Convergence of the EU Member States towards the EMU requirements, 1986 to 1993
European Journal of POLITICAL European Journal of Political Economy ECONOMY Vol. 13 (1997) 247-259 ELSEVIER Convergence of the EU Member States towards the EMU requirements, 1986 to 1993 Peter Guldager Department of Economics, The Aarhus School of Business, Fuglesangs Allg 4, 8210 Aarhus V, Denmark Accepted 1 August 1996 Abstract This paper analyzes the five convergence criteria of the Maastricht Treaty by means of factor analysis. The results show that the criteria may be summarized by two independent factors, inflation and public debt performance, with inflation figuring more prominently. The factor scores suggest that between 1986 and 1990 France, Ireland, Spain, Portugal and the United Kingdom progressed substantially towards convergence. Some backsliding, however, occurred between the years 1991 and 1993. The performance of Greece deterio- rated, while that of the other EU Member States remained largely stationary. To meet the EMU requirements by 1999, economic convergence would have to take place at an accelerated pace in Stage II of the EMU. JEL classification: F02; F15; F33 Keywords: European economic and monetary union; Maastricht criteria; Convergence; Factor analysis; Inflation and public debt performance 1. Introduction The Treaty on European Union as agreed at Maastricht stipulates that the Economic and Monetary Union (EMU) is to be achieved in three stages, i The Stage I of the Economic and Monetary Union commenced on July 1, 1990 and Stage II on January 1, 1994. 0176-2680/97/$17.00 Copyright 1997 Elsevier Science B.V. All rights reserved. PII S0176-2680(96)00038-9 248 P. Guldager / European Journal of Political Economy 13 (1997) 247-259 final stage (III) will be initiated on January 1, 1999. The authors of the Treaty originally envisaged the possibility of an EMU taking place before 1999. If, by the end of 1996, a majority of EU Member States were to have fulfilled the five criteria of economic convergence set forth in the Treaty, an earlier date could have been set. This option was abandoned by the European Council in Madrid in December 1995. A decision of the Council in early 1998 will determine the Member States qualifying for participation in the EMU. The abandonment of the early option indicates that convergence of national economic policies among a majority of the Member States was not taking place at a sufficient pace to establish the EMU before January 1, 1999. The European Commission has initiated a number of detailed country studies to assess the convergence of fiscal policies towards the EMU of all Members except Luxembourg, which had already met the criteria by the end of 1993. Based upon these studies, De Haan et al. (1994), after assessing in detail the margins for consolidating the budgets of the 11 Member States, found scope for consolidation both on the expenditure and the revenue side, but generally that more scope existed for cutting expenditure. The need for consolidation was deemed most urgent in Greece, Belgium and Italy. 2 My intention here is to analyze economic convergence among the EU Member States by specifying a simplifying aggregative framework to assess the pace of progress towards the EMU. This framework is based upon all the requirements of the Union Treaty for membership in the EMU and is established by factor analysis. A factor analysis such as I undertake here is a useful procedure in the context of EMU evaluation. The main features of the five convergence criteria can be summarized into a few common factors which offer insight into the basic common dimensions that the criteria are measuring. Understanding is thereby enhanced of the basic impediments to Member States' meeting the criteria. The assessment, which covers the period 1986 to 1993, also determines the extent to which the necessary degree of convergence envisaged in the Union Treaty for Stage I of the EMU took place. The paper is organized as follows: In Section 2 the five criteria of economic convergence set forth in the Maastricht Treaty are briefly discussed. The method of factor analysis is reviewed in Section 3. In Section 4 a factor analysis is performed to establish the independence of and the correlation among the criteria, and the factors which the criteria are measuring in common are identified. In Section 5, I calculate the factor scores for the EU Member States which form the basis for assessment of convergence towards the EMU requirements. Section 6 presents concluding comments. 