Dominant Currency Paradigm - An interesting progress in open economy macroeconomics after IS-LM-BoP

Does it matter if the exporters/importers of a country show preference for a certain currency while invoicing the trade? I had dabbled with this few years ago here, which was based on the Dominant Currency Paradigm presented by Gita Gopinath, Casas The expanded theory titled "Dominant Currency Paradigm" (DCP) was published at the AER in 2020 and the corresponding replication code with data is made available here

DCP is a significant development after Mundell-Fleming's model using IS/LM/BP curves which has been a workhorse model for decades. Mundell-Fleming (and the Devereux model) uses bilateral exchange rates to arrive at equilibrium rates, whereas DCP takes into account the additional factor that countries may use a third currency as invoicing currency while trading bilaterally. For example, traditional Mundell Fleming model for trade between India and Japan would use INR/Yen pair to understand disequilibrium whereas the actual invoicing of this trade may happen in  USD due to various factors. DCP corrects this by allowing usage of third currency, and goes on the prove that for the world that we are currently in, the USD is the dominant currency, with most third currency invoicing taking place in USD. India invoices around 86% of its imports in USD and around 10% in Euro as per the paper, the importance/pitfall of which would be clear in a moment. 

Let me summarize DCP's key takeaways: 

- The terms of trade for a country are uncorrelated with exchange rates in short and medium term: Mundell-Fleming predicts that, because prices are sticky in the producer's currency (Producer Currency Paradigm - PCP), a nominal exchange rate depreciation is associated with a depreciation of a country's terms of trade.  That is, currency depreciation would lead to increase in ratio of price of imports to price of exports. If INR/USD deprecates to Rs 100/USD, our import prices would shoot up leading to deteriorated terms of trade with the USA. A counter argument was given by Devereux that as the prices are sticky in destination country's currency (Local Currency Paradigm - LCP), a depreciation in exchange rate would lead to appreciation of country's terms of trade. With 100 Rs/USD, our exporters would now get more rupees for each dollar worth of goods exported and hence terms of trade would improve. 

DCP researchers find no such correlation. They find that the terms of trade are not correlated with bilateral exchange rates - neither PCP nor the LCP holds. This is counter-intuitive as it is usually believed that nominal appreciation/depreciation of currency would worsen/improve the terms of trade depending on whether the country imports more or exports more. Rather the terms of trade are defined in a way to align to this view and this mental model drives a lot of understanding about future direction of currency movement. If this is not true, a lot of other things that we know about forex markets appear to be standing on shaky grounds. The researchers have used non-commodities data, after excluding agri commodities (chapter 1 - 27) and certain engineering commodities (chapter 72-81) constituting almost 25% of total trade. Thus the research covers 75% of the international trade as per my estimates. (The paper says it covers 91% of trade, I am not sure how) 

- Dollar is the dominant currency and influences the international trade disproportionately: DCP takes a stand that neither PCP nor the LCP models hold true when it comes to international trade, but it is a small set of dominant currencies led by USD that determines the direction. The paper estimates that a 1% appreciation in USD against all other currencies leads to a 0.6% shrinking of global trade after controlling for business cycles.

The effect of dollar appreciation on exports outweighs bilateral currency appreciation (appreciation of recipient country's non-dollar currency).  If India is exporting to Japan, the effect of dollar appreciation would lead to greater fall in exports when compared with appreciation in Yen. This is the effect of invoicing Inso-Japan trade in dollars. 

- The price pass-through for non US economies  is directly in proportion to trade invoiced in USD, the dominant currency:  The trade mechanism for exchange rate adjustments work for all economies except the dominant currency economy, i.e. USA. This leads to two further findings. US economy is negligibly impacted by price changes in Dollar - which implies that the inflation targeting monetary policies run by other economies has insignificant impact on US. On the other hand, price changes in US arising out of monetary adjustment leads to impact on other economies through trade route (there may be other mechanisms of adjustments through capital flows which is not dealt directly). The exchange rate pass throughs are significant for all economies but US is immune to it. Any adjustment in US has to happen through other routes other than import price pass through. 

Where does that leave us? A revision is in order in the way people construct mental models about exchange rate effect on exports. A general thinking is that a depreciation in rupee is good for our exports as we become more competitive. The paper shows that it may be contrary and the trade may actually shrink when the dollar gets stronger. 

What next:

One concern I find in this paper is that there is no discussion about hedging of currencies at the time of invoicing. Usually, long term players cover their risk through currency positions they take while raising invoices in order to immunize against currency fluctuations in short and medium term. This is especially true for non-commodity trades. This, and the fraction of players who do this, may have a significant role in the way the pass through works. This needs investigation. 

Also, it is difficult to imagine that if a country mandates local currency pricing (imagine India issuing a fiat to quote all invoices in INR), it would somehow lead to lesser pass through. The exchange rates are an indirect indication of relative prices and to imagine that the way it is quoted leads to effect on trades is counter-intuitive. I will take time to get used to this new smell. To that extent Mundell-Fleming/Devereux made more intuitive sense.