Empirical Analysis of Structural Change in Turkish Exports

Abstract

Turkish exports has experience high growth rates since 2001, which is well above its historical average. While the average yearly growth rate of exports is 11.2 percent between 1948 and 2000, it reaches to 20.9 percent in the 2001-2006 period. The paper extensively examines the strong performance of Turkish exports, particularly after 2001. The results suggest that none of the elasticity parameters of Turkish export function are stable during the 1987-2006 period. There is a continuous increase in imports as well as income elasticity as opposed to the persistent decrease in the real effective exchange rate elasticity. Change in the composition of exports from traditional to non-traditional commodities played an important role in structural transformation of the Turkish export function.

Full text

Research and Monetary Policy Department
Working Paper No:07/08
Empirical Analysis of Structural
Change in Turkish Exports
The Central Bank of the Republic of Turkey
October 2007
Faruk AYDIN
Hülya SAYGILI
Mesut SAYGILI

1
Empirical Analysis of Structural Change in Turkish Exports*
Faruk Aydın
Hülya Saygılı
Mesut Saygılı
Research and Monetary Policy Department
Central Bank of the Republic of Turkey
October 2007
* The views expressed in the paper are those of the authors and should not be attributed to the Central Bank of the Republic
of Turkey (CBRT). We would like to thank Bilgehan Karabay, Ercan Türkan, Fethi Öğünç, Gökhan Yılmaz, Uğur Çıplak
and Vuslat Us for their valuable comments.

2
Table of Contents
1. Introduction
2. Overview of Turkish Economy in the Post-1980 Period
a. Exports, Imports, and Trade Balance
b. GDP Growth Rate
c. Competitiveness Indicators
i. REER, Unit Labor Cost and Productivity Developments
ii. Terms of Trade
3. Structural Change in Turkish Exports
a. Commodity Composition of Trade
i. Recent Trends in Commodity Composition of Exports
ii. Commodity Concentration Ratio
iii. Intra-industry Trade
iv. Transformation of the Factor Intensity of Exports
b. Country Composition of Trade
c. Import Dependence of Exports
d. Competitiveness Indicators and Integration of Turkey to the World Markets
i. Export Market Shares for Turkey and Emerging Markets
ii. Relative Export Performance for Individual Products: Turkey and Other
Emerging Economies
iii. The Revealed Comparative Advantage (RCA)
4. Empirical Analysis: Time Varying Parameter Estimates for Export Supply and Demand Functions
a. The Model
b. The Method: Kalman Filter Approach
c. Results
5. Conclusion
Appendix

3
1. Introduction
Turkish exports has experience high growth rates since 2001, which is well above its
historical average. While the average yearly growth rate of exports is 11.2 percent between 1948 and
2000, it reaches to 20.9 percent in the 2001-2006 period. Historically, the evolution of exports in
Turkey can be divided into five sub-periods: (i) the period in which the protective policy is the
dominant strategy in foreign trade (early 1930s till 1960s); (ii) the period of import substitution
policy (1960s and 1970s); (iii) the period of financial liberalization and export subsidy policy in
order to support export-led growth strategy (1980s); (iv) capital account liberalization episode
(1990s); and (v) the adoption of floating exchange rate regime (from 2001 onwards).
Turkey launched its customs union agreement (CU) with EU in 1996, the single most
important trade agreement of Turkey since the beginning of its liberalization in the early 1980s.
However, Turkish economy was also hit by major economic and financial crises in the 1994-2001
period. The sources of these crises vary: 1994 and 2001 crises originated from the domestic economy
and political problems, while 1997 and 1998 crises were affected from Asian and Russian crisis,
respectively. Besides, two major earthquakes disrupted Turkish economy in 1999. All of these crises
were characterized by recession and sharp currency depreciation. As a response to the 2001 crisis
Turkish policymakers have initiated an extensive reform program under the supervision of IMF. The
program primarily aimed to reduce public deficit, reform the banking sector, implement floating
exchange rate regime and decrease inflation rate to single digits. After following a modest course in
the 1990s, Turkish exports experienced the most exceptional upward trend in its history despite real
appreciation of the Turkish lira. However, the sustainability and the sources of the surge in export
growth are included in the main issues of the researchers’ and policymakers’ agenda. It becomes
important to understand the structural transformation and the sources of this change in the Turkish
exports to design effective public policies.
There are several studies on Turkish exports and competitiveness indicators that worth
mentioning. Among these studies Yükseler and Türkan (2006) is particularly very informative in
understanding the structure of Turkish production and foreign trade. They extensively analyzed
Turkish exports and provided very rich set of indicators. Among their conclusions, increase in import
dependence of Turkish exports is chiefly important for our purpose. In our study, we followed their
methodology in constructing similar figures for Turkey and few new EU member countries. A study
by Sönmez (2005) also argues that implementation of inward processing regime increased import
dependence of Turkish exports.

4
There are also empirical studies investigating the relationship between Turkish exports and
main macroeconomic variables by using different econometric approaches.1 While Şahinbeyoğlu and
Ulaşan (1999) and Saygılı et al. (1998) found that there are statistically significant relationships
between exports and real effective exchange rate as well as foreign income, Aydın et al. (2004)
disagree on the significance of real exchange rate. Instead, exports are determined by unit labor costs,
export prices and national income in their study. Sarıkaya (2004), similarly, demonstrates the
importance of real unit wage rates on determining exports instead of real exchange rates after 1999.
Thereby, he concludes that, improvement in labor productivity can compensate negative impact of
real exchange rate appreciation on attaining sustainable export growth.
Within this framework, this paper broadly examines the strong performance of Turkish
exports, particularly after 2001, by analyzing different indicators used in the literature and employing
an econometric technique. As a first step, the main trends and structure of Turkish foreign trade in
the post-1980 period are examined. Next, the structural changes in Turkish exports are analyzed
descriptively. For this purpose, the study looks at the indicators for country and commodity
composition of trade and import dependency of exports. Besides, competitiveness of Turkey in the
global economy is examined on the basis of convergence to the world market and comparative
analysis with emerging economies.
In the last section, the structural change in both export supply and demand functions of
Turkey are examined by employing Kalman filter method. Kalman filter approach, where parameters
of the export functions can be estimated as time varying coefficients, allows one to study structural
changes in these equations without imposing any predetermined breaking points in time.
The paper proceeds as follows: Section 2 broadly discusses the structure of Turkish foreign
trade. Section 3 rigorously analyses the structural change in exports by examining different trade
indicators. Finally, section 4 proceeds to Kalman filtering, while section 5 offers summary remarks
and conclusions.
1 Şahinbeyoğlu and Ulaşan (1999) and Saygılı, et al.(1998) used co-integration method. Aydın, et al. (2004)
and Sarıkaya (2004) used both co-integration and VAR methods.

5
2. Turkish Economy in the Post-1980 Period
Before proceeding with the analysis of structural change we would like to overview
developments in the Turkish economy. Thus, this section is motivated by the investigation of the
main trade and competitiveness indicators of Turkey in the post-1980 period. We also studied the
relative performance of Turkish economy compared to a selected group of developed and developing
countries. The section starts with the analysis of exports, imports, openness and trade balance of
Turkey, and continues with competitiveness indicators and real GDP growth. Lastly, the section ends
with a brief overview of changes in export and import price and quantity indices for Turkey.
2.a. Turkish Exports, Imports, and Trade Balance
Trade openness rate of Turkey has been increasing since the liberalization of the Turkish
economy at the beginning of 1980s. The ratio of trade volume to GDP increased from 15.7 percent in
1980 to 53.4 percent in 2006 (Table 2.1, Figure 2.1). A similar pattern is also observed especially for
other emerging economies. In that respect, these developments in Turkey cannot be considered as an
exception. As expected, the lowest openness rates were observed for the large developed countries
due to size of their domestic markets such as US with 20.7 percent and Japan with 22.8 percent in
2006. Some small liberal economies, such as Singapore and Malaysia have historically very high
openness rates, which were 365 percent and 191 percent respectively in 2006. Firms in these
countries uses significant amount of imported goods in their production. In fact, some firms may
import semi-finished goods just for re-exporting purposes. According to WTO figures 45.8 percent of
Singapore and 93.1 percent of Hong Kong’s exports are re-exports in 2005 (WTO, 2006).
Trade deficit figure of Turkey also follows an interesting path. Turkey suffered high trade
deficits and current account problems at the end of 1970s and early 1980s. Trade deficit accounted
more than 7 percent of the GDP in this period. Structural adjustment programs and liberalization of
Turkish trade at the beginning of 1980s were followed by improvements in trade deficit until 1988.
After the beginning of capital account liberalization in 1989, ups and downs in trade balance mostly
followed boom and burst cycles of the real GDP growth until 2001. However, 2001 crises followed
by fast economic growth and widening trade deficits. Trade deficit reached to 12.8 percent of the
GDP in 2006.
Turkey is not the only country suffering from high trade deficit problem. Countries in the EU
accession process, such as Czech Republic, Romania, and Poland, also had large trade deficit in
2000s. Yet, in some of these countries, such as Czech Republic and Poland, trade deficits either

