Turkey's economic integration with the EU: An evaluation of current status and future prognosis.

Choudhury, Askar H. ; Naidu, G.N.

INTRODUCTION

Socio-political standing of Turkey is so unique that it offers a very unique and interesting model for European Union integration. Its desire to enter Single Market Europe was intensified in recent years with the enlargement of the European Union. Turkey has been seeking to become a member of the European Union (EU) for nearly four decades. Most likely, Turkey's integration with the EU will remove the distortions in their price system, which in turn will boost the efficiency of their economy. This economic efficiency will make the country a better place for foreign direct investment. Thus, the efficiency gains from integration will be boosted by induced capital formation. However, Turkey's most recent effort to enter the EU in 2002 has been denied by the EU. Although, countries such as Poland, Hungary, Czech Republic, Greece, Cypress, Slovenia, Malta, and many others entered the EU in May of 2004. Bulgaria and Romania also got admitted to EU effective January 1, 2007, while Turkey still remains outside. Therefore, it will be beneficial to study the future prognosis of Turkey's economy using forecasted data on economic factors. Turkey has been a very faithful and loyal NATO member. During World War II, Turkey remained neutral until February 1945, when it joined the Allies. During the Cold War, Turkey allied itself with the western world. Turkey's long standing economic relationship with the western world is undeniable. Its capital city is located in Europe. In 1963, European Community signed an association agreement with Turkey that envisioned the mutual lowering of trade and migration barriers. In 1973, a protocol addendum was signed and a joint commission was established for removing migration barriers between the EC and Turkey by 1986. But, no advancement has been made to lower the trade and migration barriers until 1982 due to a military coup that took place in 1980. As a result, the EC suspended its relations with Turkey. In spite of all these negative developments, Turkey applied for EU membership in 1987 and it was rejected by the EU. Latest attempt by Turkey to join the union was blocked at the Copenhagen Summit in December of 2002. However, EU pledged Turkey that if it carries out certain criteria on human rights and democracy by December 2004, accession talks could begin soon. The basic motivation for European integration is to establish a single market for goods, services, and capital among the member nations of the EU. Since this is the premise on which the EU expansion drive is advanced, Turkey should then be examined on the economic dimensions of integration (Naidu and Choudhury, 2005). Other researchers also studied various dimensions of Turkey's country profiles (see, Lejour and de Mooij, 2005; Tausch, 2003; Harrison, Rutherford, and Tarr, 1997) for EU integration.

The real issue is that European integration and enlargement does not have the same geopolitical meaning for all of its member countries. Therefore, there is no certainty among the current EU members to accept Turkey in the union. The grounds for this disagreement are numerous. Former French president Valery Giscard d'Estaing the architect of EU constitution, objected to Turkish membership because of its "different culture, different approach, and different way of life." Blocking Turkey from entering the EU contradicts the Union's motto: "Unity in diversity." Resistance for Turkey's entry into the union comes from many directions. Turkey's cultural identity (Islam), social structures (for example, the strong role of military), and demography (projected to be the EU's most populated country in ten years) is perceived to be fundamentally disruptive for a consistent European union. At the center of this issue lies the concern about the continuing disharmony among Turks as to the cultural and social structure of their society. There were several events of social and cultural disharmonies between these two factions occurred time to time. Turkey was shaped by the ideology of Kemalism, after Mustafa Kemal Ataturk. Ataturk implemented a series of reforms that modernized the country and moved it more towards European culture. Turkey adopted a strict separation of state and religion. State strictly controls the practice of religion. However, Crack between Kemalists and Islamists continues to live and occasionally come up on the surface. According to Kemalists' Turkish society should respect secularism as practiced by their European neighbors. While the majority of Islamists interpretation of secularism is religious freedom. EU is also concerned about how Turkey deals with their minority groups in the society. Law enforcement institutions are primarily dominated by the Kemalists, while the Islamists dominate the military ranks and parliament. Despite of the appearance of these social disharmonies, Turkey upholds the basic principles of secularism and democracy.

