Properties of the fundamental equilibrium exchange rate in models of the UK economy.

Church, Keith B.

The Fundamental Equilibrium Exchange Rate (FEER) is that value of the real exchange rate that is consistent with macroeconomic equilibrium. This article uses the long-run trade equation elasticities from the models of Her Majesty's Treasury, the National Institute of Economic and Social Research and the Bank of England to examine the FEER calculation. The sensitivity of the results to changes in the key elasticities and to the possibility of a recent improvement in UK trading performance is considered. Historical comparisons are made between the FEER and the actual real exchange rate. All the results suggest that the real exchange rate was above the FEER at the time of ERM entry. The fixing of the nominal exchange rate removes a possible mechanism by which the economy might reach equilibrium and therefore convergence requires a period where UK inflation is lower than that of its trading partners.

Introduction

Assessments of national economic performance usually attempt to examine a general picture that abstracts from the effects of the business cycle. Economic commentators and politicians often refer to the |true' or |underlying' position of the economy, usually during the depths of recession, and it is always useful to judge current economic performance against an equilibrium path. In this context an indicator receiving much current attention is the Fundamental Equilibrium Exchange Rate (FEER) of Williamson (1983), which is the exchange rate which delivers a |sustainable' current account balance while the economy is growing at its |natural' rate. The purpose of this article is to examine the calculation of the FEER of the UK economy as viewed by three of the large-scale macroeconomic models, with particular attention being paid to the long-run trading performance of the economy. The models in question are those of Her Majesty's Treasury (HMT), the National Institute of Economic and Social Research (NIESR) and the Bank of England (BE) as deposited with the ESRC Macroeconomic Modelling Bureau in Autumn 1991. The HMT model was also publicly released at this time.

The method we choose for calculating the FEER is described and then undertaken for each of the G7 countries in Barrell and Wren-Lewis (1989)(1). We follow their approach closely and we use the stylised framework developed in that paper and, importantly, adopt their treatment of the supply side. The calculation depends on internal and external balance. Internal balance comes initially from the interaction of wages and prices, which determines an equilibrium level of activity in the domestic economy that is consistent with stable inflation. On the supply side of the economy a higher real exchange rate lowers import prices and hence reduces inflationary pressure in the domestic economy. As a result the equilibrium level of activity consistent with stable inflation rises. However any increase in equilibrium activity also tends to lead to an increase in imports and a worsening of the current account balance and as a result a lower real exchange rate is required to achieve a sustainable current account. External balance is defined for our purposes as occurring when the real exchange rate is such that the current account of the balance of payments is offset by |structural' capital flows, to which we return below. The FEER is that value of the real exchange rate at which external balance holds whilst simultaneously the equilibrium level of utilization of domestic capacity exerts no pressure for change on the current account. This unique value is the focus of attention of this article. Ideally it would be extracted from the large-scale models by looking simultaneously at the supply side and trading sector. Unfortunately of the three models considered only that of the NIESR has a well defined supply-side framework which allows this to be done. For this reason the determination of equilibrium activity that we use reflects the treatment of the supply side in National Institute models.

The FEER is a concept describing a possible medium-term equilibrium real exchange rate. In equilibrium the current account balance can deviate from zero if there are corresponding |structural' capital flows into or out of the economy. By this we do not mean speculative flows which move from country to country in search of high short-term rates of return, but inflows or outflows which are likely to persist for a sustained period of time. The investment of Japanese car manufacturers in setting up operations in the UK would fall into this category. Wren-Lewis (1992) suggests that structural flows could be treated as endogenous, although they are exogenous in our model. In the long-run we might expect |Purchasing Power Parity' to tell us what the underlying exchange rate actually is. For the actual exchange rate to approach that implied by PPP would require arbitrage by consumers, and they would have to buy from countries whose products are cheaper than home produce (once foreign currency prices have been converted into domestic currency). Similarly producers should locate where the price of inputs is lowest and hence the potential for high profits greatest. Clearly these concepts are extremely long-run. It is also the case that PPP might be considered an unreliable guide to the underlying exchange rate because many goods are non-traded. In any case consumers rarely have the information to arbitrage in this way and transaction costs are high. Similarly, uncertainty about the future and the high cost of moving an entire production process might deter producers from relocating. Given that we wish to examine the position of the UK economy over a relevant policy horizon it seems reasonable to concentrate on a measure of the exchange rate that reflects medium-term considerations. Williamson (1991) gives further reasons for doubting the suitability of PPP as a macro equilibrium concept.