6
decreased or turned into surplus at the end of 2006. This may indicate that any potential negative
effect of EU accession process is temporary.
Table 2.1: Ratios of Trade Volume and Balance to GDP (1980-2006)(1)
Trade Volume/GDP Trade Balance/GDP
1980 1990 2001 2006 1980 1990 2001 2006
Brazil 18.2 11.2 22.4 24.1(2) -1.2 2.3 0.5 5.6(2)
China 21.8 34.9 43.1 65.4(2) -0.4 3.5 2.1 2.6(2)
Czech Rep. -- 82.8(3) 113.0 133.3 -- -1.5(3) -5.0 1.4
Poland 52.6 47.6 47.8 70.7 -3.1 6.1 -4.0 -1.5
Romania 69.2 38.9 64.1 65.2 -4.8 -8.9 -7.4 -12.2
Russia -- 42.3(4) 50.8 47.6 -- 6.1(4) 15.7 14.1
Singapore 357 301 299 365(2) -25.4 -4.4 20.1 32.4(2)
S. Korea 62.9 51.4 60.1 71.4 -7.4 -1.0 2.8 3.3
Malaysia 96.1 125.1 179.1 191.5(2) 9.8 5.7 20.9 25.4(2)
Mexico 20.1 31.3 52.6 60.3 -1.6 -0.3 -1.5 -0.7
Germany 46.2 50.1 55.0 71.4 0.9 4.6 4.6 6.8
Japan 23.4 16.3 17.1 22.8(2) 0.2 2.3 1.7 2.1(2)
UK 40.2 40.0 42.1 44.3 0.6 -3.3 -4.1 -6.5
US 17.0 15.3 18.4 20.7 -0.9 -1.9 -4.2 -6.3
Turkey(5) 15.7 23.4 50.0 53.4 -7.3 -6.2 -6.9 -12.8
(1) USD figures. It includes goods trade only. (4) 1994 figure.
(2) 2005 figure. (5) GDP is taken from SPO database.
(3) 1993 figure.
Source: IMF IFS, SPO Databases.
Table 2.2: Average Growth Rates of Exports and Imports (1980-2006) (1)
Exports Imports
1980-89 1990-2000 2001-06 1980-89 1990-2000 2001-06
Brazil 8.5 4.4 16.5(4) 0.2 10.7 5.7(4)
China 10.8(2) 17.3 25.1(4) 16.4(2) 14.4 24.0(4)
Czech Rep. -- 10.7(3) 21.9 -- 11.8(3) 19.4
Poland -0.3 9.8 21.8 -2.0 12.8 16.8
Romania 1.2 -0.1 20.9 -2.2 3.3 25.5
Russia -- 7.7(3) 19.4 -- -1.9(3) 24.2
Singapore 12.4 11.6 8.7(4) 10.8 10.6 6.9(4)
South Korea 15.4 10.0 11.1 11.6 9.7 11.3
Malaysia 8.4 13.4 7.6(4) 10.0 12.9 7.0(4)
Mexico 11.8 15.2 7.1 9.8 15.8 6.6
Germany 7.2 4.4 13.0 5.5 5.7 11.4
Japan 10.4 4.9 5.0 6.7 5.5 7.7
UK 5.8 5.9 7.9 7.5 5.2 10.3
US 7.0 7.2 4.8 8.5 8.9 7.2
Turkey 17.8 8.2 19.8 12.0 11.9 16.1
(1) USD figures. It includes trade of goods. (4) 2001-2005 period.
(2) 1983-1989 period.
(3) 1994-2000 period.
Source: IMF IFS Database.
Export and import growth rate of Turkey can be classified in three sub-periods (Table 2.2,
Figure 2.2). The first period is 1980s where the initial effects of trade liberalization in 1980s caused
fast export and import growths. In that period, on average, exports grew by 17.8 percent while
imports grew by 12.0 percent. At the interim period (1990s) both export and import growth rates

7
slowed down. However, trade deficit widened due to faster growth of imports than exports. After the
financial crisis in 2001, with the help of the recovery in the economy, both export and import growth
rates accelerated to 19.8 percent and 16.1 percent, respectively. Yet, if we exclude the crisis year,
2001, the average growth rate of exports and imports increase to 21.2 and 26.4 percents, respectively.
The difference between exports and imports growth rates (5.2 basis points) may explain the fast
deterioration of Turkish trade balance in the post-2001 period.
Figure 2.1: Turkey’s Ratio of Trade Volume and Trade Balance to GDP
(Percent, 1980-2006)
-20
-10
0
10
20
30
40
50
60
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
percent
Trade Balance/GDP
Trade Volume/GDP
Source: IMF IFS and SPO Database
Comparison of the recent trends in Turkish exports with exports of other economies reveals a
similar pattern in new EU member Eastern European states, as well as Russia, Brazil, China, and
Germany. In particular, the coincidence of export boom in Turkey and Eastern European countries is
noteworthy and its implications may deserve further study.
2.b GDP Growth Rate
In terms of period averages, long-term growth rate of output in Turkey was roughly 4.2
percent in the 1980-2006 period (Table 2.3). However, financial crises in 1990s and 2001 created
boom-and-bust cycles in the economy such that fast growth in one year was followed by negative
growth in the consecutive year. In general, we observe significant slowdown in the growth rates of
developing economies in the sample during the 2000s with the exception of former communist
countries and China. Energy exports of Russia, investment boom in and FDI flows to China, and EU

8
accession process for the others seem to help these countries to achieve high growth in the 2000s.
Except Romania, all these economies were able to sustain their fast economic growth without
developing major balance of payments problem.
Figure 2.2: Turkey’s Value of Exports and Imports (Billion USD, 1980-2006)
0
20
40
60
80
100
120
140
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
billion USD
Exports
Imports
Source: IMF IFS Database
Table 2.3: Average Growth Rates of Output (1980-2006)
1980-89 1990-2000 2001-6 1980-89 1990-2000 2001-6
Brazil 3.0 2.4 2.2(1) Malaysia 5.8 7.3 4.7
China 9.7 9.8 9.5(1) Mexico 1.5(2) 3.6 2.3
Czech Rep. -- 3.1(3) 4.0 Germany 1.8 3.1 1.0
Poland 1.2(2) 3.1 3.5 Japan 3.8 1.3 1.4
Romania 1.4(2) -2.1 6.1(5) UK 2.3 2.3 2.5
Russia -- -3.8(3) 6.1 US 3.0 3.2 2.5
Singapore (4) 7.1 7.9 3.9(1) Turkey 4.0(6) 4.1(6) 4.6
South Korea 7.6 6.4 4.6
(1) 2001-2005 period. (4) GDP volume index is used (2000=100).
(2) 1981-1989 period. (5) 2001-2004 average.
(3) 1994-2000 period. (6) State Planning Organization data.
Source: IMF IFS Database.
2.c Competitiveness Indicators
2.c.i REER, Unit Labor Cost and Productivity Developments
In the international economic literature the real effective exchange rate (REER) is considered as a
demand-side competitiveness indicator. Appreciation of REER of a country is often interpreted as a
relative loss of price competitiveness of its domestic producers. Recent studies, on the link between

9
REER and exports, reveal that REER is not the most significant determinant of exports in Turkey.
Aydın et al. (2004) finds empirical evidence, which supports the relationship between REER and
imports rather than exports. Sarıkaya (2004) shows that, in the long run, the positive effect of
decrease in real wages dominates the negative effect of appreciation in REER. Put differently, he
points out the importance of real wages in determining exports in Turkish export growth. Within this
framework, Turkish lira preserved its strong position throughout the 2002-2006 period. Accordingly,
the average real appreciation of the Turkish lira during this period was 15.5 percent compared to
2000 in terms of PPI based REER index.2 However, analyzing REER of a country without
considering the REER developments of its major competitors may be inaccurate and misleading. In
this context, 6 of the 9 major trade competitor of Turkey experienced steady trend of real
appreciation between 2002 and 2006 (Table 2.4). In particular, the real appreciation of domestic
currencies in Brazil, Russia and Czech Republic are more evident than Turkey. Furthermore,
Singapore, Malaysia and China also experienced real depreciation in the same period while real
appreciation in the Euro area was 18.8 percent between 2002 and 2005. All in all, it appears that
Turkey did not lose its price competitiveness relative to its international rivals, since majority of
these countries also suffered from real appreciation.
Table 2.4: Real Effective Exchange Rates (2000=100)
Average 1996-
1996-9 2002-6 2001 2002 2003 2004 2005 2006* 2006
Brazil 80.6 127.0 124.0 137.1 142.1 139.8 114.0 101.8 107.3
China 99.8 95.2 104.3 101.9 95.2 92.7 92.5 93.9 98.2
Czech Rep. 95.5 121.8 106.4 118.7 116.8 118.3 125.2 130.1 108.8
Poland 89.6 103.5 112.7 108.1 96.3 96.2 107.4 109.6 99.0
Romania 85.2 108.3 101.5 102.3 99.1 101.6 119.9 118.6 98.5
Russia 128.0 138.7 120.2 123.6 127.3 137.3 149.2 156.2 129.7
Singapore 106.7 94.3 100.5 97.9 94.3 93.3 92.1 93.9 99.9
Malaysia 111.5 98.6 104.9 105.0 99.2 94.9 95.2 98.6 104.0
Mexico 127.5 103.7 91.2 88.7 104.3 112.2 107.3 106.1 79.5
Turkey 93.2 115.5 84.4 101.8 110.8 115.8 125.1 123.9 103.3
EURO Area 121.7 114.3 101.8 104.7 115.1 118.6 118.8 - 114.9
* January-December for Turkey, January-October for other countries
Source: IMF,IFS and OECD
2 The PPI based REER is obtained by deflating the nominal effective exchange rate with price indices.
According to the definition used by International Monetary Fund (IMF), the real effective exchange rate is
computed as the weighted geometric average of the price of the domestic country relative to the prices of its
trade partners. In this computation, IMF country weights based on trade in manufactures, primary commodities
and tourism services over 1988-1990 are used.

10
Figure 2.3: Competitiveness Indicators of Turkey
a) Real Effective Exchange Rates
(ULC based, 2000=100)
60
70
80
90
100
110
120
130
1995Q1
1996Q2
1997Q3
1998Q4
2000Q1
2001Q2
2002Q3
2003Q4
2005Q1
2006Q2
b) Productivity and Real Unit Labor Cost in
the Private Manufacturing Industry
(2000=100)
50
60
70
80
90
100
110
120
130
140
150
1997Q1
1997Q4
1998Q3
1999Q2
2000Q1
2000Q4
2001Q3
2002Q2
2003Q1
2003Q4
2004Q3
2005Q2
2006Q1
2006Q4
Real ULC Productivity
Source: CBRT, TURKSTAT, IFS
Labor productivity and real wages are known as supply-side competitiveness indicators. As a
result of boost in machinery and equipment investment due to the use of external resources, labor
productivity increased exponentially between 2001 and 2006. The upward trend in labor productivity
together with decreasing real wages increased competitiveness and therefore the export performance
of Turkey. According to the OECD labor productivity figures, Turkey and Slovakia have the highest
labor productivity increase in our sample countries (Table 2.5). Moreover, relative unit labor cost
index, which also measures relative competitive position of countries, points out Turkey’s vigorous
position among its trade competitors (Table 2.6).
Table 2.5: Labor Productivity for the Total Economy Indices, 2000 = 100
Average
1993-99 2001 2002 2003 2004 2005 2006*
Czech Republic 90.8 102.0 103.3 108.5 112.9 119.0 125.0
Hungary 92.1 103.8 108.3 111.3 117.5 122.5 127.0
Korea 85.8 101.8 106.0 109.4 112.5 115.4 119.6
Mexico 93.0 99.7 98.1 98.3 98.6 102.2 104.8
Poland 81.8 103.4 108.1 113.6 118.1 119.5 122.1
Slovak Republic 89.6 102.6 107.4 109.9 116.3 121.6 128.7
Turkey 91.1 92.7 100.9 107.8 114.0 121.0 128.0
United Kingdom 92.3 101.5 102.8 104.5 106.9 107.9 109.7
United States 92.9 100.9 103.7 106.3 109.3 111.2 113.3
Euro area 95.5 100.3 100.5 100.9 101.7 102.2 103.4
Total OECD 93.3 100.6 102.4 104.1 106.3 107.9 109.9
* OECD forecast
Source: OECD Economic Outlook 80 database.