2 For further elaboration of the conflict between the public finance requirements and the stabilization function of fiscal policy, see Jensen and Jensen (1995). P. Guldager / European Journal of Political Economy 13 (1997) 247-259 249 2. The convergence criteria of the Treaty on European Union The Union Treaty stipulates that a Member State aspiring to membership in the EMU must satisfy the following conditions: 3 Condition 1. Consumer price inflation over the previous year must not exceed by more than 1.5 percentage points that of, at most, the three best performing Member States in terms of price stability (article 109j with protocol). This formulation leaves some room for interpretation. It is, however, the received opinion that an average of the inflation rates in the three best performing Member States will be used as the threshold value unless one country exhibits an extreme value (e.g. negative inflation). In that case it could be disregarded. 4 One would expect that convergence in inflation to a common rate would be a major requirement for sustaining a monetary union, although theoretically this common rate need not be low or even stable. Condition 2. Long-term government bond yields, observed over a period of one year, must not exceed by more than 2 percentage points that of, at most, bond yields in the three countries with the best inflation performance (article 109j with protocol). Again, the opinion is that the threshold value will be an average of the bond yields in the three best performing Member States. Theoretically, one would expect Members of a monetary union with low rates of inflation to have low bond yields if real interest rates are assumed to equalize throughout the union (the Fisher effect). 5 Condition 3. The currency should have remained within the normal exchange rate band of the ERM for at least 2 years, without devaluing against the currency of any other Member State (article 109j with protocol). On August 2, 1993 the normal fluctuation band of the ERM was temporarily suspended by the decision to introduce the wider 15 percent band. 6 This suspension followed the removal of capital controls mandated by the Single European Act which, as Eichengreen (1993) points out, undermined the viability of the ERM and brought the Community to a choice between greater exchange rate flexibility and monetary unification. 3 An extensive discussion of criteria can be found in Gros and Thygesen (1992). 4 This notion has recently been confirmed by the Danish Ministry of Economic Affairs. 5 For elaboration, see Fisher (1896). 6 The Netherlands and Germany, however, continue bilaterally to use the normal band. 250 P. Guldager / European Journal of Political Economy 13 (1997) 247-259 Condition 4. The ratio of the planned or actual government deficit to GDP should not exceed 3 percent, unless the ratio has declined substantially and continuously, or is only temporarily above 3 percent (article 104c with protocol). Condition 5. The ratio of government debt to GDP should not exceed 60 percent unless the ratio is sufficiently diminishing and approaching the 60 percent limit at a satisfactory pace (article 104c with protocol). The qualifications to the two principles of fiscal discipline allow Members with excessive deficit-GDP and/or debt-GDP ratios some latitude and make a liberal interpretation of both fiscal criteria possible. This suggests that the convergence process towards the EMU is more than simply fulfilling a list of ironclad criteria. 3. Factoring the Maastricht criteria Economic theory postulates relationships among the five Maastricht criteria. Govemment budget deficits and public debt are related, and nominal bond yields and the stability of the intra-ERM exchange rates are affected by changes in nominal inflation rates (the Fisher effect and Cassel's purchasing power parity theory). 7 In some cases, inflation may also relieve pressure arising from budget deficits, and some EU governments could be tempted to reduce the pressure of high public debt by inflation taxation. 