III. The Rise of the Sterling Real Exchange Rate1
1. The past few years have witnessed a marked appreciation of the United Kingdom’s real exchange rate (RER), most notably against the euro-area countries, but also in trade-weighted effective terms.2 On the CPI-based measure, sterling rose by some 30 percent between mid-1996 and end-2001, while the increase in relative unit labor costs, at almost 40 percent, was even more spectacular.3 Yet, during this period, the current account deficit—which widened from 1 percent of GDP in 1996 to about 2 percent of GDP in 1999/2000 (before narrowing to 1.3 percent of GDP in 2001, partly for cyclical reasons4)—has been perhaps surprisingly modest. More formally, a “macroeconomic balances” approach suggests that only a relatively moderate real depreciation would be required to bring the saving-investment balance closer to its historical norm.5 As such, irrespective of the original causes of the real exchange appreciation, the equilibrium real exchange rate may have increased as well, making a sudden reversal of the RER less likely. This paper explores alternative theories that could account for an appreciation of the equilibrium RER.
2. The first explanation relies on downward sloping demand curves for countries’ exports. According to this theory, since growth of manufacturing productivity in the United Kingdom has been outstripped by productivity growth in the U.K.’s main trading partners, the fall in the relative supply of U.K. goods implies that their relative price can rise—that is, the equilibrium RER can appreciate. As a simple test of this theory, a cointegrating relationship between the RER and productivity differentials is used to estimate the “long-run” equilibrium real exchange rate, which indeed shows a substantial increase in recent years.
3. The second explanation, more difficult to establish econometrically, argues that a higher real exchange rate can be supported by the diversification in recent years of the U.K.’s exports into higher value-added and “knowledge-intensive” goods as well as tradeable (especially financial and business) services. This trend has limited the deterioration of the merchandise trade balance, and helps explain the resilience of the services’ balance—in turn making the current account balance and the observed real appreciation more sustainable.
4. The plan of the paper is as follows. Section 2 briefly reviews the developments in the real exchange rate and the external balance over the past few years. Sections 3 and 4 turn to alternative theories that would suggest that the equilibrium RER may have increased. Section 5 presents some further evidence on the appreciation of the equilibrium RER by examining the effects of the observed RER appreciation on exports and the manufacturing sector more generally. Section 6 concludes.
B. The Real Exchange Rate and the External Balance—Summary of Developments
5. Starting in mid-1996, sterling began to appreciate in real terms particularly against countries in the euro-area, but also in trade-weighted effective terms; Figure 1. Between mid-1996 and the third quarter of 2001, the CPI-based RER had appreciated by some 30 percent, while relative unit labor costs rose by some 40 percent.
Figure 1.United Kingdom: Real Effective Exchange Rate and Relative Unit Labor Costs
Sources: International Monetary Fund, Information Notice System; and International Financial Statistics.
6. A number of explanations have been put forward to account for the rise of the RER. One view is that the real appreciation, which has been mostly vis-à-vis the euro, represents not so much the strength of sterling than weakness of the euro nominal exchange rate.5 Another view is that an increase in domestic demand resulted in the real exchange rate appreciation. Meredith (2001) develops a model in which a decline in the equity premium spurs higher consumption and investment. The monetary tightening that this higher aggregate demand is assumed to elicit in turn appreciates the real exchange rate. More generally, a demand boom may have contributed to the real exchange rate appreciation by driving up the price of nontradables, especially housing.6
7. But while nominal and demand side factors may explain the initial real exchange rate appreciation, they provide a less satisfactory explanation for why the real appreciation has been sustained for more than five years. If the real appreciation represented nothing more than the nominal exchange rate movement, then, over time, it should be offset by inflation differentials. Likewise, the widening of the interest differential (against Germany and the United States) does, more or less, coincide with the appreciation of sterling between 1996 and 1998, but thereafter the interest rate differential narrowed sharply while the real exchange rate did not depreciate—indeed, it experienced a further (modest) appreciation (Figure 2).
Figure 2.Real Exchange Rate and Nominal Interest Rate Differentials
8. Regardless of the reasons for the RER appreciation, it has been associated with a deterioration of the external balance. But the extent of this deterioration has been perhaps surprisingly modest. Between 1996 and 2000, the current account deficit widened from 1.1 percent of GDP to 1.8 percent of GDP, reflecting a deterioration of the trade balance of about 1.4 percent of GDP, partly offset by an improvement in net services exports (0.2 percent of GDP) and higher net income and transfers (0.5 percent of GDP), Figure 3. Moreover, a “macroeconomic balance” approach suggests that only a relatively moderate real depreciation would be required to bring the saving-investment balance closer to its historical norm.
