مشکلات سیاست های پولی برای اتحادیه ارز : عدم تقارن و مساله تجمع در منطقه یورو

This paper highlights implications for a single monetary policy when key economic relationships are non-linear or asymmetric at a disaggregate level. Using a four equation model with data for the EU countries we find considerable non-linearities and asymmetries in the Phillips and Okun curves. High unemployment has relatively limited effect in pulling inflation down while low unemployment can be much more effective in driving it up. Economic downturns are both more rapid and sustained in driving unemployment up than recoveries are in bringing it down. There is considerable variety in these relationships and IS curves across countries, sectors and regions. Monetary policy reacts more vigorously outside a central corridor.

All economies face problems of aggregation in running macroeconomic policy. If economic behaviour can be approximated reasonably by a linear representation then, however, diverse the economic structure, regional and industry performance, aggregation remains a second order issue. The consequences for each region or industry of single economy-wide policy, such as monetary policy, will, however, be drastically different if the economy is very diverse. This will pose problems for redistribution of income and wealth, migration, regional, industrial and other policies aimed at addressing the consequences for those differentially affected. If behaviour is both non-linear and asymmetric, it becomes much more difficult to determine the appropriate single policy. In most mature economies this potential problem is relatively little studied, in part perhaps because policymakers readily allow for it and partly because the offsetting redistributive measures are extensive and automatic. The problem is thus less visible ex post. In the euro area, on the other hand, there was little experience of running a single monetary policy before 1999, yet there was a great deal of information and experience at the member state level. The nature of the aggregation process has thus been much more obvious as has the diversity among the component economies.1 Furthermore, since there is little redistribution across national borders the divergent consequences are also more obvious. However, while the extent of the variation in behaviour is well known,2 euro area policy simulations have typically been conducted with models that use euro level aggregated data or which handle the euro countries separately (with appropriate cross-country constraints) and aggregate the results.3 Such aggregations are usually either unweighted or based on GDP or similar weights. In this paper we show that there are strong grounds for believing that there are considerable asymmetries and non-linearities in inflationary behaviour and monetary transmission. Ignoring these could have substantial adverse effects on particular sectors, regions and member states within the euro area. In Section 3 we develop a small conventional model of the monetary transmission mechanism in the euro area and show in Section 4, using a dataset that covers all of the EU countries except Greece and Luxembourg,4 that there are good empirical grounds for asymmetry and/or non-linearity in each of the relationships. In Section 5 we show how these results pose problems for aggregation. Section 6 concludes but before going further, we explain how we use the term asymmetry and motivate the rest of the discussion in Section 2.

