Just say no to exchange rate intervention

Just in the past week, we had a meeting with the president and the economic principals and we had ruled out any currency intervention,” White House economic adviser Larry Kudlow, CNBC, July 26, 2019.

“I could do that in two seconds if I want to. I didn’t say I’m not going to do something.”
President Donald J. Trump, Washington Post, July 26, 2019

Whenever possible, policymakers should explore a broad set of options before responding to challenges they face. However, when the President and his advisers recently discussed foreign currency intervention, we hope everyone quickly concluded that it would be a profoundly bad idea.

Before we get started, it is important to explain what foreign currency intervention is and how it is done. Very briefly, an intervention is the purchase or sale of one currency for another. For example, when the Swiss authorities wished to cap the value of the Swiss Franc in 2011, they sold Swiss Francs in exchange for euros. In Switzerland (a case we will return to below), the central bank is responsible for foreign exchange policy. In the United States, the Treasury Department takes the lead, while the Federal Reserve simply executes the transactions in its role as the U.S. government’s banker.

Regardless of who is making the decisions, foreign exchange intervention alters the balance sheet of the central bank. For example, if the United States were to intervene to weaken the dollar―that’s what Mr. Kudlow says he and President Trump discussed―the Secretary of the Treasury would instruct the Federal Reserve to sell U.S. dollars in exchange for some other currency (presumably a combination of euros, Japanese yen, British pounds and possibly even Chinese renminbi). The result would be an increase in the size of the Fed’s balance sheet.

Of course, the Federal Open Market Committee could sterilize the impact of the intervention on the size of its securities portfolio by selling some of the U.S. Treasurys it currently owns. (The actual mechanics of this are a bit complex, involving something called the Exchange Stabilization Fund.) However, in light of the enormous scale of the U.S. dollar market, sterilized intervention that leaves monetary policy unchanged likely would have little sustained impact on the value of the currency. Instead, it would add to short-term currency market volatility, fueling uncertainty and risk.

So, our first two points about U.S. foreign currency intervention are: the Secretary of the Treasury makes the ultimate decision on what, when, how, and how much; and, any effective intervention has a durable impact on the Fed’s balance sheet, so it is a monetary policy operation.

Turning to our third point, over the past quarter century, market forces have determined the value of the U.S. dollar. U.S. intervention, when it has occurred, has been rare, small and aimed to help others.

Once upon a time, the U.S. government intervened frequently. From the collapse of the Bretton-Woods fixed exchange rate system in March 1973 through August 1995, the Treasury (jointly with the Federal Reserve) intervened 244 times in the yen-dollar market, and 971 times in the Deutsche Mark-dollar market: that’s an average of four and one-half times per month!  Importantly, however, these were nearly always joint operations, done in coordination with the country issuing the currency on the other side. And, they were very small – the median operation was $128 million for yen, and $40 million for the Deutsche Mark.

Since 1995, the United States intervened exactly three times. On June 17, 1998, U.S. authorities joined the Bank of Japan’s efforts to prevent further yen depreciation, acquiring $0.8 billion worth of yen. On September 22, 2000, in an effort to strengthen an extremely undervalued euro, the Fed made a small, symbolic purchase of €1.5 billion in a joint operation with the European Central Bank and the Bank of Japan. Similarly, on March 18, 2011, to counter the sudden rise in the value of the yen that followed Japan’s March 11 earthquake and tsunami, the Fed sold $1 billion of Japanese yen in a coordinated intervention involving the ECB, the Bank of Canada, the Bank of England, and the Bank of Japan. (Neely provides a detailed discussion of intervention by the G7 since 1973.)