The article proceeds as follows. In the second section the framework in which the FEER is determined is introduced. As our major objective is to compare models of the UK economy the trading sectors of the relevant models are then examined, with particular attention being paid to the volume equations for imports and exports of manufactures and services. Comparisons of the long-run competitiveness and activity elasticities are made across the models and these estimates are used in the third section to calculate the FEER for the three models. These are then contrasted with the actual historical real exchange rate, also trends are extrapolated in order to trace out the expected future path of the FEER. Differences in these expected paths are explained in terms of differing elasticities, and the most important factors in determining the properties of the FEER are identified. The following section examines the extent to which a possible recent improvement in the trading performance of the UK might change the equilibrium position. The penultimate section assesses the sensitivity of the FEER to increases in competitiveness elasticities that might not be reflected in current econometric estimates. Conclusions are drawn in the final section.

Calculating the FEER

The method for calculating the FEER is illustrated in the following stylized framework. The first step is to calculate the structural or trend balance for exports and imports of goods and services. This is simply constructed by multiplying the steady state volumes of exports and imports by the relevant prices and using the identity: -

BGS = XG.PXG + XS.PXS - MG.PMG - MS.PMS (1) where

BGS: Balance of trade on goods and services
XG: Volume of exports of goods
MG: Volume of imports of goods
XS: Volume of exports of services
MS: Volume of imports of services
PXG: Price of exports of goods
PMG: Price of imports of goods
PXS: Price of exports of services
PMS: Price of imports of services

In all the models trade volumes are explained by domestic and world activity and price competitiveness, and prices by a combination of domestic and world prices. By dividing the above identity by GDP and rearranging it is possible to express each of the components in terms of the real exchange rate and activity. When interest, profit and dividend (IPD) flows and time trends are introduced the equilibrium current account balance as a proportion of GDP can be expressed in the following form:

CBT = f(R,Y,YW,IPD,TIME ...) (2) where

CBT: Trend current account balance
R: Real exchange rate index
IPD: Interest, profit and dividend flows
Y: Equilibrium domestic GDP
YW: Equilibrium world GDP

If in the above internal balance is already being achieved then the FEER is the value of R which ensures that the current balance is equal to structural capital flows. If it is assumed for simplicity that these are zero then the FEER is obtained by solving the following:

O = f(FEER,Y,YW,IPD,TIME ...) (3)

The level of activity at any period in time is influenced by the level of the real exchange rate and hence is affected by deviations of the real exchange rate from its equilibrium value. Not only is it possible to find historical values of the FEER it is also possible to project exogenous variables into the future and solve the system for possible future combinations of the FEER and equilibrium activity. As an aid to examining the sensitivity of the FEER to different equilibrium activity paths the model can also be solved assuming that GDP takes actual values in the past and has different possible growth rates in the future. If these paths are denoted by Y then the FEER is given by:

O = f(FEER,Y,YW,IPD,TIME ...) (4)

The balance of trade on goods and services is crucial in determining the long-run prospects for the exchange rate, and we now turn to its treatment in the models. The three models featured here adopt a disaggregated approach to the explanation of the behaviour of imports and exports, typically distinguishing manufactures, services, oil and non-manufactures or |others'.

