Why Rich Countries Should Float Their Currencies

This article provides more depth to some comments I made in"MMT and Constraints". I explain why developed countries should allow their currencies to float, which is the policy stance advocated by Modern Monetary Theory (MMT). It is probably a good idea for developing countries to float their currencies as well, but they face inherently difficult policy problems that I do not know enough about to comment on. The implication of advocating a free-floating currency is that I do not see the "external constraint" as being a serious issue, or at least an issue that policy makers can hope to do anything useful about.

In that previous article, I made some quick comments in response to an assertion by Thomas Palley that the "foreign exchange market constraint" is very important for countries other than the United States. Since he did not explain that assertion, I was unable to offer a very detailed criticism. My response was too short, and drew some comments. I expand my explanation here.

I will first explain why I do not think that there is a significant external behavioural constraint on policy makers; but an accounting constraint obviously exists.

I will also note that what I am writing is my opinion, and does not necessarily reflect the views of the economists who developed MMT. I think that poorer countries face some difficulties with free-floating currencies, a view with which they may not agree.

Fiscal Constraint

It is often declared that governments face a constraint on fiscal policy if foreigners buy their bonds. As long as they are issuing bonds issued in a domestic currency, I will just assert here that foreign ownership is a red herring - foreign bond holders have no special status, and so if there is a behavioural constraint on government fiscal policy, that constraint would appear even if 100% of the government debt is owned by domestic investors. That is a big topic, and I would refer to my previous post, "What Is  A Government Bond?" to see my thoughts on the subject.

But as a quick explanation of my distinction between accounting and behavioural constraints, imagine a country where the fiscal authority (Treasury) does not issue bonds, rather it "finances" expenditures by overdrafts at the Central Bank. The Central Bank then issues bonds to drain "excess reserves" in order for the yield curve to remain at a target level. (If the central bank did not pay interest on excess reserves, the interbank rate would be driven to zero if significant excess reserves exist. This would mean that the Central Bank would lose control over the overnight interest rate.) It is clear that there is no possibility of the government defaulting with this institutional arrangement. The government can spend as much as it wants without worrying about "financing" - but it does face the real constraint that very loose fiscal policy is inflationary. Therefore, there is no "financial" constraint on government behaviour, just a real constraint (inflation).

However, government finance obeys the accounting identities that are painstakingly detailed in the article by Palley; changes in the monetary base are related to net bond issuance and the fiscal deficit. The MMT argument is that the term "constraint" is being deliberately misused by economists, blurring the distinction between the technical meaning of the word in mathematics with the common language meaning where it is a limitation on behaviour.

However, I will note that the institutional arrangement I describe above does not hold for real-world economies; there are legal constraints on government finance which make default possible (e.g., the debt limit in the United States). The question is how much those legal constraints matter; I will have to cover that topic at a later point as a result of its complexity.

Financing The Current Account

I argue that the situation for the "external constraint" is similar to my previous example. There is an accounting relationship that says that foreign entities* have to place financial inflows into a country to match the outflows corresponding to its current account deficit (ignoring small external flows like foreign aid transfers). This seems to imply that foreigners have veto power over a country's policies, and I have seen arguments that domestics are forced to borrow in foreign currencies as a result of the accounting.

However, the volume of foreign exchange transactions have been found to be an order of magnitude larger than what is needed to support trade flows. This hyperactivity is partially the result of foreign exchange trading, but it also reflects very large gross cross-border capital market flows. These flows determine the relative value of currencies. The ultimate counterparty to an importer is most likely a foreign investor who wishes to run foreign exchange risk; there is no necessity for domestics to have to borrow in foreign currencies to finance imports.

It is very possible that a fall in the currency will make a current account worse (as imports become more expensive, and quantities do not immediately adjust), a point that was made in the comments to my previous article. But since the valuation of currencies are driven by capital flows, not trade flows, this cannot go on forever. The domestic wage bill of exporters is being deflated versus international peers, and they become more competitive. (Imported input prices rise in local currency terms, but they pay the same world prices faced by competitors.) Since the exporters are more competitive, expected future profits rise, making domestic equities relatively more attractive. This effect will eventually limit the weakness of the currency. And the empirical reality is that the developed market currencies move around a lot, there appears to be a limit how far they can deviate from a purchasing-power parity fair value estimate.

