Interview with Gita Gopinath
Harvard economist on the dollar’s unique status, crises & productivity, and policy spillover to emerging markets
Published December 20, 2016
More than ever, the world economy is interlinked through massive flows of people, products, finance and information. These networks are dauntingly complex and constantly evolving. And though economists have long studied the international economy, solid answers to many of its quandaries remain elusive.
Harvard’s Gita Gopinath is at the forefront of today’s efforts to fathom this global enigma. Her work on exchange rates, sovereign debt, monetary unions, global trade, productivity and capital flows, among other areas, is expanding the frontier of research on international economics. And in doing so, it has upended conventional wisdom, established new facts and pioneered novel theory.
Economists have generally assumed that high-debt nations would prefer monetary unions with low-debt nations, but Gopinath’s work finds that they benefit more from union with a mixture of high- and low-debt countries. Most theorists argued that the “Great Trade Collapse”—the sharp, massive drop in international trade during the recent financial crisis—was caused by cost shocks specific to traded goods; Gopinath and colleagues showed that such price changes accounted for “almost none” of the drop in trade.
Raised in Mysore and educated in New Delhi prior to getting her doctorate at Princeton in 2001, Gopinath maintains her ties with India, advising the nation’s Ministry of Finance in 2013–14 and the chief minister of Kerala state in southern India currently. She also sits on the New York Fed’s Economic Advisory Panel.
Her heart remains in research, though, and those who’ve worked with her know her contributions will be enormous. “Gita was certainly one of the strongest and most promising students I ever worked with,” notes Ben Bernanke, one of her doctoral advisers at Princeton. “A very independent thinker. She has made herself a big player in international economics.”
Interview conducted October 26, 2016.
Is monetary policy too loose?
Region: You’ve just returned from Kerala state in India, where you have an honorary advisory post with the chief minister.
Gopinath: Yes, that’s right.
Region: I’d like to ask you about that role a bit later, but first want to ask for your thoughts on a broader question. As you know, Raghuram Rajan departed last month as governor of India’s central bank, and he voiced concern that central banks globally have been relying for too long on loose monetary policy and that it can be a difficult trap to escape. He said central bankers may hesitate to raise interest rates, worried that tightening might stifle growth and also that reliance on loose monetary policy might prevent policymakers from enacting other necessary reforms, fiscal or otherwise.
Do you share that concern? And do you share a related worry voiced by some economists about a new global savings glut and the risky environment that might create?
Gopinath: To decide on whether monetary policy is too loose, one needs to take a stand on what the full employment real rate of interest is. What we have witnessed over the past 20 to 30 years is a gradual decline in real interest rates, some of which has to do with demographics and an aging population and some to do with low productivity growth. A third factor is the global savings glut, the fact that there is a huge appetite—especially from emerging markets—for safe assets that in turn contributes to a deficit of demand in the world economy. Besides this, there is the cyclical downturn of the financial crisis.
Given that some of these trends are expected to persist, I don’t share Raghu’s surety that monetary policy is too loose. Inflation is still below target in all advanced economies. A couple of years ago, the tightening of credit spreads between lower-rated corporate debt and Treasuries had suggested that maybe there was excess liquidity in the United States, but those spreads have widened recently.
However, I am concerned that market expectations have coordinated on the opposite extreme that nominal rates will stay very low for the foreseeable future and that we may never see 4 percent nominal rates. This seems counter to the considerable uncertainty around the forces keeping real rates low.
Another factor that flags concerns for emerging-market central bankers like Raghu is that low interest rates in the developed world might spill over to emerging markets. With capital flowing into emerging markets, the potential for a crisis always emerges, especially when policy is reversed and interest rates rise in advanced economies. This is partly the reason Raghu has often emphasized that there should be more coordination among monetary policymakers and at least some appreciation of the fact—especially if you are a dominant currency provider like the U.S.—that actions you take can have negative spillovers.
Region: Is that the case in India? He was criticized by some in the business sector for tightening policy.
Gopinath: Indeed. It is always difficult to bring down the kind of persistently high inflation India experienced without losing out some on short-term growth. Raghu was successful but unpopular in some circles because of the high rate.
Region: That brought inflation down, as he intended. But has foreign capital flowed in?
Gopinath: Actually, in the case of India, a lot of foreign direct investment has gone in, the good kind of foreign capital. So, hot portfolio inflow is not a major concern for India. It is a concern for China, and for Brazil, but for India not so much.