11
Table 2.6: Competitive Positions: Relative Unit Labor Costs Indices
(ULC-based REERs), 2000 = 100
Average
1993-99 2001 2002 2003 2004 2005 2006*
Czech Republic 89.3 103.9 118.3 116.1 117.2 124.1 129.8
Hungary 111.5 109.5 124.8 125.0 131.1 129.4 124.0
Korea 116.1 92.7 97.8 95.6 98.9 107.1 107.5
Mexico 99.8 106.5 109.9 99.1 98.0 103.5 103.8
Poland 96.9 104.3 93.8 76.7 69.6 78.3 78.0
Slovak Republic 89.2 97.4 98.0 104.7 113.6 118.4 123.3
Turkey 76.0 73.4 72.2 71.3 79.3 87.8 83.6
United Kingdom 78.0 96.9 100.9 96.6 101.1 101.4 103.0
United States 89.2 102.4 99.1 92.4 85.7 82.5 81.6
Euro area 116.8 100.7 105.9 120.7 125.4 122.9 121.5
Note: Competitiveness-weighted relative unit labor costs in the manufacturing sector in dollar terms.
Competitiveness weights are taken the structure of competition in both export and import markets of the
manufacturing sector of 42 countries into account. An increase in the index indicates a real effective appreciation
and a corresponding deterioration of the competitive position.
* OECD forecast
Source: OECD Economic Outlook 80 database.
2.c.ii. Terms of Trade
Terms of trade (TOT) is a measure of relative price of exported goods to imported
commodities. Improvement in a nation's terms of trade is preferable since it allows countries to give
up fewer exported goods per unit of imported commodities. Our TOT measure indicates
improvement in the competitiveness of Turkish exports for the 1991-2001 period relative to 1980s
but deterioration in the 2000s (Table 2.7). During the same period not only Turkey, but also China,
Brazil, Korea, Singapore and US also experienced deterioration in their TOT. On the contrary, TOT
of Czech Republic, Poland, and Romania, in other words, TOT of the new members of the EU,
together with Japan, Malaysia, Mexico, Russia and U.K, has improved in recent years.
Year to year changes in Turkish TOT, however, remained almost constant throughout the
2000-2006 period, though; the change in quantity ratio was much visible in the same period (Figure
2.4). As a result both ratios moved in downward direction implying strong deterioration of trade
balance of Turkey in the last two years. Moreover, quantity of exports per imports decreased in the
post-2001 while real domestic GDP grew faster than its historical period averages. Decrease in the
ratio of export quantity index to import quantity index in the post-2001 period is noteworthy and may
be a byproduct of increase in import dependence of exports. But, we will postpone the discussion of
this issue to section 3.c.
We may also examine the relative contribution of quantity and price changes on steady
increase in the value of exports and imports (Figure 2.5). Increase in quantity to price ratio indicates

12
relatively greater contribution of quantity changes on the increase in the values. There is a secular
long-run increase in quantity to price ratio for both exports and imports in the post-1994 period. This
trend, however, stops after 2003 for both exports and imports signifying the importance of increase in
the contribution of prices on the increase in the value of exports and imports. Throughout the 1994-
2006 period there was a sharp fall in quantity to price ratio for imports during the financial crisis of
2001 implying a drastic decrease in the quantity of imports.
Table 2.7: Terms of Trade(1)
1980-89 1990-2001 2002-2006(3)
Brazil 1.78 1.26 1.07
China(2) 0.82 1.01 0.96
Czech Rep.(1) -- 1.00 1.05
Germany 1.00 1.07 1.06
Japan 0.81 0.97 1.00
Korea 1.50 1.24 0.85
Malaysia (2) 0.98 0.90 0.99
Mexico 1.39 0.97 1.11
Poland -- 0.93 1.04
Romania (1) -- 1.00 1.07
Russia (2) -- 0.64 0.77
Singapore 1.14 1.07 0.89
Turkey 1.04 1.09 1.00
UK 1.02 1.02 1.04
US 1.00 1.03 1.00
(1) Ratio of export unit value index to import unit value index
(2) Data from Eurostat (3) Data from WDI
(4) 2002-2005 period average for China, Germany, Malaysia, Romania, Russia and UK.
Figure 2.4: Price and Quantity Ratios
0.6
0.7
0.8
0.9
1
1.1
1.2
1994-
1999
2000 2001 2002 2003 2004 2005 2006
Price Ratio (Px/Pm)
Quantity Ratio (Qx/Qm)
Note: Px: Export unit value index, Pm: import unit value index, Qx: Export quantity index,
Qm: Import quantity index.
Source: TURKSTAT

13
Figure 2.5 Quantity and Price Ratios
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1994Q1
1994Q4
1995Q3
1996Q2
1997Q1
1997Q4
1998Q3
1999Q2
2000Q1
2000Q4
2001Q3
2002Q2
2003Q1
2003Q4
2004Q3
2005Q2
2006Q1
2006Q4
Qx/Px
Qm/Pm
Note: Px: Export unit value index, Pm: import unit value index, Qx: Export quantity index,
Qm: Import quantity index.
Source: TURKSTAT
Table 2.8 presents the recent unit value and quantity developments in trade across the
manufacturing sectors. The top panel of the table shows an improvement in the terms of trade for
manufacturing industry after 2002, due to the rise in the relative export commodity prices of food and
beverages, metal industry, machinery and equipment, electronics, motor vehicles and furniture sub-
sectors. On the other hand, quantity index ratio across the manufacturing industries has been going
down steadily since 2001. Fall in the quantity index ratio in textile and wearing apparels, electronics
and furniture together with the poor real export performance of metal industry and machinery and
equipments contributed to the fall in the overall manufacturing export to import quantity ratio.
3. Structural Change in Turkish Exports
After a brief overview of trade performance of the Turkish economy, this section analysis
structural change in the Turkish trade with cross-country comparison in order to understand the
direction and the sources of growth of Turkish exports in recent years. In this section, first
commodity and country composition of Turkish exports are investigated from different perspectives.
In this vein, the commodity concentration ratio, structure of intra-industry trade (IIT) and change in
factor intensity of Turkish exports are examined. Then, import dependency of exports is analyzed and
compared to that of the new EU members. Finally, position of Turkish exports in the world, as well
as in the emerging markets is investigated by comparing and contrasting both relative export

14
performance and revealed comparative advantage (Balassa index) of Turkey at individual product
level.
Table 2.8: Manufacturing Industries Price and Quantity Index Ratios
1994-9 2000 2001 2002 2003 2004 2005 2006 2001-6
(avg)
Price Ratio (Terms of Trade, Px/Pm)
Manufacturing Industry 1.09 0.99 0.98 0.97 1.00 1.02 1.04 1.03 1.01
Food & Beverages 1.07 1.18 1.10 0.97 1.00 1.06 1.28 1.25 1.11
Textile & Wearing 1.02 0.93 1.01 1.06 1.01 1.05 1.06 1.05 1.04
Petroleum & Coal 0.76 0.70 1.24 1.36 1.01 0.94 0.86 0.88 1.05
Chemicals 0.94 1.06 1.00 0.96 1.00 0.99 1.00 1.04 1.00
Plastic & Rubber 1.26 1.09 1.06 1.00 1.00 1.04 1.03 1.05 1.03
Other Min. 0.90 0.83 0.85 0.91 1.01 0.83 0.92 0.93 0.91
Basic Metal 1.17 1.01 1.40 1.08 1.00 1.00 0.96 0.84 1.05
Metal Industry 1.34 1.05 0.88 0.81 1.01 1.27 1.32 1.37 1.11
Mach & Equip 1.01 1.64 0.93 0.96 1.00 1.34 1.40 1.44 1.18
Electronics 0.78 1.38 0.97 0.86 1.00 0.88 1.05 1.24 1.00
Motor Vehicles 1.33 1.08 1.37 1.15 1.00 1.09 1.12 1.11 1.14
Furniture 0.83 0.91 0.80 0.84 1.00 1.03 1.10 1.10 0.98
Quantity Index Ratio (Qx/Qm)
Manufacturing Industry 0.85 0.70 1.15 1.12 1.01 0.89 0.86 0.85 0.98
Food & Beverages 0.97 0.79 1.09 0.92 1.01 1.05 0.99 0.79 0.98
Textile & Wearing 1.03 1.13 1.36 1.06 1.01 0.89 0.86 0.75 0.99
Petroleum & Coal 0.69 0.22 0.38 0.51 1.05 0.79 1.06 0.93 0.79
Chemicals 1.14 0.86 1.16 1.03 1.00 1.01 0.97 1.00 1.03
Plastic & Rubber 0.90 0.73 1.17 1.14 1.01 1.02 1.18 1.14 1.11
Other Min. 0.60 0.74 0.96 1.09 1.02 0.98 0.73 0.52 0.88
Basic Metal 1.82 1.39 2.10 1.41 1.02 1.16 1.02 1.24 1.33
Metal Industry 0.42 0.57 0.51 0.68 1.00 1.00 0.95 0.85 0.83
Mach & Equip 0.39 0.58 0.81 0.95 1.01 0.84 0.84 0.83 0.88
Electronics 0.97 0.96 1.54 1.09 1.01 0.91 0.71 0.69 0.99
Motor Vehicles 0.49 0.42 1.50 1.65 1.11 0.89 0.98 1.10 1.20
Furniture 0.71 0.68 1.27 1.24 1.03 0.80 0.90 0.60 0.97
Source: CBRT and TURKSTAT
3.a. Commodity Composition of Exports
3.a.i. Recent Trends in Commodity Composition of Exports
The list of top 10 export items of Turkey is shown in Table 3.1. As can be seen in that table,
vehicles other than railway or tramway rolling stock is placed as the first with 12 percent share in
total exports for the 2001-2006 period. It is followed by articles of apparel and clothing accessories
knitted (10.5 percent), electrical machinery and equipment (7.6 percent), articles of apparel and
clothing accessories not knitted (7.5 percent), and iron and steel items (7.1 percent). These items are
very heterogeneous in terms of their growth rate, their factor content, intra-industry trade (IIT) index,
and net trade balance.

