The United Kingdom and the stability of the Euro area: From Maastricht to Brexit

The year of 2019 was to mark the 20th anniversary of the European single currency (the euro) and the expected exit of the United Kingdom, known as Brexit, from the European Union. It is then appropriate to ask what will be the impact of Brexit on the stability of the euro area? Does Brexit undermine the euro because it undermines the EU as a political and economic project, or will it allow the rest of the EU to implement institutional changes that will help the euro work better? These are difficult and pressing questions. The contribution of this paper is to provide some of the key elements for an informed debate. Regarding the stability of the euro area, a measure of how widely recognised are the current shortcomings of the Economic Monetary Union (EMU) is the Brussels’ plan for a Genuine EMU (Begg, 2014).1 Agreement on the need for a solution coexists with an apparently stark disagreement on the causes. One view is that "design flaws" (De Grauwe, 2006; De la Dehesa, 2012) deepened imbalances, while another is that "policy mistakes" (Sandbu, 2015) hindered convergence. One of the many proposed solutions is a flexible euro (Stiglitz, 2016): a two-tier model of a Northern and a Southern euro where the latter is said to be "softer." One way to explain such proposals is that the Southern euro would not be part of the "core" or that it would be "less core." All these views, however, rely upon "asymmetries": the less asymmetric, the more synchronised, the more stable will the euro area be. Regarding Brexit, in June 2016, 52% of British voters decided that being the first country ever to leave the EU was a price worth paying despite extensive advice from economists that Brexit would make the UK permanently poorer (Campos, 2019). Moreover, Brexit is one among a constellation of crises inflicting upon the EU (populism, refugees, debt, unemployment, etc.). Although one among


| INTRODUCTION
The year of 2019 was to mark the 20th anniversary of the European single currency (the euro) and the expected exit of the United Kingdom, known as Brexit, from the European Union.It is then appropriate to ask what will be the impact of Brexit on the stability of the euro area?Does Brexit undermine the euro because it undermines the EU as a political and economic project, or will it allow the rest of the EU to implement institutional changes that will help the euro work better?These are difficult and pressing questions.The contribution of this paper is to provide some of the key elements for an informed debate.
Regarding the stability of the euro area, a measure of how widely recognised are the current shortcomings of the Economic Monetary Union (EMU) is the Brussels' plan for a Genuine EMU (Begg, 2014). 1 Agreement on the need for a solution coexists with an apparently stark disagreement on the causes.One view is that "design flaws" (De Grauwe, 2006;De la Dehesa, 2012) deepened imbalances, while another is that "policy mistakes" (Sandbu, 2015) hindered convergence.One of the many proposed solutions is a flexible euro (Stiglitz, 2016): a two-tier model of a Northern and a Southern euro where the latter is said to be "softer."One way to explain such proposals is that the Southern euro would not be part of the "core" or that it would be "less core."All these views, however, rely upon "asymmetries": the less asymmetric, the more synchronised, the more stable will the euro area be.
Regarding Brexit, in June 2016, 52% of British voters decided that being the first country ever to leave the EU was a price worth paying despite extensive advice from economists that Brexit would make the UK permanently poorer (Campos, 2019).Moreover, Brexit is one among a constellation of crises inflicting upon the EU (populism, refugees, debt, unemployment, etc.).Although one among many, Brexit differs in that it can alone ignite other crises.Brexit raises existential questions about the integration project.It asks questions about the value of membership, the dynamics and distribution of its benefits and costs, and the type of integration that can sustain the net benefits seen since the 1950s.
One of the few benefits of the Brexit debate is that it has fostered a flurry of new research addressing questions that have not been sufficiently investigated previously.One of these questions regards cohesion among euro area members, where the governance structure of the relationship between the countries that use the euro as their currency (i.e. the euro-ins) and those that do not (the euro-outs) is an important issue.The latter group includes both the countries that have negotiated the right to opt-out from participation under the Maastricht Treaty of 1992 (the euro-outs, i.e. the UK and Denmark) and those who are on the path to eventual adoption of the single currency (the pre-ins, i.e.Central Eastern EU).
The paper is organised as follows.Section 2 provides a conceptual framework.It discusses the theory of optimal currency areas, its recent developments and the centrality of the concept of synchronicity.Section 3 analyses the extent to which economic activity in the UK is synchronised with economic activity in the euro area and how this has changed over time-especially after the introduction of the single currency.Consistent with the existing literature, we find synchronisation has increased after the introduction of the euro.Section 4 introduces new empirical measures of economic symmetry among European economies.Stability depends on the degree of integration among member countries or, more specifically, on the relative distance between core and periphery countries.Using these new measures, we show that the gap between core and periphery pre-EMU has diminished after the introduction of the euro and that the UK contribution was key in the sense that it moved from the periphery before 1990 to the core.On the other hand, the UK is also shown to be the one country in which this measure post-euro has varied the most (i.e. has been the least stable).Section 5 discusses policy implications to help increase the stability of the euro area.Section 6 concludes.