8 Common elements thus influence the five convergence criteria. Factor analysis can identify these common elements. Factor analysis is a statistical method that attempts to explain correlations among a set of variables in terms of a small number of such common factors. The common factors cannot be observed directly, as the variables are subject not only to the factors themselves but also to random disturbances. If common factors of the five Maastricht criteria can be established, the convergence process towards the EMU can be assessed in a more aggregative framework than each of the five criteria individually. The factor model can be written in the form x=Af+e. (1) x (p X l) is a random vector of the variables representing the Maastricht criteria. 9 e (p X l) is a vector of random errors with zero mean and covariance matrix qr which is assumed to be diagonal. A (p x k) is a matrix of constants (factor loadings) and f (k X l) is a vector of common factors. It is assumed that all the 7 For elaboration, see Cassel (1918). 8 See Keynes (1923) on the theoretical possibilities. 9 The definition of the variables is that used by Amundsen and Guldager (1993). P. Guldager / European Journal of Political Economy 13 (1997) 247-259 251 factors are uncorrelated with one another and standardized to have variance 1 and 10 zero means. As formula (1) implies, the covariance matrix (~) of x, can be expressed in terms of A and qs as = AA' + ~. (2) The estimation of the parameters of the factor model below is based upon principal components. The Maximum Likelihood Method is not feasible as the data cannot be assumed to be normally distributed with one of the variables expressed as a dummy. As the variables are not measured in the same units, the correlation matrix (R) rather than the covariance matrix (S) is used. To interpret the factor loadings, the factors are rotated using the varimax method of orthogonal rotation. This method makes the interpretation of the factor loadings the most straightforward, as the axes are provided with a few large loadings and as many near-zero loadings as possible. This is accomplished by iterative maximization of a quadratic function of the factor loadings. Following Kaiser's rule of thumb, only factors with eigenvalues above unity are included in the analysis. ~1 This rule excludes those factors where the eigenvalues are less than the average, i.e. less than one when the correlation matrix has been used. 4. Identification of the factors Table 1 presents the results of the factor analysis of the five Maastricht Treaty criteria performed for each year between 1986 and 1993. All the eigenvalues are displayed together with the varimax-rotated matrices of factor-loadings. 12 The figures show that the first and the second factor together account for 85 and 83 percent of the total variation in 1986 and 1993 respectively, whereas the three remaining factors may be discarded, in accord with Kaiser's rule of thumb. 13 Consequently only two factors are needed to explain the data adequately. It appears that the two factors generally increase in importance until 1991, and account for 91 percent of the total variation in that year. Beyond 1991, the fit of the model becomes less precise. The communalities also vary on a year-to-year basis reaching a peak for all variables in 1991 except for bond yields, for which communality peaks in 1989. 14 Turning to the factor loadings themselves, some 10 See Harman (1976) for details of the factor model. tl For elaboration, see Kaiser (1960). 12 Note that the eigenvalue of the second factor is below one for 1993. 13 The variation explained is calculated as the sum of the eigenvalues retained by the model divided by the total number of factors, e.g. (3.29 + 0.88)/5 = 0.83 for 1993. 14 The communality of a variable is the sum of its squared factor loadings, e.g. the communality of the inflation variable for 1993 is (0.952 +0.092) = 0.91. 