Should countries specialize or diversify their economies

Diversification and specialization of trade in the European Monetary Union

Table of Contents

1 Introduction

2 trade intensity

3 Intra-Industry vs. Inter-Industry Trade

4 conclusion

5 Appendix

1 Introduction

Within the framework of the endogeneity discussion of the theory of optimal currency areas, this thesis examines the effect of the introduction of a single currency in Europe on trade and the trade structure of the participating countries.

Since the plans to create a single currency for Europe with the euro became concrete, various studies and investigations into the effect of the single currency on trade have been carried out with varying results. Two standpoints have emerged which have become established in the literature as the “Krugman View” and the “European Commission View”. On the one hand, the European Commission (European Commission (1990)) and others (Ricci (2006), Frankel and Rose (1998)) assumed increasing trade intensity and synchronization of trade structures. In contrast, Krugman (1993) and others (Kenen (1969), Eichengreen (1992)) proclaimed that although trade will increase, countries will increasingly specialize.

While there were opposing opinions on how trade would develop when the euro was introduced, adequate empirical values ​​were in short supply. In these dimensions and with such highly developed participating states, the European single currency is a unique experiment to this day.

The problem with ex-ante estimates of trade development, as they are only possible at the time a currency area emerges, is the so-called endogeneity problem, which goes back to the Lucas criticism (Lucas (1976)). Policies such as the introduction of a common currency change the model and become part of the model. There are different approaches how to deal with this problem. One can try to take it into account (Frankel and Rose (1998)), argue theoretically (e.g. Kenen (1969)) or investigate empirically afterwards (e.g. Ricci (2006)).

Even 14 years after the merger, it is still not clear which point of view is correct. In a crisis that is the first serious test for the single currency and calls for the old national currencies are getting louder, an empirical evaluation of how the “experiment” has affected trade is important.

First of all, it must be mentioned that although trade is an important criterion within the framework of the optimal currency area theory, it is not the only one. It goes without saying that for a complete analysis it is important to consider more than just the trade and the trade structure. This is left to other work.

If one looks at the effect of trade, the general assumption and also a prerequisite for a positive effect of a common currency is that trade between the states increases and that the states benefit from this increasing trade. At least the second point appears relatively undisputed. More trade leads to more product diversity on the one hand, to cheaper goods and more competition on the other. There seems to be a consensus that increasing trade should be seen as positive for the countries involved.

In the second chapter of this thesis it is first examined whether the trade intensity between the member states has increased - from which both network hypotheses are based. A single currency reduces currency risks, transaction costs and increases the transparency and comparability of prices. Nevertheless, it is not clear whether the introduction of the euro had a concrete effect on trade. The arguments are relatively trivial when one considers that those mainly concerned are economically knowledgeable.

The third chapter analyzes trade in detail and thus the second, controversial part of the two competing hypotheses. The basic structure of the work by Blanes Cristobal (2006) serves as inspiration. The European Commission underpinned its thesis with increasing economies of scale and imperfect competition. So there should be an increase in trade through more intra-branch trade. Product differentiation and fewer barriers to trade should lead to more diversified trade and production structures. The countries would adjust their production and trade structures to one another and react increasingly similarly to shocks. This would reduce the costs associated with the introduction of the euro.

With the elimination of individual monetary policy, one instrument for balancing out imbalances is less available. Individual states are no longer able to devalue or revalue their currency in order to become more competitive, to finance their state expenditures through profit of coins or to reduce their national debt through high monetary devaluation.

Figure not included in this excerpt

Figure 1.1: Source: De Grauwe (1997, p.82f.), Slightly modified

Figure 1.1 shows the conflict between the two different views. The “OCA curve” shows the combinations of divergence and trade intensity which the European Monetary Union (EMU) is indifferent to introducing and not introducing ex ante between currencies. On the left of the curve the costs are higher than the income, on the right of the curve is advantageous.

The “EC curve” is falling, as the divergence decreases with increasing trade intensity and thus the costs associated with a common currency decrease. This rather optimistic view corresponds to a self-fulfilling fact. This would then also speak in favor of the generous acceptance of further states into the monetary union.

The curves "K1" and "K2" correspond to Krugmans' point of view. Both rise. As the intensity of trade increases, so does the divergence between states. Krugman explains this with regional specialization and concentration of industries. According to him, cluster formation increases productivity, innovation and costs decrease. Thus, in his opinion, the synchronicity is reduced and the currency area would only be favorable with a higher increase in trading intensity. In this case, the member states should be selected more carefully and according to stricter criteria.

A conclusion is drawn in the fourth chapter. The appendix and the bibliography can be found from the fifth chapter.

2 trade intensity

The obvious arguments in favor of increasing trade integration among the euro countries are those already mentioned: lower currency risks, transaction costs and greater transparency. This, as well as the growing integration of the European financial markets, should have been an incentive, especially for small and medium-sized enterprises, to be more active in the European single market. Companies in small countries in particular, which naturally have a higher export share in production, should also have benefited from the fixed conversion rate of currencies and the resulting planning security.