15
Table 3.1: Top 10 Exporting Commodities (2001-2006 Average)
Export
Share
Avg. Export
Growth IIT T. Balance/
T. Volume
Vehicles other than railway or tramway rolling stock 12.1 36.9 0.923 -0.02
Articles of apparel and clothing accessories knitted 10.4 10.8 0.062 0.93
Electrical Machinery and equipment 7.7 20.7 0.754 -0.25
Articles of apparel and clothing accessories not knitted 7.3 10.9 0.134 0.86
Iron and steel 7.1 21.6 0.787 -0.25
Nuclear Reactors, boilers, machinery and mech. appl. 7.0 28.7 0.461 -0.53
Articles of Iron and steel 3.6 28.8 0.701 0.33
Edible fruits and nuts 3.3 14.8 0.100 0.90
Other make up textile articles 3.0 10.8 0.052 0.95
Plastic and articles thereof 2.3 27.9 0.465 -0.54
Total Exports 63.8(1) 17.3 0.627(2) -0.21
(1) Export share of top 10 commodities.
(2) Trade balance adjusted IIT index. This is usually greater than standard IIT index.
Source: TURKSTAT.
Fastest growing sectors are also highly heterogeneous (Table 3.2). Some of them are also
among the largest exported items, such as vehicles other than railway or tramway rolling stock, or
Nuclear Reactors, boilers, machinery and mechanical appliances, but some others have very small
market share. Generally these commodities have around or above average IIT index. Interestingly,
out of the top 10 fastest growing export items; Turkey has trade deficit in 7 of them. Whether this
may indicate change in comparative advantage of Turkey in these commodity groups or not requires
further study. In section 3.d.iii changes in revealed comparative advantage of Turkey will be
discussed in detail.
Table 3.2: Top 10 Fastest Growing Export Commodities (2001-2006 Average)
Avg. Export
Growth
Exports
Share IIT T. Balance/
T. Volume
Arms and ammunition 50.7 0.31 0.473 -0.29
Vehicles other than railway or tramway rolling stock 36.9 12.15 0.923 -0.02
Miscellaneous articles of base metal 29.3 0.36 0.810 -0.19
Wood and articles of wood 28.9 0.35 0.603 -0.44
Articles of Iron and steel 28.8 3.63 0.701 0.33
Nuclear Reactors, boilers, machinery and mech. appl. 28.7 7.02 0.461 -0.53
Furniture 28.5 1.27 0.699 0.30
Wadding, felt and nonwovens 28.1 0.22 0.776 -0.22
Plastic and articles thereof 27.9 2.30 0.465 -0.54
Stone, plasters, asbestos 27.5 0.94 0.510 0.50
Total Exports 17.3 28.6(1) 0.627(2) -0.21
(1) Export share of top 10 fastest growing commodities.
(2) Trade balance adjusted IIT index. This is usually greater than standard IIT index.
Source: TURKSTAT.

16
Table 3.3: Top 10 Trade Surplus Generating Commodities (2001-2006 Average)
T. Balance/
T.Volume
Avg. Export
Growth
Exports
Share IIT
Meat and edible meat offal 0.98 9.8 0.04 0.017
Prep. of meat, of fish or of crustaceans, mulls or oth.
aquatic invertebrates. 0.96 -12.3 0.05 0.041
Other make up textile articles 0.95 10.8 3.0 0.052
Prep. of vegetables, fruits, nuts and other parts of plants 0.94 14.0 1.56 0.056
Articles of apparel and clothing accessories knitted 0.93 10.8 10.4 0.062
Products of milling industry 0.90 13.8 0.38 0.129
Edible fruits and nuts 0.90 14.8 3.3 0.100
Articles of apparel and clothing accessories not knitted 0.86 10.9 7.3 0.134
Edible vegetables 0.79 16.5 0.88 0.218
Sugars and sugar confectionery 0.77 -3.4 0.40 0.242
Total Exports -0.21 17.3 27.3(1) 0.627(2)
(1) Export share of top 10 trade surplus commodities.
(2) Trade balance adjusted IIT index. This is usually greater than standard IIT index.
Source: TURKSTAT.
There are several approaches to measure comparative advantage. The most direct way is to
rank commodities according to their net exports. The top 10 items of Turkey that have net trade
surplus for the 2001-2006 period is given in Table 3.3. As can be seen in the table, most of these
items are traditional Turkish export commodities that are considered as agricultural and/or labor
intensive instead of capital or technology intensive goods. Besides, these commodities not only have
growth rate less than average total export growth but also very low IIT. These commodities are from
the traditional sectors that may not have high potential to grow in the future indicating that even
though there is a process of transformation in the Turkish exports after the crises period in the 1990s,
Turkey is still short of building any comparative advantage in new commodities.
3.a.ii. Commodity Concentration Ratios
Concentration of exports on few commodities is usually considered as a potential problem for
economies to sustain long run high export growth, since fluctuations in export commodity prices may
also increase volatility in export receipts of a country. In this section, we will study different
measures of commodity concentration of Turkish exports over time: namely, frequency distribution
of Turkey’s normalized exports by commodity groups, weighted spread of Turkish exports, and share
of top 10 and 20 commodities in total exports.
Frequency distribution of Turkey’s normalized exports by commodity groups is shown in
Figure 3.1 and Figure 3.2.3 These graphs show whether sectoral exports are distributed symmetrically
around its mean. In general, Turkey’s frequency distribution is skewed right due to large share of few
sectors in the total exports. As the highest point of this distribution moves toward zero, one can
3 Sectoral exports are normalized by mean and standard deviation as follows:
(
)
/x
μ
σ
−
.
x
,
μ
, and
σ
are
sectoral exports, average exports, and standard deviation of exports, respectively.

17
conclude that the frequency distribution becomes more even around its mean. Increase in skewness,
however, increases the dispersion among the sectoral exports. From 1982-1985 to 1986-1990
frequency distribution moved away from zero to the left but in 1991-1995 it went closer to zero mean
again, implying that distribution of exports improved from 1986-1990 to 1991-1995. On the other
hand, distribution of exports across commodity groups slightly deteriorated during 1996-2000 period
but it slightly improved during the 2001-2006 period.
Figure 3.1: Distribution of Normalized Exports by Commodity Groups (1982-1995)
0
5
10
15
20
25
30
-0.7
-0,5 -0.3
-0.1
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5
1.7
1.9
F
r
e
q
u
e
n
c
y
1982-1985
1986-1990
1991-1995
Source: Our calculations from TURKSTAT data.
Figure 3.2: Distribution of Normalized Exports by Commodity Groups (1991-2006)
0
5
10
15
20
25
30
35
-0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9
F
r
e
q
u
e
n
c
y
1991-1995
1996-2000
2001-2006
Source: Our calculations from TURKSTAT data.

18
Frequency distribution is a useful but an inadequate measure of dispersion of exports, since it
relies on visual judgments instead of quantifiable criteria. Alternatively, spread of exports by
commodity groups, which is measured as a ratio of standard deviation of commodity exports to its
mean, can be used to examine commodity concentration of Turkish exports.4 Increase in this ratio
implies concentration of Turkish exports across different commodity groups. In Figure 3.3 three
different episodes can be identified: 1983-1993, 1993-2001, and 2001-2006. In the first period spread
of Turkish exports increased reflecting a decrease in diversification of commodities. It is followed by
a decrease in concentration of trade in the second period. In the last period, after the financial and
currency crisis in 2001, concentration of Turkish exports raised again. Close investigation of these
graphs reveals very interesting dynamics. First of all, Turkish exports enjoyed stable positive growth
during the post-1982 period until the 1994 crisis. In this period, Turkish economy increased its
exports through success of few industries, which caused export concentration ratio to increase
(increase in spread). However, during the turbulent period (1994-2001), Turkish economy struck by
major economic crises almost every two years (1994, 1997, 1998, 1999, and 2001). In this period,
some sectors were not able to adjust to new conditions, particularly volatile exchange rates and
domestic output. Few old and new industries, however, survived the turbulence by finding new
markets abroad and the last financial and currency crisis in 2001 followed by a sharp and immediate
recovery in Turkish exports. Commodity concentration of exports increased in the same period, as
well. Contrary to the case in previous crises, recovery in Turkish exports after 2001 was long lasting
and it is realized through increase in share of new commodities in exports.
The third measure of commodity concentration of exports uses the share of top 10 and 20
commodities in the total exports (Figure 3.4). One can identify two different episodes, pre-1997 and
post-1997 in Figure 3.4. The share of top 10 commodities was roughly constant during the 1983-
1996 period while the share of top 20 commodities was declining throughout the same period.
However, shares of both commodity groups have been increasing since 1997, which is a clear
indication of commodity concentration of exports. The trend in this measure of concentration ratio
and the one in Figure 3.3 are broadly consistent particularly after 1993. Note that, the third measure
takes only the top sectors into account but the previous one uses the full sample.
4 Weighted spread at time t is calculated as:
()
()
2
/
it t t
i
x
N
μμ
⎡⎤
−
⎢⎥
⎣⎦
∑, where it
x
, t
μ
, and
N
are export of
commodity i at time t, average exports at time t, and total number of commodities, respectively.