The proponents of Turkey's membership argue that it is a key regional power with a large economy and the second largest military force of NATO that will enhance the EU's position as a global player. Turkey ranks as the seventh largest economy in the Council of Europe and 15th largest economy in the world. It is one of the G-20 industrial nations. Turkey is a founding member of OECD. Turkey's GDP growth rate averaged 7.4% annually between 2002 and 2007. In the near past, Turkey's economic trends were considerably improving. However, at present, Turkey's economy is turning towards the negative. World Bank forecast a 5.4% rate of growth in 2008 and slightly higher growth rate for 2009. Consumer inflation is running at a rate of 9% for 2008, which is higher than most of the countries in Europe. It has an ongoing negative balance in Current Account with an unemployment rate of about 10%. Turkey's budget (deficit) balance as % of GDP is about 3% and continues to rise. Foreign direct investment is falling and the value of its currency is declining. Inflation and unemployment rate remain problematic for the economy. Interest rate seems to be rising at the same time. Thus, this paper examines the integration issue of Turkey on the basis of its past and future (forecasted) economic profiles. The purpose of this paper is to compare Turkey's current and future economic profiles and observe if they are converging or diverging into the future. It is even more important for us to observe if Turkey's future economic profile will converge to those of longstanding EU member countries economic profiles to make the EU integration likely.

DATA AND ECONOMIC FACTORS

The objective of this paper is to employ economic factors to identify economic integration for the union by grouping countries together on the basis of economic similarities. Both past (2002-2006) and future (forecasted data for 2008-2012 from "The Economist") data were analyzed separately to evaluate the current situation and assess the future prognosis. Classification of groups of countries would depend on the cohesiveness of their economic profiles. Economic profile is created and evaluated for each country using the following four economic factors: 1) GDP Growth Rate, 2) Current Account Balance (% of GDP), 3) Inflation Rate, and 4) Interest Rate. Data set were collected from "The Economist" and "OECD" for this research purpose. Annual data of these twelve countries for four different economic factors were obtained and analyzed. Summary statistics for these data were reported in Table 1 for the period of 2002-2006 (past) and 2008-2012 (future). Results of multivariate analyses are reported in Table 2, Table 3, and Table 4.

RESEARCH METHODOLOGIES

We have carried out statistical analysis in two steps. First, we compare Turkey's past and future economic profiles with respect to other seasoned union member countries past and future economic profiles. This was done by applying univariate tests on means to see if Turkey's economic profile is parallel or divergent to the economic profiles of other nations, as well as on its own past and future time domains. In similar fashion, we also analyzed pair-wise correlations (not reported in the paper) between countries' economic factors to observe the proximity of countries' economic profiles. Second, proximity of these countries economy on a multidimensional scale was also tested using Mahalanobis squared distance [D.sup.2] for both past and future time domains. Mahalanobis [D.sup.2] or Hotelling's T2 may be used to observe the convergence/divergence between countries and also between past and future time domains based on four economic factors collectively; namely Growth Rate of GDP, Current Account Balance (% of GDP), Inflation, and Interest Rate. Canonical discriminant analysis was then used to group similar countries together based on economic profiles for both past and future time domains.

If, we consider [x.sub.i1], [x.sub.i2], ......., [x.sub.iNi] random samples from two multivariate normal populations, [N.sub.p] ([[mu].sub.i], [[Sigma].sub.i]) for i=1,2. Then, the test statistic Hotelling's [T.sup.2] to test the difference between two mean vectors can be defined as,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where [[mu].sub.1] and [[mu].sub.2] are the two mean