Figure 3.Current Account and Components
Sources: International Monetary Fund; and National Statistics.
9. In volume terms, growth of exports (excluding oil) slowed abruptly from some 8-9 percent in 1996/97 to less than 2 percent in 1998, but then picked up again in 1999/2000, reaching a very respectable 12 percent per year, (and about 8 percent in the first half of 2001). In addition, the geographical breakdown of the trade balance does not correlate well with the behavior of the real exchange rate. Specifically, the balance with the euro-area countries—against which sterling appreciated most strongly—has actually been improving over the past few years (Figure 4). On the other hand, the trade deficit with non-EU countries has widened by about 2 percent of GDP since end-1997.8
Figure 4.United Kingdom: Exports, Imports, and Trade Balance: 1988-2001
Sources: National Statistics; and IMF staff estimates.
10. Taken together, these observations suggest that—irrespective of the original cause of the real appreciation—the equilibrium RER may have increased as well. This could explain why the impact on the external balance has been moderate, as well as the persistence of the real appreciation despite the floating exchange rate regime. But why should the equilibrium real exchange rate have appreciated? The remainder of this paper explores two possibilities.
C. Why Has The Equilibrium Real Exchange Rate Appreciated?
Perhaps the best known model of equilibrium real exchange rate movements is the Balassa-Samuelson effect, whereby faster productivity growth in the tradables sector leads to a rise in the relative price of nontradables, and hence an appreciation of the real exchange rate. However, the ratio of manufacturing/total economy productivity (typically taken as a proxy for tradables/nontradables productivity) has grown more slowly in the United Kingdom than in its partner countries9—indeed, the ratio actually declined between 1994 and 1999 (Figure 5, top panel)—so that Balassa-Samuelson effects would suggest a depreciation of the equilibrium RER, not an appreciation.
Figure 5United Kingdom: Manufacturing and Whole Economy Productivity
1/ United Kingdom: Output Per Job Filled; United States: Output Per Person; Germany: Output Per Hour Worked.
Source: National Statistics; US Bureau of Labor Statistics; Deutschc Bundesbank; and Fund staff estimates.
11. The faster growth in manufacturing productivity in the U.K.’s partner countries (Figure 5, bottom panel), however, suggests another channel through which the equilibrium RER may be affected. Under the assumption that countries’ exports are imperfect substitutes, standard trade models imply that an increase in supply will require a lower relative price because the demand curve slopes downward. In this theory, therefore, faster productivity growth of manufactures in the U.K.’s trading partners has lowered their relative price, implying an appreciation of the U.K.’s equilibrium RER.10 In line with this theory, there has indeed been an improvement in the U.K.’s manufacturing terms of trade that mirrors the real exchange rate appreciation (Figure 6), although—as discussed below—the rise in the terms of trade is consistent with alternative theories as well.
Figure 6.United Kingdom: Manufacturing Terms of Trade and Real Oil Price
Source: National Statistics; and International Monetary Fund.
The Appendix articulates a generic three-good model (nontraded, exportable, and importable good) to identify more formally factors that could affect the equilibrium real exchange rate.11 The model is characterized by two schedules: internal balance, which requires that output equal expenditure; and (long-run) external balance, which requires that net factor payments on the stock of foreign assets (W) match the trade deficit. The internal balance schedule is upward sloping in the real exchange rate (q) and W space—a higher stock of external wealth raises consumption, so that external demand must be choked off by an appreciation of the real exchange rate. The external balance schedule is also upward sloping in (q,W) space, since a higher stock of external wealth generates greater net factor payments, and can thus support a larger trade deficit and more appreciated real exchange rate (Figure 7, panel (a)).
12. Within this framework, it is straightforward to establish the long-run implications of changes in fiscal policy or changes in productivity (relative to the partner country). An increase in government expenditure requires a jump appreciation of the real exchange rate to maintain internal balance in the face of higher aggregate demand. Since the current account deficit widens, however, over time the stock of net foreign assets declines as does the stream of net factor payments. Over the longer-run, therefore, the equilibrium real exchange rate must depreciate to reduce the trade deficit (Figure 7(b)).12
13. Productivity increases in both the traded and nontraded sectors will cause a downward shift of the external balance schedule since, at any given stock of external assets, a more depreciated real exchange rate is required to offset the deterioration of the trade balance arising from the higher income. There will also be a downward shift of the internal balance schedule, since a more depreciated real exchange rate is required to increase aggregate demand to match the higher aggregate supply. The net effect is likely to be a depreciation of the real exchange rate in the short-run, and unambiguously so in the long-run (Figure 7c). Finally, an increase in productivity in the tradable sector relative to the nontradable sector, shifts both schedules upward, leading to an appreciation of the real exchange rate, essentially due to Balassa-Samuelson effects (Figure 7d).
