Scott Sumner  

Never reason from an exchange rate change

Is Tyler Cowen's Point Well Kn... Chemo For Cats...

Suppose I had a crystal ball and saw one year into the future. At that time, the Japanese yen and the Brazilian real had both depreciated by 15% in the foreign exchange market. What would that tell me? At one level, it doesn't tell me anything for standard "never reason from a price change" reasons. On the other hand, I might make some educated guesses:

1. I would probably reduce my estimate of Brazilian economic growth. In Brazil, short run fluctuations in the exchange rate often reflect "real" factors (no pun intended) such as the state of the global commodity market, and the quality of Brazil's governance. A weaker real would most likely reflect a weaker economy.

2. Japan has a large diversified economy, with stable governance and not much susceptibility to global commodity shocks. Instead, monetary factors drive changes in the value of the yen---the stance of Japanese monetary policy relatively to other regions such as the US and the eurozone. A 15% weaker yen would probably reflect a more expansionary Japanese monetary policy and stronger growth in the short run.

I thought of this perspective when I read a recent article in The Economist:

These days a dollar buys 3.4 reais, but no one in Brazil or in other emerging markets with devalued currencies is declaring a belated victory. A cheap currency has not proved to be much of a boon. Indeed new research from Jonathan Kearns and Nikhil Patel, of the Bank for International Settlements (BIS), a forum for central banks, finds that at times a rising currency can be a stimulant and a falling currency a depressant. They looked at a sample of 44 economies, half of them emerging markets, to gauge the effect of changes in the exchange rate on exports and imports (the trade channel) and also on the price and availability of credit (the financial channel).

They found a negative relationship between changes in GDP and currency shifts via the trade channel. In other words, net trade adds to economic growth when the currency weakens and detracts from growth when it strengthens, as the textbooks would have it. But they also found an offsetting effect of currencies on financial conditions. For rich countries, the trade-channel effect is bigger than the financial-channel effect. But for 13 of the 22 emerging markets in the study, the financial effect dominates: a stronger exchange rate on balance speeds up the economy and a weaker one slows it down.

Interestingly, I reach pretty much the same conclusions about developed vs. developing countries in a simple model where neither the financial channel nor the trade channel has any effect at all on growth. In my example, monetary stimulus leads to growth in Japan due to sticky wages, even if Japan were as closed as North Korea. And in Brazil, a collapse in Chinese demand for commodities hurts growth even in a world where Brazil has no debt issues.

Of course that doesn't mean the study discussed above is wrong. Rather that you need to be careful in examining these sorts of studies. Thus when looking at the trade channel, did they account for changes in the stance of monetary policy? How did they measure the stance of monetary policy? Many economists use interest rates to measure the stance of monetary policy, but I have no confidence in those estimates, for . . . you guessed it, standard "never reason from a price change" reasons. I prefer NGDP growth as a indicator of monetary policy.

I suppose it seems like I am one of those older economists obsessed with one idea---which reminds me of a clever John Carney parody.

COMMENTS (5 to date)
Brian Donohue writes:

Excellent post.

Scott Sumner writes:

Thanks Brian.

Lorenzo from Oz writes:

What Brian said. Nicely clarifying.

Mr. Econotarian writes:

People love to complain that Greece is "tied down" by the high value of the Euro, but the truth is a Grexit would just lead to a Greece unable to get loans and with a higher level of poverty as real wages sank. Economic reforms save countries, not devaluations.

Scott Sumner writes:

Mr. Econotarian, Why not both?


Return to top