vectors from two different populations and [Sigma] is the pooled variance-covariance matrix (Johnson 1998, p.420). Also, note that Hotelling's [T.sup.2] is proportional to the Mahalanobis [D.sup.2] to measure the distance between two mean vectors. Hence, Mahalanobis [D.sup.2] measure by itself is adequate to carry out the analysis and test the difference between two mean vectors on a multidimensional basis. Therefore, Mahalanobis [D.sup.2] measure was used in our analysis to assess the convergence/divergence of economic profiles at multivariate plane. Canonical correlation measures were also designed to aid into the classification in terms of economic profiles. In canonical correlation, the canonical axes maximize the correlation between linear combinations of the two sets of multivariate variables Y and X. This is obtained by maximizing the between-variable-group correlation with respect to the within-variable-group correlation. In canonical discriminant analysis, the objects are divided in to k groups, described by a qualitative descriptor. The method maximizes the dispersion of the centroids of the k groups. This is obtained by maximizing the ratio of the between-object-group dispersion over the pooled within-object-group dispersion. Given a classification variable, such as country, and several quantitative factors, the canonical discriminant analysis derives canonical discriminant functions (linear combinations of these quantitative factors) that have the highest possible multiple correlation with groups and summarizes among-class variation in much the same way that principal component analysis summarizes total variation. It facilitates differentiation of groups by taking into account the interrelationships with the independent variables (economic factors). An important property of canonical variables is that they are uncorrelated even though the underlying quantitative variables may be highly correlated. Further details of these multivariate techniques can be found in Johnson, 1998; Johnson and Wichern, 1998; Hair et al., 1998; Chatfield & Collins, 1980).

We therefore, use these Multivariate methodologies to observe the convergence/ divergence between two time domains and also between countries in terms of their economic profiles. A dimension reduction technique known as canonical discriminant analysis was then applied to classify countries that grouped together on the basis of economic profiles for both past and future time periods. The application of this technique is to classify countries according to their economic similarities and clustered them together into one or more group(s) for each time period. Groups of countries resulted from this classification analysis should exhibit high internal (within-group) homogeneity and high external (between-group) heterogeneity. Therefore, if the classification is done properly, countries within the group will be closer economy-wise, and countries between clusters will be economically distant. These analyses were performed separately for both time periods (2002-2006 and 2008-2012) to observe the convergence/divergence of economic profiles of these countries.

