The Implications of the Looser ERM for French Macroeconomic Policy

Eric Girardin *

The turmoils of the Summer of 1993 seem by now to have left almost no trace on foreign exchange markets. The relative easiness with which the franc came back towards central parity, together with a yield curve which does not seem to indicate, in the late autumn of 1993, the come-back of inflation expectation, and a reduction to a negligible amount of the differential in intervention rates with Germany, would seem to indicate that the credibility of French monetary authorities is indeed preserved. It is then worth studying to what extent the loosening of the ERM has in any way changed the choices open to French macroeconomic policy .

A reflection on the notion of credibility is a key to appreciate whether wider margins in the ERM widen the room for manoeuvre of French monetary authorities, why the switch to floating exchange rates is not a serious alternative, and what sort of additional instruments should be made available to European governments in order to make viable the transition to EMU.

Should the room for manoeuvre be used to reduce interest rates under German ones?

The ability of further falls in interest rates to alleviate the recession in France is not to be dismissed. However we know first, that reduced interest rates will only affect output with a lag of at least six months, by which time the cyclical situation may require an opposite action, second that it will not seriously affect the unemployment situation. Anyhow, not much should be expected from the long end of the spectrum, since French long-term interest rates are quite low by international standards. Indeed, the downward sloping yield curve is generated by high short-term rates as in Germany. However, French economic agents (especially firms) are more dependent on short term borrowing than German ones. Bearing this in mind should French monetary authorities temporarily generate a negative short-term interest rate differential with Germany and use at least a substantial part of the depreciation potential offered by wide margins?

The answer to this question would be positive in case one considers that the newly independent Bank of France is now credible enough in its objective of price stability so that it can very well use monetary policy in order to stabilise real output in the short-run (and even in the long-run if shocks are persistent). This would mean that ten years of "franc-fort" policy have so well born fruit that the French authorities can afford to behave as their counterparts in Germany, the United States, or Switzerland. They would thus be able to implement a temporary expansionary-countercyclical- monetary policy, which would be reversed subsequently in order to come back to the medium-run objective so as to preserve their credibility. Recent theoretical work has indeed shown that central banks can use rules with "escape clauses". In other words, they can preserve the credibility of their commitment to price stability if they stabilise output only in response to exceptional and/or independently verifiable shocks. Recent empirical work have confirmed the existence of some such room for manoeuvre for a credible central bank. The latter can use a hybrid of rules and discretion, applying rules to its medium and long-run policies and responding in a discretionary way to short-run movements in economic activity. Moreover such hybrid monetary strategies seem compatible with low inflation rates in as much as the bank commits itself to reversing short-term deviations from target over a long period.

A look at past experience is instructive to assess the potential effects of an attempt at a temporary isolated easing of French monetary policy. On several occasions the French authorities tried to put their intervention rate under or equal to the German one. This occurred already in the autumn of 1990, the spring and autumn of 1991. However, on every such occasion, retreat was extremely quick, in front of the unsustaniability of the situation. In the early summer of 1993, by contrast, retreat was overdue. It seems that the drop of intervention rates under the German level in the last two weeks of June was (wrongly) interpreted as a signal that the French authorities were ready to give up the high interest rate strategy. The franc remained around central parity and started to depreciate only about a week after the official interest rate differential changed sign. The latter happened on June 18th, while the former occurred on June 24th. The first signal was reinforced by a second one, i.e. the announcement on June 23rd of the 6% interest rate attached to the four-year Balladur bond, while at the end of May the Economy minister had projected a rate of 7%. Moreover the key feature of the Balladur bond is that in case interest rates rose the price of the bond would be kept at par. Indeed, when exchanged for equities of privatized firms, the State guarantees their nominal value. The launching of such a bond could be interpreted by the markets as showing the intentions of the French authorities who seemed to tie their own hands by rendering costly any move to raise interest rates anew in order to defend the currency. Moreover, the decrease in the intervention rate under the German one, while negligible at first, was then accentuated in early July (reaching a peak of 200 basis points) well after the franc had started to depreciate vis-à-vis the mark, thereafter the monetary authorities were unable to resist the depreciation of the franc which ultimately led to the widening of the bands. This experience teaches us how quickly the painfully acquired credibility can be threatened whenever French monetary authorities try to go it alone, independently of the path followed by the anchor currency central bank.

In June 1993, with a franc starting from around central parity, a slight negative interest differential with Germany triggered its free fall. Is there any reason to believe that with wider margins market behaviour would be any different? In other words, has the nature of policy choices changed with the widening of the band? In theory, wider exchange rate bands increase the autonomy of domestic monetary authorities in the setting of interest rates, even with free capital mobility and asset substitutability. Wider margins mean that the room for expectations of currency appreciation inside the band enable the domestic interest rate to be lower than the foreign one. Of course, this is conditional on central parity being credible. The lowering of domestic interest rates under foreign ones would initially generate a depreciation of domestic currency. With the currency thus near its floor, the increased probability of an intervention to prevent it weakening further should give rise to expectations of mean reversion of the exchange rate, i.e. of a substantial appreciation inside the band. This expected subsequent appreciation back to the central rate would compensate for the lower interest rate in the meantime.