Exchange rates and inflation
Region: Last year at the Fed’s Jackson Hole symposium, you presented a paper showing that the impact of exchange rate fluctuations on inflation varied widely among countries. You looked at over 40 countries, developed and emerging, and the factor that had the greatest explanatory power was the degree to which imports were denominated in the local currency versus foreign currency. So, if a country invoices very little of its foreign trade in its own currency, it will experience greater inflation when its currency depreciates against other currencies.
What mechanism or friction accounts for this “pass through” of exchange rate changes to domestic prices?
Gopinath: Countries rely on imported goods for consumption and production; consequently, inflation depends on how the prices of imported goods react to fluctuations in the exchange rate. This in turn depends on how exporting firms set prices.
What we have known for a while is that the dollar has a unique role in international trade since most of trade invoicing is done in dollars. More recent research shows that these dollar prices tend to be sticky—that is, these dollar prices are far more stable than exchange rates. For non-U.S. economies, therefore, a depreciation of their currency relative to the dollar leads to almost a one-to-one increase in the price of imported goods in their own currency and, therefore, the pressures on inflation are high. On the other hand, because dollar prices of traded goods are relatively stable, the inflationary pressures on the U.S. economy are weak.
This, of course, begs the question, why do most exporters rely on dollar pricing? There are a couple of explanations. Firstly, as world trade has become more competitive, firms that wish to preserve their market shares prefer to keep their prices stable relative to their competitors’. If your competitors price in dollars and keep these prices sticky, this gives the firm an incentive to also price in dollars.
Think of the following example. I’m a Japanese firm selling to the U.S. market, and I have to compete with a Chinese firm also selling to the U.S. market. If China prices everything in dollars and that dollar price isn’t changing much, then I don’t want to price in yen because if I do, then whenever the exchange rate moves around, the dollar price at which I sell relative to the Chinese firm will fluctuate a lot and that will be costly in terms of market share. Because I compete with other countries in the world market, I would prefer to keep my prices stable in whatever currency the world market uses.
Secondly, most exporters also use imported inputs for their production. If these inputs are priced in dollars that are relatively stable, then the exporters’ cost of production is relatively stable in dollar terms and, consequently, they prefer to price in dollars.
So, it’s very sensible why you would not expect to see multiple currencies being used in trade, but just a few. Of course, neither of these explanations tells you why it is the dollar that has this special status as opposed to the euro. To explain this, one would need to understand the historical origins of the dollar’s role in trade and capital markets.
Region: So, pass through is going to be much greater for most nations than it is for the U.S.
Gopinath: Exactly. Outside of the U.S., most nations are far more exposed to inflation driven by exchange rate fluctuations. What this also points to is that from most countries’ perspective, the exchange rate that matters is the value of their currency relative to the dollar.
Think of, say, Brazil selling to Japan. You might think that how much Brazil sells to Japan and how much it buys from Japan depends upon the Brazilian real/Japanese yen exchange rate, but it doesn’t. What matters is the dollar/real exchange rate. Even between third-party transactions where the U.S. has no role, what really matters is the dollar exchange rate. But that kind of makes sense; when countries talk about the fluctuations in exchange rate, they’re mostly concerned about fluctuations relative to the dollar.
Region: In your paper, if I recall, you provided the example of Turkey invoicing just 3 percent of imports in their lira. Is the figure in Brazil similarly small for invoicing imports in their real?
Gopinath: Yes, absolutely, in the single-digit numbers.
Pass through and monetary policy
Region: What does this imply for monetary policy in smaller nations that tend to invoice most of their trade in foreign currencies, particularly the U.S. dollar? Their central banks presumably are trying to target some kind of inflation rate, but they would seem to have very little control.
Gopinath: One of the reasons I wrote my Jackson Hole piece was because when I heard Janet Yellen talking about what a weaker dollar meant and I heard Raghuram Rajan talking about what a weaker rupee meant, they kind of said the same things. Janet Yellen would say that a weaker dollar is going to be inflationary for the United States and it’s going to improve U.S. competitiveness in world markets. That is exactly the same thing that Raghu would say with regards to a weaker rupee.
But the truth is that it’s very different. A weaker dollar has very little effect on inflation in the U.S., has an expansionary effect on exports and has a negligible effect on imports. On the other hand, a weaker rupee is highly inflationary for India, has a significant impact on its imports and has a negligible impact on exports. This expectation that a weaker currency for a non-U.S. economy is good for a country’s exports does not line up with the facts.