EMPIRICAL RESULTS

This paper, examined the convergence/divergence of economic profiles of countries for two different time periods. First, we calculated country specific mean vectors of size four for our economic factors and their correlations (pair-wise) matrix (not reported in the paper) to identify possible convergence/divergence between two time periods in-terms of economic characteristics (see Table 1, Table 3A, and Table 3B). Results from these preliminary analyses motivated us to further our analysis on a multivariate level. Therefore, multivariate analyses of Mahalanobis [D.sup.2] and canonical correlations were done using SAS programming software and the results were reported in Tables 2-4. Mahalanobis squared distance [D.sup.2] in Table 2A & Table 2B is consistently higher for Turkey compared to other countries for both past and future time periods. It is also interesting to observe that Greece is predictably moving further away from rest of the countries into the future by exerting higher Mahalanobis [D.sup.2]. Canonical discriminant analysis is then applied to group countries together according to their economic profiles for each time periods. Univariate mean comparison tests (top half of Table 3A & Table 3B) by economic factors reveal that all four economic factors are significantly differentiating country specific means with significant p-value < 0.0001 for both time periods. Country differences are found to be most widely separated according to their economic profiles by the first canonical function (Can1). This is a linear combination of economic factors as follows: 0.4104511577 GDP growth rate -0.4862015290 Current Account Balance + 0.2327864762 Inflation + 0.0218072319 Interest Rate with a high [R.sup.2] of 0.935925 between this (Can1) canonical variable and the country classification variable for the period 2002-2006 (Table 4A). For the period 2008-2012 (Can1), the linear combination of economic factors is 1.492526685 GDP growth rate !1.353566067 Current Account Balance !0.670461200 Inflation + 1.485200361 Interest Rate with a high [R.sup.2] of 0.991587 between this (Can1) canonical variable and the country classification variable (Table 4B). To determine the dimensionality of the canonical space, eigenvalues and likelihood ratio test statistics were used. Although, the first three canonical functions are statistically significant, first two functions alone account for 94.35% and 96.05% of the total variability and the eigenvalues of these two functions are greater than one for both (past and future) time periods respectively (Table 4A & Table 4B). Therefore, these twelve countries means seem to fall into a two-dimensional subspace within the four-dimensional space of economic factors. These functions of economic profiles were then calculated for each country for both time periods and plotted in Graph 1A and Graph 1B to observe the clustered outcome. Results suggest that there are two different distinct clusters formation by the first canonical function for the current data (2002-2006). One formed by Turkey and the other cluster formed by rest of the EU countries (Graph 1A). This result is also supported by higher Mahalanobis [D.sup.2] for Turkey as reported in Table 2A. Interestingly, second canonical function may have created another new cluster by separating Portugal from rest of the countries. It is interesting to note that the first canonical variable which discriminates Turkey with respect to other countries accounts for 73.67% of the total variation. Also note that the distance between clusters formed by the first canonical function (Can1) is much greater than the distance between clusters formed by the second canonical function (Can2) as can be seen with respect to axis (x and y) in Graph 1A. However, our analysis on the forecasted (future) data (2008-2012) does not support any improvement on the economic status of Turkey (Graph 1B). This suggests that clustering result of Turkey's economy stays parallel into the near future. That is to say, if Turkey continues on this current course of economic transformation, the prediction of Turkey's economic profile will be divergent from the EU countries (except for Greece). This result has two possible implications. First, this probably indicates that Turkey may not be able to achieve the necessary convergence in their economic profile to enter into the union in the near future. Second, since Greece's economic profile is also diverging from the EU countries in the future and creating more heterogeneous economy among EU nations in foreseeable future, Turkey may not have to attain economic similarity to enter the union.

CONCLUSION

This paper examined Turkey's current economic status and evaluated the prognosis of its future economy for EU integration. Economic relationship of Turkey with the western world is remarkably long and mutually beneficial. Since, the basic motivation for integration is to establish a single market; Turkey should be examined on the basis of economic dimension for integration. However, statistical analysis on Turkey's economic position with respect to other EU members do not support for economic integration hypothesis. Results of Mahalanobis [D.sup.2] found Turkey's economic profile is to be distantly situated with those of longstanding EU member countries on a multivariate dimension. Further, canonical discriminant analysis that clustered countries together separated Turkey from the group on the basis of economic profiles on a multidimensional scale. These results were highly statistically significant for both current (2002-2006) and future (2008-2012) time periods. Even though, Turkey is very motivated and challenging themselves to join the union. These research analyses do not reveal any sign of promising future for Turkey's accession to the union.

In this study, we applied the criterion of economic profiles of countries to observe the convergence/divergence of their economy between past and future by classification and clustering to assess the integration eligibility. Statistical analysis that incorporates multivariate test reveals that Turkey at present may not be classified as a member of the union nor it can achieve economic parallel in the near future on the basis of economic profiles that we have considered in this paper. However, given that Greece's future economic profile is also moving further away from the other union member countries provide an opportunity for a favorable argument for Turkey's accession to the union. Although, all member countries must unanimously agree on Turkey's membership for its EU accession to be successful. There is no certainty among the current EU members to accept Turkey in the union. Nonetheless, Turkey would like to place a considerable emphasis on the longterm stabilization of their economy for a more cohesive union. As the policy makers concentrate on these economic factors and implement necessary policy changes, Turkey may eventually achieve economic similarity for the well being of future Turkey and the European Union.

REFERENCES

Barber, L. (1998). Birth of the Euro. Financial Times, April 30, 1998.

Bekmez, S. (2002). Sectoral Impacts of Turkish Accession to the European Union. Eastern European Economics, 40, 57-84.

