Department of Economics
Has the eurozone moved closer to an optimum currency area?
Abdoul Amadou
10805451
Supervisor:
C. Sahin
1
Statement of Originality
This document is written by Student Abdoul Amadou who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
2
Abstract
In 1999 the was introduced and the eurozone was created. For all it’s in varying degrees beneficial to be part the eurozone. The optimum currency area (OCA) theory tries to identify characteristics countries should have to be regarded as an OCA and if is an good decision to enter a monetary union. what appears is that the characteristics of the OCA theory. With the knowledge about the
endogeneity of the OCA, this paper researcher if the euro has made the original 12 eurozone
countries (EZ12) more likely to be an OCA, and if the euro has increased the convergence towards an OCA.
These question are answered looking at the development of four OCA criteria and analyzing there linear time trend after and prior to the introduction of the euro. The four criteria are: business cycle synchronization, intra-industry trade, degree of openness and export diversification.
The main result are that in terms intra-industry trade and export diversification the EZ12 has come closer to an OCA. Only in export diversification the result show that convergence to an has increased. In terms business cycle synchronization and degree of openness the EZ12 has moved further away from an OCA.
3
Table of contents
1 Introduction
4
2 Theory and literature review
5
3 Methodology
8
4 Data
9
5 Results and analysis
11
6 Conclusion
17
4
1
Introduction
On 1 January 1999 the third stage of The Economic and Monetary Union (EMU) began and the countries involved introduced the euro as its official currency. The euro then replaced the European Currency Unit (ECU) on parity. The ECU was a basket of European currencies, which was used as an anchor for the European currencies to fluctuate around with a certain margin (Mongelli, 2008, pp. 11-12). Three years after the introduction of euro in 2002 the physical notes and coins were introduced and the euro became the legal tender in 12 European countries. These 12 countries formed a monetary union, the eurozone. Since then 7 more countries adopted the euro, Lithuania which adopted the entered the eurozone in 2015 is the most recent case.
The decision to enter or to let one enter a monetary union is not one that is easily made. Past decisions to form a monetary union have been made, based more on political than economic reasons (Mongelli, 2008, p. 1). Abandoning one’s own currency and losing monetary independence come with costs. On the other hand, being involved in a monetary union brings many benefits in terms of bilateral trade. The optimum currency area (OCA) theory deals with the macroeconomic properties a group of countries or a region should have for it to be optimal to form a monetary union.
Several authors have found evidence that the OCA theory is endogenous, forming a
monetary union affects the likelihood of that a region would qualify for a OCA. The endogeneity can either increase or decrease the likelihood. Perhaps according to some OCA criteria some countries weren’t fit to be part of the eurozone. And maybe these countries score better on some criteria after the adoption of the euro.
This paper tries to find if the original 12 countries of the eurozone have come closer to an OCA after the introduction of the euro. In the remainder of this paper these twelve countries will be referred to as EZ12. The development of four OCA criteria will be analyzed, namely: business cycle synchronization, intra-industry trade, degree of openness and export diversification.
This paper is setup in the following way. In section 2 there is a literature review and a elaboration on the OCA theory and its endogeneity. In section 3 the used methodology is explained, followed by the data collection in section 4. The analysis is in section 5 and the conclusion will be in section 6.
5
2
Theory and literature review
2.1 theory
When a country chooses to participate in a currency union and adopt a common currency with a common central bank, it loses a potential stabilization tool in economic turbulent times. The absence of this tool does the most harm when a country experiences asymmetric shocks. The common central bank then has to intervene in way that is sub-optimal and maybe even detrimental for the countries in the union (Mundell, 1961, p. 658).
The OCA theory states characteristics which I will categorize in roughly three groups. No characteristic solely belongs to one category, one can easily argue that a characteristics should also belong to another category. The first category are characteristics countries should have to cope with asymmetric shocks without the intervention of a central bank. Price and wage flexibility and factor mobility are such characteristics. When a country faces an asymmetric shock which decreases economic growth and decreases inflation, these characteristics should help to restore the equilibrium.