(Update: Another, probably more important factor is the fact that the investors in developed countries have considerable foreign currency assets as a result of diversification. Since the investors have local currency liabilities (actuarial as well as financial), a rapid fall in the currency is an opportunity to repatriate those foreign assets. This asset repatriation will also limit the downside to the currency. I did not use the term "rich" in the the title by accident; having large financial asset balance sheets aids these countries weather out volatility.)

Although currency volatility is disruptive, companies can use currency hedges to limit the impact of short-term volatility. In a country like Canada, where currency volatility is expected, business managers have learned the hard way that external currency economic exposures need to be controlled. (For example, the next year's expected foreign currency revenue may be hedged, giving time to react to forex moves.) Conversely, what we we saw in Asia in 1997 was that businesses had come to rely on central banks stabilising the currency, and they engaged in speculative cross-currency exposures (such as borrowing in U.S. dollars because "interest rates were lower"). To paraphrase Minsky, instability is stabilising.

In any event, I argue that a bid for a developed market currency always exists at some price, because of the potential demand for local currency financial assets. It would require the currency to essentially cease to exist in order for there to be no demand for the currency. This could result from the government repudiating its debt, or regime change (war, revolution). Additionally, it could result from a mass default by the domestic banking system. The latter possibility is very real, and it explains why that it is necessary for regulators to prevent domestic banks from building up foreign currency exposures (as seen in Iceland). This implies that there is a constraint on regulation - banks must be regulated in a fashion that is coherent with a free float in the currency. Many countries have failed to regulate their banks properly (e.g., foreign currency mortgages are commonplace in many countries), but their incompetence does not mean that it is impossible to run a banking system properly.

Under the assumption that there is always a bid for the currency, it will always be possible to finance a current account deficit. The only question is the price at which the financing occurs.

(Since "emerging markets" lack the deep financial markets to absorb these capital flows, they face an inherently difficult problem. I do not have any advice for countries in that position, other than they appear to face a choice: either attempt to develop behind capital controls and manage their currency, or let the currency float and be extremely vigilant about currency mismatches.)

What About Inflation?

One expected response is: OK, there will always be a price for the currency, but currency weakness could be highly inflationary. This high inflation will be unacceptable to authorities, and they will be forced to react in response.

I would agree, but only to a very limited extent. A weaker currency is presumably stimulative for an economy, and the mainstream response is that interest rates would have to be adjusted in response. That said, the effect is fairly weak, and it is only one effect amongst many that central banks react to. The sensitivity of the price level to the currency is low; a drop of 20% in the currency may only raise the price level by about 1%. (The U.K. appears to be more sensitive than other countries, but the recent experience was clouded by the decision to raise consumption taxes, which had a large mechanical upward effect on the price level. Nick Edmonds raised the example of the U.K. in the comments to my previous article; but given the issue of the VAT hike, I skipped over that case in the original text.) Given that the effect is so weak, it is very difficult to get a good sensitivity estimate, since there are other variables changing at the same time.

The Bank of Canada used to use the "Monetary Conditions Index", a rule of thumb in which a 10% depreciation the currency was equivalent to cutting rates by 1%. However, they abandoned that concept by the late 1990's once it was clear it did not offer any useful information.

The sensitivity of the price level to the level of the currency appears to be smaller than the sensitivity to energy prices. As the chart above shows; there are relatively large deviations from CPI headline from the ex-food and energy ("core") measure.**  And the reaction of sensible central banks to rising energy prices is to do nothing. (The ECB would disagree, as they hiked rates in 2008. I am in the camp that was a policy error.) An energy price spike represents a one-time upward shock to the price level;  under current circumstances where indexation in the economy is limited, this does not lead to persistent inflation. (In the 1970's, wages tended to be indexed to CPI, and so anything that raised the price level tended to lead to "second round" effects. Luckily for policymakers, that tendency  has generally faded, although the U.K. appears to retain relatively more inflation persistence.)

I argue the situation for currency moves is similar to the case for energy. It takes a very large move to generate a measurable impact on CPI inflation. But once that move is finished, the expectation should not be that it will repeat, rather the currency will reverse or at least stabilise (on asset relative valuation grounds). It makes little sense for the central bank to react to such a one-off move.

The situation would be different if the currency was steadily losing value due to higher domestic inflation than developed market trading partners. In this case, the reasonable expectation would be that the fall in the currency will continue. This is not the case for any developed market for the past 20 years; inflation rates are all near 2% (except for Japan, which has achieved price level stability). However, the currency is not the problem, the problem is inflationary domestic policy settings; the currency is just one channel for inflationary pressures.