| INTEGRATION, SYMMETRY AND STABILIT Y
Sharing a currency deepens integration.The main research question driving the optimal currency areas (OCA) scholarship regards the costs and benefits of sharing a currency (Alesina & Barro, 2002).The main cost is the loss of monetary policy autonomy.Benefits are mostly in terms of reduction of transaction costs and exchange rate uncertainty, and increasing price transparency, trade and competition.Glick and Rose (2016) summarise the econometric evidence on the trade effects of currency unions.
One insightful way of framing the OCA issue is proposed by De Grauwe and Mongelli (2005).They study the interactions between symmetry, flexibility and integration.The more changes in the levels of economic activity across countries happen in unison; that is, the more synchronised are their business cycles, the more integrated will countries be.Particularly, they show there exists a minimum combination of, for example, flexibility and integration that countries must observe for a monetary union to generate positive net benefits.De Grauwe and Mongelli (2005) place the Eurozone (EU) within (to the outside) of the OCA line, suggesting those countries are (not yet) sufficiently integrated to generate efficiency gains that can compensate for the macroeconomic costs of the union.They also note how the degree of economic integration and symmetry may change over time.
Before the EMU, there was an intense debate about the extent to which a monetary union affects symmetry (Krugman, 1993).Focusing on the symmetry-openness dimension, one can see that increased integration may raise business cycle correlation.De Grauwe and Mongelli (2005) argue the EU would move in this way: they predict specialisation will bring about less symmetry.
There are at least two recent developments in OCA theory that should be noted.The original OCA formulation stressed labour mobility, product diversification and trade openness as key adjustment criteria and explored the possible endogeneity of currency unions (Arestis & Phelps, 2016;Frankel & Rose, 1998).Recent work calls attention to the role of credibility shocks.If there are varying degrees of policy commitment (furthering time inconsistency problems), countries with dissimilar credibility shocks should find it convenient to join a currency union (Chari, Dovis, & Kehoe, 2019).A second relevant recent strand highlights that, although OCA criteria are often thought of as independent, they should instead be considered jointly, for example, by focusing on the interactions between openness and mobility (Farhi & Werning, 2015).
The optimality of a currency area is a function of the distance between its members.If relative distances are large, it is common to speak of a core and periphery gap.It is expected that core countries would be those more closely meeting the OCA criteria (Basse, 2014;Belke, 2006).Given its importance for OCA, it is not surprising there have been various attempts of classifying countries into core and periphery sets.A basic way of distinguishing these methods is whether the authors pre-impose membership, or they allow the data to determine whether a country is a member of the core or the periphery at a certain point in time.Artis and Zhang (2001), for instance, investigate actual and prospective membership of the EMU by applying clustering techniques to a set of variables suggested by OCA theory: the extent of synchronisation in business cycles (symmetry in output shocks), volatility in the real exchange rate, synchronisation in the real interest rate cycle, openness to trade, inflation convergence and labour market flexibility.Their analysis reveals that the member countries may be divided into three groups: those belonging to the core (Germany, France, Austria, Belgium and the Netherlands), those part of a Northern periphery (Denmark, Ireland, the UK, Switzerland, Sweden, Norway and Finland) and those belonging to a Southern periphery (Spain, Italy, Portugal and Greece).Bayoumi and Eichengreen (1993) put forward a more theory-based approach focusing on business cycle synchronisation embedded in a standard Aggregate Demand and Aggregate Supply framework that classify Germany, France, Belgium, the Netherlands and Denmark as core countries pre-EMU, and Greece, Ireland, Italy, Portugal, Spain and the UK as the pre-EMU periphery.Bayoumi and Eichengreen (1997) also offer an "optimum-currency-area index for European countries."They identify the determinants of nominal exchange rate variability, which reflect OCA characteristics and support predictions of which countries pertain to which sets.Conceptually, they make the point that OCA focuses on criteria that ultimately make exchange rates more stable and monetary unification less costly.In their model, bilateral exchange rate variability is a function of GDP, trade, economic structure dissimilarity and a measure of output synchronisation.Using 1973 to 1992 data, they find all these determinants carry expected signs and are of statistical significance, so they use these to forecast exchange rate variability in 1987, 1991 and 1995.Their econometric analysis allows three groups: in the first "rapidly converging" group are Germany (the numeraire), Austria, Belgium, the Netherlands, Ireland and Switzerland.The second group is characterised as one that has experienced little convergence and is composed of the United Kingdom, Denmark, Finland, Norway and France.The third group is a set of countries that are "gradually converging" to the EMU and includes Sweden, Italy, Greece, Portugal and Spain.They conclude that economic integration has thus increased countries" readiness for monetary integration (Bayoumi & Eichengreen, 1997, p. 769).