252 P. Guldager/ European Journal of Political Economy 13 (1997) 247-259 Table 1 Eigenvalues and principal factor solution for factor loadings, Varimax rotation, 1986-93 1986 1987 1988 1989 1990 1991 1992 1993 Eigenvalues A1 2.91 2.83 3.05 3.18 3.34 3.42 3.30 3.29 A2 1.32 1.43 1.41 1.34 1.20 1.15 1.14 0.88 A3 0.48 0.50 0.30 0.26 0.30 0.32 0.35 0.55 A4 0.24 0.21 0.20 0.19 0.12 0.09 0.17 0.23 A5 0.05 0.04 0.04 0.03 0.04 0.02 0.03 0.04 Inflation fl 0.94 0.93 0.96 0.96 0.93 0.96 0.95 0.95 f2 0.08 0.08 0.07 0.09 0.14 0.11 0.10 0.09 Bond yields fl 0.90 0.95 0.95 0.96 0.95 0.92 0.91 0.90 f2 0.09 0.09 0.17 0.18 0.22 0.24 0.29 0.27 Exchange rate stability fl 0.82 0.79 0.89 0.90 0.90 0.89 0.88 0.75 f2 0.16 0.14 0.01 0.00 0.00 0.01 0.03 0.14 Government deficit fl 0.49 0.41 0.50 0.55 0.64 0.70 0.63 0.72 f2 0.81 0.86 0.80 0.76 0.72 0.67 0.69 0.56 Government debt fl -0.06 -0.08 -0.13 -0.10 -0.01 0.02 0.01 0.13 f2 0.96 0.96 0.96 0.96 0.98 0.99 0.98 0.97 caution should be applied in comparing the loadings, as the matrices employed in the rotation are not identical. However, the weighting between the first factor and the dummy representing exchange rate stability is substantially lower in 1993 than in the previous years. This means that the dummy is explained less well by the first factor in 1993. The loadings for government budget deficit also appear to vary over time. In 1991 and 1993 the first factor is most important in explaining the variation in government budget deficit, whereas up to 1990 it is the second factor which is consistently the most important. A pattern is clear from the rotated loadings. The criteria of inflation rates, bond yields and exchange rate stability are heavily loaded on the first factor, the criterion of government debt is heavily loaded on the second, while the criterion of government deficit is loaded on both. 15 Thus, the five criteria may be reduced to 15 The loading of the budget deficit on factor 1 may be partly explained by seigniorage. As Gros and Vandille (1995) point out, revenue from seigniorage is lowered by disinflation. It may also be explained by the positive relationship between budget deficits and nominal interest rates found by Knot and De Haan (1995). P. Guldager / European Journal of Political Economy 13 (1997) 247-259 253 Factor I ( f l) and Inflation 4 3 2 1 En I -1-2 -1 k I 1 2 Factor 2 (f2) and Debt Fig. 1. EU Member States performance, 1993. fl/f2 Inflation/Debt 1993 two common factors: fl representing inflation rate, bond yields, exchange rate stability and government deficit, and 12 representing government debt and govern- ment deficit. To strengthen the identification of the two factors, a 1993 plot for the EU Member States with fl on the vertical axis and 12 on the horizontal axis is shown in Fig. 1. To establish the connection between the two common factors and the inflation and debt rates, the standardized inflation/debt coordinates of the Member States have also been plotted into Fig. 1. 16 The Maastricht requirements are expressed by the zero-lines. Fig. 1 reveals that the factor scores (11) generally reflect the coordinates calculated from the data concerning inflation rates and government debt ( ) . Greece's performance is better than its inflation/debt position would indicate. The United Kingdom, on the other hand, performs slightly worse on the factor scores and better on inflation/debt. The other members who show some discrepancy between the factor scores and the inflation/debt coordinates trade off good performances on some criteria against poor performances on others. The overall results of Fig. 1 make it reasonable to identify the first factor as inflation performance and the second factor as public debt performance. The importance of the two factors varies over time. According to Table 1, the first factor accounts for 58 percent of the variation in 1986 and for 66 percent in 1993. 17 The second factor accounts for 26 percent in 1986 and only 18 percent in 1993. This indicates that the inflation performance is not only the dominant factor, but becomes increasingly dominant over time, although it seems to have peaked at 16 The standardized inflation/debt coordinates are obtained by scaling the inflation rate and debt ratio into unit variance and zero mean. t7 The variation explained by the first factor is (3.29/5) = 0.66 for 1993. 254 P. Guldager / European Journal of Political Economy 13 (1997) 247-259 68 percent in 1991. The debt factor continues to display decreasing importance from 1987 onwards and may be discarded in 1993, again, in accord with Kaiser's rule of thumb. 