Its advocates argue that such monetary autonomy allows the central bank freedom to stabilize output by, for instance, lowering the domestic interest rate in a recession. We saw earlier that successful central banks have long term credibility in the form of an expected long-run mean reversion of the monetary variables these banks are concerned with. The monetary autonomy arises here since the central bank would be able to exploit the mean reversion of the exchange rate to central parity. With long-run credibility, wider bands may thus give rise to a combination of long-run mean reversion and short-run discretion.

This argument assumes that a negative interest rate differential will not damage the credibility of central parity. If, by contrast, devaluation expectations are a function of the position of the exchange rate inside the band, the lowering of interest rates might generate a sharp depreciation of the domestic currency, leading it towards its floor. The monetary authorities would then be faced with the difficulty of defending the floor. Indeed the reserves required for such a defense increase with the width of the band. One may then argue that the authorities could gain credibility by not exploiting the potential monetary autonomy provided by the exchange rate band. The experience of the Netherlands is worth considering here. During the upheavals which started in September 1992, the Dutch Guilder was always above or around central parity while money market interest rates were consistently under German ones. This was already the case ten years before in March 1983 when the government overruled the central bank and devalued the Guilder by 2% against the deutschemark. The negative impact of this devaluation on confidence was very substantial. Indeed, until 1992 Dutch money market rates were hardly ever lower than German ones. This illustrates how easily confidence is lost and how long it takes the monetary authorities to restore it. Such an experience implies that the apparent short-run gain in monetary autonomy offered by the band has to be balanced against the long-run loss represented by the risk premium that the markets will subsequently demand if the domestic currency drifts towards its floor.

The Dutch guilder was not hit by the attacks over the last eighteen months because the exchange rate policy remained credible. This credibility was preserved because the Dutch monetary authorities have for a long time accepted constraints. They were thus always very reluctant to use the flexibility that might be thought to be offered by the band (even the narrow one). They generally considered the credibility of their fixed rate to be at risk if they allowed the exchange rate to depart more than marginally (about 0.5% ) from the central rate vis à vis the mark. They thought that capital flows are equilibrating in the ERM only so long as fluctuations in the exchange rate are kept within very strict limits. This seems to imply that in a fixed rate system the widening of the band provides only limited additional room for manoeuvre in monetary policy.

Why flexible exchange rates are not a serious alternative?

It may seem that a switch to floating exchange rates, as in Britain, would, by contrast, completely free the hands of French monetary authorities. With high capital mobility a negative interest rate differential with Germany would in the very short run exert a negative pressure on the franc. However, the performance of fundamentals, especially inflation, being better in France than in Germany, the franc should appreciate in the medium run vis a vis the mark.

However we know that when short term interest rates are reduced substantially, this may increase inflation expectations and raise long term interest rates. The British example is indeed relevant here since long term interest rates are now higher there than in France. In as much as the French private sector is more dependent on long term financing than the British one, this would be a mixed blessing. Moreover, the ability of interest rate reductions to stimulate output is less strong and less quick in France than in Britain since borrowing at variable interest rates is much less widespread among French economic agents.

Besides, the switch to flexible exchange rates in France would require a nominal anchor alternative to the exchange rate. The British type of inflation target has not yet proved a realistic alternative to the search for a nominal anchor. At any rate, the Canadian experience over several decades shows that, even with flexible exchange rates, monetary autonomy is not all that large in a country with a powerful neighbour.

Whatever the potential short term macro-economic benefits of flexible exchange rates in France, they are outweighed by their structural medium run drawbacks. Indeed, an ERM with wide margins is definitely better than floating in at least three respects.

First, the recourse to floating may have encouraged a coalition of interests in France to develop its protectionist leanings. Floating may only have been the first step towards a series of measures aimed at reducing European integration (and may also have prevented the fragile compromise reached in GATT). This may be important since movements in exchange rates could increase the adjustment difficulties associated with the implementation of the single European market. If domestic industries under pressure from the lifting of barriers to intra-European trade found their competitiveness eroded by sudden and erratic movements in exchange rates, the resistance to the deepening of the single market would intensify and a demand for protection could arise.

Second, wider margins enable the old machinery of the ERM to be kept in place. This means that the franc is still a member-currency of a multilateral target zone, with the possibility of using the short term financing facility in case an undesirable drift in the exchange rate started to appear, and of compelling Germany to keep being somewhat concerned with the support of partners' currencies. This is what makes the wide-band ERM different from the experience of shadowing the deutschemark by Britain after 1987, or of the unilateral peg of Scandinavian currencies to the ecu.