Secondly the OCA theory states characteristics which should prevent asymmetric shocks to happen. Criteria in these category are: similarity of inflation rates, synchronized business cycles and diversification in export, production and consumption (Kenen, 1969). Lastly there are characteristics which must make countries able to reap the most benefits of common currencies. Most of these criteria are trade related, such as trade intensity and degree of openness.
It is important to note that not all characteristics have to present. For instance if prices and wages could be lowered, this country could come back to equilibrium without much factor
movement to happen. Prices and especially wages have proven to be downward rigid, in this case factor mobility could work as stabilization mechanism by moving factors of production from low to high inflation countries.
2.2 Literature review
It were Mundell (1961), McKinnon (1963) and Kenen (1969) who primarily contributed to the OCA theory. In their work they discussed characteristics which a group of countries should have, so that it’s optimal to form a currency union. An OCA is a region, either a group of countries or regions within an country, for which it’s optimal for the participating currencies to irrevocably fix exchange rates or to adopt a common currency. Their early contributions were mainly theoretical and not so much empirical
Business cycle correlation, trade intensity, degree of economic openness, diversification in production, fiscal federalism, price and wage flexibility and factor mobility, these are some of the
6 criteria of the OCA theory (Mongelli, 2008, pp. 2-3). Countries must score high on these criteria for them to be regarded as an OCA. But some of the criteria are actually interrelated, namely trade intensity and business cycle correlation. Depending on the composition of bilateral trade, the correlation of business cycles can either be positively or negatively affected. Moreover, being part of currency area increases trade intensity. Because irrevocably fixed exchange rates or a common currency, reduces or eliminates the exchange rate risk and increases price transparency. This takes away a burden for consumers and firms to consume or invest abroad. This is called the endogeneity of the OCA criteria, because of the endogeneity it’s possible that a currency area is ex post more likely to qualify as an OCA than ex ante (De Grauwe, & Mongelli, 2005).
Firdmuc (2004) tested the hypothesis of endogeneity of the OCA criteria. He found a weak relation between trade intensity and business cycle correlation. His result did show a stronger positive relation between intra-industry trade and business cycle correlation. These findings support the presence of endogeneity in the OCA criteria.
The way he performed his analysis is by regressing the correlation of the business cycle of on the trade intensity and the intra industry trade, measured by the Grubell-Lloyd index. This regression is done with both regressors together and with and separately. Knowing that bilateral trade levels are already being influenced by a common, exogeneous variables were used as instrument to determine the trade intensity. These instrumental variables were provided by the Gravity model of trade. First when regressing business cycle correlation on trade intensity, he found positive slope coefficients close zero, but yet significant different from zero. When using both regressors together, the coefficient for trade intensity doesn’t significantly differ from zero and the coefficient for intra-industry trade was significant different from zero and bigger than the coefficient when trade
intensity is the only regressor. And when using intra-industry trade as a sole regressor, the coefficient becomes more significant, while still maintaining its size.
The effect of increasing trade relation on business cycle correlation is ambiguous. Namely, if most trade is inter-industry, then increasing trade relations will lead to more specialization and less correlated business cycle. When most of the trade is intra-industry, then increased trade relations will increase business cycle correlation (Frankel and Rose, 1998, p. 1014).
To prove the endogeneity of the OCA criteria Frankel and Rose (1998) regressed the business cycle correlation on trade intensity. As a proxy for trade intensity they used total trade between country i and j divided by the total trade of both countries. As a second proxy they used total trade between country two countries and it divided by the GDP of the two countries. They find a strong positive relation between trade intensity and correlation of business cycles.
Krugman (2001) uses theory and the case of the United States to argue that increased economic integration may lead to desynchronization of business cycles because of specialization.
7 Krugman argues that economic integration leads to overall lower transaction costs for firms and consumers. This allows firms to expand their market and compete in other regions and benefit from economies of scale. Where previously small firms could survive, they now have to compete against more firms, and the firms with the greatest economies of scale will remain. This results in the geographical clustering of an industry and divergence of industrial structure of regions. Moreover this can be magnified by increased capital mobility, because it becomes easier for capital to flow to the places with highest return, causing structural damage to regions with low economic growth.