19
Figure 3.3: Weighted Spread of Exports by Commodity Groups (1982-2006)
0
1
2
3
4
5
6
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: Our calculations from TURKSTAT data.
Figure 3.4: Shares of Top 10 and 20 Sectors in Total Exports (1982-2006)
50
55
60
65
70
75
80
85
90
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
percent
TOP 10
TOP 20
Source: Our own calculations from TURKSTAT data.
Throughout the 1982-2006 period, there were also changes in the commodity content of top
10 and top 20 groups. Especially mid 1990s marks a structural break in the content of top 10 export
commodities. Among the commodities in the top 10 of 2006, only few were in top 10 of the 1980s
and first half of 1990s (Table 3.4). Top 10 commodities in 1982, 1985, and 1990 includes only 3, 4,
and 5 goods from the top 10 in 2006, respectively. On the other hand, 8 and 9 commodities in the top
10 in 2006 were also in the top 10 in 1995 and 2000, respectively. Commodities, which were at the
top of the list before the financial turmoil period of 1990s, were mostly replaced by new ones

20
afterwards. Rising commodities were the ones, which were flexible enough to adapt to the new
unstable economic environment. However, some others failed to cope with the difficulties. In other
words, these new rising sectors were more resilient to financial and currency crises than others.
Combining this with the previous discussion one may argue that change in the content of top
exporting sectors is accompanied with the rise in the top 10 and 20 sectors in the total exports for the
post 1997 period.
Table 3.4 Top 10 Export Items of Turkey (1982-2006)
1982 1985 1990 1995 2000 2006
1 55 73 61 61 61 87
2 08 55 72 62 62 61
3 73 08 62 72 85 84
4 25 61 08 08 72 85
5 01 42 42 85 87 72
6 24 84 52 84 84 62
7 27 27 85 87 08 73
8 07 24 24 55 63 39
9 61 60 07 63 52 08
10 58 07 25 20 73 63
01: Live animals.
07: Edible vegetables.
08: Edible fruits and nuts.
20: Prep. of vegetables, fruits, nuts and other parts of
plants.
24: Tobacco and manufactured tobacco substitutes.
25: Salt, sulphur, earths and stone plastering materials.
27: Mineral fuels, mineral oils and production of their
distillation.
39: Plastic and articles thereof.
42: Articles of leather.
52: Cotton, cotton yarn and cotton fabrics.
55: Man-made staple fibers.
58: Special woven fabrics.
60: Knitted or crocheted fabrics.
61: Articles of apparel and clothing accessories knitted.
62: Articles of apparel and clothing accessories not knitted.
63: Other make up textile articles.
72: Iron and steel.
73: Articles of Iron and steel.
84: Nuclear Reactors, boilers, machinery and mechanical
appliances.
85: Electrical Machinery and equipment.
87: Vehicles other than railway or tramway rolling stock.
Items that were also in the top 10 of 2006 were written in bold.
Source: TURKSTAT
3.a.iii. Intra-industry Trade
Another measure of commodity composition of trade is IIT index. IIT arises when countries
simultaneously export and import similar commodities. In a sense, this index measures similarity of
export and import content of countries. While some economists argue that IIT is simply a matter of
aggregation error of commodities of different features and factor intensities into broad categories,
some others supports the existence of IIT on the basis of product differentiation and/or economies of
scale. Oligopolistic competition may also cause two-way trade in identical or similar commodities
since competing local monopolies may find it profitable to penetrate their international competitors’
market. There is a huge literature on this topic and interested readers may find it useful to read
Bhagwati and Davis (1994) and Helpman and Krugman (1985).

21
Our measure of IIT is standard Grubel-Lloyd IIT index adjusted for trade imbalances. According to
this index, there is a secular long run increase in the IIT of Turkish exports (Figure 3.5). IIT may
increase due to many reasons. Generally, trade in differentiated manufactured goods is considered as
a source of IIT. Secular increase in this index may also indicate decrease in income differences
between Turkey and its main trade partners (Linder hypothesis) and/or increase in the share of
manufactured goods in total trade. A study by Gönel (2001) argued that IIT index of Turkey is still
low and 1996 Turkey-EU customs union (CU) did not have any significant effect on Turkey’s IIT.
However, the data she used in her analysis covers the 1992-1997 period, which does not reflect the
full impact of CU on the Turkish exports. Our study with longer time span, shows that there is a
significant upward trend in IIT of Turkey. As it is discussed earlier, there is also a significant
transformation within the Turkish export industries during the 1994-2001 crises episode. The rising
industries of the post crises period can be considered as relatively more capital and high technology
intensive commodities compared to the top 10 export commodities of the 1980s and early 1990s,
which might increase the intra-industry content of the Turkish exports.
Figure 3.5: Adjusted Intra-Industry Index (1982-2006)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
Source: Our calculations from TURKSTAT data.
3.a.iv. Transformation of the Factor Intensity of Exports
Among other factors affecting export performance, technological competitiveness is
frequently mentioned in the literature. The classification of exports in terms of factor
intensity5 reveals that concentration occurs in the high technology products in the world
5 The classification of exports in terms of factor intensity is based on the Standard International Trade Classification (SITC)
3-digit level. The SITC codes of these classifications can be found in Annex.

22
export market (Table 3.5). The same is also true for the emerging market economies (Table
3.6). In the case of Turkey, exports were dominantly relied on labor, agriculture and raw
materials-intensive products during 1980s. However, the picture has dramatically changed in
recent years. Regarding technological competitiveness, research and development (high and
leading-edge technology) intensity in total manufacturing exports in Turkey tripled from
1980s to 2000s while the share of the raw materials and agriculture-intensive sectors is
substantially fell (Table 3.7). Although the share of R&D-intensive product exports is well
below that of the world average and 12 out of 20 emerging market economies in our sample,
Turkey ranks the first in the growth rate of R&D products among emerging market
economies over the period of 2001-2004 (Tables 3.8 and 3.9). Put differently, this rapid
transformation in the factor intensity signals the “take-off” in the technological
competitiveness of Turkey.
Table 3.5: The Classification of Exports According to Factor Intensity
(World, % share in total exports)
High tech-
intensive
Raw material-
intensive Labor-intensive Capital-
intensive
Agriculture-
intensive
1980-1989 29.7 19.9 8.4 8.7 9.1
1990-1996 36.9 12.8 9.3 8.1 7.4
1997-2000 41.8 11.4 8.7 7.6 6.1
2001-2003 42.8 12.8 8.2 7.3 5.7
Source: UNCTAD and our calculations
Table 3.6: The Classification of Exports According to Factor Intensity
(Emerging Economies, % share in total exports)
High tech-
intensive
Raw material-
intensive Labor-intensive Capital-
intensive
Agriculture-
intensive
1980-1989 16.4 24.8 15.9 6.4 13.2
1990-1996 26.9 15.0 17.6 6.4 7.9
1997-2000 35.7 11.9 14.9 6.2 5.9
2001-2004 38.8 12.6 13.3 6.3 4.8
Source: UNCTAD and our calculations
Table 3.7: The Classification of Exports According to Factor Intensity
(Turkey, % share in total exports)
High tech-
intensive
Raw material-
intensive Labor-intensive Capital-
intensive
Agriculture-
intensive
1980-1989 6.0 16.9 30.6 9.3 24.2
1990-1996 6.9 5.5 42.7 14.8 17.7
1997-2000 12.0 3.7 44.3 12.8 13.0
2001-2004 18.0 3.9 39.4 16.0 8.8
Source: UNCTAD and our calculations

23
Foreign direct investments (FDI) might have a significant role on the specialization of
countries on high technology intensive products. Indeed, FDI is seen as an important channel of
transferring technologies to emerging countries. South East Asian countries, which have the largest
share of high technology intensive exports among other emerging markets, received substantial
amounts of FDI. Ng (2006) found evidence that the level of FDI inflows causes technological upturn
in selected Asian countries. The technology transfer by multinational firms may boost the
technological capacity of local firms of host countries. As a result, we found positive correlation
between FDI stock (inward) as a share of GDP and the export share of high technology intensive
products for selected emerging countries (Table 3.10). Moreover, this finding can be generalized for
the 20 emerging countries in our sample.6
Table 3.8: The Classification of Emerging Economies Exports According to Factor Intensity
(2001-2004, percent share in total exports)
High tech-
intensive
Raw material-
intensive Labor-intensive Capital-
intensive
Agriculture-
intensive
Philippines 73.5 2.4 9.0 3.4 4.1
Singapore 66.5 9.0 2.5 1.6 1.2
Malaysia 58.3 12.6 4.4 2.3 2.1
Korea 51.9 5.3 9.7 6.6 1.2
Mexico 51.8 10.4 8.6 10.0 4.0
Hungary 45.9 3.3 7.4 6.9 6.6
Hong Kong 44.8 1.3 21.4 1.2 0.9
Thailand 39.4 7.4 11.6 5.4 14.0
Czech Republic 39.2 5.6 10.8 14.4 2.6
China 37.4 3.9 23.3 3.8 4.1
Poland 22.1 7.6 10.7 12.5 7.4
Brazil 18.6 20.4 3.5 12.8 19.6
Turkey 18.0 3.9 39.4 16.0 8.8
Indonesia 15.8 31.8 14.1 3.9 6.1
India 13.5 9.7 29.3 5.8 11.0
Bulgaria 12.8 12.9 24.9 12.3 7.0
Argentina 10.2 25.8 1.8 7.7 32.1
Russia 6.5 57.5 1.0 10.4 1.5
Chile 3.0 27.9 1.1 29.7 22.1
Source: UNCTAD, IFS, own calculations.
6 Direction of the relationship deserves further study.