5. Convergence Among the EU Member States The results for economic convergence among the EU Member States over the period 1986 to 1993 are presented in Figs. 2-4. In these figures, economic convergence is measured by the factor scores relative to the Maastricht Treaty requirements, which are expressed by the zero-lines. The factor scores illustrate how the factors depend upon the observed variables; that is, based upon a particular set of observed criteria, statements may be made about the country's overall inflation and debt performance. Fig. 2 presents the factor scores for Germany, Italy, Ireland and the United Kingdom for each of the years 1986 to 1993. Germany consistently met both requirements, although its economic performance has been slightly less satisfac- tory since 1989. Italy has not met any of the requirements since 1986. In 1993 it faced a substantial government deficit of 9.6 percent of GDP. Ireland displays movement in a south-west direction, indicating a marked improvement in its inflation as well as debt performance over the period. The deterioration of Ireland's position from 1992 to 1993 is mainly due to a 10 percent devaluation of its currency, the Punt, on February 1, 1993. The figure also shows that the United Kingdom is consistently outside the EMU threshold for inflation performance. This is mainly due to failure to meet the Exchange Rate criterion. The UK meets the debt requirement for the whole period, and the figure reflects an improvement Factor 1 1,5 1 0,5 0 -0,5 -1 -1,5 iI= ~8 :e4, ')3 (I,2 #'8- 7 8~u 2 ~3 Germany Italy @ Ireland United Kingdom s 1 = 1986 2 = 1987 3 = 1988 4 = 1989 5 = 1990 I I I 6 = 1991 -1 -0,5 0 0 ,5 1 1,5 2 7 = 1992 Factor 2 8 = 1993 Fig. 2, Factor scores for Germany, Italy, Ireland and the United Kingdom. P. Guldager / European Journal of Political Economy 13 (1997) 247-259 255 Factor 1 3 ~6 5 ~8 3~ ~ ~7 1 6 .7 8 3 11111 i14 115 16 117 q i -0,5 0 0,5 8 -1 I I i -1 1 1,5 2 2,5 Factor 2 Denmark Ii Belgium 4, Greece Portugal 1 = 1986 2 = 1987 3 = 1988 4= 1989 5 = 1990 6 = 1991 7 = 1992 8 = 1993 Fig. 3. Factor scores for Denmark, Belgium, Greece and Portugal. in its debt position in the period 1986 to 1990 partly due to its privatization programmes. Growing fiscal deficits in 1992 and 1993 led to an increase in its debt-GDP ratio. The factor scores for Denmark, Belgium, Greece and Portugal are presented in Fig. 3. Over the period 1986 to 1993 Denmark shows an improvement in inflation performance at the expense of public debt achievement, which indicates that it has traded off inflation against public debt. In 1993 Denmark had the lowest inflation rate in the EU but its debt position has continuously worsened since 1987. Belgium scores well on the inflation factor for the whole period and is consistently Factor 1 2 1,5 1 0,5 0 -0,5 -1 -2 =:1 5~7 26 81 q.1 2 ~s I I i -1 -0,5 0 i i -1,5 Luxembourg Netherlands 4, France Spain i!= 1 =1986 2=1987 3=1988 4=1989 5=1990 I 6 = 1991 0,5 7 = 1992 Factor 2 s = 1993 Fig. 4. Factor scores for Luxembourg, the Netherlands, France and Spain. 256 P. Guldager / European Journal of Political Economy 13 (1997) 247-259 below the EMU requirements, but has the worst debt position of all EU Member States (137.2 percent of GDP in 1993) and shows no significant improvement. The worst performer of all Members is Greece. Since 1986 it has moved in an eastward direction, indicating a worsening debt position without any improvement in its poor inflation performance. In 1993 it faced acute fiscal problems with a govern- ment deficit of 13.2 percent of GDP and a fast growing public debt ratio. Portugal shows an improvement in its inflation performance but devaluations in 1992 and 1993 place her outside the inflation threshold. The strong performance of Luxembourg is noted in Fig. 4. On the inflation side, Luxembourg barely meets the Treaty requirements, but it displays an extremely low public debt rate compared with the other EU Member States. France demonstrated an improved inflation performance until 1990, moving in a southward direction with a constant debt rate, and has met both requirements since 1989, although its debt position has deteriorated since 1991. Spain also has a public debt rate within the EMU limit, but high inflation and devaluations of the Peseta in 1992 and 1993 place it outside the inflation threshold. The Netherlands, on the other hand, falls comfortably within the boundary set for inflation perfor- mance, but does not meet the debt rate limit. Its high debt ratio has been stabilized and the 1993 government budget deficit ratio of 3.3 percent is close to the Maastricht criterion. 