Benink, H.A. (1993). Financial Integration in Europe, London: Kluwer Academic Publishers.

Bilefsky, D. & H. Pope (2004). EU Panel to Back Invitation to Turkey. Wall Street Journal, Oct 6, 2004, pp. A3-A17.

Bodurtha, J. N., D. C. Cho, & L.W. Senbet (1989). Economic Forces and the stock market: An International Perspective. Global Finance Journal, 1, 21-46.

Christopher, C. (2004). Turkeys test for the Union. Financial Times, Oct 2/3, 2004.

Campbell, J. Y. &Y. Hamao (1992). Predictable stock returns in the United States and Japan: A study of long-term capital market integration. Journal of Finance, 47, 43-70.

Campbell, J. Y. & P. Perron (1991). Pitfalls and opportunities: What macroeconomists should know about unit roots, in: O.J. Blanchard and S. Fischer, eds., NBER macroeconomics annual. (MIT Press, Cambridge, MA) 141-201.

Chatfield, C. & A.J. Collins (1980). Introduction to Multivariate Analysis. Chapman & Hall.

Dombey, D., T. Buck & V. Boland (2004). Islomization warming clouds Turks' EU drive. Financial Times, Sept 8, 4.

Financial Times, Special Report June 10, 2008.

Foreign Affairs, Turkey's Dreams of Accession, Sept/Oct 2004.

Gapper, J. (1997). European Stock Exchanges. Financial Times Survey, February 28, 1997.

Hair, J.F., R.E. Anderson, R.L. Tatham & W.C. Black (1998). Multivariate Data Analysis, Prentice Hall.

Harrison, G.W., T.F. Rutherford & D.G. Tarr (1997). Economic Implications of Turkey of a Customs Union with the European Union, European Economic Review, 41, 861-870.

Johnson, D. E. (1998). Applied Multivariate Methods for Data Analysis, Duxbury Press.

Johnson, R.A. & D.W. Wichern (1998). Applied Multivariate Statistical Analysis, Prentice Hall, New Jersey.

Lejour, A.M. & R. A. de Mooij (2005). Turkish Delight: Does Turkey's Accession to the EU Bring Economic Benefits? Kyklos, 58 (1), 87-120.

Lejour, A.M., R.A. de Mooij & R. Nahuis (2004). EU Enlargement: Economic Implications for Countries and Industries, in: H. Berger and T. Moutos (eds), Managing EU Enlargement. Cambridge: MIT Press.

Naidu, G.N. & A. Choudhury (2005). Turkey's Accession to the EU: An Assessment of Economic Integration. Journal of the Academy of Finance, 3(2), 55-67.

Rossant, J. (2004). Turkey at the EU's Door: It has key allies, but opposition to its entry is rising. Business Week, October 11, 2004. p. 66.

SAS/STAT User's Guide : SAS Institute, Inc., 1990. Cary, North Carolina.

Stone, N. (2004). Guess Who's Coming to Europe ... Wall Street Journal, Oct.6,2004, p. A18.

Tausch, A. (2003). Social Cohesion, Sustainable Development and Turkey's Accession to the European Union: Implications from a Global Model. Turkish Journal of International Relations, 2(1),1-41.

Teitelbaum, M. & P. Martin (2003). Is Turkey Ready for Europe? Foreign Affairs. 82(3) (May/June) 97-111.

The Economist, Turkey, America and Europe: Who is losing Turkey?, September 20, 2006.

The Economist, Looking to Europe: A Survey of Turkey, March 19, 2005.

Togan, S. (2004). Turkey: Toward EU Accession, The World Economy, 27(7), 1013-1045.

World Stock Exchange Fact Book, Meridian Securities Markets: 2001, Vol. 1 pp 2003-214.