Figure 7.Response of the Real Exchange Rate to Alternative Shocks 1/
q: real exchange rate (increase is an appreciation)
W: stock of foreign assets
Pre-shock saddle path:
Post-shock saddle path:
14. As outlined above, the theoretical model has certain implications for the effects of fiscal policy and productivity growth differentials on the equilibrium RER.13 To test these implications a co-integrating relationship for the real exchange rate is specified. If a co-integrating relationship exists, it is natural to identify it with the long-run equilibrium real exchange rate. The estimated relationship can then be used directly to establish whether the equilibrium RER has indeed appreciated.14
15. Two variants of the model are estimated. The first uses the differential between the United Kingdom and its trading partners (an average of the United States and Germany) of the government balance, productivity in the traded goods sector (proxied by the manufacturing sector), and the ratio of productivity in traded/nontraded goods (proxied by the ratio of productivities in manufacturing and the whole economy). The second variant of the model augments this vector with the terms of trade, real oil prices, and the real long interest rate differential.15 It is not clear, however, that the terms of trade is a legitimate variable for explaining the real exchange rate movement. If, contrary to the assumption of the model, the United Kingdom were small in the world market for its exports, the terms of trade could be treated as exogenous and might help explain the real exchange rate dynamics. But if, as seems likely, the United Kingdom has at least some market power in its exports of manufactures, then the terms of trade will not be exogenous, and there is a risk of circularity in using the terms of trade to explain movements of the real exchange rate.
16. Table 1 presents standard augmented Dickey-Fuller (ADF) statistics. Once significant lags are included, the ADF statistics cannot reject a unit root in any of the variables; as such co-integration techniques are appropriate.
|Real exchange rate 2/||-2.58|
|Oil price 3/||-2.37|
|Terms of trade 4/||-3.03|
|Manufacturing Productivity 5/||-1.33|
|Government balance 6/||-2.23|
|Ratio Manufacturing/Whole economy Productivity 7/||-0.86|
|Real interest rate 8/||-2.20|
|Residual from cointegrating vector, Model 1||-4.22 **|
|Residual from cointegrating vector, Model 2||-3.58 *|
17. Table 2 presents the relevant Johansen statistics for each model. In each case, the Johansen technique finds a single co-integrating vector (except when the real interest differential—which is probably borderline stationary itself—is included), suggesting that the procedure has low power. But the co-integrating vectors, reported in Table 3, accord with the theoretical model: an improvement in the fiscal balance is associated with an appreciation of the (long-run) real exchange rate, as is an increase in the ratio of traded/nontraded productivity; an increase in productivity in traded goods sector (holding constant the traded/nontraded productivity ratio) is associated with a long-run real depreciation. In the augmented model, a terms of trade improvement is associated with a real appreciation as is an increase in real oil prices, while an increase in the real interest rate differential has a marginally negative effect.
|Horrank = ρ||-Tlog(1-μ)||using T-nm||95% critical level||-T\Sum log(.)||using T-nm||95% critical level|
|ρ==0||31.73 *||30.20 *||27.10||50.63 *||48.19 *||47.20|
|ρ = o||50.64 **||47.59 **||33.50||96.58 **||90.76 **||68.50|
|ρ==0||58.73 **||54.48 **||39.40||137.10 **||127.10 **||94.20|
|ρ<=l||36.45 *||33.82 *||33.50||78.32 **||72.66 *||68.50|
|Ratio Manufacturing/Whole economy Productivity||0.86||0.55||1.10||0.42|
|Terms of trade||…||…||2.21||0.36|
|Real interest rate||…||…||-0.01||0.01|
Figure 8 plots the actual and estimated equilibrium RERs for the two models. A number of points are noteworthy. First, the actual and equilibrium RERs generally track quite closely. Second, the equilibrium RER shows a substantial appreciation starting in mid-1996 so that, at end-2000, the actual RER was only slightly above the estimated equilibrium RER. Third, given that the ability of the two models to track the real exchange rate is roughly equal, the additional variables included in the second model do not appear to contribute much to the explanatory power. Indeed, a decomposition of the co-integrating vectors into their Table 1 components shows that, in the first model, the bulk of the movement is accounted for by shifts in productivity (with the improvement in the fiscal balance making a minor contribution), while in the second model, it is again mostly shifts in productivity and the terms of trade that matter (Figure 9). Finally, it is noteworthy that between 1990 and 1992 (the eve of the ERM crisis) the RER was substantially above its estimated equilibrium level.