We concentrate on the manufacturing sector but also look at services, as these are the two most important components of the current account balance. Oil trade volumes are important during the early part of the 1980s but then decline. While the actual trend of the FEER does depend on the oil sector, cross-model differences between the estimates do not, and so our treatment reflects that of the NIESR model throughout. We first have to describe the long-run behaviour of these sectors. There is broad agreement between the modelling teams on the determinants of the quantities of exports and imports and this allows simple cross-model comparison in the framework set out below.

The main determinants of exports and imports are changes in competitiveness and activity although the measures used across the models vary slightly. In the framework used below competitiveness is defined in terms of the movement of domestic wholesale prices relative to world export prices. However in the three models the competitiveness term is sometimes defined as the ratio of import to export prices in a specific sector, but this can be converted to our definition by using further equations in the models which determine for example the price of exports and imports of services as a function of world and domestic prices. These equations allow the calculation of the elasticity given our definition of competitiveness. In certain cases the key domestic price variable is not wholesale prices but we assume that other price measures are related to this central index by a constant and time trend factor. For simplicity all different measures of world trade and domestic activity across the models are assumed to move together and are treated equivalently.

These models differ in their description of the current account. Two of the models (HMT and NIESR) have an explicit trend specialisation term which measures the additional propensity for the UK to import manufactures over time as the ratio of world trade to world GDP increases, whilst BE model the same effect as a simple time trend. In order to ensure comparability we calculate the equivalent coefficient on a specialisation term. For all the models the appropriate long-run competitiveness and activity elasticities are taken from econometrically mated equations in the models. These are used in the construction of long-run relationships that describe the data. The residual part of the behaviour of each of the trade volumes which is not explained by the long-run competitiveness and activity effects is absorbed by a constant and time trend. Further differences between the models involve services, where HMT are alone in having a feedback from service imports into exports and vice versa. We can describe the final model as:

Volume equations

XMF = [gamma.sub.0][RX.sup.gamma1][S.sup.gamma2][e.sup.gamma3T]
 Manufacturing
 exports
XS = [theta.sub.0][RS.sup.theta1][YW.sup.theta2][MS.sup.theta3][e.sup.theta4
T] Service
 exports
MMF = [delta.sub.0][R.sup.-delta1][Y.sup.delta2][SPEC.sup.delta3][e.sup.delta
4T] Manufacturing
 imports
MS = [omega.sub.0][R.sup.-omega1][Y.sup.omega2][XS.sup.omega3][e.sup.omega4T
] Service
 imports

where

RX = WPXG/rPXG: Export competitiveness
WPXG: World export price of manufactures
r: Nominal exchange rate
PXG: Price of domestic exports
S: World trade weighted by importance of markets
RS: Real exchange rate for services
R = WPXG/rPD Real exchange rate
PD: Domestic prices
SPEC: Trend specialisation

Multiplication of each of these volumes by the appropriate price term expressed in domestic and world prices and division by nominal GDP enables a reparameterisation to take place, with prices being translated into terms of the real exchange rate, R. The definition of R used here means that a rise is a depreciation of the exchange rate in the conventional sense. The results presented in this paper reverse the definition above to ensure that a fall in R or the FEER is a depreciation. Having constructed the balance on goods and services the introduction of IPD flows gives the expression for the current account as a proportion of GDP shown by relationship (2).

Disparities between the results across the models are clearly going to arise from differences in the long-run elasticities in the volume equations, shown in Table 1.

[TABULAR DATA OMITTED]

Taking the treatment of manufactures first we can see a degree of consensus across the models, particularly with respect to activity. The NIESR export activity elasticity of 1.02 is in fact the only departure from unity. The spread of competitiveness elasticities is greater, more so for imports than exports, with the NIESR estimate being considerably larger than the view taken by both HMT and BE. The disagreements across the service sector are far greater. HMT are alone in having a feedback from imports of services into exports and vice-versa. The NIESR model has the lowest activity elasticity on exports and the highest on imports. The BE model has a competitiveness elasticity on exports which is considerably lower than the other models and a higher value on imports. Differences in these key elasticities affect the implied long-run position of the current account and hence the value of the FEER consistent with external balance.

