Recent Posts

Sunday, October 20, 2013

Currency Regimes Matter If Policymakers Understand Them

In this article, Antonia Fatas argues that exchange rate regimes (like the euro) have limited power to explain differences of economic outcomes. It is based on an article by Andrew K. Rose, which looks at the currency regimes of smaller (mainly developing) economies during the global financial crisis.  

Paul Krugman responded here, noting that bond yields only rose due debt concerns in the euro countries. From the point of view of the bond markets, that is a crucial point: a country that does not control the currency of its debt emissions is just another credit market borrower, and can end up facing prohibitive default risk premia. 

Since his article illustrates that point well, I will discuss here the non-interest rate aspects of this debate. The currency regime is a critical component of Modern Monetary Theory (MMT), and so this debate is very important for understanding MMT.

Antonia Fatas’ main point is this:

I have written before my views that run contrary to the conventional wisdom. Many believe that while the Euro might make sense as part of a political process of European integration, it has had clear negative consequences on economic performance, consequences that are obvious when one looks at the effects of the current crisis in the Euro periphery countries. In several blog posts (here or here) I have provided anecdotal evidence that this conclusion is not supported by the data by comparing the performance of countries in and out of the Euro area and also by looking at the consequences of previous crisis when some of the Euro countries still had their exchange rate.

But what happens if one goes beyond the anecdotes and tries to systematically analyze the difference in performance of different exchange rate regimes? Unfortunately this is a difficult task. My own reading of the literature was that the evidence is mixed and inconclusive, there are no strong empirical results that prove that the exchange rate system has a significant effect on economic performance. But most of these academic papers were quite old.

My view is that membership of the euro is a bad idea, and will ultimately prove damaging to the economies of the countries concerned. However, he is correct that these problems may not show up in standard statistical tests. For example, it is possible that you could run a panel regression on GDP growth rates versus currency regimes and find no statistically significant relationship (this may change if we get a few more centuries of data). Why? A statistical test has an embedded underlying mathematical model of the relationship between the variables analysed. However, you need to look at the underlying economic dynamics, and see whether they are compatible with those implicit models.

The analysis by Andrew K. Rose is mainly of a panel of developing economies, and I do not find it controversial that he saw no effect of currency regimes in those economies during the Global Financial Crisis (GFC). The GFC was centered in the developed economies, and the emerging markets were unusually not a source of financial instability. This was the consensus opinion in published financial market research during the crisis. Therefore, I believe this particular analysis has little applicability to the situation of the euro member countries.

Returning to the question of euro members versus the developed economies with free floating currency regimes, I believe that the statistical differences will show up only in the following two areas:

  1. Bond yields (as discussed in Paul Krugman’s article). Only euro zone markets have seen central government bond yields decouple from cash rates.
  2. Countries with unusually high youth unemployment and underemployment rates (>40%). Although unemployment is too high across almost all the developed economies, the unusual total breakdowns in the employment market have only occurred along the Eurozone periphery.

That said, I find it unsurprising that statistical tests on GDP growth rates (for example) could appear inconclusive when testing whether Eurozone membership is a negative. To see why, we need to look at the economic dynamics as to why this could be so.

Exchange Rates Don’t Matter, Currency Regimes Do

My first point is that exchange rates (i.e., the price of a currency) have a limited impact on developed economies, but the currency regime (e.g., a fixed currency peg versus a floating currency does) does matter in terms of policy options that are open. (I have a bias towards exchange rate movements having limited impact on domestic economic variables; not everyone will agree with that.)

The euro is only a fixed exchange rate system within the member countries. The euro has weakened versus Asian trade competitors, so the euro does not pose a problem for euro zone countries in aggregate when analysing the impact of trade. The periphery faces a too high exchange rate, but they are outweighed in the aggregates by the larger German and French economies that are benefitting from a weaker currency. Therefore, any statistical panel of a trade effect should show a mixed effect, and the sign will depend upon how the countries are weighted within the panel.

However, the currency regime matters in how it restricts government policy (this is emphasised by the MMT theorists). Under normal circumstances, a fixed currency regime presents no binding constraints upon a government. The only time the peg matters is if the markets believe the government is unwilling to tolerate a too-strong exchange rate which is leading to current account deficits. Speculation against the currency will show up in financing problems for the government – it has to have higher interest rates to attract capital to finance a current account deficit. (Conversely, a too strong exchange rate poses little problems; the government can “sterilise” capital inflows without negatively impacting the domestic economy.) Therefore, economic differences only show up when a fixed peg is under attack. This is only a small subset of most observed data sets, and so if we do a statistical test on the entire observed data set, this will probably not show up as being statistically significant.