Fatas (1997) investigated how the start of the EMS 1979 has changed the synchronization of business cycles is the EU. He does this by computing the correlations of employment growth for every country with the EU12. The EU12 are the twelve members of the European union in 1992. He also computes the average correlations of regions within countries with the EU12 and with the country the region is in. The sample runs from 1966 till 1992, with 1979 as breakpoint. The findings were that between countries business cycles synchronization has increased after the EMS. What also is found is that the synchronization of regions within a country with the country itself has decreased and that the synchronization with the EU12 has increased. These result show that specialization has not increased for the EU12 but it has in in the regions within a country.
8
3
Methodology
The sample period is 1979 to 2016. 1999, the start of the euro is the breakpoint and cuts the sample in two. The period 1979-1998 will occasionally be referred to as pre-euro and the 1999-2016 as post-euro. 1979 is chosen as starting because it was the year in which the European Monetary System (EMS) was established, the EMS’s main objective was to prevent large exchange rate fluctuations between the members countries.
The business cycle analyses will be done in 2 ways for both measures. The first one is a comparison of the pre-euro correlation coefficient and the post-euro coefficient. Firstly the employment and GDP figures will be detrended using the Hodrick–Prescott filter. Leaving us with only the cyclical component of the two measures. For the period pre-euro the correlation of the business cycle between country I and the sum of the countries will be calculated, the same will be done for the post-euro period. To compare the correlation coefficient of the two periods, the Fisher transformation first had to be applied to them and then they could be compared.
The second measure is the variance of the relative measure of economic activity, again GDP and employment will be used as measures. This is the country’s figure divided by that of the whole monetary union, the variance for both periods will then be calculated. Using an F-test on the variance will be determined whether the variances of the two periods differ significantly.
The other three criteria - intra-industry trade, degree of openness and export diversification - will be analyzed in the same way. Further specification on the measurement of these criteria will follow in the next section. The measure will first be regressed on the time. The slopes of the linear trend line from the pre- and post will be compared with each other using a t-test. Before calculating the t-statistic, first must be determined if the variances of the respective slope coefficients are equal or not. This will tell what the appropriate standard deviation is to use in a t-test.
The relevance of looking at the slope coefficients, is that it resembles a linear trend and it is a measure for the speed of convergence. And a positive or negative sign of the slope coefficient tells us if this country is either converging or diverging from a OCA in that respective criterion. And by comparing the slopes of the two the two periods, I want to find out if the introduction of the euro has changed the speed of convergence.
9
4
Data
The countries in this research are the twelve first countries that had introduced the euro banknotes and coins in 2002, namely: Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal and Spain. The aggregation of the twelve countries will be referred to as the EZ12. The period that will be observed is 1979 till 2016. For some criteria the data of some countries wasn’t completely available, these countries would then be dropped from the analysis of the criterion. For some criteria, before 1999 only data on the Belgium–Luxembourg Economic Union (BLEU) was available, and not of the countries separately. Therefore the data of Belgium and Luxembourg was aggregated to create on entity for the whole sample period. The same has been done with the former German Democratic Republic, Federal Republic of Germany and the current German state.
For the business cycle analyses two different measures of economic activity will be used, namely GDP and employment. The GDP data was collected from the OECD database. The sample contains annual nominal GDP of twelve countries over the period 1979-2016. GDP data are measured in million US dollars at current prices and PPPs. The GDP of the separate countries are summed up for every year, to create a figure for the EZ12 so that each countries relative GDP can be calculated. The other measure of economic activity that is used, is employment. This data is also collected from the OECD database, employed people are those aged 15 or over who had a job during the year. Employment is measured in thousands of people, and it’s seasonally adjusted. The employment data for the period 1979-2016 of Luxembourg was not available in the OECD database, therefore
Luxembourg has been left out of this sample. The sample will then contain annual employment data of eleven countries over the period 1979-2016. Again the annual employment of the separate countries are all summed to calculate each country’s relative employment.