Furthermore, the location theory predicts that the integration of the economic area will make a potential choice of location within the euro zone more attractive. For example, more foreign companies should have invested and built up production in the euro area, and the outsourcing of production by local companies to the so-called low-wage countries should have been reduced.

However, it is right to question whether the euro was actually the cause of any structural changes or whether these can be traced back to increasing integration within the EU. Certainly the southern countries in particular received a great gain in planning security and credibility through a price stability-oriented monetary policy of the European Central Bank.

Data and methodology, discussion

In order to examine the effect of the European Monetary Union on trade between the member countries, a gravitational model of international trade is set up below. In the context of a gravitational model, Tinbergen (1962) checked in his simplest model only the economic size of the two acting states, as well as the distance and a constant factor. More recent regressions (e.g. Glick and Rose (2002)) contain significantly more factors. In addition to the gross domestic product and the distance, the population, language, common border, free trade agreement, geographical features, colonial past, country sizes, common currency area, and much more are also included. a. controlled. It is an extensive undertaking to record and / or quantify all influencing factors on trade (especially cultural and political aspects) and thus correctly map all special features in bilateral relationships in order to isolate the influence of the single currency.

This work makes use of the fact that most of the peculiarities do not change over time. The interesting factor - membership in the European Monetary Union - is not constant over time, but has changed for EMU-111 1999, with its establishment, as well as for the new states2 later changed. In this way you can carry out a “fixed effects” regression and neglect all constant effects.

The dependent variable was created from bilateral trade data (import and export) from 34 countries (all OECD countries, as of March 2013) [OECD (2013a)] (see below). The trade flows are quarterly and range from 1960 to 2012. The data availability varies considerably. For the states of the former Soviet Union in particular, there are no reliable data prior to 1992. This is how the regression was for the full data set3, for the countries with good4 and for the countries with the best5 Data availability carried out. The bilateral trade flows used correspond to a share of at least 50% to 95% of the total trade (2000) of the countries examined, whereby a large part of the total trade is recorded for most countries (see Table 5.3).

The empirical regression model looks like this:

Figure not included in this excerpt

trade stands for the sum of bilateral exports and imports of land i and country j at the time t. gdp as pop denote the gross domestic products and the population of the countries i or. j at the time t. Often the per capita gross domestic product is used instead of the population, the result is the same.

With EMU, EU and FTA is for a common membership in the European Monetary Union (EMU), The European Union (EU) and a free trade agreement (FTA) - bilateral or multilateral - controlled. The value of this dummy variable assumes the value 1 if both countries are at the point in time t Are members, otherwise zero. A positive effect on trade is expected here in each case.

"Time" contains a battery of time dummy variables6 and controls for particularities in the respective quarter, as well as seasonal particularities, "Trend" controlled for a time trend7such as increase in trade due to increasing globalization, falling transport costs, etc.

τ is the constant and indicates the average value of the fixed effect in a fixed effects regression.

In Table 2.1, the model is also expanded to include an interaction term between the EU or EMU and the trend Variables expanded. This interaction controls for additional trend effects from the merger. In Table 5.2 the interaction terms EMU * Country i8 expanded. This shows the effect of the European single currency Country i individually compared to the other countries. In addition, a dummy variable crisis for the crisis years (1982, 1983, 2001, 2008-2011) included.


Figures 5.1 and 5.2 show the development of trade in EMU-11 + Greece. The values ​​are normalized to 1 for the year of entry (EMU-11: 1999, Greece: 2001) into the Union. It should be noted that trade has expanded significantly in all countries since the introduction of the currency and despite the global economic crisis. We're talking about doubling and tripling the volume of trade in 12 years. In most countries, total trade increased faster than trade within the monetary union. The exceptions are Ireland, Greece and France. Here trade in the member states has increased more than in the rest of the world. The question now needs to be clarified how trading would have developed without the introduction of the euro.

The results of regression 2.1 are shown in Table 2.1. The first 3 specifications contain the full data set, (4) and (5) only those with good data availability and in (6) and (7) only those with the best data availability.

The addition of time and trend variables improves the estimate and was therefore always included from (2) onwards. Are expected β 1 and β 4 consistently significantly positive. The control variable population is inconsistent, sometimes positive, sometimes negative, never significant. This is also due to the fixed effects methodology, since here it is not the absolute size of the economies but only the deviations from the average that are considered. Since the population changes (weakly) over time, the variable remains in the estimate. β 5 is still consistently significantly positive for the larger data sets, it becomes insignificant for (6) and (7). The proportion of country pairs that have a free trade agreement is between 28% and 30% in all data sets, ie not significantly different. The good data availability in the EU compared to some other OECD countries means that 17.5% of the country pairs in the full data set are still in the EU together, while the value rises to 18.5% and 21.8% with increasing data quality . It results

Figure not included in this excerpt

Table 2.1: own presentation, data: OECD (2013a)

So there is an increasing collinearity with the variable FTAwhich can explain the loss of significance.