6. Concluding comments The study has shown that the five criteria of the Maastricht Treaty can be basically summarized by two independent factors, inflation and public debt performance. These factors can serve as an aggregative framework for analyzing convergence towards the EMU of the EU Member States. Measured by these two factors, five EU Member States (France, Ireland, Spain, Portugal and the United Kingdom) show some convergence towards the EMU requirements in the period 1986 to 1990, although some of the ground gained was lost between 1991 and 1993. During 1986 to 1993 Denmark improved its position with regard to inflation performance, but at the expense of its public debt. The economic performance of Greece deteriorated, whereas that of the other Member States remained largely stationary. The overall result confirms insufficient convergence in Stage I of the EMU for a monetary union comprising all Member States by January l, 1999. Poor debt performance is likely to inhibit some Member States from participat- ing in the EMU. Italy and Greece in particular appear to be in a non-sustainable long run situation, with debt-GDP ratios that are steadily deteriorating. Nor is the situation comfortable in Belgium and Ireland, where the debt rate has stopped increasing, but remains at a high level. In 1993 Greece, Italy and Belgium faced a serious two-fold problem of excessive public deficit and public debt. An excess in both is a particularly uncomfortable combination; the one compounds the other. P. Guldager / European Journal of Political Economy 13 (1997) 247-259 257 Countries with this two-fold problem face serious obstacles on the road to the EMU. However, all may not be so bleak. De Haan et al. (1994), in their study of available options to consolidate revenue and expenditure categories, find that Greece, Italy and Belgium could improve their budgetary position in areas such as income and corporate taxes, and reduce expenditure on social security and defence. The large public debt of these countries generally makes the redress of public finances through reduced expenditures difficult, as interest payments, which in all three Members amount to more than 10 percent of GDP, cannot be cut. Thus, when the day comes for the European Council to vote on the inclusion of Greece, Italy and Belgium in the EMU, a very liberal interpretation of the two principles of fiscal discipline will be required, if they are to pass the test. Belgium, with its good inflation performance, is a better candidate than Italy and Greece where prospects seem dim due to poor performance on the other criteria. My results thus suggest that there is a core group of Members that, based on inflation and debt performance, is ready for membership in the EMU. This group comprises Luxembourg, France, Germany, the Netherlands and Ireland. Denmark also belongs to the core group but has obtained a waiver enabling it to opt out of the Monetary Union. The other Members are in a group which may have to wait beyond 1999 to join the EMU, unless economic convergence takes place at an accelerated pace in Stage II. These results compare well with von Hagen and Neumann (1994) who found that a core group consisting of Germany, her smaller neighbours and France had prospects for membership in the EMU and who proposed a 'Europe in Two Speeds'. De Grauwe (1996), however, warns that placing excessive emphasis on the convergence criteria causes a 'Great Divide' of the EU Member States. He proposes instead that future monetary institutions be strengthened. My study shows that, if less emphasis is to be placed on the convergence criteria, the main features for requisite criteria can be captured by considering inflation and public debt performance of the Member States. The diminishing importance of debt performance over time suggests a prominent summary role for inflation perfor- mance as determining membership eligibility in the EMU. Acknowledgements I am grateful to Rolf Amundsen, Jane Cook, two anonymous referees and the Editor of this Journal whose suggestions have improved the substance of this paper. Remaining inadequacies are my own. Appendix A. Data sources and definition of variables The 1986 data for all variables except the dummy representing exchange rate stability are