The bilateral trade data for the intra-industry trade analysis, are retrieved from the Comtrade database of the United Nations. The Standard International Trade Classification (SITC) at the 1-digit level is used to distinguish commodities. Commodities are categorized in one of the ten different commodity sections. The annual bilateral trade data in ten different commodities of 11 countries over the period 1979 till 2016 is collected. For all countries the annual import and export in
commodity k to partner country j is collected. Using this data the Grubel–Lloyd index per commodity is calculated as in equation 1, Which is the Grubel–Lloyd index for country i in commodity k in year t. To calculate an overall Grubel–Lloyd index for each country, a weighted average is calculated as in equation 2. The share of trade of commodity k in total commodity trade is used as weight. K is the number of different commodities which is ten in this case. The sample contains overall Grubel–Lloyd
10 indices between all countries, and between all countries and EZ12. Data of the BLEU and Germany has been edited as described above.
𝐺𝐿𝑖𝑘𝑡= 1 − |𝑋𝑖𝑘𝑡− 𝑀𝑖𝑘𝑡| 𝑋𝑖𝑘𝑡+ 𝑀𝑖𝑘𝑡 (1) 𝐺𝐿𝑖𝑡 = ∑ (𝐺𝐿𝑖𝑘𝑡((𝑋𝑖𝑘𝑡+ 𝑀𝑖𝑘𝑡) ∑(𝑋𝑖𝑘𝑡+ 𝑀𝑖𝑘𝑡) 𝐾 𝑘=𝑡 ⁄ )) 𝐾 𝑘=1 (2)
To analyze the openness criterion two data sets are used. The first one is trade as percentage of GDP, this data is collected from World Development Indicators database of the World Bank. Trade is measured as the sum of annual exports and imports of goods and services and this will be
expressed as a percentage of GDP. The sample contains annual trade (% of GDP) data 12 countries from the period 1979-2016.
The second data set is the trade with EZ12 partners as percentage of total trade. For this, the data is collected from the UN Comtrade database. Bilateral import and export on a SITC 1-digit level, from the EMU-12 countries was retrieved. For each country the export and import in all commodities to the other union countries are summed up, to end with the total trade of commodities with the EZ12. For every EZ12 country in the sample will contain total trade in commodities with the EMU-12 as percentage of total trade of commodities to the world. Data of the BLEU and Germany has been edited as described above.
The Herfindahl-Hirschman-index (HHI) will be used to analyze the export diversification criterion. It is the sum of squared shares of each commodity in total commodity export. Commodities are distinguished using the SITC at the 2-digit level. The data of the export of all the different
commodities comes from the UN Comtrade database. Data of the BLEU and Germany has been edited as described above. This sample will contain the HHI index of 11 countries from 1979 to 2016.
𝑠𝑘= 𝑋𝑖𝑘 ∑ 𝑋𝑖𝑘 𝐾 𝑘=1 ⁄ (3) ∑ 𝑠𝑘2 𝐾 𝑘=1 = 𝐻𝐻𝐼 (4)
Formula 3 and 4 above gives an mathematical presentation of how the HHI is calculated. X is the total export of country i in sector k. s subsequently is the share of sector k in the total export. By summing up all shares squared this then gives us the HHI of a country in a certain year.
11
5
Results and analysis
5.1 Business cycle synchronization
For the correlation analysis of the business cycle analysis the hypotheses are stated below. Subscript 1 and 2 respectively corresponds to the period 1979-1998 and 1979-1998. This will consistently be case for the other criteria.
𝐻0∶ 𝑟1= 𝑟2
𝐻1∶ 𝑟1≠ 𝑟2
When the null hypotheses is rejected, a positive or negative sign of the z-statistic tells us respectively if the correlation has decreased or increased. Presented in table 1.1, with employment as measure of economic activity, the correlation of three countries has increased significantly opposed to two countries of whom the correlation has decreased significantly. It shows that the countries who were underperforming in the pre-euro period experience the biggest increase. These three countries are Austria, Greece and Ireland. Moreover Austria’s employment was negatively correlated with that of the EZ12 partners. The unweighted average correlation has increased from 0,616 to 0,666.
