ADJUSTMENT TO ASYMMETRIC SHOCKS
|I. Optimum Currency Areas|
Temporary and permanent shocks
I. Optimum Currency Areas
The international monetary system is characterised by a multiplicity of different currencies. For the most part, these are based on sovereign nation states - a situation which John Stuart Mill famously described as "barbarism"; and, as the structure of modern government has developed, the economic segregation between currency areas has often become acute, notably through the imposition of exchange controls and inconvertibility.
Even during the period of "monetary union" provided by the Gold Standard before 1914, some countries maintained the gold link more successfully than others (see Flandreau, Le Cacheux and Zumer 1998).Little attention, however, was paid until relatively recently to the economic, as opposed to the political, basis for separate currencies.
The initial formulation of what should determine the geographical coverage of a currency is widely attributed to a paper published in 1961 by R. Mundell (see Box 1). Since the purpose of money was to be "a convenience", he argued, the ideal currency area was "the world, regardless of the number of regions of which it is composed".
Given the practical need for stabilisation policies in existing economies, however, an area needed a separate currency if, given some macroeconomic shock, the economic costs of adjustment through changes in wage and price levels, or through factor mobility (labour and capital), would be higher than those of altering the exchange rate. Mundell's analysis was shortly afterwards elaborated on by McKinnon (1963), and by Kenen (1969), since when optimum currency area (OCA) theory was developed by a growing number of studies, both theoretical and empirical.
The case for separate currency areas clearly holds good only if the impact of a shock varies between areas: i.e. is asymmetric. If the impact were to be the same on all, the exchange-rate changes needed for adjustment would be the same for all, in which case separate currencies would serve no purpose. OCA theory indeed implies that any two countries generally experiencing symmetric shocks, and trading significant proportions of their GDP bilaterally, should fix their exchange rates.
As Milton Friedman (1953) had already observed:
"A group of politically independent nations all of which firmly adhered to, say, the gold standard would thereby in effect submit themselves to a central monetary authority, albeit an impersonal one. If, in addition, they firmly adhered to the free movement of goods, people and capital without restrictions, and economic conditions rendered such movement easy, they would, in effect, be an economic unit for which a single currency would be appropriate."
In a short "taxonomy of shocks" contained in a recent Commission Economic Paper (1997a), four distinctions are made:
Temporary and permanent shocks
Perhaps the most useful distinction to be made is between shocks likely to have only transitory effect - for example an unanticipated fall in aggregate demand - and shocks which entail a permanent decline in competitive position. Shocks of the first kind can be corrected by counter-cyclical changes in fiscal and/or monetary policy, or by borrowing. Shocks of the second kind, however, can in general only be met by a decline in comparative real incomes and prices; by labour force migration; or by major long-term restructuring.
The distinction is important if only because confusion between them can lead to action which aggravates rather than improves the situation. More particularly, treating shocks with a permanent, structural effect as if they were temporary may only serve to entrench the underlying loss of competitivity and make necessary reform more difficult. The Commission (1997) observes:
"If, as the evidence shows, structural change is already too slow in Europe compared to, say, the US, slowing it down further through monetary cushioning may be exactly what policy makers should want to avoid".
This, in essence, is also the argument underlying the "no bail out" provisions of the Treaty in relation to sovereign debt; and, more contentiously, resistance to any extension of EU structural and cohesion payments. As in the wider global context, easy access to borrowing or transfer payments may create a "moral hazard": it will be easier for an economy in trouble to wait for the rescue package than to put its house in order.
It is probably a mistake, however, to make the distinction between the temporary and the permanent too rigid. Most real-life shocks are likely to have elements of both, implying that both short-term policy adjustments and long-term structural reforms are required. Moreover, inter-regional financial transfers can be useful, not merely to iron out cyclical problems, but also to promote structural change. The problem is to ensure that the funding is used in the correct way. (3)
Box 1: Mundell's Theory of Optimum Currency Areas
The 1961 paper by Mundell examined possible mechanisms of adjustment when countries or regions face exogenous country-specific shocks, with particular reference to the US and Canada. He concluded that exchange-rate changes between the US and Canadian Dollars did not provide either country with a satisfactory means of adjustment, since the main asymmetry was not between the countries themselves, but between the eastern and western parts of both. Mechanisms were therefore required to adjust relative prices between east and west rather than between north and south.