Because the NIESR is the only one of the three models used that has the necessary well defined supply side our treatment is to use the NIESR approach and model domestic GDP as a function of the real exchange rate, world prices and a time trend. We utilise the NIESR elasticity estimates, the most important of which is the elasticity of supply with respect to the real exchange rate of 0.20.

Results and findings

Any projected path for the FEER depends on a number of assumptions about the behaviour of the relevant exogenous variables. It is assumed that there are no structural flows which might allow the equilibrium current account to deviate from zero. Over the historical part of the analysis, exogenous variables are set to their actual values, partly to aid interpretation of the results, partly because of the difficulty in deciding the equilibrium path of variables such as world oil prices. Over the forecast period we assume future growth rates of 3 per cent for world GNP, 4 per cent for world trade and a constant real oil price. Large current account deficits in the; early-1990s have led to a fall in net overseas assets and hence we project that net IPD receipts will fall, but the decline in domestic interest rates assumed to follow ERM entry is expected to reduce the payments on deposits made in the UK and hence improve net IPD flows.

Chart 1 shows the actual and trend current accounts for the three models, the trend current account being the value that emerges from equation (2). In this partial equilibrium the real exchange rate is set equal to its actual and forecast values and consequently the trend current account balance differs from zero. The general path is agreed upon by all three models. The actual current account lies above the trend at the start of the 1980s and by a lesser amount from the beginning of the 1990s. The general path is downwards reflecting the trend increase in manufacturing imports that is a feature of all three models.

To consider the full equilibrium, the trend current account is set equal to zero. Chart 2 shows the resulting path of the FEER for each of the three models. The differences between these paths may be explained in terms of the different model elasticities. Over the historical period of the exercise, comparisons can be drawn between the actual and fundamental exchange rate. One striking feature is the similarity of their relative positions at the start and end of the 1980s. The actual real exchange rate is above the FEER in both cases. Both periods are associated with a tight monetary policy with a high nominal exchange rate and high interest rates being used to try to reduce demand and hence inflation. These policies improve the current account position by reducing imports. During the mid-1980s the actual real exchange rate is below the FEER, but both are fairly stable. The actual real exchange rate falls from its level at the start of the decade through a combination of a depreciation in the nominal exchange rate and an improvement in UK inflation relative to its competitors. The FEER is stable during the first half of the 1980s as the positive balance from oil trade and the high propensity to import offset one another. As the former effect becomes less important in the later years of the decade, so the FEER declines.

It is noted that for each model the actual rate is some 5 to 10 per cent above the equilibrium rate at the time of the UK's entry into the ERM. Subsequently this gap widened in the first quarter of 1991 due to an appreciation in the UK effective exchange rate, especially against the dollar. If in the long-run it is argued that large current account deficits are in fact unsustainable then the actual value must converge to the equilibrium. Barrell, Gurney and In't Veld (1992) argue that a current account disequilibrium leads to changes in real wealth which impact on output through consumption so that departures from current account equilibrium are self-correcting. The policies needed to reach equilibrium and their associated costs are not discussed in the present paper but are assessed in the discussion of optimal entry rates for ERM in Wren-Lewis et al. (1991).

The steady downward trend in the forecast is partly due to assumptions on the speed of adjustment of the economy to equilibrium. If the UK continues to have large current account deficits into the future then foreign acquisition of claims on the economy leads to a further downward pressure on the current account through IPD flows, and hence a larger depreciation in the FEER is needed to achieve external balance. By contrast, if the actual real exchange rate converges rapidly to equillibrium then this deterioration of IPD flows does not occur and the decline in the FEER is less pronounced.