1.1 Correlation with the EZ12
Employment GDP 79-98 99-16 z-value 79-98 99-16 z-value AUT -0,4019 0,2914 -2,0496* 0,8411 0,7253 0,8642 BEL 0,9159 0,3492 3,3836** 0,9412 0,5666 3,1218** DEU 0,8983 0,6700 1,8423 0,7914 0,7174 0,4876 ESP 0,9167 0,8854 0,4738 0,8905 0,8619 0,3491 FIN 0,5426 0,6319 -0,3860 0,3302 0,8831 -2,955** FRA 0,9039 0,5468 2,4819* 0,9492 0,8049 2,0076* GRC 0,1671 0,7775 -2,4568* 0,7922 0,6308 0,9442 IRL 0,4165 0,8746 -2,5653* 0,5777 0,3604 0,7952 ITA 0,8617 0,8011 0,5604 0,8975 0,8718 0,3349
LUX N/A N/A N/A 0,8011 0,7907 0,0805
NLD 0,7166 0,7085 0,0468 0,8367 0,9181 -1,0355
PRT 0,8345 0,7873 0,3911 0,9214 0,7711 1,6242
Average 0,6156 0,6658 0,7975 0,7418
** significant change at α = 0,01 * significant change at α = 0,05 Bold z-statistic supports the case of EZ12 moving closer to an OCA
12 When GDP is used a measure of economic activity, only one country’s correlation increased significantly, whereas there are two countries of which it significantly decreased. Finland being the only country having a significant increase, was the country with the lowest pre-euro correlation and became the country with the second highest correlation after the introduction of the euro. Just as with employment as measure of economic activity, it are Belgium and France that have a significant decrease of the correlation. What appears is that in the pre euro period, Belgium and France are with coefficients higher than 0,9 among the countries with the highest correlation. The unweighted average correlation in this case has decreased from 0,798 to 0,7418.
𝐻0∶ 𝜎12= 𝜎22
𝐻1∶ 𝜎12≠ 𝜎22
The hypotheses for the analysis of variance of relative economic activity is stated above. The way the F-statistic consistently is calculated, is by dividing the variance of the pre-euro period by that of the post-euro period. Therefore a F-statistic greater or smaller than zero respectively corresponds to a decrease or decrease in variance.
1.2 Percentage change variance of relative economic
activity
Employment GDP
% change F-value % change F-value
AUT -47,36 1,900 847,34 0,106** BEL 47,96 0,676 193,99 0,340* DEU 320,99 0,238** 219,17 0,313** ESP 574,30 0,148** 572,41 0,149** FIN -98,24 56,931** -81,94 5,538** FRA 130,84 0,433* 34,31 0,745 GRC 1025,88 0,089** 202,94 0,330* IRL -20,81 1,263 323,17 0,236** ITA -94,05 16,813** 583,17 0,146**
LUX N/A N/A 99,03 0,502
NLD -98,19 55,307** -63,13 2,712*
PRT 530,57 0,159** -45,07 1,821
Average 206,53 240,45
** significant change at α = 0,01 * significant change at α = 0,05 Bold F-statistic supports the case of EZ12 moving closer to an OCA df = 19 & 17
As can be seen in table 1.2, these result are very significant, but an important caveat must be made. Because the periods 79-98 and 99-16 are unequal in length, they also have a different number of data points. And because of the fact that variances decrease when increasing the sample size, and
13 because of the relatively few data points in both periods. It’s likely that the difference in sample size partly drives this increase in variance. With that said, Finland, Italy and the Netherlands both have a significant decrease of their variance with employment as measure. But it has significantly increased for five other countries, and the unweighted average has more than tripled.
When GDP is the measure of economic activity, it shows an even more striking result. A significant decreases in two countries and a significant increase in seven countries. And a unweighted average increase of 240%.
5.2 Intra-industry trade
The slope coefficients must be interpreted as a measure of speed of convergence in the respective period. For the remaining criteria the hypotheses are stated below. If the null hypotheses proves to be false, we reject it and accept the alternative hypotheses which is that the linear trends in both period are unequal. Looking at the sign (positive or negative) of the t-statistic we can tell if there was decrease or increase of the slope coefficient.