A simple version of the theory assumes two regions - A and B - each producing a good. A demand shift caused by a change in preferences from the goods produced by A to the goods produced by B (i.e. an asymmetric shock), will lower demand in A, raising unemployment and causing a trade imbalance; while inflation will increase in B (see Figure 1). In such a situation, a common monetary policy cannot solve the problems of both economies at the same time. A restrictive monetary policy (S up) might reduce inflation in B, but worsen the unemployment problem in A. An expansionary monetary policy (S down) would reduce unemployment in A, but worsen inflation in B.
The disequilibrium caused by a shock will therefore require a change in relative prices to restore the previous equilibrium. If the two regions have separate currencies, this can be achieved by altering the exchange rates: i.e. by a devaluation of currency A vis à vis currency B. Country A would then recover its competitive position through lower real wages and prices (though nominal wages and prices would remain constant). Demand would rise (D upshift) and unemployment fall.
If, however, the two regions have a common currency - or maintain a fixed exchange rate - production and employment in A must be restored through other means: for example
Mundell´s analysis therefore suggested that:
Further refinements by Kenen (1969) pointed towards a high degree of product diversification - the more a group of countries or regions specialized in the production of particular goods, the more likely it would be that external shocks would have asymmetric effects. The different parts of a currency area should therefore produce a similar mix of goods.
McKinnon´s (1963) main criterion was the degree of openness in an economy, linked to the relative importance of traded to non-traded goods. A high degree of openness reduced the effectiveness of an autonomous monetary policy, and limited the usefulness of exchange-rate changes as a means of restoring competitivity, since devaluation rapidly fed through into domestic prices.
Country specific and sector specific shocks
The Commission observes that changes in monetary policy or in the exchange rate - which will have a general effect on the whole of an economy - are the wrong instruments to meet a shock which affects only one sector or region of that economy. This may seem obvious. Yet the point is of considerable significance in the light of the empirical evidence which has been collected in relation to the creation of EMU.
Briefly, this shows that only a small proportion of the shocks hitherto experienced by the EU have been country-specific: i.e. specific to an existing currency area. A "significant proportion" have been industry-specific; while, measured in terms of differential employment effects, some 80% have been either non-specific (that is, have been common to the whole EU area); or region-specific. The implication is that the loss by EU Member States of the ability to alter exchange and interest rates is likely to have only minimal consequences for the handling of the actual shocks that are experienced.
In the context of OCA theory, a purely sectoral shock can only be of any real relevance if a particular area is overwhelmingly dependent on the industry in question. In these circumstances it becomes identical to a regional shock, and opens up the question of whether the region should create its own currency; or, alternatively whether it should reduce its degree of specialisation.
Real and financial shocks
Even when shocks are country-specific, the Commission argues, variations in exchange rates are only an appropriate remedy if the effect is on real aggregate demand. If, on the other hand, the shock is purely financial - an example quoted by the Commission is a shock to the domestic money supply process - the correct answer is fixed exchange rates, or a single currency, which minimises the impediments to money flows across national borders.
Studies into the effect of shocks upon employment have reached similar conclusions.
"Financial shocks can occur as a result of currency-specific portfolio adjustments. In response to financial shocks the exchange rate should be kept constant and the shift in the portfolio composition should be satisfied by varying the supply of assets denominated in specific currencies. Asymmetrical financial shocks can be better dealt with in a monetary union than in a system with adjustable exchange rates and are thus not an additional source of unemployment". (Ochel 1997)
Exogenous and policy-induced shocks
The Commission's final distinction is between shocks which are caused by outside events over which the authorities in a particular economy have no direct control (i.e. which are exogenous) and shocks arising from internal policies. The Commission emphasises the dangers of confusion: whereas many shocks appear at first sight to be "exogenous phenomena with which policy authorities are suddenly faced", they can turn out, on more careful examination, to be the consequences of their own political activities.
Examples quoted are a rise in wages as a result of unions' and employers' expectations that the rise will be "accommodated" by monetary or fiscal expansion; and shocks caused by the political cycle itself: i.e. artificial stimulation of an economy before an election. Both of these have, in the past, been seen as important elements in the idiosyncratic behaviour of the UK economy.
Bayoumi and Eichengreen (1994) have expressed the useful additional insight that a shock which has exogenous origins can become policy-induced.