Figure 8.Actual and Estimated Long-run Real Exchange Rate
Source: Author’s calculations.
Figure 9.United Kingdom: Decomposition of Estimated Equilibrium Real Exchange Rate
Source: Author’s calculations.
18. To summarize, theory implies that if countries face downward-sloping demand curves for their exports, a decrease in relative productivity in the tradables sector will be associated with a higher price—that is, a real exchange rate appreciation. Estimates based on a parsimonious co-integrating vector suggest that the equilibrium RER has indeed appreciated.
D. Alternative Explanations
19. An alternative explanation for why the equilibrium RER may have appreciated does not rely on a downward-sloping demand curve for aggregate exports.16 Rather, it suggests that a shift toward higher value-added exports in recent years has resulted in an improvement in the terms of trade and an appreciation of the equilibrium RER.17
20. Why would this shift to higher value-added exports not be reflected in the productivity data? Traditional manufacturing productivity measures may not be able to capture price and quality increase as the product mix shifts, and of course completely ignore the services sector—yet the United Kingdom has seen a marked expansion of “tradable” services, most notably the financial and business services sector.
Is there any evidence in support of this view? While direct evidence is difficult to obtain, there are a few indicators. First, over the past few years, many low value-added manufactured exports—food, beverages, clothing—have had sluggish or no growth, while exports of chemicals, machinery and transport equipment, and scientific and photographic equipment have enjoyed robust growth (Figure 10). Second, after a sharp fall in 1998, profitability of manufacturing firms has been recovering (until the recent cyclical downturn) (Figure 11). Indeed, net rates of return were comparable to levels seen in 1994-95 (prior to the sharp appreciation) and higher than levels in the late 1980s (before the United Kingdom joined the ERM). Third, crude measures of the “knowledge-content” of products suggest that U.K. exports rank among the highest in the G-7, and have been increasing over time (Figure 12). Finally, there has been an expansion of the business, communications, and financial services sectors. Overall, these indicators suggest that there has been a shift toward tradable services and higher value-added exports, which would imply an appreciation of the equilibrium RER, though it is not possible to quantify by how much.
Figure 10.United Kingdom: Export Growth by Sector
Source: National Statistics.
1/ Value index, 1990Q1=100
Figure 11.United Kingdom: Net Rates of Return of Private Non-Financial Corporations Figure 12.United Kingdom: Exports of high and medium technology goods, G7 comparison: 1981-1999
Source: National Statistics.
E. Further Evidence on the Equilibrium Real Exchange Rate Appreciation
21. Taken as a whole, the evidence above suggests that the U.K.’s equilibrium RER may have appreciated in recent years.18 Beyond such purely statistical concepts of equilibrium, however, the RER may be considered close to equilibrium if it does not lead to an unsustainable external deficit. As noted in Section B, to date the impact on the trade balance seems to have been relatively moderate. But does this simply reflect lags in the effects of the real exchange rate appreciation on exports? If so—and if the lagged effects were sufficiently large—it could undermine an assessment that the equilibrium RER has appreciated. This section therefore bolsters the analysis by examining whether the effects of the observed real exchange rate appreciation have completely fed through to the economy.
22. Thus far, the real appreciation has lasted more than five years, so presumably much of the effects on production decisions should have been manifested already. On the other hand, inasmuch as the real appreciation affects investment decisions, there may be substantial gestation lags. Since these are difficult to pin down, the focus here is simply on the impact on exports and the share of manufacturing in the economy.
23. Figure 13 relates the actual log level of the volume of exports to its long-run co-integrating vector, where the latter depends on activity in the partner countries (Germany, the United States, and Ireland) and the real exchange rate.19 The actual value of exports, as of 2000, is very close to the value implied by the co-integrating relation, suggesting that—barring any further real appreciation—most of the effect on exports has already been observed.
Figure 13.United Kingdom: Exports-Actual and Long-run Fitted (1979-2000) (volume, in logarithms)
Source: Author’s calculations.