𝐻0∶ 𝛽1= 𝛽2
𝐻1∶ 𝛽1≠ 𝛽2
A positive or negative t-statistic respectively corresponds to an increase or a decrease in the slope. For each country the Grubel-Lloyd index is regressed on the time, which gives us an linear time trend. The Grubel-Lloyd index is a measure that ranges from 0 to 100 and the slope coefficient tells us how much the index annually on average changed during that time period. The slope of the trend line are presented in table 2. The annual growth trend of intra-industry trade has decreased in five of the eleven countries in two of those countries the growth trend has increased. The two countries in which the growth trend has increased are Greece and Ireland, a thing that characterizes both these countries is that they are the two countries with the lowest convergence in the pre euro are period. With numbers lower than zero, they were actually the only two countries that show a divergence before the introduction of the euro. And in 1999 at the introduction of the these two countries had the lowest GL-index. Because of this it’s arguable that Greece and Ireland had some catching up to do, hence the increase in the convergence. During the period 79-98 the unweighted average growth in intra-industry trade was 0,333 per year. In the period after the introduction of the euro the unweighted average growth decreased to 0,154.
14
2 Linear trend Grubel–Lloyd index
79-98 99-16
slope slope t-value
AUT 0,608 (0,063) -0,017 (0,061) -2,156* BLEU 0,221 (0,051) -0,119 (0,111) -2,786** DEU 0,486 (0,062) 0,264 (0,07) -1,025 ESP 0,696 (0,108) 0,537 (0,128) -0,740 FIN 0,619 (0,087) -0,247 (0,203) -3,920** FRA 0,488 (0,106) -0,562 (0,08) -4,942** GRC -0,401 (0,135) 1,2 (0,206) 6,500** IRL -1,298 (0,169) -0,088 (0,169) 4,148** ITA 0,489 (0,042) 0,252 (0,053) -1,356 NLD 0,986 (0,107) -0,013 (0,115) -3,583** PRT 0,768 (0,144) 0,489 (0,131) -1,117 Unweighted average 0,333 0,154
** significant change at α = 0,01 * significant change at α = 0,05 Bold t-statistic supports the case of the EZ12 moving closer to an OCA Standard deviation between the parentheses
5.3 Openness
Looking at the trade (% of GDP), it shows that almost all countries have experienced a significant increase, and the unweighted average trend has increased quit allot. This means that trade has become a larger portion of the GDP and therefore the countries have become more open to the world.
Table 3.1 shows that most of the EZ12 countries have become more open when openness is measured as export + import as a percentage of the GDP. A drawback of that measure is that it uses total export and import, so that it measures openness to the world instead of openness to (potential) currency union partners. When you participate in a currency union, one would do that to reap the benefits of increased trade with the union partners because of lower transaction costs. It is therefore not of absolute relevance what the degree of openness is to countries outside of the monetary union, it’s more relevant to know a countries degree of openness to the other monetary union countries.
When looking at trade with the EZ12 partners as fraction of total trade as presented in table 3.2. Before the start of the euro. For all but the EMU-12 increasingly became a more important trade partner. But after the start of the euro almost all countries switch sign and the EMU-12 becomes a less important trade partner. Moreover the average trend has decreased from 0,563 to -0,382. Ireland and The Netherlands are the only countries for which the EZ12 has become a more important partner in trade.
