Milton Friedman on exchange rates #4

Always floating exchange rates?

The theoretical literature often distinguishes between completely fixed exchange rates on the one hand and freely floating exchange rates on the other. Milton Friedman has pointed out, however, that this sharp distinction often does not apply to the exchange rate regimes that are used in practice. As well as the two “extremes” (completely freely floating exchange rates in which the central bank never intervenes, and a firmly fixed exchange rate with no fluctuations allowed), a common system is to have fixed but adjustable exchange rates – or rather exchange rate bands. The Danish krone, for example, can swing freely within a band of +/- 2¼% around the “fixed” euro exchange rate of 7.44 kr. per euro.

The three global majors, the US dollar, the Japanese yen and the European euro do float freely against each other – as do a number of smaller currencies, such as the Swedish krona, the British pound, the Korean won and the New Zealand dollar. However, even such in principle freely floating exchange rates do not prevent the central banks of these countries from being active in the FX markets from time to time.

A system with fully fixed exchange rates is in practice the same as a monetary union and involves the complete abolition of any form of monetary independence. One example is Hong Kong. The Hong Kong Monetary Authority is obliged at all times to exchange US dollars for a fixed amount of Hong Kong dollars (7.8 Hong Kong dollars per US dollar). This means in essence that Hong Kong is in a monetary union with the USA – the only difference is that Hong Kong has its own banknotes. A second example is the European Monetary Union, where all members have given up monetary independence and left all monetary policy decisions to the European Central Bank.

An example of a system with fixed but adjustable exchange rates is the European fixed exchange rate mechanism, the EMS. Members of the EMS pursued a mutual fixed exchange rate policy – or more correctly, exchange rates were allowed to float within a narrow band and the various central banks were obliged to ensure (via for example changes in interest rates or intervention in the FX market) that they remained there. Denmark, Latvia and Lithuania currently follow a fixed exchange rate policy within the framework of a similar system, ERMII.

According to Friedman, however, a system of fixed but adjustable exchange rates is the worst of all worlds. Such a system means that the country abandons the option of an independent exchange rate policy. However, at times the need to use the exchange rate policy for “domestic purposes” – for example to tackle an asymmetric shock – will be irresistible, and the country will then either have to adjust exchange rates (devalue or revalue), or completely abandon the fixed exchange rate policy. This will, meanwhile, cause uncertainty in the FX market about just how “fixed” the policy is in reality. Thus a system with fixed but adjustable exchange rates will always be a potential “target” for speculative attack: one has so to speak closed the door, but not locked it. In a monetary union with irrevocably fixed currencies one has, in contrast, closed the door, locked it and thrown away the key – there is simply no doubt about how solid the fixed exchange rate policy is and thus speculation in exchange rate movements will therefore cease.

Hence for Friedman the choice is between either a freely floating exchange rate or some form of monetary union. Friedman has over the years presented the criteria by which to choose between the two exchange rate regimes. Basically there are six criteria that a small country (A) should consider when deciding its exchange rate policy in relation to the “rest of the world” (country B):

1.     How important is foreign trade for the economy of country A?

2.     How flexible are wages and prices in country A?

3.     How mobile is labour across national borders?

4.     How mobile is capital?

5.     How good is monetary policy in country A and the “rest of the world”?

6.     How are political relations between country A and the “rest of the world” ?

These criteria in fact define what in modern economics literature is termed an optimal currency area[1]. If there are close trade ties, high wage and price flexibility, and high capital and labour mobility between country A and the “rest of the world”, there is, according to Friedman, no reason why the two countries should not form a monetary union with a common currency.

Friedman stresses, however, that a country should not abandon its monetary policy independence to another country if that country is expected to pursue a poorer monetary policy than the first country itself would have done. Friedman places greatest emphasis on this criterion.