[GRAPHIC 1A OMITTED]

[GRAPHIC 1B OMITTED]

Askar H. Choudhury, Illinois State University

G. N. Naidu, Illinois State University TABLE 1: Summary Statistics:2002-2006 and 2008-2012. Variable Country Mean Mean 2002-2006 2008-2012 GDP Growth Austria 1.9000000 2.0400000 Curr-Balance 0.9016000 3.2600000 Inflation 1.7800000 2.0200000 Interest Rate 4.0910002 4.1000000 GDP Growth Belgium 1.9200000 1.8000000 Curr-Balance 3.3600000 0.5200000 Inflation 1.9800000 2.0400000 Interest Rate 4.0541666 4.1000000 GDP Growth France 1.8200000 1.8800000 Curr-Balance -0.4522000 -1.2400000 Inflation 1.8200000 1.8800000 Interest Rate 4.0595000 4.1000000 GDP Growth Germany 1.4600000 2.3000000 Curr-Balance 4.3140000 2.9000000 Inflation 1.7000000 1.8600000 Interest Rate 4.0011666 4.1000000 GDP Growth Netherlands 1.8800000 2.2000000 Curr-Balance 7.5920000 5.2800000 Inflation 1.5800000 1.8200000 Interest Rate 4.0526666 4.1000000 GDP Growth Ireland 5.0000000 3.0400000 Curr-Balance -2.6800000 -3.1400000 Inflation 3.3400000 2.4800000 Interest Rate 4.0568332 4.1000000 GDP Growth Italy 0.7000000 1.2200000 Curr-Balance -1.4160000 -2.5400000 Inflation 2.3000000 2.0800000 Interest Rate 4.2380478 4.1800000 GDP Growth Finland 3.1000000 2.6800000 Curr-Balance 6.1740000 3.0400000 Inflation 1.0200000 2.2600000 Interest Rate 4.0716668 4.1000000 GDP Growth Portugal 0.6400000 1.8800000 Curr-Balance -8.2060000 -7.1200000 Inflation 2.9400000 2.2800000 Interest Rate 4.1366502 4.1000000 GDP Growth Greece 4.3000000 3.0000000 Curr-Balance -6.4240000 -9.6200000 Inflation 3.3600000 2.8000000 Interest Rate 4.2601666 10.6800000 GDP Growth Spain 3.4400000 1.8000000 Curr-Balance -6.7100000 -5.1400000 Inflation 3.0600000 2.7000000 Interest Rate 4.0716834 4.1000000 GDP_Rate Turkey 6.6800000 4.5200000 CB_P_GDP -5.9720000 -5.9600000 INF 13.0000000 6.1200000 IntRate 31.0600000 13.8600000 Variable Country Std Dev Std Dev 2002-2006 2008-2012 GDP Growth Austria 0.9137833 0.2073644 Curr-Balance 1.5426812 0.3646917 Inflation 0.3898718 0.3834058 Interest Rate 0.5824679 0.0707107 GDP Growth Belgium 0.9121403 0.2121320 Curr-Balance 1.0734757 0.1303840 Inflation 0.5019960 0.7368853 Interest Rate 0.5586389 0.0707107 GDP Growth France 0.4764452 0.2387467 Curr-Balance 1.0370796 0.2073644 Inflation 0.2683282 0.3492850 Interest Rate 0.5336601 0.0707107 GDP Growth Germany 1.3164346 0.2738613 Curr-Balance 1.4479744 0.7842194 Inflation 0.3674235 0.3577709 Interest Rate 0.5227919 0.0707107 GDP Growth Netherlands 1.0329569 0.1581139 Curr-Balance 1.2475857 0.4207137 Inflation 0.4207137 0.2387467 Interest Rate 0.5572630 0.0707107 GDP Growth Ireland 0.7615773 0.6949820 Curr-Balance 2.2603318 0.5029911 Inflation 1.0454664 0.4381780 Interest Rate 0.6091289 0.0707107 GDP Growth Italy 0.7582875 0.4658326 Curr-Balance 0.7108657 0.1516575 Inflation 0.2915476 0.5540758 