"Even if countries experience large, asymmetric disturbances, it need not follow that policy autonomy is useful for facilitating adjustment. Policy makers may systematically misuse policy rather than employ it to facilitate adjustment One interpretation of asymmetrically distributed aggregate demand shocks is that the countries concerned are poor candidates for monetary union, because policy makers can use demand-management instruments to offset demand shocks emanating from other sources. But, if domestic policy itself is the source of the disturbances, monetary unification with a group of countries less susceptible to such pressures may imply a welfare improvement".
"Where do macroeconomic shocks come from?"ask Belke and Gros (1997). "It is likely that policy itself is a source of shocks".
The significance of this factor was illustrated during the hearing into asymmetric shocks held by the European Parliament's Monetary Sub-Committee on 2nd. September 1998. A member of the Sub-Committee, Professor Katiforis, posed the question: given that Member States within EMU will no longer have an independent monetary policy, and will be limited by the Stability and Growth Pact in their ability to use fiscal policy, would the ability to alter exchange rates be a more significant mechanism of adjustment than at present, were it still to exist?
The reply from Professor Dr. von Hagen was that it would (although other consequences of EMU, notably the creation of area-wide financial markets, would improve adjustment). That of Dr. Ansgar Belke was that exchange-rate flexibility would become less significant, since it would no longer be needed to adjust for the consequences of national monetary and fiscal decisions.
The way in which different economic areas respond to macroeconomic shocks may be analysed in a number of ways. A distinction might be made, for example, between shocks which, by their very nature, will have asymmetric effects; and those which in principle should affect all similar economies in similar ways.
In One Market, One Money (1990)the Commission makes the obvious point that country-specific shocks are by definition asymmetric in the European context. Major natural disasters are likely to fall into this category, as will the immediate effects of political events like German re-unification or Portuguese de-colonisation. Yet where such events are, in effect, Acts of God, there can be few significant implications for OCA theory. Since they are by definition unpredictable, it would not be logical to establish separate currency areas in order to deal with them. More arguable is the case for a separate currency in an area particularly prone, for example, to earthquakes - though it is no means obvious that adjustment via repeated devaluations would be less costly in such cases than other adjustments within a larger currency area.
One Money, One Market also makes the point that common shocks - for example, a rise in world commodity prices or a technological innovation - can have either symmetric or asymmetric effects on different economies, depending on their structure. As Caporale (1993) observes, the Commission's finding that
"sectors producing homogeneous goods with few trade barriers mainly experience symmetric shocks" while "in other sectors there appears to be an inverse correlation between the existence of trade barriers and the degree of symmetry of the shocks"
has the "obvious implication" that completing the Single Market should decrease asymmetry.This may or may not be true (see later, "Convergence, Divergence and Diversification").
Cyclical and structural asymmetry
A more useful distinction can perhaps be made, as in the case of shocks, between short-term and long-terms causes. A rise in short-term interest rates may, for example, have differing effects in different areas because they are at different stages in an economic cycle. But they may also be due to long-term differences in financial structure: the relative importance of banking finance, for example, or of fixed-rate mortgages; and to differences in monetary transmission mechanisms.
The relative importance of cyclical and structural factors in causing asymmetric responses to monetary policy plays a large part in the current debate on possible EMU membership in the UK. The Government, for example, has cited the fact that the UK economy is currently completing an economic upswing, while most of the EU core economies are just emerging from a trough, as the main reason for non-participation in Stage 3 of EMU in 1999. It has nevertheless indicated that membership is likely in the early years of the next century.
It can be argued, however, that the cyclical misalignment of the UK economy is not accidental, but reflects more long-term structural differences. Eltis (1998) emphasises in particular the ties between the UK economy and that of the US, partly through financial links, partly because "the UK, like the US but unlike any other EU country, is a significant oil producer", and partly because "the UK also resembles the US in the extent of its high-tech industries such as biochemicals, aircraft, scientific instruments and telecommunications". If these turn out to be decisive, full cyclical alignment of the UK economy with that of the euro area may never take place. (4)
However, research by the IMF in 1987 found that asymmetric responses to monetary policy were not confined to the UK. Interest-rate sensitivity varied considerably between EU countries. In Germany, the Benelux, Austria, Finland and the UK, a 1% increase in interest rates led to a decline in production of between 0.7 and 0.9%, which flattened out after nearly three years. In other countries the decline was only between 0.4 and 0.6%, and lasted only one and a half years.