AREA?
During the negotiations for the 1992 Maastricht Treaty, Denmark and the UK secured an opt-out, that is, the right not to join the European Monetary Union (EMU).Every one of the other current 26 European Union members is legally committed to adopting the euro as its currency when ready (De Grauwe, 2016).
In 1997, the new Labour government in the UK decided to reconsider the decision to stay out of the euro.The Treasury was charged with the policy analysis, which focused on the so-called "five tests" involving synchronisation of business cycles, labour mobility, investment, competitiveness of the financial system, and growth and stability.Despite several studies showing convergence between the euro area and the UK (e.g.Canova, Ciccarelli, & Ortega, 2005;Giannone, Lenza, & Reichlin, 2010), the final verdict from the Treasury was that long-term convergence of UK and euro area business cycles had not reached satisfactory levels and that "despite the risks and costs from delaying the benefits of joining", a decision to join was not "in the national economic interest." Since the introduction of the euro on 1 January 1999, among the EU15 the UK and Sweden adopted a free float exchange rate regime, while Denmark decided to participate in the ERM2 with the krona pegged to the euro.The high levels of business cycle synchronisation and a large share of exports to the euro area suggest the costs of adopting the euro remain small for Denmark (Holden, 2009).Pesaran, Smith, and Smith (2007) provide econometric evidence, suggesting that both Sweden and the UK would have benefited had they joined the euro in 1999.Saia (2017) estimates trade flows between the UK and its main trading partners if the UK had joined the euro.He finds that that aggregate flows between the UK and euro area members would have been as much as 13% higher and that similar results obtain for trade with non-euro area Member States.
In order to understand the extent of synchronisation between the euro area and the UK, we carry out a correlation analysis of the cyclical components (i.e.gap) in industrial production.Figure 1 shows the co-movement between the UK and euro area business cycles (Engle, 2002;Harding & Pagan, 2006).It shows, for instance, both the consequences of the 1992 exit of the British pound from the F I G U R E 1 Conditional correlation: UK and euro area cycles .
Notes: In the figure, we generate a measure of this correlation that is conditional on cyclical features.We use the exponential smoother from Engle (2002) and obtain cycles using a Kalman filter (Harvey, 1989).Given possible structural breaks, the specification for the trend-cycle decomposition is augmented with standard interventions.To detect influential residuals, we use the Harvey and Koopman (1992)  EMS and the 2007-09 run-up to the crisis.In line with existing studies, we find that there has been an overall increase in synchronisation (De Haan, Inklaar, & Jong-A-Pin, 2008).Accordingly, the average correlation coefficient between industrial production growth in the UK and euro for the full period (1991-2015) is 0.54, which is line with most of the evidence (Campos, Fidrmuc, & Korhonen, 2019), but it started from 0.37 in 1991-98, increased to 0.77 in 1999-2006, and again to 0.81 in 2007-2015 during the Great Recession (Figure 1).Our estimates show that, after the introduction of the euro, the UK and euro area business cycles became substantially more synchronised.This result has important and yet still poorly understood implications in terms of a possible exit from the EU, that is Brexit.Here, three observations are in order.One is that the net benefits from the increases in synchronicity since 1999 are not irreversible.They can be reduced by policy inconsistencies and delays, but irreversibility should not be taken for granted.
The second regards the consequences of this upsurge in synchronisation.Our results suggest a euro-out such as the UK became somehow more integrated even when not adopting the euro as its currency.All else equal, an upsurge in synchronisation leads to an increase in the net benefits of currency union membership and raises the costs of leaving the EU.
The third remark is this standard analysis has two main limitations.The first is that it only allows relative comparisons of symmetry based on individual estimates when interdependence or country groupings is the main issue of interest.The second is that synchronisation is an important (measurable) part of the explanation of symmetry adjustment within an OCA but surely not the only one.

MAASTRICHT TO BREXIT
The seminal paper by Bayoumi and Eichengreen (1993) establishes the existence of a core-periphery pattern in the run-up to the EMU.Using pre-EMU data to estimate the degree of business cycle synchronisation, Bayoumi and Eichengreen convincingly argue that there is a core (Germany, France, Belgium, the Netherlands and Denmark) where supply shocks are highly correlated and a periphery (Greece, Ireland, Italy, Portugal, Spain and the UK) where synchronisation is significantly lower.This is mostly based on the degree of supply shock synchronisation as they note that demand shock correlations are substantially lower, even for those countries in the "core."Yet, they consider, correctly, that this pattern would undermine the EMU project if persistent.