derived from European Economy, No. 58, 1994, (published by the 258 P. Guldager / European Journal of Political Economy 13 (1997) 247-259 European Commission), and for the years 1987 to 1993 from No. 60, 1995. The table numbers referred to below are the same for both volumes. Information about the dummy representing the criterion of normal fluctuation margins for at least two years within the Exchange Rate Mechanism is obtained from the description detailed for the individual EU Member States in Exchange Arrangements and Exchange Restrictions, Annual Reports, 1987 to 1994, from the IMF. Consumer price inflation is reported in Table 27, Price deflator private con- sumption (annual percentage change in national currency). For each year the three best performing countries are defined as the three EU Member States reporting the lowest inflation rates in Table 27. Long-term government bond yields are recorded in Table 54, Nominal long-term interest rates, (yields on government bonds). The Maastricht criteria for inflation rates and long-term government bond yields are calculated annually as the average of the inflation rates (Table 27) and bond yields (Table 54) in the three best performing countries plus 1.5 and 2 percentage points respectively. General government budget deficit is measured by Table 62, Net lending or net borrowing; general government (percentage of GDP at market prices). The ratio of gross government debt to GDP is found on line 14: General government gross debt (percentage of GDP) in the country Tables (68 to 82). The criterion of normal fluctuation margins for at least two years within the ERM is measured at the end of each year by a dummy. The dummy is assigned a value of zero if the Member State has remained within the normal band of the ERM without changing its central rate for the whole of that year and the preceding one, and is otherwise assigned a value of one. Due to the introduction of the wider band, the dummy for 1993 receives a zero value when the Member State has not realigned the central rate for the preceding two years and a value of one otherwise. Specific adjustments for individual Members are: Germany: All German data relate to West Germany except for public deficit and public debt rate of GDP for the period 1991 to 1993. Greece: Nominal long-term interest rates are not reported after 1988. Instead nominal short-term interest rates are used (Table 53). References Amundsen, R. and P. Guldager, 1993, Om Maastricht-traktatens konvergenskrav til den konomiske og moneta~re union, NationalCkonomisk Tidsskrift 131, 211-220. Cassel, G., 1918, Abnormal deviations in international exchanges, Economic Journal 28, 413-415. De Grauwe, P., 1996, The economics of convergence: Towards monetary union in Europe, Weltwirtschaftsliches Archiv 132, 1-27. Eichengreen, B., 1993, European monetary unification, Journal of Economic Literature 31, 1321-1357. Fisher, I., 1896, Appreciation and interest, AEA Publications 3, 331-442. Gros, D. and G. Vandille, 1995, Seigniorage and EMU: The fiscal implications of price stability and financial market integration, Journal of Common Market Studies 33, 175-196. P. Guldager / European Journal of Political Economy 13 (1997) 247-259 259 Gros, D. and N. Thygesen, 1992, European monetary integration: From the EMS to EMU (Longman, London). De Haan, J., C.G.M. Sterks and C.A. de Kam, 1994, An assessment of margins for consolidation in the Member States of the European Union, In: European economy, Reports and studies, no. 3: Towards greater fiscal discipline (European Commission, Brussels). von Hagen, J. and M.J.M. Neumann, 1994, Real exchange rates within and between currency areas: How far away is EMU?, Review of Economics and Statistics 76, 236-244. Harman, H.H., 1976, Modern factor analysis, 3rd ed. (The University of Chicago Press, Chicago, IL). Jensen, S.E.H. and L.G. Jensen, 1995, Debt, deficits and transition to EMU: A small country analysis, European Journal of Political Economy 11, 3-25. Kaiser, H.F., 1960, The application of electronic computers to factor analysis, Educational and Psychological Measurement 20, 141 - 151. Keynes, J.M., 1923, A tract on monetary reform (Macmillan, London). Knot, K. and J. De Haan, 1995, Fiscal policy and interest rates in the European Community, European Journal of Political Economy 11, 171-187.