With sterling entering the ERM it is no longer possible for the actual to converge to the equilibrium rate by adjustment of the nominal exchange rate. Instead, all the movement has to occur through relative prices. This implies that if the FEER were a constant rate somewhere below the current effective rate then British inflation would have to be held below that of competitors until the FEER is achieved, and from that point onwards be at the same level as those nations. This would seem to be the path that France and the Netherlands are currently following in relation to Germany, with UK inflation also forecast by many to fall below that in Germany in the coming months. Even after the initial misalignment is corrected the trend of the FEER is still downwards, so a continued inflation differential is necessary. Barrell and In't Veld (1991) estimate that the UK inflation rate must be 1.54 percentage points below that in Germany in order to maintain equilibrium. When the UK came out of recession in the early 1980s much of the convergence of actual real exchange rate to equilibrium levels was brought about by a large fall in the nominal exchange rate. With this option seemingly ruled out a prolonged bout of sub-German inflation is required by the UK in order to reach the FEER, and this has to be sustained if the economy is to remain in equilibrium.

Cross-model differences in the FEERs shown in Chart 2 are the consequence of the differing long-run elasticities in the trade volume equations. The sensitivity of the results to changes in these elasticities can be examined by substituting parameter values from one model into the FEER calculation from another and observing the difference this makes. The first exercise takes the most optimistic FEER, that of the NIESR, and replaces the service elasticities with those of the BE. Despite a considerable disagreement over the effects of activity, particularly on exports of services, it is found that changing these elasticities makes little difference to the path of the FEER. Competitiveness elasticities do however have an important effect as shown in Chart 3. The FEER falls when the lower BE export competitiveness value is substituted in, as a greater depreciation in the rate is needed in order to maintain current account balance. Conversely when the higher import competitiveness elasticity is substituted the exchange rate need do less to discourage imports, and the FEER appreciates. The combination of both changes leads to a virtually unchanged FEER.

In 1989 exports of services accounted for approximately 25 per cent of the value of all exports and imports of services 18 per cent of the value of all imports. The corresponding manufacturing shares were 65 per cent and 68 per cent respectively. This would indicate that the UK's long-run performance in manufacturing trade is by far the most important determinant of the trend current account position and hence the FEER. Chart 4 decomposes the variation between the HMT and NIESR FEERs which is due to the manufacturing volume elasticities. There is virtual agreement on activity elasticities (NIESR differ from unity with 1.02 for exports) so no difference in the calculation emerges from this source. Small differences to the FEER path result from replacing the HMT estimate of the manufacturing exports competitiveness elasticity with the slightly larger NIESR value. The combination of this and the lower trend specialisation response in the NIESR import equation leads to a FEER that depreciates less to achieve current account balance. However the most important determinant of differences between estimates is the import competitiveness elasticity for manufactures. Comparison of Charts 2 and 4 suggests that the difference in this elasticity accounts for virtually all the difference between the HMT and NIESR FEERs. The same elasticity also explains why the projected path of the BE FEER is lower than that for the NIESR.

The current account constraint on economic growth

One of the frequent debates conducted between different schools of economic thought has been that of the effect on the UK economy of |Thatcherism'. Did the 1980s herald a new era of fitter more competitive industry or were the relatively prosperous times of the mid-1980s merely a short-term |blip' before the economy returned to |normal'?

Some evidence has been produced to suggest that there has been a halt in the long-run decline of the UK's share of world exports. Landesmann and Snell (1989) argue that the income elasticity of demand for UK manufacturing exports rose during the 1980s to a value higher than that characteristic of the 1970s after a huge fall at the start of the decade. However our results seem robust to changes in this elasticity. By way of contrast a London Business School study (Holly and Wade, 1989) does not find an increase in this parameter in a recursive estimation exercise. On the other hand they conclude that an improvement has taken place on the supply side, so that when faced with an expansion of demand there is now a tendency for UK firms to increase output, whereas previously most of the increase might have been in prices.