24
Table 3.9: The Classification of Emerging Economies Exports According to Factor Intensity
(2001-2004, percent growth with respect to 1997-2000 )
High tech-
intensive
Raw material-
intensive Labor-intensive Capital-
intensive
Agriculture-
intensive
Turkey 50.7 5.9 -11.0 25.2 -31.9
Indonesia 43.1 4.4 -4.1 48.6 -11.4
China 41.1 -21.8 -18.3 -2.9 -25.5
Hong Kong 29.8 -37.5 -14.7 -24.6 -35.5
India 14.0 70.4 -11.2 24.9 -20.5
Czech Republic 13.8 -16.4 -16.7 2.7 -16.3
Hungary 13.4 -17.7 -20.5 8.3 -24.2
Korea 13.1 -1.6 -28.7 0.6 -30.6
Poland 7.4 -10.6 -27.1 5.7 -21.3
Thailand 3.5 17.8 -13.7 37.1 -15.0
Mexico 3.4 8.5 -11.6 6.8 -15.8
Philippines 3.0 -5.2 -10.1 33.0 -0.2
Bulgaria 0.5 -3.8 38.9 -5.4 -3.7
Brazil 0.1 26.5 -9.0 -15.2 -2.4
Russia -0.5 14.4 -22.2 -17.7 50.6
Singapore -1.6 1.5 -17.2 -6.0 -22.8
Malaysia -5.1 15.4 -19.2 17.4 7.5
Venezuela -15.1 3.5 -40.2 1.7 -50.0
Chile -16.8 9.9 -29.3 -5.3 -1.9
Source: UNCTAD, IFS, own calculations.
Table 3.10: FDI Stock and the Share of High-technology Intensive Exports for
Selected Emerging Economies (2005, %)
FDI stock
(inward, ratio to GDP)
Share of high-technology
intensive exports
Coefficient of correlation
(1990-2005)
Turkey 8.1 18.0 0.86
Czech Republic 43.6 39.2 0.93
Malaysia 58.4 58.3 0.90
Singapore 161.5 66.5 0.63
Source: UNCTAD, IFS, own calculations.
3.b. Country Composition of Trade
As it is discussed earlier, concentration of exports on few commodities and markets are
usually considered as potential dangers for trading nations. Diversification in markets is especially
important in minimizing the risks of business cycles. Economies with diversified markets would be
able to adjust to changing macroeconomic conditions in a fast and flexible way thanks to the
existence of variety of trade channels between economies.
In this section, two different concentration measures will be examined in detail. First of these
measures is the spread of Turkish exports by countries. Due to emergence and dissolution of some
countries in the former Soviet Union and Eastern Europe, post 1993 period is used in the analysis.

25
Throughout the period of 1993-2005 there was a continuous increase in country concentration ratio
of Turkish exports even after the Turkey-EU Customs Union agreement in 1996 (Figure 3.6)
signaling concentration of Turkish export markets. In 2006, however, this trend suddenly turned
downward. Yet, it is too early to conclude whether this fall in concentration ratio is permanent or
temporary.
The second measure of country concentration of exports tells us a similar story. In
Figure 3.7 the export share of top 5, 10 and 20 countries in the total is shown. It is evident
that there is a secular increase in concentration of trade in 10 and 20 countries throughout
our sample period. The share of top 20 countries in total increased from 89 percent in 1994
to almost 92 percent in 2006. On the contrary, the increasing trend in the share of top 5
countries reversed after it reached its peak in 2000. Previously we noted deterioration in
Turkey’s commodity concentration ratio especially for the post 2001 period. Yet, country
concentration of Turkish exports has been deteriorating throughout the whole period, except
2006, with seemingly no relationship between crises period and these statistics. Secondly,
there are relatively small changes in the members of the top 10 export destinations over our sample
period (Table 3.11). Among the economies that penetrate into the top 10 of the list in 2006, 7, 8, and
8 of them were also in the top 10 in 1993, 1995, and 2000, respectively. Yet, among these 10
countries the number of EU members increased from 5 to 7 from 1993 to 2006.
Figure 3.6: Weighted Spread of Exports by Countries (1993-2006)
16.5
17
17.5
18
18.5
1993 1995 1997 1999 2001 2003 2005
Source: Our calculations from TURKSTAT data.

26
Figure 3.7: Shares of Top 5, 10 and 20 Countries in the Total Turkish Exports
(1993-2006)
80
82
84
86
88
90
92
1993 1995 1997 1999 2001 2003 2005
%
70
71
72
73
74
75
76
%
TOP 10 TOP 20 TOP 5 (right scale)
Source: TURKSTAT
Table 3.11: Top 10 Export Destinations for Turkish Commodities (1993-2006)
1993 1995 2000 2006
1 Germany Germany Germany Germany
2 US US US UK
3 UK Italy UK Italy
4 France Russia Italy US
5 Italy UK France France
6 Saudi Arabia France Netherlands Spain
7 Netherlands Netherlands Spain Russia
8 China Saudi Arabia Israel Netherlands
9 Russia Belgium-Luxembourg Belgium-Luxembourg Romania
10 Taiwan Spain Russia U. Arab Emirates
Note: Countries that were also in the top 10 of 2006 were written in bold.
Source: TURKSTAT
3.c. Import Dependence of Exports
The recent performance of imports is attributed to the rise in import dependency of exports.
There are different approaches to analyze this question. In this section, we investigate the question by
comparing the changes in the share of exports in total production with changes in the share of exports
in total supply. The analysis is done for overall economy, manufacturing industry and sub-sectors of
the manufacturing industry. A detailed analysis can be found in Yükseler and Türkan (2006). Here,
we calculated respective ratios for Czech Republic, Hungary, Poland and Slovakia for comparison
purpose.

27
3.c.i. Overall Economy
The ratio of exports to total production measures the share of exported goods and services in
domestic production. Therefore, increase in the ratio indicates a rising tendency in production for
exports. In order to calculate this indicator, total production from the 1998 input-output tables is
updated by using growth rate of industrial production index for the years from 1999 to 2005. The
same indicator is also calculated for Czech Republic, Hungary, Poland and Slovakia to compare and
contrast the export tendency of production in these countries with Turkey. Input-output tables for
these countries are available in Eurostat for few years. The same updating procedure is applied to
proxy total production for the missing years for these countries as well.
Several points arise from the examination of Table 3.12. First of all, it is found that Turkey
exported about 16 percent of its domestic production during the 1998-2001 period, but 15 percent if
we exclude 2001. The ratio increased to 17 percent during the 2002-2005 period. Cross-country
comparisons show similar increasing trend in exports share of production since 1998 in all of
countries in our sample. The rise is slightly lower in Turkey than Eastern European countries,
particularly after 2001, implying relatively higher domestic use of production in Turkey (Table 3.12).
Note that, even if the ratio of exports to domestic production is almost the same for Turkey and
Poland during the pre-2001 period, it increased much faster in Poland in the following years.
Ratio of exports to total supply is calculated to analyze the contribution of imports to exports.
Total supply is calculated as sum of imports and domestic production. In order to obtain real import
values, imports from 1998 input-output tables are updated by using growth rate of import quantity
indices. Next, real imports are added to total production to find out total supply. The ratio of exports
to total supply for the overall economy, which was 14 percent in 1998, increased to 16 percent in
2001 and then decreased back to 13 percent at the end of 2005 (Table 3.13). Meanwhile, the share of
exports to total supply for the new EU member countries increased steadily throughout this period.
Comparison of the ratio of exports to total supply to the ratio of export to total production
would show us how supply for exports changes with increase in imports. Sharp fall in the share of
export in total supply compared to the share of exports in total production implies increasing demand
for imported goods to produce exported goods. The difference between percentage point change in
export to total production and export to total supply ratios relative to 1998 is presented in Table 3.14.
Share of exports in production stayed roughly the same (16 percent) from 1998 to 2005. On the other
hand, export share in total supply decreases from 14 percent in 1998 to 13 percent in 2005. As a
result, export share in total supply decreased by 0.002 basis point more than the export share in

28
production from 1998 to 2005, implying an increasing tendency to use imported commodities for
exported goods.
Table 3.12: Ratio of Exports to Total Production (in purchasing prices, 1998-2005)
Czech Rep. Hungary Poland Slovakia Turkey
1998 0.23 0.26 0.14 0.25 0.16
1999 0.24 0.27 0.13 0.27 0.13
2000 0.26 0.33 0.14 0.30 0.14
2001 0.26 0.33 0.14 0.31 0.19
2002 0.25 0.31 0.15 0.30 0.17
2003 0.25 0.30 0.18 0.32 0.16
2004 0.28 0.31 0.20 0.31 0.17
2005 0.28 0.32 0.20 0.32 0.16
Note: Following procedure is used to calculate bold rates: 1.The output value in output-input tables is updated by using
industrial production index. 2. Exports, which include both goods and services, are obtained from OECD statistics web site.
Cross country comparisons show that while the average exports to total supply ratio for the
overall economy stayed the same in Turkey from 1998 to 2005 (except 2001 crisis period), it slightly
increased in Hungary, Czech Republic, Poland and Slovakia, indicating an increase in share of
exports in total supply for Eastern European new EU member countries. Table 3.14 suggests that
import dependency of exports in Czech Republic, Hungary, Poland and Slovakia increased more than
that of Turkey. It is also worth to note that even if the rate of import dependency is almost the same
in Poland and Turkey; the rate had increased much faster in Poland during the 1998-2005 period.
Table 3.13: Ratio of Exports to Total Supply (in purchasing prices, 1998-2005)
Czech Rep. Hungary Poland Slovakia Turkey
1998 0.18 0.20 0.12 0.20 0.14
1999 0.19 0.21 0.11 0.21 0.12
2000 0.20 0.25 0.12 0.23 0.12
2001 0.21 0.25 0.12 0.23 0.16
2002 0.20 0.23 0.13 0.22 0.14
2003 0.20 0.23 0.15 0.24 0.13
2004 0.22 0.23 0.16 0.24 0.14
2005 0.22 0.24 0.16 0.24 0.13
Note: Following procedure is used to calculate bold rates: 1. The last output value in output-input tables is updated by
using industrial production index. 2. Exports, which include both goods and services, are obtained from OECD statistics
web site.
3.c.ii. Manufacturing Industry
Sectoral export to production ratio is presented in Table 3.15. First of all, the ratio of exports
to production for total manufacturing industry has been in an increasing trend in Turkey since 1998.
Meanwhile, the manufacturing exports to production ratio has been increasing in Czech Republic,
Hungary, Poland and Slovakia, as well. Indeed, among these economies exports to production ratio
for the total manufacturing industry is the lowest in Turkey during the 1998-2005 period, implying