Region: And what does this mean for monetary policy? To what degree do small countries have any real control over their own inflation rates if exchange rate fluctuations over which they have no control affect domestic prices?
Gopinath: What it basically means is that central bankers in developing countries have to respond pretty aggressively to exchange rate depreciations.
For a commodity-exporting country, a collapse in commodity prices is clearly recessionary and calls for a cut in interest rates. However, for most of these countries that target inflation, the central bank has to raise interest rates to contain the impact of the weaker exchange rate on domestic prices. The possibility of facing high inflation and a recession is a real problem for these countries.
Region: Let me raise another question related to currency devaluation, but on a very different tack. The eurozone crisis presented its members with a quandary. A nation’s ordinary response to reduced trade competitiveness would be an exchange rate devaluation. But, of course, members of the Economic and Monetary Union (EMU) can’t unilaterally devalue the euro.
In research with Farhi and Itskhoki, you provide one of the first formal analyses of so-called fiscal devaluations that might substitute for actual devaluations. You discover the circumstances in which an ad valorem tariff plus a uniform subsidy for exports (as Keynes proposed) or value-added taxes plus lower payroll taxes (as suggested by economists more recently) would have an equal impact as an exchange rate devaluation, and when it would require additional policy change.
Could you tell us about your results, which seemed quite promising? Do you see potential political obstacles to implementing fiscal devaluations? What kind of policy traction has your research had?
Gopinath: This particular paper we wrote, as you said, in the context of the eurozone crisis … Well, I guess it’s still in crisis! So, in the beginning of the crisis, a lot of commentary said, “If only these countries had their own exchange rates, they could have devalued and that would have helped their economies.”
So what we analyzed in our paper is a set of fiscal instruments that would deliver the same outcomes as a currency devaluation. This idea goes back to Keynes, as you said, who proposed import tariffs and export subsidies as a substitute for currency devaluation. Given the illegality of using tariffs of this nature, we instead explored the role of value-added taxes and payroll subsidies or, more specifically, raising value-added taxes and cutting payroll taxes. What we found, surprisingly, is that this form of intervention did extremely well in mimicking the outcomes of a currency devaluation, not approximately but exactly.
The mechanism works as follows. When a country raises its value-added tax, foreign exporters selling to the country increase their prices, making them less competitive. Domestic firms that sell locally also face the high value-added taxes, but because these prices are slow to adjust and because the government provides them with a payroll tax cut, their prices do not rise as much. That makes them more competitive in the same way that an exchange rate devaluation would. Moreover, a fiscal devaluation has the same effect on inflation and the same effect on redistribution as an exchange rate devaluation.
In terms of policy traction, France did implement a partial fiscal devaluation. They cut payroll taxes but did not raise value-added taxes, despite the announcement that they would. So they went only halfway there.
Despite the virtues, there are political challenges to implementing a large fiscal devaluation. Countries live through a 10 percent exchange rate depreciation without immense anxiety, but if you raise value-added taxes by 10 percent, that would be very salient and likely politically infeasible.
But the broader point we made was that there are instruments other than exchange rate devaluations that a country can use to gain trade competitiveness.
Region: So, loss of competitiveness is not necessarily an impediment to monetary union.
Gopinath: Exactly. It’s not as big an impediment to monetary union.
Monetary unions and sovereign debt
Region: You’ve done quite a lot of other work on monetary unions, much of it with Aguiar, Amador and Farhi. In “Coordination and Crisis in Monetary Union,” you examine the incentives that high-debt nations face when joining a monetary union. Standard theory suggests that the low-inflation credibility of an austere monetary union might be optimal for high-debt countries, since the union’s low inflation won’t tempt high-debt nations to borrow more in the hope that debts will be inflated away.
But your research suggests otherwise. You find that in some cases, high-debt nations would want a mix of nations, some with similar high-debt profiles as well as some low-debt nations.
Could you explain that result? Why would a nation like Greece benefit from being with other high-debt countries like Italy, as well as low-debt countries like Germany?
Gopinath: The way to understand the result is to recognize that while there are some debt crises that are based on fundamentals—such as when the government spends excessively relative to output, or output growth collapses—there are others that arise from market failures like coordination failures among lenders. In the case of coordination failures, debt crises arise when panicking investors refuse to roll over debt at low enough interest rates; that panicked response pushes a country into default.