Such findings seem to augur badly, not just for possible UK membership, but for the EMU itself.
The "Lucas critique"
This conclusion, however, ignores the possibility that what appear to be long-term structural differences may rapidly disappear when circumstances change. The "Lucas critique" (5)
of much existing research into asymmetries within the euro area is precisely that shocks will have a far smaller asymmetric effect than hitherto because EMU will in itself change behaviour. As the editor's introduction to EMU: Prospects and Challenges for the Euro (1998) puts it:
"Institutions that evolved in response to history (for example, a history of variable and uncertain inflation) may adapt quickly to a new environment if it is believed to be here to stay".
Eltis (1998), for example, makes much of the differences between prevalence of mortgages and other variable interest rate liabilities in the UK compared to other EU countries.
"..aggregate mortgage debt is 60% of GDP in the UK but only 40% in Germany, 25% in France and less than 10% in Italy. The variable interest rate liabilities of the UK personal sector total 64% of GDP. They are only 16% in France, 3% in Germany and 2% in Italy."
Yet variable-rate mortgages are the legacy of a past "inflation culture", which should no longer exist within EMU. There are indeed already signs, observed by Eltis himself, that the low level of UK inflation in recent years has encouraged a switch to fixed-rate borrowing. Currently observed cyclical and structural asymmetries may also both be subject to the "Lucas critique". Christodoulakis, Dimelis and Kollintzas (1995) observed that:
"..a widespread belief in the EC goes as follows: if shocks are asymmetric, common institutional arrangements and policies will tend to exacerbate business cycles, because national governments will lose parts of their stabilization toolboxes."
In their own paper, however, they provide
"support for the opposite conclusion. In fact, our findings suggest that observed differences in shocks and business cycles will tend to melt down as common institutions and policies start to emerge."
When it is stated that a shock has "asymmetric effect", it is useful to know exactly what is being measured. One approach is simply to measure exchange-rate changes: if some event is followed by the realignment of parities between any two currencies, the effect of the event may be said to have been asymmetric, and the realignment to have been the mechanism of adjustment to it.
There are at least two problems with this approach. First, in the real world, exchange rates change for a variety of reasons (see "Exchange rates"later in this study), making it difficult to isolate the effects of a particular event. Secondly, only asymmetric effects between currency areas can be measured, not those within them. Where a currency union is being created from separate areas, as in the case of EMU, it cannot be presumed that the effect of shocks will be as before, for the reasons outlined above.
The effects of a particular shock can also be measured in terms of differential inflation rates, in the case of separate currency areas; or differential movements of consumer prices, unit labour costs or asset prices both between and within currency areas. Changes in GDP can also be measured both in relation to separate currency areas and to some extent within them, where comparable national or regional statistics are available.
A more precise statistical base is provided by rates of unemployment, which are available even at the very local level of travel-to-work areas. It is probably for this reason that a number of recent studies have defined asymmetry on the basis of differential employment effects. A fixed relationship has often been assumed between changes in GDP and changes in the rate of unemployment.
One of the most recent studies of this kind is that of Belke and Gros (1997). They conclude that
"... we have not been able to detect a robust and statistically significant link between unemployment and external shocks..... hence we would argue that EMU is unlikely to lead to the serious unemployment problems that have often been predicted."
Evaluating the effects of a particular shock is not, therefore, a simple matter. An event may have asymmetric consequences for employment, but not for inflation, or vice versa. The effects of, say, a reduction of interest rates may be asymmetric for asset prices, but not for wage rates or consumer prices. The statistical data may not be comparable for different areas - even the best efforts of Eurostat were not able to make the harmonised index of inflation inclusive of all consumer prices in time for the EMU evaluation in March 1998.
Finally, the effects of a particular event must be disentangled from the "background noise" of other events - including, as already observed, the actions of public authorities in response to them.
3. This problem often expresses itself in the context of EU funding as the "additionality" issue. If funding from the EU budget is merely used to substitute for national expenditure which would have taken place anyway, the effect will be more cyclical than structural.
4. Except, perhaps, in the paradoxical sense of the analogue watch that has stopped, and which is absolutely accurate twice every 24 hours. The same could be true of the UK and euro area economies as their upswings and downswings momentarily cross.
5. Formulated by R.E.Lucas in 1976: if something new happens, there may be little to be learned from past experience.
|© European Parliament: September 1998|