Their methodology (1993) extends the Blanchard and Quah's (1989) procedure for decomposing permanent and temporary shocks.Based on the standard Aggregate Demand-Aggregate Supply (AD-AS) model (Figure 2), supply shocks have permanent, while demand shocks have temporary effects on output.Both have permanent (but opposite) effects on prices.In other words, following a demand shock, the long-run effect on output should be zero (i.e. the economy would return to its natural rate Y), whereas a supply shock would instead permanently increase demand as expectations are revised and the short-run aggregate supply keeps shifting until Y ′′ , as illustrated below.In order to fully reflect the structure of the underlying theoretical model, we impose an additional over-identifying restriction in the VAR that captures the theoretical implication derived from the AS-AD model that supply shocks have permanent effects on output.
The main reason this has not been done before, we believe, is convenience.Without this additional restriction, the model is perfectly identified and can be estimated without major identification issues.Despite both the growth in computing power and the massive improvements in econometric techniques we have witnessed since the late 1980s and early 1990s, this convenient assumption has remained standard in the empirical literature.Yet, this has happened at the cost of reflecting more fully the underlying theoretical model.Our approach precisely addresses this gap.
Adding the fifth restriction as the cornerstone of our identification strategy allows us to put forward a test that produces a theory-consistent measure of the extent to which a country can be classified as periphery or core.Our indicator is the frequency with which such a hypothesis of symmetry is rejected.
In addition to more closely reflecting the underlying theoretical model, our approach has at least two other important advantages.One is that the indicator can be calculated yearly as opposed to an average for 10-or 20-year windows.Another advantage is that it does not depend on the adoption of a specific country as the numeraire (such as Germany).
The way we operationalise this idea is by first estimating a SVAR model that is fully consistent with Bayoumi andEichengreen (1993, 2017), in terms of lag length and identification of the demand shock.This means that the number of lags is set equal to Bayoumi andEichengreen (1993, 2017), as we want to use their results as the relevant benchmark.
Differently from the huge literature that follows Bayoumi and Eichengreen, we bootstrap the original VAR residuals in a i.i.d.fashion and generate K = 10,000 data sets.For each, we impose an additional parameter restriction on the response of output to supply-side shocks.
As the model becomes over-identified once this restriction is imposed, this generates an over-identifying restriction test where the appropriate information criterion is a likelihood-ratio test. 2eeping the bootstrap replications constant at 10,000, we record the number of rejections of the over-identifying restriction test (such that the response to a supply-side shock has a permanent effect on demand) and then calculate the percentage number of rejections (we call this measure NORD).
There are, of course, alternative identification schemes.For example, sign restrictions can also be used to identify demand and supply-side shocks.As discussed above, while we acknowledge these alternatives, we strive to keep Bayoumi and Eichengreen's approach as a benchmark for comparability of our results.In addition to trying to implement the AS-AD model more fully, another reason we adopt the over-identifying restriction approach is that inflation differentials are often considered a "normal feature of currency unions" (see, e.g., ECB, 2011).We pay particular attention to modelling the effect of permanent (supply) shocks on output, on top of the usual demand-side one.Since the proposed over-identifying restriction is sufficient to generate structural disturbances in line with AD-AS dynamics, any additional long-run restriction may be redundant in this setting.
Theoretically, testing for the "symmetry" of shocks also reflects the idea that the distinction between permanent and temporary shocks matters when it comes to adjustment within a currency union.According to De Grauwe (2016), when shocks are permanent, the slope of the usual existing trade-off between flexibility and symmetry is likely to depend on the nature of the shocks.Particularly, when permanent shocks dominate, this trade-off is likely to be steeper.Conversely, when temporary shocks dominate, the tradeoff will be flatter.We extend this idea further and suggest that it is not only the nature of the shock that matters but also the direction and extent by which shocks are pushing an economy "out of sync." 4  The test (and the indicator we calculate from it) is interpreted as that the lower (higher) the percentage of rejections, the more a country is said to be part of the centre (periphery).
In order to quantify how countries have become entrenched since the euro, or to put it differently to assess whether the EMU has strengthened or weakened the core-periphery divide, we first revisit Bayoumi and Eichengreen (1993) using the same sample and time window (25 years) to extend and replicate their results for 1989-2015, yet using the estimation approach explained above (i.e. with the additional restriction).
Our results suggest that the introduction of the euro weakened the original core-periphery pattern and even countries not using the euro as their currency have become progressively more synchronous (Figure 3). 5 These results broadly confirm the endogenous OCA hypothesis (Frankel & Rose, 1998).Note also that comparing pre-and post-EMU in Figure 3 reveals that although the range on the demand side remains the same, it has increased in terms of the supply shocks (with minimum values of −0.7 after as opposed to −0.3 before).