The classical argument applied to a world where such crises did not arise. As all crises were assumed to be driven by bad decisions of governments, the argument went that a country like Greece with high debt should be in a union with Germany because that made it able to credibly commit not to inflate away its debt. Greece would also want to be in such a union with Germany because then it could borrow at lower interest rates.
But what if it’s not just about the problems created by the government, but you also have problems created by the markets: financial contagion or self-fulfilling crises, coordination failures among investors, problems along those lines? What happens in that world? Well, in that world, you actually want a central bank who is able to credibly say that it will intervene in the events of markets going haywire to prevent the price of your bonds from collapsing, to prevent a crisis from occurring. This credibility arises when the union has a sufficient number of high-debt members.
Region: Mario Draghi provided that kind of reassurance in his July 2012 London speech, no?
Gopinath: Exactly! “Draghi’s put,” which we can always interpret in many different ways, but his strong suggestion was that the European Central Bank would intervene to buy sovereign debt and prevent its prices from collapsing.
Region: It was simply the promise of “whatever it takes.”
Gopinath: Simply the promise.
Region: Regardless of its actual policy implementation, and it never was—it was rendered unnecessary because it calmed markets.
Gopinath: Exactly, and those are the situations where you actually know it’s a panic-driven crisis. If it were a fundamentals-driven crisis, then if the ECB said, “OK, we’re going to bail out these guys,” and the fundamentals were actually bad, then they would have needed a bailout.
But if it is driven by market panic, then just reassuring markets kills the panic and prevents a debt crisis.
Another reason it is important is that when they were having discussions way back in the beginning about monetary union, they spent a lot of time talking about what happens if we bring Germany and Greece together and they have very different growth rates and inflation rates. If we have just one instrument, just one interest rate for both those countries, that can be a problem. That was one of the standard concerns.
What they spent a lot less time thinking about is, what if you bring countries with different levels of debt together? Who’s going to play the lender of last resort?
Region: But the Stability and Growth Pact addressed that to some degree, didn’t it? And your research provides support for the debt ceiling set in the pact.
Gopinath: The Stability and Growth Pact does not help to deal with self-fulfilling debt crises. What it does help with is addressing excessive debt accumulation in the union. An individual country in the union may think, “I’m a small part of the whole union, so if I on the margin raised my debt by a little bit, I’m not going to have much of an effect on the central bank’s temptation to inflate.” But if every country does that, you will end up having an effect on inflation. That’s the fiscal externality, and that’s why you would need a Stability and Growth Pact.
The fact that it was forbidden for the European Central Bank to be the lender of last resort was, in the original rules of the game for the ECB, meant to prevent moral hazard problems of countries borrowing excessively. I think it’s a perfectly good argument, but you have to recognize that there will be times when it’s not about a government behaving badly, but about market panic.
Our paper was basically arguing that—while not ignoring issues in the debt market—monetary policy has a legitimate role to play. After the Draghi put, the ECB was taken to court as it was argued that this was not part of the original [EMU] agreement. This is a very reasonable question for people to ask: Is there a role for monetary authorities in debt crises? And I think our paper basically says yes. You certainly do not want to be the central bank that always bails out governments in trouble, but it absolutely has to give itself the possibility of doing it—once in a while.
Region: Let me turn to work you’ve done with Itskhoki and Neiman on trade collapse. International trade dropped dramatically during the financial crisis and the Great Recession, falling far more than production and GDP. Many theories have tried to explain the phenomenon through cost shocks related to trading, like trade financing. But you took a close look at U.S. export and import prices at the height of the financial crisis, and that evidence suggests that such explanations are faulty.
You write that “variation in price changes account[s] for almost none of the variation in quantity changes in differentiated sectors,” and you note that “the sharp reduction in differentiated goods trade was entirely a quantity-drive phenomenon.”
Why is the distinction between differentiated and nondifferentiated products so important to understanding why trade dropped so dramatically?
Gopinath: By “differentiated” products, we mean products with unique characteristics as opposed to goods that are homogenous and traded on exchanges. Oil or other commodities—those are homogenous products. Differentiated goods are, say, a particular style of shoe.
Region: So, you used the Rauch index to distinguish between the two classes of goods: differentiated and homogeneous commodities.