We have argued that there are clear asymmetries in the relationship between demand pressure, inflation and employment in the European Union and the euro area in particular. These asymmetries exist at the sectoral and regional levels as well. As a result, using arithmetic weights to add effects across countries in order to determine area wide monetary policy could produce erroneous results.40 This is exacerbated by the fact that there is considerable variation across the EU countries in their responses. It therefore matters which part of the area is experiencing which shocks. Average values can be misleading. However, if business cycles among the EU countries are becoming relatively co-ordinated, as Artis et al. (1999) indicate, then the problem is reduced.41 Differences in the speed and extent of the transition mechanism within countries will matter rather than differences according to where they are in the cycle as well. The problem still will not disappear if the shock falls on countries such as Spain where the impact of disinflationary policy is slower and milder than elsewhere. Much of the literature on asymmetry in EMU is misconceived for our purposes as it focuses on the idea that individual countries will vote for the policy that would be best suited to their own needs and that the compromise or majority position may be suboptimal.42 Our implicit assumption is that all those deciding on monetary policy are trying to so from the point of view of what is best for the euro area as a whole.43 Our concern is simply that if arithmetic weights are used in a non-linear and asymmetric world there is a danger of generating inefficient outcomes.44 We primarily focus on asymmetries stemming from the behaviour of the labour market.45 Rapid downturns in the economy appear to have more than proportionate downward effects on unemployment, partly because of mismatch between the sectors and regions where the jobs and unemployed lie. This effect is likely to be greater in the EU where labour mobility is lower than in the US where the phenomenon is already clear. A slower response to adverse shocks makes recovery phases longer and unemployment persistent. However, these forms of asymmetry have a rather different impact on the inflationary process. The straightforward asymmetry, inherent in the convexity of the Phillips curve, is that excess demand in product or labour markets has a significant upward effect on inflation while deficient demand has little or no effect on lowering inflation. The process is, however, more complex as the dynamics suggest that big differences between sectors and regions distort the picture. It is the existence of tightness in parts of the labour market that affects overall inflation and average unemployment and by analogy probably tightness in sectors of the product market that tends to intensify the inflationary pressure. Thus our findings indicate that in each example we have considered, ignoring the disaggregated problem will tend to result in misleading policy conclusions. The asymmetry is not restricted to demand shocks, as supply shocks, particularly through the exchange rate and foreign sector, can have sharply differing impacts both across the member states of the EU and across the sectors of industry. The traditional implication for policy set out in Laxton et al. (1995) is that monetary policy will need to be set somewhat more restrictively than is implied by linear symmetric models. However, it is also likely that any ‘new economy’ effects, where faster non-inflationary and higher unemployment growth develops, may occur in the areas of high demand and relative labour shortage (Oliner and Sichel, 2000). Hence the implications of the asymmetric effects, observed in data from the past may need to be rethought if major sectors in the economy are undergoing structural change in their responsiveness and flexibility. None of this argument implies that running a single monetary policy is inappropriate. However, it does have two other major implications. First, it implies that in setting monetary policy the Eurosystem needs to take account of the problems of asymmetry and aggregation. Second, it entails that the governments of the member states both individually and jointly need to consider what other policy changes are needed in order offset the blunt nature of the impact of monetary policy. Structural and fiscal policies can be far better tuned to have detailed impacts on parts of the economy. This second message is not new and is not the focus of this paper. Our concern is to highlight the first implication, that for the setting of monetary policy. However, it would be mistaken to assume that the effects will be wholly negative in terms of reducing the bite of monetary policy at low levels of inflation or negative output gaps. Much of the point of EMU is to change macroeconomic behaviour for the better in the member states. The more rapidly developing economies will be facing looser monetary policy than would have been the case without the union (except for countries that were closely targeting the DM, where there will be rather less change).46 We can expect, for example, that the countries with the positive output gaps will in fact try to hold down prices more than they would previously out of fear that their competitive position would worsen now that they have no independent exchange rate to offset the worsening in inflation. Indeed there are signs in both Finland and Ireland, which have been growing rather faster than the rest of the euro area that recent growth-inflation combinations have become more favourable. In other words that the sustainable rate of growth has increased and that calculations of the output gap need to be revised (downwards). It is, however, problematic to infer from this evidence that the response of the authorities to these observed asymmetries should itself be asymmetric in order to compensate. Past policy is part of the adjustment process and hence its results are incorporated in the estimated relations, especially those of a strongly reduced-form nature. Asymmetries in policy may themselves have contributed to the observed asymmetries over the course of the cycle. Haldane and Quah (1999) note that the aggressive policy response to inflation in recent years may make the short-run Phillips curve look near horizontal. Our estimates of a monetary reaction function for the euro as part of the model suggest that policy has indeed been asymmetric. However, it is not simply that the authorities have reacted more strongly to high rather than low inflation but that they have reacted more strongly still to deflation. Thus what we see is a ‘corridor’ pattern, where within the corridor the response is relatively mild and outside the response is much more aggressive. Given that the Eurosystem has adopted a target range for inflation over the medium term and not a single midpoint value, this corridor may turn out to be a better description of future policy than some simple linear reaction function