In specification (2) the coefficient we are interested in is (EMU) still insignificant (t value: 1.45), while the others produce significantly positive values.

Checked for the above-mentioned factors, as well as all factors constant over time, we thus obtain a positive correlation between common membership in the European Monetary Union and trade between the states. Specifically, it can be said that the introduction of the euro increased trade within the member states by 4.9% (insignificantly) or 8.4% - 17.5%. The expected, or hoped for, both Krugman and the European Commission and thus the first part of the two joint hypotheses is fulfilled. The interaction terms EM U t / EU t show a tendency towards negative development in the course of the existence of the unions. In part significant, the values ​​are an inconsistent indication that there was a kind of “euro boom” in the first few years and that there has been increasing disillusionment in the last few years. Furthermore, the coefficients may be distorted by the last few years (global economic crisis from [2007]).

Table [5] [2] shows the specific effects of the introduction of the euro on state trade. Compared to the other euro countries, there are significant negative effects on trade in Greece, Finland and the Netherlands, while trade in Germany, Spain and Italy benefited more.

Up to here the EMU Dummy with the year 1999 or 2001 the time of the merger of the euro zone. However, with the signing of the Maastricht Treaty (1993), the introduction of the euro was decided by 01.01.1999 at the latest. It was true that the states had to meet the convergence criteria, but as early as 1993 it seemed a foregone conclusion that practically all states would also introduce the euro (see Vaubel (1993)). Expectations have been formed and the structures have already been adapted to the future situation. Here, too, however, the question remains whether the effect will actually be on the agreed introduction of the euro, on z. B. the EU internal market reform (1992) or other integration processes are falling behind.

Table 5.1 shows the results of this specification. You can see convincingly positive values ​​for the newly created one EMU (93) Variable. Here, too, there are negative time effects that suggest a strong anticipation in the first few years.

3 Intra-Industry vs. Inter-Industry Trade

The following section deals with the second part of the two hypotheses tested. Did the introduction of the euro lead to diversification or specialization in trading? By introducing a single currency, the participants give up an economic policy instrument, namely individual monetary policy. From now on, this instrument will be used centrally by the common central bank for all countries and can no longer be used adequately in the event of asymmetrical shocks.The European Commission (1990) wrote in “One Money, One Market”:

”Inthecontext of EMU, thequestionthereforeariseswhetherasymmetrieswill have a tendency to decrease or to increase relative to the present situation. The cost of losing the nominal exchange rate (...) will be diminished if asymmetries tend to disappear. ”(European Commission ([1990], p. [140])

Asymmetries mean differences in trade and production structures in the individual countries. Decreasing asymmetries reduce costs. Increasing symmetry in trade and production structures means more intra-industrial trade. Intra-industrial trade describes the trade between two countries with the same or similar goods, while the counterpart, inter-industrial trade, describes the trade in different goods. If every country has the same industries in similar proportions, they are hit by shocks in a similar way and a uniform central bank policy can be optimized for all countries. In its analysis, the European Commission comes to the following conclusion:

"TheresultsfromthissectionindicatethatasymmetricshocksintheCommunity, even though they exist, are likely to diminish with the disappearance of trade barriers through the completion of internal market." (European Commission ([1990], p. [147])

So you assume that the asymmetries in these structures disappear or decrease.

Krugman (1993) argued in the other direction in his book "Geography and Trade". He substantiated his thesis of increasing divergence with the empirical finding that specialization in the European states (at this point in time without a single currency) is less pronounced than in the examined regions in the United States. He measured this on the basis of the workers employed in the various industries. In conclusion, he writes:

"(...) iftheevidencepresentedinthelastfewtablesisright, Europeannationsare likely to become less similar, not more (...) and they will in this respect become less suitable as an optimum currency area as a result of increased integration." (Krugman ([1993]) , p. [83])

In his opinion, primarily inter-industrial trade is increasing. Intra-industry trade is falling as a result of increasing integration.


1 Belgium, Germany, Finland, France, Ireland, Italy, Luxembourg, Netherlands, Austria, Portugal, Spain

2 Greece (2001), Slovenia (2007), Slovakia (2009) and Estonia (2011); Malta and Cyprus (2008) are not included

3 all 34 OECD countries

4 OECD countries except Chile, Estonia, Slovenia, Slovakia and Israel

5 EMU-11 excluding Belgium and Luxembourg, Australia, Canada, Greece, Japan, Norway, Sweden, Switzerland, Turkey, Great Britain, USA

6 D1960Q1, D1960Q2, ..., D2011Q4

7 trend = 1,2, ...; trend 2 ; trend 3

8 Dummy variable. 1 if country = Country i, 0 otherwise