Interest Rate 0.5339420 0.2387467 GDP Growth Finland 1.4747881 0.1788854 Curr-Balance 3.5858444 0.5366563 Inflation 0.5932959 0.4827007 Interest Rate 0.5990036 0.0707107 GDP Growth Portugal 0.8234076 0.3114482 Curr-Balance 1.4317577 0.9576012 Inflation 0.5683309 0.0836660 Interest Rate 0.5692919 0.0707107 GDP Growth Greece 0.5099020 0.2549510 Curr-Balance 1.8531676 1.4669697 Inflation 0.3049590 0.6363961 Interest Rate 0.5558113 0.1095445 GDP Growth Spain 0.3130495 0.8306624 Curr-Balance 2.2928476 1.0502381 Inflation 0.4335897 0.8124038 Interest Rate 0.5795121 0.0707107 GDP_Rate Turkey 1.4788509 0.8584870 CB_P_GDP 1.7914017 0.3974921 INF 6.9274815 2.4783059 IntRate 18.2283570 2.5145576 TABLE 2A: Mahalanobis Squared Distance- D2: 2002-2006. Country AU BE FI FR AU 0 2.11997 10.68059 0.60673 BE 2.11997 0 4.30944 4.94622 FI 10.68059 4.30944 0 16.15136 FR 0.60673 4.94622 16.15136 0 GE 4.02929 0.63405 4.02794 7.58554 GR 26.7067 39.23423 56.89549 21.44127 HO 14.70313 5.88088 2.33775 21.2455 IR 19.27507 25.34656 34.72302 17.51349 IT 3.33552 9.28635 25.19277 1.66593 PO 28.97209 45.66528 74.46353 21.32783 SP 22.95647 36.32294 56.19359 17.24937 TU 89.03946 104.6546 122.0362 82.18638 Country GE GR HO IR AU 4.02929 26.7067 14.70313 19.27507 BE 0.63405 39.23423 5.88088 25.34656 FI 4.02794 56.89549 2.33775 34.72302 FR 7.58554 21.44127 21.2455 17.51349 GE 0 49.22816 3.73023 33.98282 GR 49.22816 0 72.81678 5.403 HO 3.73023 72.81678 0 48.69153 IR 33.98282 5.403 48.69153 0 IT 11.48023 24.97737 28.25363 25.13067 PO 52.19834 17.60809 83.46459 34.34597 SP 45.31793 1.05542 70.43894 8.87974 TU 119.2871 34.56721 145.5522 41.88836 Country IT PO SP TU AU 3.33552 28.97209 22.95647 89.03946 BE 9.28635 45.66528 36.32294 104.6546 FI 25.19277 74.46353 56.19359 122.0362 FR 1.66593 21.32783 17.24937 82.18638 GE 11.48023 52.19834 45.31793 119.2871 GR 24.97737 17.60809 1.05542 34.56721 HO 28.25363 83.46459 70.43894 145.5522 IR 25.13067 34.34597 8.87974 41.88836 IT 0 15.09294 18.60386 89.19872 PO 15.09294 0 10.15134 70.91952 SP 18.60386 10.15134 0 40.881 TU 89.19872 70.91952 40.881 0 TABLE 3A: Univariate and Multivariate Tests on Equality of Means: 2002-2006. Univariate Test Statistics Variable Total Pooled Between Standard Standard Standard Deviation Deviation Deviation GDP Growth Rate 1.9713 0.9681 1.8305 Current Balance % of 5.4474 1.8413 5.3734 GDP Inflation 3.5976 2.0598 3.1908 Interest Rate 8.9015 5.2897 7.7832 Multivariate Statistics and F Approximations Statistic Value Wilks' Lambda 0.00570255 Pillai's Trace 2.32365858 Hotelling-Lawley Trace 19.82715627 Roy's Greatest Root 14.60667444 Univariate Test Statistics Variable R-Square R-Square / (1-RSq) GDP Growth Rate 0.8038 4.0965 Current Balance % of 0.907 9.7583 GDP Inflation 0.7333 2.7497 Interest Rate 0.7127 2.4807 Multivariate Statistics and F Approximations