15
3.1 Linear trend trade (% of GDP)
79-98 99-16
slope slope t-value
AUT 0,395 (0,127) 1,255 (0,212) 3,477** BEL 0,509 (0,265) 2,163 (0,294) 6,809** DEU 0,231 (0,108) 1,885 (0,187) 7,652** ESP 0,771 (0,162) 0,382 (0,182) -1,416 FIN 0,3 (0,289) 0,383 (0,247) 0,239 FRA 0,053 (0,106) 0,555 (0,112) 1,807* GRC -0,222 (0,102) 0,804 (0,206) 4,466** IRL 2,459 (0,389) 3,537 (0,722) 1,315 ITA -0,042 (0,161) 0,6 (0,123) 2,224* LUX 1,73 (0,475) 8,756 (0,896) 6,925** NLD 0,415 (0,212) 2,635 (0,329) 5,674** PRT 0,309 (0,15) 1,075 (0,181) 2,999** Unweighted average 0,576 2,003
** significant change at α = 0,01 * significant change at α = 0,05 Bold t-statistic supports the case of the EZ12 moving closer to an OCA Standard deviation between the parentheses
3.2 Linear trend EZ12 trade (% of total trade)
79-98 99-16
slope slope t-value
AUT 0,333 (0,087) -0,587 (0,06) -4,66152** BLEU 0,072 (0,104) -0,411 (0,066) -3,91302** DEU 0,075 (0,105) -0,548 (0,039) -5,59479** ESP 1,657 (0,175) -0,891 (0,08) -13,2156** FIN 0,525 (0,102) -0,187 (0,077) -2,30997* FRA 0,336 (0,092) -0,325 (0,038) -6,61151** GRC 0,712 (0,178) -1,056 (0,153) -8,25292** IRL 0,31 (0,064) 0,078 (0,069) -0,78974 ITA 0,448 (0,138) -0,434 (0,133) -3,18199** NLD -0,065 (0,107) 0,186 (0,112) 0,758322 PRT 1,784 (0,152) -0,028 (0,207) -7,31205** Unweighted average 0,563 -0,382
** significant change at α = 0,01 * significant change at α = 0,05 Bold t-statistic supports the case of the EZ12 moving closer to an OCA Standard deviation between the parentheses
16
5.4 Export diversification
The Herfindahl-Hirschman-index (HHI) is the sum of squared fractions of the export sectors in a country. So a lower index means that a country’s export is more diversified. The results are
presented in table 4. All countries but one actually have decreased their linear trend, although only four of them were actually significant. In addition 5 countries slope coefficient’s sign has switched from positive to negative, meaning that the HHI of these countries was effectively decreasing after the introduction of the euro, whereas it was increasing in the pre euro period.
The only country that experienced a significant increase is Greece. When plotting their indices over time as in done in the graph on the next page, It then appears that in 2009 the index has steeply increased. This has very much likely to do with the financial crisis and the European sovereign debt crisis that followed it up. Greece being more or less the locus of the European debt crisis, have seen export competing firms in their country struggling because of the lack of available credit. From 1999 until 2009 the index was actually steadily decreasing in Greece. But because of the shock that the European debt crisis has caused, the overall linear trend of the 1999-2016 has become positive.
4 Linear trend Herfindahl–Hirschman Index
79-98 99-16
slope slope t-value
AUT 0,116 (0,012) -0,033 (0,012) -0,619 BLEU 0,026 (0,01) -0,027 (0,006) -4,592** DEU 0,033 (0,018) -0,044 (0,019) -0,247 ESP 0,259 (0,026) -0,212 (0,026) -3,183** FIN 0,162 (0,028) -0,401 (0,026) -3,868** FRA 0,075 (0,013) -0,054 (0,018) -0,544 GRC -0,021 (0,038) 0,559 (0,13) 4,291** IRL 0,297 (0,035) 0,039 (0,045) -1,007 ITA 0,081 (0,01) -0,068 (0,032) -4,475** NLD -0,024 (0,026) -0,113 (0,028) -0,306 PRT 0,087 (0,024) -0,218 (0,021) -1,583 Unweighted average 0,099 -0,052
** significant change at α = 0,01 * significant change at α = 0,05 Bold t-statistic supports the case of the EZ12 moving closer to an OCA Standard deviation between the parentheses
17
6
Conclusion
The goal of this paper was to research if EMU have moved closer to an optimum currency area (OCA). More specifically the question was if the euro has increased convergence to an OCA. This is plausible because of the endogeneity of the OCA criteria, also authors as Firdmuc (2004) and Frankel and Rose (1998) have provided empirical evidence that the OCA theory is endogenous.
Several criteria of the OCA theory have been looked at in this paper. The first criteria that was looked at was the correlation of the business cycles. GDP and employment were used as measures of economic activity. And they were analyzed by looking at the correlation between each country and the EMU-12. This showed an increase in the unweighted average correlation for employment, but at the same time the same measure for GDP has decreased by approximately the same amount. Only Finland showed a significant increase.