Our approach allows us to track how the core and the periphery changed over time.The distance between the core and the periphery could well have increased post-EMU.The periphery could have fully converged with the core, or it could have moved towards the core, or both could have moved towards each other.The asymmetry could also have decreased by core and periphery converging by large changes in demand and small changes in supply correlations or the other way around. 6Actually, 4 One concern is the possible effects from any regime changes between 1960 and 2015.Using a dummy saturation approach, first proposed by Hendry et al. (2008), Johansen and Nielsen (2009), we attempt to detect model selection problems and correct the original series for possible regime changes.In the light of the above, we generally identify three regimes for both GDP and inflation, broadly consistent across countries: 1960:I-1969:I; 1984:I-1992:I; and 2008:I-2015:I.These results are available upon request from the authors. 5Here, one could argue that non-parametric methods provide a useful alternative since they could be used to determine unknown trends, while not relying on any specific functional form.For instance, a kernel-weighted local polynomial regression of ΔY on , would display a graph of the smoothed values of GDP growth, making no assumptions about the functional form (Cerqueira, 2013;Cerqueira & Martins, 2009).However, as we are interested in the impact on demand purely from a supply-side shock, the structural identification of the shocks is a crucial step (Bayoumi & Eichengreen 1992, 1997;Campos & Macchiarelli 2016, 2018).This results in a reduced two-variable model that is represented by a moving average process. 6A high business cycle correlation is often seen as a key criterion for an optimum integration area.However, it can be argued that the elasticity with which countries react to the "common cycle" is equally important.This raises the possibility that the main problem might not be a de-synchronisation of business cycles, for instance, between core and periphery, but instead that individual countries have cycles that are tightly correlated, but of very different amplitudes, thus sometimes requiring different policies at the peak than at the trough (Belke et al 2017a).
we find that the periphery experienced a decrease in demand correlations and an increase in supply, while the core experienced a decrease in both.
For the EU12, we find that the periphery is composed of Ireland, Greece, Portugal and Spain, while the core contains the UK, Denmark, Germany, France, the Netherlands, Belgium and Italy.Our results are comparable with the results that Bayoumi and Eichengreen and others have produced for the pre-EMU period (Di Giorgio, 2016).This snapshot covers the post-EMU period but says little about such dynamics evolved over time.
Using this methodology and conditioning the first sample to 1960-85, we can generate a time-varying measure, estimated, each time, on a fixed 25-year window (see Appendix A for details).In Figure 4, higher (lower) values of the index indicate a higher probability of a country being classified as periphery (core).The results suggest that while Germany has been safely below the threshold, the UK has been moving in and out of the core (using a 50% admittedly arbitrary cut-off).This is not surprising ex-post and consistent with the (summary) of previous business cycle synchronisation analyses.
We identify three groups of countries (Figure 4): a core that becomes more homogenous over time; a periphery that changes little over time; and a mixed set of countries with interesting trajectoriesthe index for Denmark is almost constant, Greece and Sweden becomes systematically less core over time, Spain becomes systematically more core over time, and the UK is in and out of the core set of countries.
In order to understand the dynamics of this measure, we consider a set of variables suggested by OCA theory.We examine four main groups of possible explanatory variables: fiscal (debt-to-GDP ratio, cyclically adjusted budget balance), financial (corporate bond spread, 10-year government bond spread, 3-month interbank interest rate spread, interest on the average on consumer loan spread, return For each country, a bivariate SVAR is estimated using (log) real GDP and the (log) deflator, both in first differences.The structural identification of the shocks also follows Bayoumi and Eichengreen (1993) and controls for changes in regimes.Red arrows denote movements of the so-called "core" countries, and blue arrows, movements of the "periphery.on equity differential, a set that is consistent with the European Central Bank's definition of financial integration, ECB, 2011), external (FDI and real effective real exchange rate), structural reforms (employment protection legislation, EPL, and product market regulation, PMR) and a dummy variable on euro area membership.The estimation includes, besides the UK, Austria, Belgium, Germany, Denmark, Spain, Finland, France, Greece, Ireland, Italy, the Netherlands, Portugal, as well as EU non-euro area countries such as Sweden and non-EU countries such as Switzerland and Norway.We present the results for the period 1991-2015, but these are robust when we stop our estimation before the financial crisis in 2007.
In order to address possible endogeneity problems, we use a GMM approach where the weighting matrix is set to equal the 2SLS.The results are qualitatively the same if instead we use a Huber/Eicker/ White heteroscedasticity-robust variance-covariance matrix (these are available upon request from the authors).In the estimation, the number of endogenous variables equals the number of instruments, where the instruments are selected to be the lagged dependent variables, plus the constant and the Eurozone (EZ) membership dummy.
