Sterling, the euro and the dollar.

Blake, Andrew P. ; Byrne, Joseph P.

In the period of floating exchange rates a distinct dollar-D-mark polarisation was noted, whereby periods of weakness for some European currencies relative to the dollar coincided with periods of strength relative to the Deutsche mark and vice versa. The relationship continued to be observed for those currencies such as sterling which remained outside the Exchange Rate Mechanism for most of the period. The observation, although more 'an empirical regularity in search of a theory' than a model (Buiter et al., 1998) seems reasonably robust; recently the dollar has fallen sharply against sterling while the euro has risen. An analysis of this relationship and its stability is important because it provides a guide as to how the euro might be expected to move against sterling as a part of any adjustment against the US dollar. In turn it sheds some light on the degree to which a dollar depreciation might affect Britain inside the Euro Area.

Haldane and Hall (1991) investigated sterling's relationship with the dollar and the D-mark from the mid1970s to the end of the 1980s. Their primary purpose was to see if the 'EMS-effect' operated, so that the sensitivity of the DM/[pounds sterling] bilateral exchange rate to the DM/$ rate is reduced by European Monetary Integration. Since then, of course, the euro has replaced the Deutsche mark, but the same method can be applied.

In essence, they suggest estimating:

[([[euro].sup.*]/[pounds sterling]).sub.t] = [[alpha].sub.t] + [[beta].sub.t] [([[euro].sup.*]/$).sub.t] + [e.sub.t] (1)

and allowing both [alpha] and [beta] to vary over time. All variables are in natural logarithms. We denote the euro-sterling exchange rate by [[euro].sup.*]/[pounds sterling], because before European Monetary Union we use the straight DM/[pounds sterling] rate and afterwards the euro rate converted into DM. This is to give exact comparability with the Haldane and Hall (1991) results.

We approach this by the simplest of time-varying parameter models, where we assume a random walk for the coefficients. This implies a model of the form:

[y.sub.t] = [e'.sub.t][z.sub.t] + [[epsilon].sub.t] (2)

with [e.sub.t] a vector of the euro/dollar rate (the chosen explanatory bilateral exchange rate) and a constant set at unity, [y.sub.t] is the euro/sterling rate and [[epsilon].sub.t] is a random error with [[epsilon].sub.t] ~ N (0, [Q.sub.t]), all at time t. [z.sub.t] is a potentially time-varying parameter vector such that:

[z.sub.t] = I[z.sub.t-1] + [[omega].sub.t] (3)

so that the coefficient vector reflects the random walk, with [[omega].sub.t] ~ N (0, [R.sub.t]). Technically, this is a simple statespace model with observation equation (2) and transition equation (3). Estimation can be approached using the Kalman filter (see Harvey, 1989).

We use daily (noon New York buying rates) data from 4 January 1971 to 28 June 2002 from the US Federal Reserve. We find that there was a diminishing relationship between sterling and the dollar during the 1980s, as shown by the fall in [[beta].sub.t]. Our results parallel those reported by Haldane and Hall (1991) and their analysis over the period is similarly appropriate. Their evidence supports the view that an 'EMS effect' on sterling did exist to a certain extent, with much of the relationship between sterling and the dollar having diminished previous to 1978. Their results also indicated that sterling's link to the dollar declined after the official announcement of the UK'S 'shadowing the Deutsche mark' policy in the Spring of 1987 and generally up until 1989. Additionally, Haldane and Hall concluded that membership of the full European Monetary System would be much easier by 1990 than if sterling had still been largely determined by its relationship with the dollar.

After 1992 (post-dating the previous study), we estimate that sterling's relationship with the dollar strengthened considerably. This linkage is mainly due to the fact that the exchange rate with the Deutsche mark/euro has been floating since September 1992. There has also been a considerable change in monetary policy regimes throughout the Euro Area and, indeed, the United States, with a universal move to inflation targeting. With successful control of inflation, the sterling exchange rate should reflect the trading position with respect to each partner, which is confirmed by the movement in [[beta].sub.t]. Income flows may also be important in determining this relationship. The peak of the relationship between sterling and the dollar coincides with the operational independence of the Bank of England in May 1997. Since the inception of European Monetary Union, and the irrevocable fixing of bilateral Euro Area exchange rates, there has been a decline in the dollar-sterling correlation, but at a much slower ra te than at the inception of EMS in 1978.