Gopinath: Yes. The Rauch index is very useful to tell apart products for which a firm has some market power and those for which firms are price takers. If higher financing costs raise the cost of production for firms, then we should see firms that set prices raising those. But we see very little of that, at least where they’re selling to U.S. markets. We found that prices barely change, but quantities move a lot.
Region: Yes, you found that the 30 percent decline in trade value was entirely quantity driven.
Region: So, if the collapse isn’t price driven, what is going on? What’s the mechanism?
Gopinath: The collapse in quantities could have arisen from a rise in prices and/or a decline in spending. In a recession like the one that followed the financial crisis, overall consumption and investment spending fall as people lose jobs and firms do not invest. What our paper points out is that the decline in quantities was primarily driven by the decline in spending and had very little to do with prices, as prices barely moved.
Productivity, financial crises and capital allocation
Region: I’d like to ask you about two papers you’ve co-authored on productivity declines, one analyzes how capital flows impact productivity and the other examines the relationship of productivity to import collapses during financial crises.
In the latter paper, you look at Argentina’s financial crisis at the turn of the century, when the dollar value of its imports fell 69 percent and manufacturing productivity fell about 11 percent.
Your analysis found that the usual explanations—entry and exit of firms, and imported product varieties—explained no more than 23 percent of the import decline. By contrast, churning of inputs within firms appears to play a very sizable role.
Could you explain the mechanism at play here? You provide some really striking examples—purchasing patterns for cooling fans and lathes by large firms. Purchases plunged from quite high levels to zero almost immediately and didn’t recover until late in 2003, I think.
Region: How does that internal churn at individual firms generate such a large impact on the entire nation’s productivity?
Gopinath: Firms in Argentina rely heavily on imported inputs for their production. These are inputs for which there are few domestic substitutes: Domestic substitutes are inferior in quality; even more extreme, nobody domestically is producing them in the short run.
This ties back to one of the first questions we talked about, the currency that imports are invoiced in. Whatever Argentina imports is invoiced in dollars, so if Argentina’s peso depreciates 70 percent against the dollar, its prices at the dock go up 70 percent.
So manufacturers that should be using this particular imported input now have to look for a domestic substitute. If they don’t find it, that’s going to affect their ability to produce. The firm might have the same amount of labor, the same amount of other kinds of capital, but if it doesn’t get the right imported inputs, it is not going to be able to produce much.
That was the point we were trying to make: If you look within firms, you do see this kind of churning. They stopped using a lot of imported inputs that they were previously relying on, and that lowered productivity substantially.
Productivity and capital allocation, post-euro
Region: One last question, about your recent research on productivity in Southern Europe. You look at Spain in depth, but more broadly at a few other countries. And you look at productivity relative to capital allocation after the introduction of the euro in ’99 as nations joined the EMU.
You find that productivity losses increased significantly as did dispersion in returns to capital. And it’s partly because capital was increasingly misallocated to large, high-net-worth companies, rather than productive companies. Why was that? Was it a question of realpolitik, connections rather than profitability?
Gopinath: I don’t think we managed to answer your last question. What we did document was the fact that large (high-net-worth) firms were able to leverage up more relative to small firms. Whether this was because large firms were more politically connected or because banks were just more comfortable lending to large firms—we don’t know. We don’t have the data to be able tell these stories apart.
An interesting insight of the paper is that while in a setting of efficient financial markets, lower borrowing rates for a country is a good thing, that isn’t necessarily the case with inefficient markets. When financial markets are imperfect, as they tend to be, lower rates can harm the productivity of a country as resources get misallocated to firms that are not necessarily the most productive. Interestingly, this ties back to the first question we discussed about the costs of too-low interest rates. Misallocation would be one such cost.
Region: In addition to Spain, you looked at Italy, Portugal, Germany, Norway and France. What did you find for the Northern European countries, where presumably credit markets are better developed?
Gopinath: We found that over the period 1999 to 2013, while measures of misallocation deteriorated for countries in the south —Spain, Italy and Portugal—it remained mostly unchanged for countries in the north like France, Norway and Germany.
A policy role?
Region: Thinking back to your work as an adviser to Kerala’s government, is a policy role something you want to pursue in the future?
Gopinath: I am always eager to contribute towards India’s development. However, I don’t see myself permanently transitioning to a policy space. There is something to be learned from interacting with policymakers to understand how things actually function, and I find that quite interesting.