29
greater share of production for domestic consumption in Turkey. This is not a surprising outcome as
Turkey is being the largest economy in our sample of countries.
Table 3.14: Difference Between the Change in Export to Total Production and Export to Total
Supply with Respect to 1998
Czech Rep. Hungary Poland Slovakia Turkey
1998 .. .. .. .. ..
1999 0.003 0.003 -0.002 0.002 -0.005
2000 0.012 0.032 0.000 0.012 -0.002
2001 0.013 0.028 -0.001 0.021 0.006
2002 0.007 0.019 0.001 0.014 0.002
2003 0.008 0.016 0.007 0.020 0.000
2004 0.019 0.021 0.013 0.018 0.004
2005 0.018 0.025 0.011 0.024 0.002
Note: Calculated from Table 3.12 and 3.13.
Secondly, share of exports in total production tend to increase in all manufacturing sectors
during the 1998-2005 period, in Turkey. Share of exports in production is the highest in textiles and
wearing sectors followed by motor vehicles. More than a half of production is devoted to exports in
these sectors. Meanwhile, in addition to textile and wearing apparels, some capital-intensive sectors
such as coke, motor vehicles and semi-trailers, electrical machinery and apparatus, and machinery
and equipments increased their production for exports rapidly after 2001. Comparing exports to
production ratios, there is a shift in the exports performance across the sectors in favor of capital-
intensive sectors. These findings are also consistent with Yükseler and Türkan (2006). Cross-country
comparisons reveal that export share of capital-intensive sectors tends to increase in our sample of
countries after 2001.
Furthermore, analysis of export to total supply across the manufacturing sectors in Table
3.16 shows that Turkey has the lowest exports to total supply ratio for the manufacturing industry
compared to the other countries in our sample. While other countries use about 30-40 percent of their
total manufacturing supply for exports, the rate fluctuates around 20 percent in Turkey. The ratio
tends to increase in all countries and increase is the highest in Czech Republic and Poland. When we
compare the change in export to total production to the change in export to total supply for
manufacturing industries across these countries we observe that increase in import dependency of
manufacturing exports is the highest in Slovakia in 2004 and Czech republic in 2005 (Table 3.17).
Investigating the sub-sectors of the manufacturing industry in terms of the import
dependency rate, we found high difference between exports to total supply and exports to production
ratios in electrical machinery, motor vehicles, textile and wearing apparels, and machinery and

30
equipment in Turkey, implying relatively high import dependency of exports in these sectors (Table
3.17). Cross country analysis shows that the difference between exports to total supply and exports to
production ratio is higher mostly in textile, chemical and basic metal in most of the countries in our
sample. Hungary, similar to Turkey, had an increase in the use of imported products in sectors of
machinery and equipment, and electrical machinery and apparatus.

31
Table 3.15: Sectoral Exports to Production Ratio
Note: Following procedure is used to calculate bold rates: 1. The last output values from the input- output tables are updated by using sectoral industrial production index. 2.Sectoral
exports are obtained from OECD statistics web site.
Manuf.
Industry
Food prod.,
beverages &
tobacco
Textiles,
Wearing appeals
& furs
Coke, refined
petroleum prod. &
nuclear fuels
Chem., chem..
prod. & man-made
fibers
Basic metals Fabricated metal
p
rod. except mach.
& equip.
Machinery &
equipment
Electrical
machinery &
apparatus
Motor vehicles,
trailers & semi-
trailers
Turkey
1998-2001 0.22 0.09 0.54 0.07 0.14 0.30 0.12 0.20 0.23 0.31
2002-2003 0.33 0.10 0.70 0.11 0.18 0.39 0.27 0.37 0.33 0.59
2004 0.37 0.11 0.72 0.15 0.20 0.41 0.37 0.33 0.49 0.53
2005 0.39 0.12 0.85 0.24 0.19 0.39 0.31 0.38 0.47 0.55
Czech Rep
1998-2001 0.50 0.16 0.66 0.25 0.56 0.48 0.42 0.71 0.63 0.63
2002-2003 0.55 0.16 0.68 0.27 0.61 0.51 0.44 0.84 0.69 0.64
2004 0.61 0.19 0.78 0.21 0.68 0.57 0.48 0.95 0.76 0.68
2005 0.72 0.21 0.81 0.24 0.73 0.65 0.47 0.93 0.73 0.64
Hungary
1998-2001 0.60 0.25 0.72 0.25 0.57 0.55 0.40 0.63 0.84 0.88
2002-2003 0.66 0.22 0.83 0.32 0.62 0.59 0.40 0.82 1.16 0.88
2004 0.73 0.26 0.85 0.41 0.76 0.65 0.42 0.98 1.27 0.90
Poland
1998-2001 0.33 0.11 0.73 0.18 0.28 0.42 0.27 0.34 0.50 0.44
2002-2003 0.46 0.12 0.79 0.17 0.29 0.50 0.30 0.46 0.63 0.55
2004 0.56 0.16 0.84 0.32 0.36 0.72 0.32 0.56 0.64 0.56
Slovakia
1998-2001 0.68 0.18 1.16 0.47 0.84 0.71 0.50 0.77 0.90 0.94
2002-2003 0.75 0.25 0.88 0.64 0.79 0.68 0.61 0.84 0.88 1.05
2004 0.85 0.30 0.96 0.81 0.92 0.91 0.59 0.87 0.93 1.13

32
Table 3.16: Sectoral Exports to Total Supply Ratio
Manuf.
Industry
Food prod.,
beverages &
tobacco
Textiles,
Wearing appeals
& furs
Coke, refined
petroleum prod. &
nuclear fuels
Chem., chem..
prod. & man-made
fibers
Basic metals Fabricated metal
p
rod. except mach.
& equip.
Machinery &
equipment
Electrical
machinery &
apparatus
Motor vehicles,
trailers & semi-
trailers
Turkey
1998-2001 0.16 0.08 0.46 0.06 0.07 0.17 0.10 0.09 0.14 0.17
2002-2003 0.22 0.09 0.54 0.09 0.08 0.17 0.22 0.17 0.17 0.32
2004 0.22 0.10 0.54 0.13 0.08 0.17 0.29 0.15 0.18 0.26
2005 0.23 0.11 0.60 0.19 0.08 0.16 0.25 0.16 0.15 0.27
Czech Rep
1998-2001 0.33 0.14 0.40 0.16 0.28 0.30 0.33 0.39 0.40 0.44
2002-2003 0.35 0.13 0.39 0.17 0.27 0.31 0.33 0.45 0.44 0.44
2004 0.39 0.15 0.42 0.14 0.30 0.34 0.36 0.51 0.47 0.46
2005 0.43 0.17 0.42 0.16 0.32 0.36 0.36 0.54 0.46 0.46
Hungary
1998-2001 0.37 0.22 0.43 0.20 0.28 0.30 0.25 0.27 0.49 0.54
2002-2003 0.39 0.20 0.45 0.23 0.31 0.32 0.25 0.38 0.61 0.52
2004 0.42 0.22 0.45 0.27 0.37 0.31 0.25 0.45 0.59 0.51
Poland
1998-2001 0.24 0.10 0.49 0.15 0.14 0.30 0.21 0.17 0.32 0.27
2002-2003 0.31 0.11 0.47 0.14 0.14 0.31 0.23 0.22 0.40 0.33
2004 0.54 0.16 0.79 0.31 0.33 0.68 0.31 0.52 0.61 0.54
Slovakia
1998-2001 0.41 0.14 0.55 0.41 0.41 0.52 0.33 0.34 0.45 0.52
2002-2003 0.41 0.20 0.49 0.53 0.38 0.48 0.34 0.38 0.46 0.53
2004 0.48 0.23 0.49 0.62 0.38 0.61 0.34 0.43 0.50 0.60
Note: Following procedure is used to calculate bold rates: 1. The last output values from the input- output tables are updated by using sectoral industrial production index.
2. Sectoral exports are obtained from OECD statistics web site.

33
Table 3.17: Difference Between the Change in Sectoral Export to Total Production and Export to Total Supply with Respect to 1998-2000
Manuf.
Industry
Food prod.,
beverages &
tobacco
Textiles,
Wearing
appeals &
furs
Coke, refined
petroleum
prod. &
nuclear fuels
Chem.,
chem.. prod.
& man-made
fibers
Basic metals Fabricated
metal prod.
except mach.
& equip.
Machinery &
equipment
Electrical
machinery &
apparatus
Motor
vehicles,
trailers &
semi-trailers
Turkey
2002-2003 0.044 0.002 0.075 0.005 0.027 0.090 0.034 0.101 0.075 0.134
2004 0.081 0.004 0.105 0.018 0.043 0.109 0.061 0.078 0.218 0.135
2005 0.095 0.006 0.169 0.038 0.042 0.108 0.041 0.126 0.221 0.141
Czech Rep
2002-2003 0.022 0.000 0.023 0.002 0.060 0.025 0.010 0.073 0.017 0.012
2004 0.057 0.010 0.097 -0.020 0.102 0.060 0.026 0.130 0.057 0.032
2005 0.119 0.016 0.123 -0.010 0.128 0.122 0.017 0.075 0.038 -0.003
Hungary
2002-2003 0.038 -0.001 0.090 0.043 0.025 0.026 0.012 0.083 0.190 0.010
2004 0.080 0.014 0.114 0.090 0.109 0.088 0.018 0.171 0.332 0.039
Poland
2002-2003 0.051 0.001 0.079 0.003 0.006 0.073 0.006 0.072 0.048 0.054
2004 -0.076 -0.007 -0.194 -0.02 -0.11 -0.081 -0.055 -0.124 -0.154 -0.147
Slovak
2002-2003 0.062 0.008 -0.226 0.047 -0.024 0.013 0.099 0.031 -0.036 0.098
2004 0.098 0.028 -0.143 0.134 0.103 0.118 0.082 0.01 -0.027 0.096
Note: Calculated from Tables 3.15 and 3.16.