Figure 14 undertakes a similar exercise for the share of manufacturing in total value added, which is related to the real exchange rate and a time trend. There has been a secular Figure 13 share of manufacturing is about ½ percentage point above the implied long-run value, suggesting that the full effects of the real appreciation to date are likely to entail a further (albeit modest) decline in the share of manufacturing in the United Kingdom economy.
24. The persistence of the real exchange rate appreciation observed in the United Kingdom over the past few years (despite the floating exchange rate regime) suggests that the equilibrium RER may have appreciated as well. This paper has explored two alternative theories for why the equilibrium RER could have appreciated. The first theory focuses on the price effects of productivity differentials across countries. The second explanation is based on the shift toward higher value-added exports and tradable services. In practice, both factors are likely to have been at play.
25. The assessment that the equilibrium RER has appreciated is bolstered by the relatively moderate impact on the external balance of the observed real exchange rate appreciation. Moreover, simple econometric exercises suggest that much of the effect of the RER appreciation on exports has already been realized, and while there may be further effects on the manufacturing sector more generally, they are unlikely to be terribly large.
A Theoretical Model of Real Exchange Rate Determination
26. This Appendix outlines a three good (non-traded, exportable, importable) model of real exchange rate determination.
27. The home country uses labor to produce a nontraded good (N) and an exportable good (T). Since labor is mobile, the marginal product of labor will be equated across sectors. Let 0 denote labor productivity, then assuming flexible prices, the price of nontraded good relative to the home country (QN) depends on relative productivity of the traded good:
Or, taking logarithms:
28. Using the envelope theorem, the value of output in terms of the home country’s exportable may be written:
B. Real Exchange Rate
29. Let qN denote the price of the home country’s non-traded/traded good price ratio relative to that of the foreign country:
which depends only on the relative productivity differentials. To simplify, in what follows, the foreign country’s productivities are assumed to be constant and their logarithm normalized to zero.
30. Let qx denote the relative price of the home traded good:
31. The CPI-based real exchange rate q = p - (p*+e) then captures both the nontraded/traded price ratio (relative to the foreign country) qN, and the price of the home traded to the foreign traded good (qx).
where αNαT are the shares of nontraded and imported goods in the CPI.
C. Internal Equilibrium
32. Internal equilibrium requires that the value of income equal the value of expenditure:
where W is wealth (foreign assets), r, is the real interest rate, Y is GDP, and Cw > 0,C, < 0,0 < CY < 1;TBq < 0,TBy < 0
33. Real interest parity requires that
34. Totally differentiating (7) and using (8) gives the schedule for internal equilibrium:
35. The internal equilibrium is upward sloping in (q,w) space: a higher real exchange rate deteriorates the trade balance and reduces aggregate demand. This fall in aggregate demand requires higher consumption and therefore higher wealth.
D. External Equilibrium
36. External equilibrium states that foreign wealth is accumulated as long as there is a current account surplus:
37. The external equilibrium is also upward sloping in (q,w) space: higher stock of foreign assets yields a greater net factor payments which finances a trade deficit (in turn associated with a more appreciated real exchange rate).
Alberola, Enrique; Cervero, Susana G.; Lopez, J.Humberto; Ubide, Angel (1999), “Global Equilibrium Exchange Rates - Euro, Dollar, “Ins,” “Outs,” and Other Major Currencies in a Panel Cointegration Framework” IMF Working Paper No. 99/175.
Barrie, Robert and RobertJukes (2001), “Learning to Fly,” Economics Research, Credit Suisse First Boston (Europe), October5, 2001.
Barker, K. (2001), “The Policy Implications of Economic Imbalances,” Speech delivered at the Royal Bank of Scotland,/Scottish Economic Society Annual Lecture, October23, 2001, Edinburgh, available from Bank of England website: http://www.bankofengland.co.uk/speeches/speechl47.htm
Church, Keith B. (1999), “Properties of the Fundamental Equilibrium Exchange Rate in the Treasury Model,” National Institute Economic Review No. 169 pp. 96–104.
Hansen, Jan and WernerRoeger (2000), “Estimation of Real Equilibrium Exchange Rates,” European Commission, Economic Papers No. 144
Isard, Peter; Faruqee, Hamid; Kincaid, G. R.; Fetherston, Martin J. (2001), “Methodology for Current Account and Exchange Rate Assessments,” IMF Occasional Paper No. 209.
Krugman, Paul (1989), “Differences in Income Elasticities and Trends in Real Exchange Rates,” European Economic Review No. 33, pp 1031–54.
Lipschitz, Leslie; McDonald, Donogh (1991), “Real Exchange Rates and Competitiveness -A Clarification of Concepts, and Some Measurements for Europe” IMF Working Paper No. 91/25.