As a result, I do not see imported inflation posing a particular problem for policy making. Even if the central bank reacts to every wiggle in inflation, fiscal policy can safely ignore the currency.

This may be less true from some countries, such as Iceland, that have considerable dependence upon consumer imports. In such a case, the external constraint is a real economic constraint that has to be taken seriously. But even then, the question arises - what can policymakers do about it?

"There Is No Alternative"

My final point is that there is almost no sensible policy space involving currencies. We did not arrive at a free-floating currency system as the result of a fad in economic thought; we are here as the result of the collapse of the Bretton Woods system. Yes, the Europeans are attempting to keep yet another failed currency peg system together, because that is what European policy elites do. Most other countries realise that attempting to peg the currency is doomed. (Even the continental Europeans allow the euro to float versus the yen, dollar and sterling).

The British government was repeatedly humiliated during the Bretton Woods era in its attempts to keep the pound at overvalued levels. And this was a period when the capital account was closed, government intervention in the economy was at its peak, and the culture of amoral bypassing of regulations with derivatives did not yet exist. They were defeated by "leads and lags" - international traders lengthening and shortening the terms of accounts receivable and payable. This mechanism to defeat currency pegs will always exist, and it would be supplemented by capital market players using derivatives. As soon as the ink dries on a law limiting currency trading, investment banks would hire away the technocrats who helped draft the law. Unless the financial sector in all of the concerned countries are completely crushed under the heel of government interference, attempting to peg currencies would be doomed.

(As the Japanese and the Swiss have shown, it is possible to limit the upside in a currency, as a government has an unlimited ability to issue money to buy foreign currency. That only provides control in one direction, and it also presumes that the other country whose currency is being purchased does not object. As a result, this type of intervention will only happen in rather special situations.)

Policymakers could attempt to "coordinate" policy to guide currencies, "guiding" markets without attempting to peg their levels.*** This currency guidance was attempted during the 1980s and 1990s, but the studies that I have seen showed that these policies were not particularly effective. (The last coordinated currency intervention was the effort to raise the value of the euro shortly after it was introduced.****) All that policymakers can do is help nudge the currency back towards "fair value" once it was clear that the trend of moving away from fair value has broken. And the question remains: who is going to agree on the direction of intervention? The Europeans appear to be the only masochists who value a strong currency; not many other policymakers would be willing to intervene to strengthen their currencies in the current environment of low inflation and chronic overcapacity.

And attempts at "guiding" the currency is a dangerous policy. As we saw in Asia during the 1997 crisis, business leaders are willing to run massive currency bets if they think "the central bank has their back". Currency volatility should be welcomed, as it helps keeps imbalances under control.

In conclusion, for most rich developed economies, I do not see very large costs associated with floating the currency. Currency levels can move a lot, but those end up being self-limiting, one-time shocks. And realistically speaking, there is no chance of implementing policies to control the level of the currency. However, the situation for some smaller countries and for poorer "emerging markets" appears to be more difficult, but once again, they have limited ability to do anything about the currency without closing the capital account.

Footnotes

*Which may include pools of foreign capital that are owned by domestic investors, which is an extremely important point when discussing Japan.
** I am skating over two objections to this example. Firstly, there is the issue that oil prices are a world price and so currency weakness increases oil prices, but what we see in practice is that energy prices are much more volatile than developed market currencies. Secondly, the U.S. has relatively low taxes on gasoline, so the energy component of the CPI is more volatile than other countries.
***Somewhat similar to their stance towards long-dated bond yields; central banks comment on bond yields if they do not like where they are, but they are not currently attempting to set them.
**** The currency swap arrangements the Fed entered into during the financial crisis were not done to address the levels of foreign currencies; rather they were there to deal with the funding mismatches of foreign banks. These banks originally piled into U.S. dollar assets, using funding from the U.S. money markets. As money market players drew away from those banks, the banks were in danger of losing the funding for those assets. The banks were able to get funds from their local central banks, but this funding would be in the local currency, not U.S. dollars. The banks could not use that funding, as they are not allowed to run the currency mismatch. They thus turned to use currency swaps with the Fed, in which the Fed swaps U.S. dollar funding in exchange for the foreign currency.

(c) Brian Romanchuk 2014