We consider the possibility of a structural break in the 1980s by examining whether the residual part of the behaviour of imports and exports of manufactures which is not explained by the long-run elasticities has changed. Instead of fitting the trend part of the volume equations over the period 1975-91 they are fitted over 1983-91. The recalculated coefficients show that the trend tendency for UK exports to decrease over time has turned into a small increase when the models are fitted over the later sample period. A small reduction in the import trend is also found, improving the future trend current account balance further. Under the |old regime', UK growth rates in excess of 1 per cent per annum would have had to have been associated with inflation rates significantly below those in competing economies. Models fitted over the more recent period would suggest that the key growth rate for triggering deflationary pressures is around 2 per cent. With changes in long-run parameters being picked up by a single trend term no conclusion can be drawn about whether it is a greater responsiveness to increases in world trade or to relative movements in world and domestic prices which has led to the improved export performance. If the improvement is in the competitiveness elasticities then the real exchange rate would have to do less work in order to make a given trend current account balance |sustainable'. Obviously for completeness it would be necessary to re-estimate the model equations over the new sample period, as any recent change in the structure of the economy cannot be reflected in estimates made over previous regimes. The trend current accounts from the |old' and |new' regimes are shown in Chart 5. The improved trading performance is reflected in a current account that still declines, but at a much slower rate.

As might be expected this improvement in the current account is reflected by a similar strengthening in the FEER, as shown in Chart 6. Over the historical part of the exercise the FEER is lower than in the original estimate, giving an overvaluation of the actual real exchange rate of nearly 20 per cent at the start of the 1980s. The FEER appreciated gradually until the end of the decade, when a gradual decline commenced. This would seem to be in line with the findings of Barrell and In't Veld (1991).

Sensitivity to competitiveness

One of the problems in predicting the future development of the economy is that the way the economy behaves is subject to change. Indeed there is usually more than one explanation proposed for that which has already occurred which shows the difficulty in stating what is about to arise. In passing comment on the present state of the economy and the prospects for the future it is assumed that the model which describes the recent past still holds and will continue to hold. In the previous section it was found that the trend export of manufactures increased when the analysis was carried out over a more recent sub-sample of the data. This trend increase might however merely reflect changes in other elasticities which, because of a lack of data from the |new regime', econometric estimates cannot show.

If it is believed that the UK economy has undergone a change in its structure then it might be expected that the competitiveness elasticities are now higher than those given by the econometric estimates of the modelling teams. There are various reasons why these elasticities might be higher than current estimates in the future. The first of these reflects continuing increases in world trade. It would be expected that, as barriers to trade are removed, given relative price movements across countries will generate greater trade volumes. The advent of a single European market in 1993 should ensure that competitiveness elasticities are indeed higher in the future. Part of the single market programme is the laying down of common standards for goods. This will lead to a more homogeneous product, a greater degree of substitutability and again higher competitiveness elasticities. Another distinction that can be made between the present and past is the UK's membership of ERM. The consequence of this is the removal of some of the uncertainty associated with movements of the nominal exchange rate which might discourage trade. Previously, changes in competitiveness might be hidden by movements in the exchange rate. In cases where ordering and payment for goods were some distance apart the possibility existed that a decision made on the grounds of a relative price comparison might prove to be the wrong one by the subsequent movement of the nominal exchange rate. With the UK now in ERM it might be expected that any improvement in UK price competitiveness will be reflected by a greater change in trade volumes because of the reduction in uncertainty.

All the results seem to demonstrate that there is a tendency for the FEER to fall over time. Given the possibility of alteration to the structure of the economy as described above, and the small changes to the FEER that occur through cross-model differences, the effects of more radical changes to competitiveness elasticities are shown in Chart 7. The two variants show the effect of first an increase in the export competitiveness elasticity ([NI.sup.*]) and then the additional benefit from raising the import competitiveness estimate ([NI.sup.**]). The downward drift of the FEER is still present but the rate of decline is slightly slower with the higher elasticities. This result is consistent with the sensitivity analysis presented in Barrell and In't Veld (1991). Tle major divergence between the paths starts in the mid-1980s with the decline in North Sea oil revenues, as already explained. The chart indicates the degree to which substantially higher elasticities prevent such a large fall in the FEER over the latter part of the decade. The implications of this are that if competitiveness elasticities are higher than current econometric estimates then convergence of the real exchange rate to the FEER will not require such a severe reduction in UK inflation relative to that of competitors as that implied by estimates in the current models. However even the substantial changes shown here do not alter the need for UK inflation to be lower than that of countries such as Germany for a period of time to facilitate this convergence.