What does this mean for the value of the dollar? In the following chart, we plot the real broad effective exchange rate of the U.S. dollar starting when intervention effectively ceased. This is the trade-weighted average of the dollar’s value in relation to a set of 40 countries plus the euro area, adjusted for changes in domestic prices (see here). An increase in this index represents an appreciation. The chart shows both the monthly average level of the index, the average over the entire period (109.8), and both a one- and two-standard-deviation band. (The standard deviation is 9.1.) We draw two very simple conclusions. First, the real exchange rate is relatively volatile, frequently moving by 5 percent up or down over a period of a few months. Second, the current reading of 117.2 is less than one standard deviation above the mean of the past quarter century. Furthermore, while the dollar has appreciated by nearly 8 percent since early 2018, it remains below its early-2017 level, and far below its peak in the early 2000s.

U.S. real broad effective exchange rate, September 1995-June 2019

Source: FRED, and authors’ calculations.

Source: FRED, and authors’ calculations.

Our fourth point is that the dollar’s apparent “overvaluation”—estimated at about 10 percent in the IMF’s recent 2019 External Sector Report (see page 95)—reflects the U.S. economy’s recent outperformance.

First, the U.S. economy is expanding faster than the average advanced economy growth rate of less than 2 percent. Second, despite an unemployment rate at a 50-year low, the record-long U.S. expansion exhibits few imbalances. As a result, recession probability indicators remain subdued (see here and here). Third, at 2½ percent of GDP, the U.S. current account deficit has been stable for a decade. Fourth, the U.S. financial system is more robust than that of the euro area, Japan and China, all of which have been struggling for years. As a result, despite low national savings and high government debt, U.S. financing costs remain very low. Put simply, investors still find the United States a very desirable destination for their funds, helping to buoy the dollar.

Against this background, the long-run harm from U.S. currency intervention would be very large, while the short-run gains would be very small (if any). We see two enormous negative consequences from any action aimed at weakening the dollar. The first concerns the Federal Reserve. To be effective, intervention requires sustained easing of monetary policy. In pursuit of an exchange rate goal, the Federal Reserve would be compelled to sacrifice monetary control, undermining the credibility of its commitment to price stability and maximum employment. As if this weren’t bad enough, by treating international trade as a zero-sum game, intervention would serve as a beggar-thy-neighbor policy employed explicitly to harm our trading partners, almost certainly triggering retaliation.

Starting with monetary control, keep in mind that the daily volume in the foreign exchange market exceeds $5 trillion per day, and that the dollar is one side of nearly 90 percent of transactions (see here). This means that, to have any sustained impact, the volume of intervention would have to be very large and persistent―our guess is that it would need to be unprecedented in scale (measured in the hundreds of billions of dollars). Should others retaliate, the scale would need to be even bigger.

To give some sense of the magnitude of the actions that effective U.S. dollar intervention might require, consider the Swiss experience. On September 6, 2011, with the Swiss Franc trading at near parity to the euro, the Swiss National Bank (SNB) announced their intention to keep the exchange rate from falling below 1.20 using unsterilized intervention. The press release announcing this policy shift contained the following sentence: “The SNB will enforce this minimum rate with the utmost determination and is prepared to buy foreign currency in unlimited quantities.” (Our emphasis.) And, unlimited it was.

The following chart shows the Swiss Franc/euro exchange rate (in red on the left-hand scale) and the size of the SNB’s balance sheet, as a percentage of Swiss GDP (in black on the right-hand scale). The policy was explicitly in effect until January 15, 2015. During those three-plus years, the SNB’s assets rose by something like 45 percent of GDP. And, from 2015 to 2018, assets continued to increase rapidly. At this writing, the SNB holds assets equal to nearly 120 percent of Swiss GDP. (We discuss this episode in more detail in a post from January 18, 2015.)

Swiss Franc-euro exchange rate and Swiss National Bank Balance Sheet (monthly), 1999-2019

Source: Swiss National Bank, FRED and authors’ calculations.

Source: Swiss National Bank, FRED and authors’ calculations.