Taylor (2002) has also tested the relationship between sterling and the dollar in a recent paper considering the fall we can expect in sterling's effective exchange rate if the UK joins EMU. With regard to the recent relationship between sterling, the euro and the dollar, he examines monthly data between 1975 and 2001 and estimates the correlation between the dollar--sterling and dollar--DM rates (the direct inverse of the relationship between DM--sterling and DM--dollar). He also estimates over sub samples to describe the evolution of the relationship between the D-mark and sterling. He finds that the relationship between the D-mark and sterling was at a sustained peak between April 1987 and August 1992, which is consistent with our results. We also agree with his finding that sterling's dollar affinity has approximately equalled its euro/DM affinity since the UK left the ERM. Taylor suggests that the level of affinity is based on the UK'S external business (including income flows). The changing relationship between the UK, US and Euro Area currency is interesting since it indicates that exchange rate policy has been clearly influential in determining sterling's relationship to the dollar and the D-mark.

We can consider the implications for exchange rate behaviour between our three currencies on the basis of our estimates of [[beta].sub.t]. According to chart 1, approximately half of any move in the euro--dollar rate will be reflected in the sterling--dollar rate. Between January and June of this year, the dollar and sterling depreciated by 9.8 per cent and 5.6 per cent respectively against the euro. This is in line with our Kalman filter estimates of [[beta].sub.t]. Some further policy implications of the estimated coefficient in chart 1 are considered by Barrell (2002).

The current value of [[beta].sub.t] is 0.52. This implies that if the dollar falls by 10 per cent against the euro it is also likely to fall by 4.8 per cent against sterling. In an enlarged Euro Area including Britain, the UK would amount to about 16 per cent of the total economy. Weighting the exchange rate changes by the relative sizes of the two economies, a 10 per cent fall of the dollar against today's euro would be likely to be replaced by a 9.2 per cent fall of the dollar against a Euro Area which included the UK. Thus membership of an enlarged Euro Area which included the UK would be likely to result in the UK experiencing almost twice as much volatility against the dollar as it does at the moment. If, however, Taylor's view that the affinity depends on income flows is correct, then the effect of the UK on the relationship between the expanded euro and the dollar is likely to be greater than the above calculation suggests, because UK currency flows outside the Euro Area are probably more than 16 per c ent of the overall flows of the expanded Euro Area. This suggests that UK membership of the euro could be a help with the reduction of overall currency volatility, and not just the volatility between the UK and the current Euro Area.

The impact of government policy with respect to the euro for the sterling--dollar correlation needs to be considered. If the rate with the euro was announced with entry in, say, three years' time, there is a possibility that, if the policy were perfectly credible, the relationship between sterling and the dollar would quickly diminish (as would be expected if the Bank of England began issuing euros immediately) and [[beta].sub.t] would fall towards zero. It may be argued that currently there is not a strong market expectation of the UK joining Monetary Union since sterling still has an affinity with the dollar. It also seems likely that this aspect of 'convergence' is directly linked to UK policy, suggesting that there is little sense in treating the stability of the [pounds sterling]/[euro] rate as an exogenous test of convergence. These points, it should be stressed, are an interesting area for future research.

In summary, whether there is a formal or informal, pegged or floating exchange rate, policy is clearly important in influencing the relationship between sterling and the euro or dollar. In the 1990s the relationship between sterling and the dollar increased. At the moment, since UK monetary policy has no exchange rate target, the sterling--dollar rate is largely determined by the UK and dollar block's trade pattern and income flows.

[GRAPH OMITTED]

REFERENCES

Barrell, R. (2002), 'Equity markets, block realignments and the UK exchange rate', National Institute Economic Review, 181, July.

Buiter, W.H., Corsetti, G. and Pesenti, P.A. (1998), Financial Markets and European Monetary Cooperation: The Lessons of the 1992-1993 Exchange Rate Mechanism Crisis, Cambridge, Cambridge University Press.

Haldane, A.G. and Hall, S.G. (1991), 'Sterling's relationship with the Dollar and the Deutschemark: 1976-1989', The Economic Journal, 101, pp. 436-43.

Harvey, A.C. (1989), Forecasting, Structural Time Series Models and the Kalmon Filter, Cambridge, Cambridge University Press.

Taylor, C. (2002), 'Sterling volatility and European Monetary Union', NIESR Discussion Paper No. 197.

Joseph P. Byrne *

* National Institute of Economic and Social Research. E-mail: a.blake@niesr.ac.uk; j.byrne@niesr.ac.uk