Third, within the perspective of the functioning of the single market, wide margins still offer an anchor to the expectations of firms, enabling them to plan ahead without being too much dependent on costly hedging activities. Excess volatility may lead to resource misallocation if agents cannot distinguish justified movements from excessive ones. Moreover, recent work on target zones implies that their main function is to stabilize expectations on the foreign exchange market. The aim of such a zone would not be to induce rational stabilizing speculation but to keep exchange rates from fluctuating enough to generate irrational speculative selling.

The wide band ERM thus comes closer than the former narrow band ERM to the kind of system envisaged by the initial advocates of a target zone system, and implemented in an implicit way by G7 officials after the Louvre accord. The latter was motivated by the fear of unstable market behaviour as illustrated by the bubble on the dollar followed by its free fall which exemplified that floating exchange rates, instead of delivering on the promise of monetary autonomy for domestic authorities, can greatly complicate domestic economic management.

With the second stage of EMU, i.e. the coordination of monetary policies and the convergence criteria, the wide band ERM starts to look like the second version of the target zone proposal (the so-called "blueprint"). Indeed the exchange rate stability criterion together with the requirement of small and stable inflation differentials comes close to defining a real exchange rate target. However the blueprint proposal also contained two additional elements: i.e. compensatory adjustments of fiscal policy aimed at GDP targets, and "soft buffers" which would allow countries to let their exchange rate drift outside the zone, when major transitory shocks occur. These two elements should be taken into account in order to strengthen the Maastricht apparatus.

The need for a temporary softening of the fiscal policy convergence criterion.

In as much as the widening of the band does not seem to enlarge substantially the autonomy of monetary authorities, the low level of economic activity has to be remedied through other means. The persistence of high unemployment in France, as in Europe, is a long lasting phenomenon (dating back to the early eighties). However the lack of concern for this very unemployment problem testifies that the negotiators of Maastricht were still conceiving of a Europe with a lasting expansion (GEMU was a recent expansionary shock). This means they conceived of EMU in stages two and three as needing to combat inflation through an appropriate monetary policy setting and fiscal discipline. However they did not conceive of the necessity of keeping an instrument for the discretionary management of the employment level.

Experience indicates that countries in which monetary authorities are credible enough are able to combine temporary expansionary fiscal policies and tight monetary policies so as to mix a struggle against inflation with a mastering of output developments. Another teaching of the escape clause approach thus seems to be that the fiscal deficit criterion of the Maastricht treaty should be somewhat softened. In time of recession European governments' budget deficits should be allowed to breach the 3% norm not only because of the normal play of automatic stabilizers but also because these governments should be able to use their discretionary expenditure or tax instruments in order to stimulate economic activity. The 3% target would represent a medium term reference, implying a commitment by a government which is temporarily allowed to enlarge its deficit above the limit, to reduce it under the target over a specified period in order not to affect the ratio of public debt to GDP in the medium run. Such short-run flexibility granted to fiscal policy should be part of a wider coordination of fiscal policies between member countries.

A non-discretionary exchange rate escape clause as a condition for a return to narrow margins.

Wide margins in the ERM can be regarded as an emergency measure in front of mounting speculative pressures, suppressing the former one-way bet. The significance of preserving the system in this form is that, by doing so, participating countries indicate their intentions to return to a more effective exchange rate mechanism when circumstances allow this. The temporary suspension of the narrow margin ERM is thus another instance of the use of an escape clause. Combining the rule of fixed exchange rates with the escape clause offered by the widening of the band was preferrable to the previous simpler rule since the costs of sticking to the latter became too large.

This preserved the spirit of the exchange rate stability criterion of the Maastricht Treaty. The very fact that central parities were kept unchanged in August 1993 is thus very important. It may be that in the fashion of the G7 Louvres accord, European monetary authorities have agreed to maintain implicit narrow margins. In as much as they succeed in this endeavour, they will be able to claim ex post that exchange rates have de facto stayed inside the "normal"-band ERM during the required two years before the formation of EMU. It would be conceivable in this way to go directly from wide margins to stage three.

However the letter of the Treaty seems to refer to "normal" margins as the de jure narrow band. This would require the narrowing of the margins in three years time at the latest. The experience of last summer should lead us to be cautious about this last period of the transition to EMU. The lesson of these events is that it is not reasonable to leave a narrow margin ERM without an escape clause. A non-discretionary exchange rate escape clause explicitly providing for a temporary widening of the bands in case of exceptionally large and independently verifiable shocks seems necessary to enable a narrow margin ERM to survive during these two crucial years. This would of course be much easier for supply-side shocks and large-scale real demand shocks. As for financial demand shocks, like increased currency substitution destabilizing the demand for money in individual countries, the acceleration of the move to an EMU-wide monetary policy would be the only appropriate remedy.