International interdependence and business cycle transmission between Turkey and the European Union.

• Author: Sayek, Selin

1. Introduction

   Because of globalization and the Asian financial crisis, interest in the international linkages and economic comovements between different regions of the world has recently heightened. One economic relationship of particular interest is the relationship between Turkey and its European neighbors. This relationship is important because of Turkey's European orientation and because of Turkey's long-held ambition of joining the European Union (EU).

   In view of this situation, we examine the linkages and comovements of outputs, prices, interest rates, and money supplies between Turkey and Europe. The main question we address here is, Do the Turkish and European economies move together through time? That is, are their business cycles synchronized? Three related questions are, How or why does this synchronization or comovement take place? Is there any evidence of business cycle transmission between the European economies and Turkey? Have these comovement relationships changed over time?

   To address these questions, we gathered statistics on macroeconomic variables, trade linkages, capital flows, and historical economic events. For most of the analysis, Germany, Turkey's largest trading partner, is used as a proxy for the EU nations. Because of the limited period of availability of the quarterly data, we examine the relationship using both annual and quarterly data. Using annual data for 1956-1998 and quarterly data for a shorter period, from the first quarter of 1981 through the fourth quarter of 1998, we calculate correlations between macroeconomic variables for Turkey and Germany and perform principal-components analysis, tests of cointegration and Granger causality, and a structured vector autoregression (SVAR)--impulse response analysis using the method of Bernanke (1986) and Sims (1986). The synchronization issue is investigated through the use of the graphs, correlations, and principal-components analysis. The transmission issue is investigated through the use of the Granger causality t ests and SVAR--impulse response analysis.

   Our results show that the business cycle patterns of Turkey and Europe are quite different; indeed, in some cases they appear to be almost 180[degrees] out of phase. In many cases the correlations between the real growth rate of Turkey and those of the EU countries are statistically insignificant. Principal-components analysis reveals a large difference between fluctuations in Turkey and those in the EU countries. Associated regressions also show a surprisingly insignificant relationship between the real GDP growth rates for Turkey and the EU and between those for Turkey and Germany. However, the SVAR-impulse response analysis reveals a modest positive transmission from Germany to Turkey.

   Our conclusion is that although Turkey and Europe will obviously both benefit from participation in a free trade relationship with each other (see Harrison, Rutherford, and Tarr 1997), because of the absence of synchronization between Turkey and the EU, it appears that Turkey would incur serious stabilization costs if it were to participate in the European Monetary Union. In addition to the transaction benefits of participating in an optimal currency area, there are economic stability losses incurred by a country joining an optimal currency area. Economic costs arise because a country joining an optimal currency area loses its ability to implement an independent exchange rate and monetary policy for income and unemployment stabilization (see Krugman and Obstfeld 2000). Hence, countries out of sync with the rest of the countries in the group will be continuously affected by inappropriate monetary policy. Turkish business cycle fluctuations do not display the similarity with European business fluctuations that would make it beneficial for Turkey to be a part of the European Monetary Union at this time. (1) On the other hand, dissimilarity in business cycles might make it profitable for European investors to invest in Turkey for purposes of diversification.

   A brief literature review is presented in section 2, and a simple theoretical model is given in section 3. In section 4, the data and the econometric analysis are described. In section 5, the results of the analysis are presented and an interpretation of these results is offered. A summary and conclusions are presented in section 6. A data appendix is also included.

2. Literature Review

   Since long before Wesley C. Mitchell (1927), economists have been interested in the phenomenon of international business cycle comovements and the possible transmission of fluctuations between countries. The researchers of Project LINK attempted to systematically estimate the strength of international linkages and transmissions using large-scale linked structural econometric models. Hickman and Filatov (1983), working on Project LINK, concluded that the linkages they found were not strong enough to bring about the synchronization of business cycles. The international linkages proved to be much noisier and more elusive than previously thought. Recently, economists have turned more to time series techniques for investigating these international comovements (Swoboda 1983; Burbidge and Harrison 1985; Dellas 1986; Ahmed et al. 1993; Selover 1997; Canova and Marrinan 1998; Kim 1999; Kwark 1999; Selover and Jensen 1999).