For the variance analysis part, it showed big significant increases in variance. As has been noted before, these big differences could have been the caused by the used small unbalanced dataset. But variances that have more than tripled both for employment and GDP is quit striking. All in all the correlation part showed a slight increase in employment measure and a slight decrease in the GDP measure. And the variance analysis showed a significant increase in variance in both
measure. Altogether in terms of business cycle correlation the EZ12 has moved further away from an OCA. 0 2 4 6 8 10 12 14 16 18
Herfindahl-Hirschman-index Greece
GRC18 The second criteria was intra-industry trade. The trend of intra-industry trade before the introduction of the euro is compared to the period after the introduction of the euro. And what is found is that after the introduction of the euro the unweighted average of the trend has been halved. There is still convergence happening but it has decreased heavily. When looking at the individual countries, four countries show a significant decrease in the intra-industry trade trend, at a significance level of 0,01. And two countries show a significant increase at the same significance level. Overall the level of intra-industry trade has convergence of has decreased after the start of the euro. Still the convergence in positive so that it can be said that in terms of intra-industry trade the EZ12 does have moved closer to an OCA.
The third criteria was degree of openness. The trend of trade as percentage of GDP has increased in almost all countries and the EZ12 average trend has almost quadrupled. Trade (% GDP) measures openness to the world, but more important is the openness to the monetary union partners. Therefore EZ12 as percentage of total trade was. This actually showed that the EZ12 hasn’t become a more important trading partner for any country. It shows the exact opposite, the
unweighted average trend decreased from 0,563 to -0,382. The convergence of trade (% GDP) has strongly increased for the EZ12. But more important for the OCA theory is the openness to monetary union partners. Here convergence has decreased strongly, and with can be said that in term
openness the EZ12 has moved further away from an OCA.
The last criteria used was export diversification. It measured using the Herfindahl-Hirschman-index. This showed that four countries saw their trend significantly decreasing after the introduction of euro. Furthermore all but one country has had a decrease in the linear trend, although not significant for all of them. The only country whose trend has increased is Greece. This significant increase is due to the shock that the European sovereign debt crisis has caused. The convergence of the HHI has increased in the most almost all countries, which means that exports of the EZ12
countries have become more diversified. It also means that in term of export diversification the EZ12 has come closer to an OCA and that convergence towards an OCA has increased after the start of the euro.
They way is this research could have been inproved for future research is by using a larger sample period and to use monthly data instead of yearly. Still available data will be limited because it’s nearly 20 years since the euro was introduced. What also could be an improvement is trying to explain patterns using historical events.
19
References
Fatas, A. (1997). EMU: Countries or regions? Lessons from the EMS experience. European Economic Review, 41(3-5), 743-751.
Fidrmuc, J. (2004). The Endogeneity of the Optimum Currency Area Criteria, Intra‐industry Trade, and EMU Enlargement. Contemporary economic policy, 22(1), 1-12.
Fidrmuc, J. (2004). Is Accession to EMU More Justifiable ex post than ex ante?. In The Euroarea and the New EU Member States (pp. 23-38). Palgrave Macmillan, London.
Frankel, J. A., & Rose, A. K. (1997). Is EMU more justifiable ex post than ex ante?. European Economic Review, 41(3-5), 753-760.
Frankel, J. A., & Rose, A. K. (1998). The endogenity of the optimum currency area criteria. The Economic Journal, 108(449), 1009-1025.
Grubel, H. G., & Lloyd, P. J. (1971). The Empirical Measurement of Intra‐Industry Trade. Economic record, 47(4), 494-517.
Kenen, P. (1969). The theory of optimum currency areas: an eclectic view. Monetary problems of the international economy, 41-60.
Krugman, P. (2001). Lessons of Massachusetts for EMU'. International Library of Critical Writings in Economics, 134, 41-61.
McKinnon, R. I. (1963). Optimum currency areas. The American Economic Review, 53(4), 717-725. Mongelli, F. P. (2008). European economic and monetary integration, and the optimum currency area
theory (No. 302). Directorate General Economic and Financial Affairs (DG ECFIN), European Commission.
Mundell, R. A. (1961). A theory of optimum currency areas. The American economic review, 51(4), 657-665.
De Grauwe, P., & Mongelli, F. (2005). Endogeneities of optimum currency areas: What brings countries sharing a single currency closer together?.