The overall results (last column in Table 1) suggest that a strong role is played by the strictness of product market regulation whereby a high PMR increases the likelihood of country being in the periphery.This is in turn not surprising given that the index is based on supply-side dynamics and the extent to which those prompt similar GDP reactions among Member States.A second factor is the level of debt-to-GDP, again, in reducing the likelihood of a country being in the core, albeit the statistical evidence is not strong.
Membership to the currency union, for the countries in our sample, suggests an important role in making countries less "peripheral", impacting the probability of being classified as periphery by as much as 16 percentage points (Figure 4).
These findings are in line with the idea that one of the main concerns for monetary union membership would be represented by the costs of adjustment in order to deal with asymmetries.In the absence of sufficient labour flexibility, and equally of fiscal transfers at the euro area level, many countries would suffer from severe adjustment problems.Note: '*,**,***' denote significance at the 1 5 and 10 percent, respectively.In the table, our measure of persistence (NORD) is regressed on a set of possible explanatory variables: where X 1it = fiscal (debt-to-GDP ratio, cyclically adjusted budget balance), X 2it = financial (corporate bond spread, 10-year government bond spread, 3-month interbank interest rate spread, interest on the average on consumer loan spread, return on equity differential, a set that is consistent with the European Central Bank's definition of financial integration, ECB, 2011), (c) X 3it = external (FDI and real effective real exchange rate), X 4it = structural reforms (employment protection legislation, EPL, and product market regulation, PMR).In all columns, a constant term and a dummy variable on euro area membership are included.In the GMM estimation, the number of endogenous variables equals the number of instruments, where the instruments are selected to be the lagged dependent variables, plus the constant and the Eurozone (EZ) membership dummy.
Even before the launch of the EMU, the concern about entrenched asymmetries spurred an alternative approach to European integration: the possibility of a two-tier or "multi-speed Europe."From an economic viewpoint, it is true that smaller groups of countries may be better candidates for forming an OCA given that they may be more homogenous (see also De Grauwe, 2016).Looking at the early evidence on the degree of synchronisation of shocks across countries before the EMU (1963-88), compared to the same pattern 25 years after the EMU, however, suggest that a new, smaller, periphery has emerged (Spain, Portugal, Ireland and Greece).Thus, the EMU has weakened the core-periphery pattern, resulting in countries being more integrated over time.
The UK, with its mixed experience shown above, represents a much lesser threat to euro area stability than the absence of concerted and forceful action from euro area Member States themselves.This is particularly true if Brexit occurs in an orderly and gradual manner.In this respect, we argue that while the hypothesis of a "multi-tier" Europe cannot be dismissed on the basis of the available evidence, our results support the view that a viable alternative to a "multi-speed" scenario is a serious process of coordinated reform.This is indeed the spirit of the Five Presidents' Report (Junker, Tusk, Dijsselbloem, Draghi, & Schulz, 2015).
As mentioned above, there has been considerable thinking and planning on how to make the EMU more effective, that is how to ensure the stability and integrity of the EMU.The first clear attempt at addressing this matter was the so-called Four Presidents' Report (those of the European Council, European Commission, European Central Bank and Eurogroup) that in 2012 put forward as an explicit goal the need to move towards a Genuine Economic and Monetary Union (Macchiarelli, 2016).The choice of words (i.e.Genuine) is indicative of the extent of the consensus about the need for EMU reform.The Five Presidents' 2015 report provides a roadmap for further deepening of the EMU in order to ensure the stability and smooth functioning of the EMU.It stipulates a detailed range of actions and a clear timetable (in three phases) to bring progress in four main areas, namely economic, financial, fiscal and political union.
Our analysis documents that the introduction of the single currency preceded a substantial increase in symmetry among Member States, thus improving an important dimension in most considerations about the stability of the euro area.The main policy implications we derive hence complement those put forward by the Five Presidents' Report.This Report indicates what is to be done and when, while our analysis suggests countries that should receive special attention in order for these policy actions to be more effective.
Our results suggest Sweden is a crucial country in order to fulfil the goal of increasing the stability of the euro area.After the UK, the trajectory of Sweden's index since 1990 is worrisome.It indicates that this is one of the few countries that continue to leave (or it continues to increase its distance from) the core and has done so in a systematic and sustained way.No other country exhibits such a trajectory.Moreover, Sweden is an important trade partner to the Baltics, which also show surprisingly (despite their euro membership and relatively low levels of product market regulation) to have large distances from the core.Third and finally, without the UK (post-Brexit), Sweden will become the country closer to the border with the periphery.For all these reasons, the EU should focus on Sweden to foster the stability of the euro.