MacDonald, Ronald (1997), “What Determines Real Exchange Rates? The Long and Short of it.” IMF Working Paper No. WP/97/21.
Meredith, Guy M. (2001), “Why Has the Euro Been So Weak?” IMF Working Paper No. 01/155.
Søndergaard, Jens (2002), “Real Exchange Rate Determination in a New Open Macroeconomics framework—A Refinement of the Balassa-Samelson Hypothesis.” Mimeo, Georgetown University.
Wadhwani (1999), “Currency Puzzles,” speech delivered at the London School of Economics, September16, 1999, available from Bank of England website: http://www.bankofengland.co.uk/Links/setframe.html
Prepared by Atish R. Ghosh.
Some implications of the real exchange rate appreciation are discussed in a previous Selected Issues paper for the 2000 Article IV Consultation (SM/01/145). See Wadhwani (1999) for a comparison of models explaining the appreciation of sterling; and Barker (2001) on possible implications of the exchange rate for the United Kingdom economy.
Lipschitz and MacDonald (1991) provide a useful discussion on interpreting various indicators of competitiveness.
Projection based on data through the third quarter of 2001; the current account deficit was also smaller because of two “one-off” effects: a change in the accounting of income flow from interest rate swaps, and lower than expected transfers to the EU.
Meredith (2001) explores some of the reasons that the euro nominal exchange rate has been weak.
See Selected Issues paper on “Why Has United Kingdom Household Consumption Been So Strong?”
One explanation is that, in order to cut costs, U.K. firms have started out-sourcing the manufacture of more basic components to other countries, particularly in Eastern Europe and East Asia, while focusing their own production on higher value-added goods. The desire to maintain market share in Europe may also have been a factor, making it difficult to correlate bilateral trade balances to exchange rate movements.
For convenience, the U.K.’s “partner country” is taken to be an unweighted average of the United States and Germany. An alternative would be to construct a trade-weighted measure, in line with the effective exchange rate, although this is not explored in the paper. Productivity measures are the “preferred” measures for each country: output per job filled (United Kingdom), output per person (U.S.), output per hour worked (Germany).
At first blush, it may appear counter-intuitive that faster productivity growth in the tradable sector could depreciate the equilibrium real exchange rate. It bears emphasizing, however, that this mechanism is quite distinct from the Balassa-Samuelson effect, whereby higher productivity growth in the tradables sector relative to the nontradables sector results in a higher price of the nontraded good, and hence an appreciation of the RER. Here, faster productivity growth in the foreign country’s exportable sector leads to a fall in its relative price, and hence an appreciation of the home country’s equilibrium RER. A recent paper by Sondergaard (2002) provides empirical evidence for such an effect in a panel of OECD countries. Another possibility is that higher productivity growth, by presaging greater wealth, leads to a consumption boom, which to the extent that it falls on nontraded goods, appreciates the RER (through an increase in the price of the nontraded good). While this might explain an appreciation of the actual real exchange rate, it would not necessarily represent an appreciation of the equilibrium RER.
In fact, econometric estimates for the United Kingdom suggest that a fiscal expansion depreciates the real exchange rate even in the short-run; see the Selected Issues paper on “The Macroeconomic Effects of United Kingdom Fiscal Policies: An Empirical Exploration.”
What follows is not a full test of the theoretical model. Such a test would require specifying the dynamics toward the long-run equilibrium, as well as the inclusion of other variables such as the stock of external wealth. The net international investment position (in percent of GDP) can be included in the cointegrating vector (estimated below), but in practice it makes only a marginal explanatory contribution and little difference to the other coefficient estimates.
See MacDonald (1997) for a similar approach.
All variables are in logarithms, except the government balance differential (percent of GDP) and the interest rate differential (percent per year).
Indeed, Krugman (1989) argues that countries need not experience a real depreciation even if their export growth rate is higher than that of their partner countries. In his model, the expansion of exports takes the form of the production of new varieties for which there is corresponding demand. The alternative explanation for the appreciation of the United Kingdom equilibrium RER presented here is in the spirit of this model, though the specific details differ.
It is an open question whether this shift is a response of the economy to the real exchange rate appreciation, or whether it reflects underlying technologically-drive trends that are independent of the exchange rate.
For a similar conclusion, see Barrie and Jukes (2001).
The activity elasticities are around 0.5, while the real exchange rate elasticity is about- 0.4. The ADF t-statistic on the residual from the cointegrating relation is -4.5.