   There are three major competing theories that attempt to explain apparent international business cycle comovements and synchronization. The first theory is the locomotive hypothesis, whereby it is assumed that business cycles are transmitted from one country to another, usually from a large country to a smaller one, via the transmission mechanisms of trade flows and capital movements. (2) These transmissions are often examined in different dimensions; for example, they have been examined for income shocks (Dellas 1987; Ahmed et al. 1993), for price shocks (Darby et al. 1983; Fukuda and Kano 1997), and for interest rate shocks (Glick 1987; Kirchgassner and Wolters 1987).

   In the 1980s, some researchers, including Dellas (1986, 1987) and Gerlach (1988, 1990), found business cycle synchronization but failed to find statistically significant evidence of business cycle transmission. These researchers therefore proposed an explanation of international business cycle synchronization based on common shocks and real business cycle theory. Dellas (1986) based his common shocks hypothesis on the existence of common world shocks, such as technology shocks or commodity supply shocks.

   Other researchers, such as Selover and Jensen (1999), found the common technology shock explanation implausible and developed a new nonlinear mode-locking explanation of business cycle synchronization. Mode-locking is a phenomenon whereby systems with a tendency to oscillate, such as economies, will, even when weakly linked, affect the timing of each other's oscillations in such a way as to bring about synchronization. This mode-locking hypothesis is intriguing because it means that linkages between economies do not have to be very strong to bring about the cycle timing shifts necessary for comovements and synchronization. Selover and Jensen (1999) found prima facie evidence for possible mode-locking among the major economies of Europe, North America, and Japan.

   In other applied econometric research concerning business cycle transmission and synchronization, Selover (1999, 2000) found evidence of a world business cycle and various regional business cycles using principal-components analysis and vector autoregressions. A similar strategy is followed for the analysis presented in the present paper. Important for the analysis is some background on recent Turkish economic history.

   Background on the Turkish Economy

   Throughout its postwar history, the Turkish economy has been marked by a series of economic booms punctuated by economic and financial crises. The economy has been buffeted by political conflicts, three military coups, an armed minority insurrection, two regional wars that significantly diverted trade flows, significant economic policy errors, debt crises, and two and a half worldwide oil price shocks. All of these events occurred while Turkey was making the transition from an agrarian to a modem industrial economy and from a heavily state-owned economy to an increasingly privatized one. (3)

   The main events impacting the Turkish economy are listed in a chronology given in Table 1. The graphs of real Turkish GDP growth and of annual Turkish consumer price index (CPI) inflation, shown in Figure 1, can be examined for possible correspondences with the events listed in the chronology.

   In the first two decades of the Turkish Republic (1923-1943), the economy was largely agricultural, and the government built up the public sector to act as an engine of growth and to industrialize the economy. From 1932 to 1945, the government set up state economic enterprises (SEEs) in key industries, such as the textile, paper, ceramic, glass, chemical, and iron and steel industries.

   Following the elections of 1950, Prime Minister Menderes steered the economy in a more market-oriented, private-enterprise direction. From 1950 to 1953, there was rapid growth, but gradually, growth slowed and inflation accelerated, and by 1958-1960, the country was entering into a financial and economic crisis. The chronically unprofitable state economic enterprises were forced to borrow from the central bank, causing the money supply to grow and inflation to accelerate. Under a fixed-exchange-rate regime, inflation caused the real exchange rate to appreciate, slowing exports, increasing imports, and causing foreign debt to grow so quickly that it soon became impossible to service. Although Menderes's administration enacted a stabilization program in August 1958 that would eventually prove successful, he was thrown out of office by a coup d'etat in May 1960.

   The new government, supported by the coup...