Despite Milton Friedman typically – and rightly – being labelled as the standard bearer for floating exchange rates, he often stresses that the choice is not easy, and he has repeatedly emphasised that countries have achieved both good and bad results with fixed and floating exchange rates. He points out for example that in 1985 Israel successfully implemented a fixed exchange rate policy against the dollar that helped cut inflation without causing any negative long-term economic repercussions.

By way of contrast, Chile implemented a fixed exchange rate policy against the dollar in 1976. Results were good for the first year following the implementation. However, when US monetary policy was seriously tightened between 1980 and 1982, causing the dollar to surge, monetary policy in Chile also had to be tightened: Chile suffered a serious economic setback, and in 1982 it abandoned its fixed exchange rate policy.

Friedman used the two cases above to underline that identical exchange rate policies can lead to different results. The outcome of the fixed exchange rate policy depends on how “lucky” one is with regard to the monetary policy in the country whose currency one has fixed to. Israel was lucky to introduce a fixed exchange rate policy at a time when monetary policy was relatively accommodative in the USA, while Chile was unlucky to fix just before US monetary policy had to be vigorously tightened. Or as Friedman says:

“Never underestimate the role of luck in the fate of individuals or of nations.”[2]

[1] The theory on optimal currency areas can be traced in particular back to Robert Mundell, see eg, Mundell, R. A., “A Theory of Optimal Currency Areas”, American Economic Review, Vol. 51, No. 4, September 1961, pp 657-665.

[2]”Money Mischief”, page 241.

Leave a comment


  1. Blake Johnson

     /  October 31, 2011

    Really enjoying the series on Friedman Lars. I’ve read a decent amount of Friedman, but wit such a prolific academic, there is always more to learn. Keep up the good work.

  2. I turned Friedman’s argument against fixed exchange rates against NGDP level targeting. I hope someone can tell me why my argument is wrong.

  3. Blake Johnson

     /  November 1, 2011

    I am going to take a crack at your argument Peter.

    The NGDP target works through two channels, the first is the expectations channel, which you highlight, and the second is through a satisfaction of excess demand for money, which you seem to gloss over. Further, one of the principles of Market Monetarism is announcing your target, and credibly assuring markets that you will do what is necessary to meet that target. It seems unlikely to me that after many years of successful NGDP targeting that a 1% drop in NGDP would shake the credibility of central banks claims to continue targeting NGDP. Even if the shock were bigger, if the Fed announces that it is targeting NGDP at level X, and have demonstrated that they will do what is necessary to meet that target over a period of time as you suggest, the circularity problem should not come in to effect.

    Beyond this point, if the Fed is targeting NGDP correctly, and keeps things even relatively close to being on trend, the effect via the channel of satisfying excess demand for money should still be strong.

    It would seem to me that your argument might be better suited against a NGDP growth rate target, but I must admit I have not read Selgin’s work on NGDP growth rate targeting closely enough to comment further on that possibility.

  4. Blake Johnson

     /  November 1, 2011

    One last comment to tie this back in to the exchange rate example.

    Friedman’s criticism seems to center around the temptation to manipulate the exchange rate to deal with asymmetric shocks. The knowledge of this temptation is what causes speculators to doubt credibility in how “fixed” the fixed exchange rate is.

    Think Kydland and Prescott’s Time Inconsistency model, people know that if they expect inflation of X, the fed is tempted to surprise them with X+Y% inflation to lower unemployment. I am not sure what comparable temptation you are imagining in abandoning NGDP targeting, especially if you assume it has worked successfully for many years. If you have something in particular in mind, I would appreciate if you could elaborate on that point for clarity.

  5. Blake, thanks for your reply. I tried to clarify my argument here:

  1. Milton Friedman on exchange rate policy #6 « The Market Monetarist
  2. Mises was clueless about the effects of devaluation « The Market Monetarist
  3. The luck of the ‘Scandies’ « The Market Monetarist
  4. Celebrating Friedman and Hetzel | The Market Monetarist

Leave a Reply