Statistic F Value Num DF Wilks' Lambda 11.32 44 Pillai's Trace 6.05 44 Hotelling-Lawley Trace 19.71 44 Roy's Greatest Root 63.74 11 Univariate Test Statistics Variable F Value Pr > F GDP Growth Rate 17.88 <.0001 Current Balance % of 42.58 <.0001 GDP Inflation 12 <.0001 Interest Rate 10.82 <.0001 Multivariate Statistics and F Approximations Statistic Den DF Pr > F Wilks' Lambda 174.11 <.0001 Pillai's Trace 192 <.0001 Hotelling-Lawley Trace 117.06 <.0001 Roy's Greatest Root 48 <.0001 TABLE 3B: Univariate and Multivariate Tests on Equality of Means: 2008-2012. Univariate Test Statistics Variable Total Pooled Standard Standard Deviation Deviation GDP Growth Rate 0.9303 0.4624 Current Balance % of 4.6178 0.6999 GDP Inflation 1.3766 0.8633 Interest Rate 3.2054 0.7324 Multivariate Test Statistics and F Approximations Statistic Wilks' Lambda Pillai's Trace Hotelling-Lawley Trace Roy's Greatest Root Univariate Test Statistics Variable Between R-Square R-Square Standard / (1-RSq) Deviation GDP Growth Rate 0.8612 0.799 3.9746 Current Balance % of 4.7378 0.9813 52.5001 GDP Inflation 1.1757 0.68 2.1254 Interest Rate 3.2487 0.9575 22.544 Multivariate Test Statistics and F Approximations Statistic Value F Value Num DF Wilks' Lambda 0.00004912 48.96 44 Pillai's Trace 2.98261443 12.79 44 Hotelling-Lawley Trace 143.9956252 143.17 44 Roy's Greatest Root 117.8634584 514.31 11 Univariate Test Statistics Variable F Value Pr > F GDP Growth Rate 17.34 <.0001 Current Balance % of 229.09 <.0001 GDP Inflation 9.27 <.0001 Interest Rate 98.37 <.0001 Multivariate Test Statistics and F Approximations Statistic Den DF Pr > F Wilks' Lambda 174.11 <.0001 Pillai's Trace 192 <.0001 Hotelling-Lawley Trace 117.06 <.0001 Roy's Greatest Root 48 <.0001 TABLE 4A: Canonical Discriminant Analysis: 2002-2006. Canonical Adjusted Approximate Squared Correlation Canonical Standard Canonical Correlation Error Correlation 1 0.967432 0.96013 0.008342 0.935925 2 0.896633 0.877854 0.025523 0.803951 3 0.713964 0.668433 0.063826 0.509745 4 0.272099 0.078629 0.12055 0.074038 Test of H0: The canonical correlations in the current row and all that follow are zero Likelihood Ratio Approximate F Value 1 0.00570255 11.32 2 0.0889979 5.79 3 0.45395792 2.53 4 0.92596203 0.48 Eigenvalues of Inv(E)*H = CanRsq/(1-CanRsq) Eigenvalue Difference Proportion Cumulative 1 14.6067 10.5059 0.7367 0.7367 2 4.1008 3.061 0.2068 0.9435 3 1.0398 0.9598 0.0524 0.9960 4 0.08 0.004 1.0000 Test of H0: The canonical correlations in the current row and all that follow are zero Num DF Den DF Pr > F 1 44 174.11 <.0001 2 30 135.7 <.0001 3 18 94 0.002 4 8 48 0.8645 Raw Canonical Coefficients Variable Can1 Can2 DP Growth Rate 0.410451158 0.821453965 Current Balance % -0.486201529 0.287934091 of GDP Inflation 0.232786476 0.266521119 Interest Rate 0.021807232 -0.042297343