The Swedish Statistical Office monitors public opinion towards the single currency, and the levels of rejection have been above 70% in recent years.Yet, there is a clear economic explanation for this.Campos, Coricelli, and Moretti (2016) argue that Sweden benefited relatively little from EU membership after it joined in 1995 (in large part because it was already a high-income country with highly developed institutions) and benefited substantially from avoiding joining the euro (partly because its | 1803 CAMPOS And MACCHIARELLI largest trading partners are not Eurozone members).Indeed, the evidence suggests that while around year 2000 the benefits from not joining were relatively difficult to estimate (or close to zero), a decade later these have become substantial and significant (Gyoerk, 2017).
In terms of the actual UK withdrawal from the EU, Brexit will certainly challenge both internal and external equilibria, with some EU non-euro area Member States such as Poland, Denmark and indeed Sweden, but also the "pre-ins", feeling they will lose grip in shaping Eurozone policies (Oliver, 2016), especially against an enhanced role of Germany and the other euro area Member States.This may trigger further scepticism, should the EMU fail to provide an attractive alternative model for integration.We argue that deeper integration should carry on to the point of making euro-outs be eager to join, as anticipated in phase 3 of the Five Presidents' Report.

| CONCLUSIONS
What is the impact of Brexit on the stability of the euro area?This paper argues that if one is concerned about stability and cohesion, asymmetry and imbalances, one way of thinking about these issues is offered by the notion of the probability of a country being classified as periphery in a core and periphery framework.
Before Maastricht, the seminal contribution of Bayoumi and Eichengreen (1993) generated a clear picture.Looking at correlations between demand and supply shocks, one could see two distinct groups of countries: a core and a periphery.It is a seminal paper because, inter alia, it is one of the first to point out the risks of an entrenched core-periphery to the then-nascent EMU.Their influential diagnostics was based on data covering 25 years from 1963 to 1988.Using the same methodology, sample and time window, we replicate and extend their results for 1989-2015.We ask whether the EMU strengthened or weakened the core-periphery pattern.Our results suggest the EMU has weakened the original pattern; that is, the number of countries in the periphery (core) decreased (increased).How did these groups (core and periphery) change over time?We find the UK belongs to a mixed set of countries.While Denmark's index is basically flat, that is it changes little over time, Greece and Sweden become systematically less integrated over time, while Spain shows the opposite pattern.Unsurprisingly perhaps, the UK goes in and out of the core.
Finally, we ask the question of what drives symmetry (and thus stability).Our estimates show that euro membership and Product Market Regulation are key.We find the probability of a country of being classified as core is driven chiefly by euro membership and product market regulation.This finding provides renewed and direct support for the endogenous OCA hypothesis and its interpretation in a broader sense.
In terms of future research, our analysis identifies at least four main fruitful areas.One is incorporating more EU member countries so as to provide a fuller picture of the behaviour of the core and of the periphery over time.This should be also useful to assess the more recent experience of accession to the euro area and inform the policies of various EU Member States that are considering joining.The second is that future research would benefit from a regional-level perspective, complementing the country-level one.Third, this paper and the majority of the scholarship that precedes it focus on the real side, not on the monetary.Future research would do well to bring in the financial sector and study how it interacts with the real sector in this framework and what sort of policy implications can be derived from such an analysis.Fourth, and specifically related to future research on the stability of the euro and Brexit, the features of the trade agreement should come into play in the future when conducting similar research stressing trade and financial integration in addition to business cycle correlations.

TECHNICAL ASPECTS OF THE SYMMETRY MEASURE
In what follows, we summarise the methodology used to construct our measure of summery.The underlying methodology is that of Bayoumi and Eichengreen (1993), which is an extension of the Blanchard and Quah (1989) procedure for decomposing permanent and temporary shocks.
Let us consider a system where the true model is represented by an infinite moving average of a (vector) of variables, X t , and shocks, t .Using the lag operator L, a bivariate VAR featuring real GDP and its deflator can be written as an infinite moving average representation of demand and supply disturbances: where X t = Δy t ,Δp t and the matrices A represent the impulse response functions of the shocks to the elements of X.It follows that: where y t and p t represent the logarithm of output and prices, and t are i.i.d.disturbances, which identify supply and demand shocks (Ramey 2016, forthcoming).For the ith country, a 11i represents element a 11 , in matrix A i and so on.This framework implies that supply shocks have permanent effects on output, while demand shocks have temporary effects.Both have permanent (opposite) effects on prices.The cumulative effect of demand shocks on the change in output must be zero: The system can be estimated using a VAR.Each element can be regressed on lagged values of all the elements of X.Using B to represent these estimated coefficients: where e t represents the residuals from the VAR equations.Using the standard relation between the VAR's residuals (e t ) and structural disturbances-that is demand and supply shocks-that is e t = C t , it is clear that, for each country, the exact identification of the C matrix requires four restrictions.Two are normalisations, which define the variance of the shocks dt and st .The third restriction is from assuming that demand and supply shocks are orthogonal to each other.The fourth is that demand shocks have only temporary effects on output.