Conclusion

This paper has shown that there is a high level of cross-model agreement on the long-run position of the current account and hence on the FEER. The result is a reflection of the similarity in the key elasticities which itself is symbolic of a general convergence in the approach to modelling these areas. The differences that do exist between the FEERs presented here come from the manufacturing imports equation. The methodology is the same, with imports being determined along with domestic production and domestic demand but different competitiveness elasticities result, with that of NIESR being slightly higher than those of BE and HMT. These differences do not alter the overall picture that emerges from the exercise. The FEER is projected to depreciate in the forecast. This particular feature depends on the speed of convergence of the real exchange rate to the equilibrium rate. If convergence is rapid the trend depreciation will be lower. However if it is believed that current account deficits will persist for some time into the future, this worsens the position of the UK IPD flows and hence the FEER must depreciate by more to achieve balance.

The results are also sensitive to the period over which the trade equations are fitted. Recent experience has shown that export equations in the large-scale models tend to underpredict and require residual adjustments in forecasting (see Bray et al., 1992). Fitting the equations over a more recent sample improves the trend estimates and prevents much of the decline in the forecast of the FEER. Even with this improved performance the actual real exchange rate was about 5 per cent above the FEER in September 1990 when sterling entered the ERM. This does not imply that the rate at which the UK joined was too high but adjustment of actual to the FEER must take place through movements in relative prices because the nominal exchange rate is fixed. Barrell, Gurney and In't Veld (1992) estimate that to eliminate three-quarters of a 10 per cent overvaluation of sterling might require up to five years of lower UK inflation relative to competitors. All the results presented in this article suggest that the UK FEER is depreciating, which in turn means that the gap between actual and equilibrium actually widens if UK inflation is not below that of competitor countries. The implications are that the longer the period before adjustment commences, the longer the time that UK inflation will have to be below that of other countries to achieve equilibrium.

NOTES

(1) We are grateful to these authors for providing us with the data and a computer program for the calculation of the FEER. The forecast data is taken from the May 1992 National Institute Economic Review.

REFERENCES

Barrell, R. and Wren-Lewis, S. (1989), |Fundamental equilibrium exchange rates for the G7', Centre for Economic Policy Research Discussion Paper No.323. Barrell, R. and In't Veld, J.W. (1991), |FEERS and the path to EMU', National Institute Economic Review, No.137, 51-58. Barrell, R, Gurney, A. and In't Veld, J.W. (1992), |The real exchange rate, fiscal policy and the role of wealth: an analysis of equilibrium in the monetary union', Journal of Forecasting, forthcoming. Bray, J., Hall, S., Meen, G., Westaway, P. and Whitley, J.D. (1992), |UK policies, non-price competitiveness, and convergence to an EMU', ESRC Macroeconomic Modelling Bureau Discussion Paper No.28. Holly, S, and Wade, W. (1989), |Focus: Supply-side improvements in UK exporting of manufactures', Economic Outlook 14, 35-37. Landesmann, M. and Snell, A. (1989), |The consequences of Mrs Thatcher for UK manufacturing exports', Economic Journal, 99, 1-27. Williamson, J. (1983), The Exchange Rate System, Cambridge, Mass: MIT Press. Williamson, J. (1991), |FEERs and the ERM', National Institute Economic Review, No.137,45-50. Wren-Lewis, S. (1992), |On the analytical foundations of the Fundamental Equilibrium Exchange Rate', In C.P. Hargreaves (ed), Macroeconomic Modelling of the Long Run, pp.323-338, Aldershot: Edward Elgar. Wren-Lewis, S., Westaway, P., Soteri, S. and Barrell, R. (1991), |Evaluating the UK's choice of entry rate into the ERM', Manchester School, 5 9, 1-22.