We don't know much about what the exchange rate ought to be

Many people are confidently saying that the market's price of the rupee-dollar exchange rate is wrong. They think they know what the exchange rate should be. There are many pitfalls along this path.

The simplest international trade perspective

To fix intuition, let's think there are only two countries: India and the US. Suppose there was no capital account. Suppose there was only import and export of goods and services on the current account. In the absence of a capital account, in every time period, the current account would have to work out to zero. The currency market would clear to yield an appropriate exchange rate from the viewpoint of export competitiveness. We would find an exchange rate at which we earn enough export proceeds to be able to pay for for our imports (that are also a function of this exchange rate).

Suppose we had a base year in which things were square. Now, from that point onwards, one could look at inflation in the two countries and the exchange rate and get a sense about how export competitiveness was changing. As an example, inflation in India is 10% and inflation in the US is 2% so we might think that the exchange rate should depreciate steadily at 8% per year in order to keep the CAD at 0.

The Real Effective Exchange Rate (REER) goes one step further and implements this calculation while taking into account the fact that India trades with many countries. A note of caution: The best REER measure for India is that made by the BIS.

Problem 1: The export basket is different from the CPI

The CPI is, indeed, the best measure of inflation. But the things that we export and import are not the things that the average household consumes. Hence, this crude estimate -- that the rupee should depreciate by roughly 8 per cent per year on average -- is wrong.

There is really no way to solve this problem. In an ideal world, we would have price indexes that are specific to the import and export basket. But these indexes are essentially impossible to make. Think of all the goods and services that India imports or exports, and the problems of obtaining a sensible price for each of them.

Problem 2: Productivity is not constant

All this assumes that nothing else is changing. But productivity is changing, slowly in the US and dramatically in India. In our backyard, there are a million microeconomic mutinies, through which production is becoming more efficient. Individuals learn. Firms improve their processes. Indian firms get internationalised, and global firms start operating here, all of which drives a ferment of improving knowledge. Our indirect tax policy & administration has slowly gotten better. State of the art equipment gets brought in. We learn how to utilise better and more specialised raw materials (which are often imported). Infrastructure gets better, thus linking up productive capacity in difficult locations (e.g. 100 km away from Bombay, 1000 km away from Bombay, 4000 km away from Bombay) to global markets.

The pair of graphs here is blindingly obvious and yet revealing. From 1994 till 2010, the Indian REER fluctuated between 85 and 110. Over this period, merchandise exports grew from \$2 billion a month to \$15 billion a month. What was going on? Superior Indian productivity growth.

Qingyuan Du, Shang-Jin Wei and Peichu Xie have a fascinating new paper Roads and the real exchange rate, which finds that changes in transport infrastructure have important implications for the real exchange rate.

Problem 3: The exchange rate is actually mostly about finance and not trade

So far, we have played in the simplest world where there was no capital account, and all that was going on was imports and exports (of goods and services). That's not the world that we live in anymore. We have a mostly open capital account, which gives each country the convenience of not having to achieve a CAD of 0 in every time period. We obtain microeconomic gains from international capital flows.

In 2012-13, India had inflows on the current account of \$452 billion and outflows of \$583 billion. If this was all that was going on, on the currency market, we would have trading volume of \$583 billion per year. The outflows would require buying \$583 billion. On the other side would be exporters selling \$452 billion and capital inflows filling the gap. This translates to (one way) turnover on the currency market of \$2.3 billion per day. In truth, the currency market for the rupee trades between \$40 billion and \$70 billion every day. The overall activity on the currency market dwarfs the minor business of sorting out the current account. The currency is now a financial product.

Bottom line: Exchange rate assessment is a mugs game

We started out with a simple and intuitive world, but ran into three problems. We can't quite use CPI in India and in trading partners, as the CPI is the basket consumed by the household and not a reflection of prices of the goods and services that are traded. The big story is all about productivity growth in India, which we know little about. And, the exchange rate is a financial object and not driven in the short run or medium term by trade considerations.

Policy makers should be cautious when thinking there is something going wrong with the market's exchange rate

When so little is known about exchange rate assessment, we should be cautious before determining that something is wrong with the exchange rate that the currency market has made.

None of this is new (1997, 2007). But every few years, we seem to come up against a situation where policy makers have strong views on what the exchange rate ought to be. We may all have our own personal opinions about what the exchange rate ought to be. A Hippocratic oath should prevent us from undertaking interventions in the real world that have manifestly visible costs [link, link], until we are really sure that there are corresponding gains.