the current level with. But the point is that if you believed that a lot of black money is stored in cash (as against circulated as working capital and laundered before storage) you should have expected a much higher level of chaos.

A full model of how much chaos to expect from the demonetization would consider:

It looks like Modi has a similar mental model of demonetization. The government has flipped and flopped multiple times, but it has been consistent in one thing, which is that it had imposed ever tighter restrictions on changing money over the counter.

This makes no sense if you believe that the folks changing old 500 notes are the unbanked poor, but makes complete sense if you believe that it’s a channel to launder demonetized notes into remonetized ones AND if you believe that those notes will go back to black money storage. The problem though is that anything that hurts black money storage (#3) will also hurt the circulation of black money (#2) and this puts pressure on #4 even more.

I don’t know if this experiment will succeed. As I write this, it is the end of November and the ATMs around me are empty. If people are unable to withdraw money when they get their salaries in the first week of December, I expect hell to break loose.

But what this experiment has taught us is not, as one article put it, how much India runs on cash, but how much of India’s unaccounted money is stored in cash.