The lesson is clear: if authorities adopt an exchange rate goal, they give up control of the size of the central bank’s balance sheet. Put differently, there is no escaping the open-economy trilemma: in a world of free cross-border capital flows, a fixed exchange rate and an independent monetary policy aimed at domestic stabilization are incompatible (see, for example, here). The Swiss chose the fixed exchange rate, giving up the discretionary monetary policy. And, the SNB intervention was absolutely massive!

We see the Swiss experience as a warning, rather than a guide, to U.S. policymakers. First, the United States is the world’s largest economy, nearly 40 times larger than Switzerland, so the economic impact of efforts to weaken the dollar on our trading partners would be significant.

Indeed, unilateral intervention by the U.S. authorities (as opposed to the coordinated, multinational actions they took part in during the 1970s and 1980s) would be a distinctly hostile act. Falsely treating global trade as a zero-sum game, the goal of currency intervention would be to seek a U.S. advantage in global trade at the expense of others. For the United States, where the sum of imports and exports remains a modest 30 percent of GDP (versus nearly 120 percent in Switzerland), the benefits are likely to be small. But in the rest of the advanced world―the euro area is nearly twice as open as the United States―the potential impact would be large, fueling pressure to retaliate. That is, a decision to intervene in the dollar probably would trigger a currency war.

Second, given the governance structure in the United States, where the Secretary of the Treasury is in charge of foreign exchange policy, a U.S. shift to an exchange rate target would result in a loss of Federal Reserve independence. While we doubt that the balance sheet of the Fed would rise to 100 percent of U.S. GDP―that would be an increase to $21 trillion from the current $3.8 trillion―it would have to rise substantially. This would be a form of fiscal dominance, sacrificing the Fed’s legally mandated objectives of low inflation and low unemployment for the executive branch objective of reducing the value of the currency. If, as we expect, the central bank were to resist such a loss of monetary control, open conflict between the Treasury and the Fed could itself undermine the attractiveness of U.S. assets and the dollar.

It would be possible for the Fed to keep policy rates above zero using the current toolkit (see here), but only at a large cost. To see why, note that the Fed would continue to pay a positive interest rate on banks’ reserve deposits (that is the interest on excess reserves, or IOER), but it would be purchasing German and Japanese bonds that currently pay negative interest rates. The “negative carry”—the gap between the U.S. and foreign interest payments—of roughly 2½ percent would diminish the profits that the Fed turns over to the U.S. Treasury. This translates directly into a larger federal budget deficit.

Finally, the dollar’s dominant role in international trade and finance arises partly from U.S. authorities’ willingness over decades to let the dollar’s value be market determined. As we have written before, banks outside the United States have approximately $15 trillion of U.S. dollar liabilities. Foreigners hold more than $6 trillion of U.S. Treasury securities. And, roughly 30 percent of world trade involves non-U.S. entities pricing their products in dollars. The U.S. benefits from this are a combination of reduced financing cost and the ability to sustain a large current account deficit. On the first, the current consensus is that the U.S. financing benefit is in the range of 0.5% of GDP per year (see here).

Perhaps most important, foreigners supply funds to Americans at attractive rates in part because they trust the U.S. government will not interfere in markets. If that were to change, would the dollar remain the currency of choice? Would the dollar continue to the leading currency for trade invoicing? Would banks still be willing to issue liabilities in dollars (and would their customers want them)? Would the cost of financing the U.S. Treasury remain so low? Would interest rates paid by U.S. firms and households continue to be so favorable?

We don’t know the answer to any of these questions. What we do know is that that the risks for the United States are all to the downside. That is, should the authorities decide to directly influence the value of the dollar, we could very well see a reduction in the use of the dollar and an increase in the cost of financing for anyone borrowing in dollars. Why would anyone sacrifice these significant long-term benefits for what are sure to be slim and transient gains?

Today, the United States is the anchor and foundation of the international financial system. A decision to intervene in the dollar market would signal that the United States is now an exchange rate manipulator that can no longer be trusted to play that fundamental global role. We can only hope that the President takes this self-destructive option completely and permanently off the table.