Observed Economic Behaviour Is Driven By Automatic Stabilisers

When we look at observed economic data, we are not observing how a capitalist economy operates without intervention. The data is the result of the interactions between the private sector and the public sector, including policymakers (fiscal and monetary) who are attempting to steer the economy in some direction. Within economics, this idea is known as Friedman’s Thermostat (as discussed by Nick Rowe in this article). (In my old field of control theory, this is the difference between an “open loop” system, and a “closed loop” system that has control feedback applied to its operation.) Nick Rowe presumably attaches a great deal of significance to the operation of monetary policy, but I emphasise the role of non-discretionary aspect of fiscal policy. In other words, the “automatic stabilisers” of the Welfare State.

My simulation of an arbitrary debt limit shows what would probably happen if the automatic stabilisers were suddenly turned off: the economy goes into freefall. This is effectively what happened during the pre-Welfare State era, when the government sector was too small to stabilise the economy. However, the Eurozone periphery did not turn off the welfare state; they allowed it to operate. Their austerity plans were largely based on cutting program spending. This meant that their deficits exploded above planned levels, and so the economy was eventually stabilised despite the planned austerity. This was made possible by ECB interventions. Since the euro is not pegged to a commodity, it is possible to finance their trade deficits with a fiat currency, and so the periphery did not face truly hard financing constraint.

Policymakers Need To Understand Their Options Within The Currency Regime

In order for there to be a difference in economic outcomes between nations with currency pegs and those with free-floating currencies, policymakers in the free-floating nations have to understand the constraints they face. However, after the worst of the crisis was over, many policymakers in the developed countries acted as if they faced the constraints of a fixed currency regime. The most well-known example is the decision by U.K. policymakers to voluntarily impose the same sort of austerity policies that Eurozone policymakers were forced* to impose. U.K. economic performance suffered as a result.

As a result, the economic performance of the “free floating” currency nations was punished by policies that replicated pegged currency regime performance. Therefore, it should be no surprise that a country like Germany, which did not get hit directly by the credit excesses before the GFC, and for whom the currency peg does not bind policy, could have a “better” performance than a badly-managed “free floating” currency nation. That said, no free-floating nation is likely to replicate the disastrous performance of the Eurozone periphery with their 40+% youth unemployment rates. Therefore, the statistical difference will have to be of the form of looking for extremely bad economic outcomes, and see which currency regime produced them.

* One could argue that the ECB could have allowed policymakers to avoid austerity policies; however, the ECB argued that it was following European law. I have no idea whether there was a legal way to avoid austerity policies, but it was clear there was no political will to do so.

(c) Brian Romanchuk 2013


  1. Using Andrew Rose’s paper “Surprising Similarities: Recent Monetary regimes of Small Economies” to argue that exchange rate regimes (like the euro) have limited power to explain differences of economic outcomes is a bit of a stretch.
    Ross excludes from his sample “the five systematically important economies of China, the Euro-zone, Japan, the UK and the USA.” These countries (consider the Euro-zone a country) accounted for 63% of Global GDP in 2012 (PPP, World Bank). Of the next 23 largest countries by GDP, which accounted for 32% of global GDP, only Saudi Arabia had a “hard-fix” currency regime. Two additional hard-fix regimes, Egypt and Pakistan, switched to floating rate regimes early in the rather short 2006-2012 sample period.
    That leaves countries representing just 5% of global GDP to populate the bulk of Rose’s paper, the goal of which was to “be as comprehensive as possible.” Granted, he emphasized that the focus of the study was small economies. But how realistic is it to expect a study of tiny-country currency regimes to improve our understanding of the effects of different currency regimes on economies?
    My perspective, from 29 years of trading currency markets, is that restricting the free movement of one’s currency against the dollar (volatility suppression) will eventually result in an extreme “catch-up” move. In the 1990-1996 period leading up to the Asian financial crisis, the Thai Baht moved in a restricted 7% high-to-low range, the Malaysian Ringgit moved through a small 14% range and the Indonesian Rupiah was restricted to an 18% range. Of the 28 countries that currently represent 95% of global GDP, only the Saudi Arabian Rial (2% range) and the Egyptian Pound (18% range) have restricted their currencies to such small high-to low ranges. There is little need to change a currency regime when a currency is allowed to adjust (either through floating or frequent re-pegging) to economic conditions.
    Of the 60 small economies that maintained fixed currency regimes continuously from 2006 through to 2012, only Saudi Arabia is of any economic significance (measured by GDP). And one could argue that their economy is dollar-based.
    Rose’s study – comparing currency regimes to economic indicators - is as significant to the Euro-zone as quantifying aviation safety for the large carriers by examining the crash rate of ultralight aircraft.

  2. This comment has been removed by a blog administrator.

  3. This comment has been removed by a blog administrator.

  4. This comment has been removed by a blog administrator.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.