The standard AD-AS model implies that demand shocks should raise prices in both the short and long run, while supply shocks should lower prices and increase demand permanently.In order to achieve that, it suffices to impose the additional over-identifying restriction in the VAR that supply shocks have permanent effects on output.We need to impose this restriction in our sample for the demand and supply shocks to be theory-consistent.This differs from Bayoumi and Eichengreen (1993) because they do not impose this last restriction, which leaves the model exactly identified.One reason we adopt the proposed over-identifying restriction is that inflation differentials are often considered a "normal feature of currency unions."Therefore, we pay particular attention to modelling the effect of shocks on demand.The role of co-movements in output's cyclical fluctuations is further in line with the business cycle literature.Since the proposed over-identifying restriction is sufficient to get structural disturbances in line with AD-AS dynamics, any additional long-run restriction may be redundant in this setting.
We test for the above over-identifying restriction, by imposing ∞ ∑ i=0 a 12i = , where  > 0. Under the latter assumption, demand across each country is restricted to respond qualitatively (sign) and quantitatively (size) in the same way to supply shocks.In terms of the structural VAR analysis, this implies: .The value we chose to report, consistent with Campos and Macchiarelli (2016), is = 1.
In order to construct a test for the over-identifying restriction described above, we estimate the SVAR model consistent with Bayoumi and Eichengreen (1993).Differently from the latter, we bootstrap the original VAR residuals in a i.i.d.fashion and generate K = 10,000 data sets.For each of the kth samples, we proceed with a structural analysis and test for the over-identifying restriction based on a LR test.We record the number of rejections of the over-identifying restriction test at each bootstrap replication, and calculate: where L u and L r are the maximised values of the (Gaussian) log-likelihood function of the unrestricted and restricted regressions, respectively.Under H 0 , the lr statistic has an asymptotic distribution with de- grees of freedom equal to the number of long-run restrictions (q) minus n 2 − n ∕2, where n is the var dimension (in this case, n = 2).
The dynamic version of the index is obtained by letting T be larger than before where denote the width of a subsample or window and define the rolling sample "metrics."Here, we define: The windows are rolled through the sample one observation at a time, so there the procedure returns T − + 1 rolling estimates of the NORD (Campos & Macchiarelli, 2016, 2018).
The basic intuition for our NORD measure is that it reflects the percentage of times we observe the rejection of the key restrictions needed to estimate the Aggregate Demand-Aggregate Supply model.The higher the percentage of rejections (or the more often they happen), the higher is the value of NORD.As such, NORD values range between 0 (perfect, the probability of a country of being classified as periphery content) and 100 (i.e. the probability of a country of being classified as periphery content implying a perfect periphery).
two-step auxiliary regression procedure.In the first step, the focus is on outliers and break detection.The second step involves estimating the model with those interventions, which were found significant in the first step.Source: Authors calculations.Data on industrial production from Datastream [Colour figure can be viewed at wileyonlinelibrary.com] 3

F
I G U R E 2 Aggregate Demand and Aggregate Supply Framework [Colour figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 3The dynamics of the correlation of supply and demand disturbances between pre-EMU and post-EMU.Notes: The figure compares estimates from pre-Maastricht based onBayoumi and Eichengreen (1993), covering the period 1963-88, withCampos and Macchiarelli (2016) equivalent estimates for the period 1991-2015 (post-EMU).

F
I G U R E 4 Measuring the probability of a country being classified as periphery over time.Notes: The figure compares our measure of persistence (NORD) across countries from a SVAR estimated using (log) real GDP and the (log) deflator in first difference, as explained in the Annex.The first sample is 1960-85.The measure we obtain is estimated, each time, on a fixed 25-year window and 2 lags.Source: Authors' calculations [Colour figure can be viewed at wileyonlinelibrary.com]Determinants of the probability of a country being classified as periphery: GMMPanel estimates How to cite this article: Campos NF, Macchiarelli C. The United Kingdom and the stability of the Euro area: From Maastricht to Brexit.World Econ.2020;43:1792-1808.https ://doi.org/10.1111/twec.12919APPENDIX A 2 +…+B n X t−n + e t = (I − B (L)) −1 e t = I + B (L) + B (L) 2 +… e t = e t + D 1 e t−1 + D 2 e t−2 + D 3 e t−3 , restrict a priori; instead, we vary in the interval [0.1, 2]