34
3.d. Competitiveness Indicators and Integration of Turkey to the World Markets
3.d.i. Export Market Shares of Turkey and Emerging Economies
Historically, there are two milestones in the export market share of Turkey in the world: the
early 1980s and 2001-2006 period (Table 3.18 and Figure 3.8). After the failure of the import
substitution policy in boosting exports during 1970s, the export market share increased rapidly as a
result of the liberalization process and export subsidy policy in the 1980s. In value terms, Turkey
accounted for 0.36 percent of world exports of goods during the 1980-1990 period doubling the
1970s figures. While the export market share of Turkey remained relatively stable, displaying only a
small increase over the 1991-2000 period, it experienced decent performance over the 2001-2006
period. However, the structures of these two successful periods are quite different. The success in the
1980s was mainly due to macroeconomic policies that not only intentionally led the Turkish lira to
depreciate but also allowed certain privileges, such as enormous amount of subsidies to export-
oriented firms. Therefore, we may say that the success in export performance during 1980s was
basically policy-driven. In other words, the export-oriented firms had gained artificial price
competitiveness through the application of these policies. On the other hand, the success in export
performance during the 2001-2006 period was not policy-driven. Policymakers had no intention to
boost the exports by depreciating the Turkish Lira. In this period, the export-oriented firms suffered
from loss of price competitiveness. However, firm-driven factors (technological progress, integration
to the world export market, attaching importance to quality…etc) compensated the detrimental effect
of real currency appreciation.
Figure 3.8: The Share of Turkish Exports in the World Total
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
Import substitution policy period Liberalization and export
subsidy policy period
Floating
exchange rate
p
eriod
Source: WTO, TURKSTAT

35
Table 3.18: The Share of Turkish Exports in the World Market
World Export (billion US
dollar)
Turkish Export (billion
US dollar) Share (%)
1990 3,493.6 13.0 0.37
1991 3,506.2 13.6 0.39
1992 3,760.3 14.7 0.39
1993 3,774.7 15.3 0.41
1994 4,313.3 18.1 0.42
1995 5,168.9 21.6 0.42
1996 5,397.5 23.2 0.43
1997 5,577.5 26.3 0.47
1998 5,493.8 27.0 0.49
1999 5,705.9 26.6 0.47
2000 6,435.7 27.8 0.43
2001 6,177.4 31.3 0.51
2002 6,465.2 36.1 0.56
2003 7,490.3 47.3 0.63
2004 8,975.6 63.2 0.70
2005 10,120.0 73.5 0.73
2006* 11,334.4 85.5 0.75
1951-1960 100 0.3 0.33
1961-1970 206 0.5 0.22
1971-1980 1,027 1.7 0.17
1981-1990 2,380 8.5 0.36
1991-2000 4,913 21.4 0.44
2001-2006 7,378 44.4 0.60
* Estimate
Source: WTO, TURKSTAT
When compared with the emerging economies, although there were significant increases in
the emerging market countries’ exports over the 2001-2004 period, Turkey had experienced a
significant rise in its market share in the emerging markets over the whole period (Table 3.19).
Table 3.19: The Share of Turkish Exports in the Emerging Markets
Emerging Market Export
(billion USD)
Turkish Export
(billion USD) Share (%)
1993 645.1 15.3 2.38
1994 769.0 18.1 2.35
1995 936.6 21.6 2.31
1996 987.8 23.2 2.35
1997 1050.3 26.3 2.50
1998 994.3 27.0 2.71
1999 1,027.3 26.6 2.59
2000 1,249.1 27.8 2.22
2001 1,209.1 31.3 2.59
2002 1,314.4 36.1 2.74
2003 1,609.2 47.3 2.94
2004 2,073.6 63.2 3.05
1993-1996 834.6 19.6 2.35
1997-2000 1,080.3 26.9 2.49
2001-2004 1,551.6 44.5 2.87
Source: UNCTAD, TURKSTAT

36
3.d.ii Relative Export Performance of Individual Products: Turkey and Other Emerging
Economies
This section aims to find out the recent relative export performance and competitiveness of
emerging economies (including Turkey) in individual products. A country’s export performance in
individual products relative to world exports is considered as a good indicator of its competitiveness
in those products.
Disaggregated 3-digit Standard International Trade Classification (SITC) data is used to compute
export performance statistics for Turkey and emerging economies. The UNCTAD database (2006)
provides the three-digit SITC product code of annual exports comprising 239 types of products. The
basic analysis is conducted according to the growth rate of each product in 20 emerging economies.
In this framework, if the growth rate of the export value of the product i in country j is higher than
the growth rate of the world export value of that commodity, then country j is considered as highly
competitive in that commodity. Reverse is also true:
(2001 2004 ) (1990 2000 )
(2001 04 ) (1990 2000 ) 1 1
(1990 2000) (1990 2000)
1
nn
ij ij
ij ij j j
n
ij ij
j
XX
XX
XX
−−
−−= =
−−
=
−
−>
∑∑
∑⇒ Highly competitive
(2001 2004 ) (1990 2000 )
(2001 04 ) (1990 2000 ) 1 1
(1990 2000) (1990 2000)
1
nn
ij ij
ij ij j j
n
ij ij
j
XX
XX
XX
−−
−−= =
−−
=
−
−<
∑∑
∑ ⇒ Less competitive
where subscript i denotes products and subscript j denotes countries. Left hand side of these
inequalities refers to growth rate of product i in country j and the right hand side refers to growth rate
of product i in the total world market. The products with export value less than 1 million USD are
omitted from this analysis.
Analyzing Table 3.20, the emerging economies as a whole showed successful performance in
the 2001-2004 period. The share of highly competitive products in total exports is greater than fifty
percent in all 20 emerging economies, which implies sound performance of emerging economies
relative to world export market. Turkey is one of the best among the emerging economies in terms of
both the number of highly competitive products and the share of these products in total exports.
Turkey has 164 of 239 products with the growth rate higher than world export and they account 88.1
percent of total export.

37
Table 3.20: Competitiveness of Emerging Economies in the World Market (2001-2004)
Rank Countries
Number of Prod. with
High Level of
Competitiveness Share
Number of Prod.
with Low Level of
Competitiveness Share
1 China 199 98.0 30 2.0
2 Czech Republic 152 91.5 64 8.5
3 Chile 110 91.4 64 8.6
4 Poland 174 90.6 46 9.4
5 Turkey 164 88.1 42 10.9
6 Hungary 135 88.0 66 11.0
7 Venezuela 67 86.8 83 13.1
8 India 182 85.0 35 15.0
9 Philippines 99 83.7 73 16.3
10 Mexico 138 82.6 74 17.4
11 Argentina 108 79.7 97 20.3
12 Brazil 137 78.2 81 21.8
13 Russia 119 77.8 97 22.2
14 Bulgaria 131 77.4 52 22.6
15 Korea 130 76.3 75 23.7
16 Malaysia 141 73.9 72 26.1
17 Thailand 150 68.1 63 31.9
18 Indonesia 156 61.1 58 38.9
19 Hong Kong 74 57.9 141 42.1
20 Singapore 72 53.5 138 46.4
Source: UNCTAD and our calculations
Table 3.21: Competitiveness of Emerging Economies in the Total Emerging Market
(2001-2004)
Rank Countries
Number of Prod. with High
Level of Competitiveness Share
Number of Prod. with
Low Level of
Competitiveness Share
1 China 196 97.1 33 2.9
2 Poland 160 86.3 61 13.7
3 Venezuela 50 84.9 101 15.1
4 Turkey 151 84.7 55 14.3
5 Chile 79 82.2 96 17.3
6 India 168 81.6 50 18.4
7 Hungary 105 80.6 98 19.3
8 Philippines 82 80.0 92 20.0
9 Czech Republic 113 75.4 104 24.5
10 Bulgaria 118 73.2 65 26.5
11 Russia 87 71.5 130 28.5
12 Brazil 100 63.6 119 36.4
13 Korea 100 61.3 106 38.7
14 Argentina 69 57.3 136 42.7
15 Indonesia 137 54.7 78 45.3
16 Mexico 110 49.9 104 50.1
17 Singapore 50 42.8 160 57.2
18 Thailand 111 41.3 103 58.8
19 Malaysia 107 38.5 106 61.5
20 Hong Kong 37 37.1 178 62.9
Source: UNCTAD and our calculations
Table 3.21 limits the analysis to total emerging export markets rather than world economy in
order to assess the relative competitive position of Turkey more accurately. Turkey is again among
the best performing emerging economies in terms of the number of highly competitive products and

38
the share of these products in total exports. Turkey has 151 of 239 products with the growth rate
higher than the total emerging export and they accounted for 84,7 percent of total exports. These
results together with the jump in high–technology-intensive product exports signal fast integration of
the Turkish economy to the world export markets.
In order to sustain integration to the world markets, countries need to introduce new products
into the world market’s “rising-star” product spectrum. Table 3.22 shows the number of emerging
economies’ competitive products that penetrates into top 10 and 25 export product which performs
the highest increase in the world market. In this context, among Turkey’s major competitors Turkey
ranks the second, just behind China, in the number of the competitive products in the top 25 list.
Moreover, between 2000 and 2004 Turkey introduced 8 “rising-star” products from the top 10 list to
the world markets.
Table 3.22: Competitiveness of Emerging Economies (2001-2004)
Rank Countries
Number of Competitive
Prod. in Top 10
Number of Competitive
Prod. in Top 25
1 China 7 21
2 Turkey 8 17
3 Czech Republic 7 17
4 Poland 5 17
5 Malaysia 7 15
6 Indonesia 7 15
7 Mexico 9 15
8 India 6 14
9 Hungary 6 13
10 Brazil 4 12
11 Singapore 5 12
12 Thailand 6 12
13 Argentina 3 11
14 Russia 3 11
15 Hong Kong 5 11
16 Chile 5 10
17 Korea 4 10
18 Bulgaria 3 9
19 Venezuela 4 8
20 Philippines 3 8
Source: UNCTAD and our calculations.

39
3.d.ii. The Revealed Comparative Advantage (RCA) Analysis of Turkey with Respect to
Factor Intensity
The purpose of this section is to identify the Turkish industries that have revealed
comparative advantages in exports. Export specialization index of Turkey has been calculated in
terms of factor intensity measured by the revealed comparative advantage (Balassa) index over the
1994–2004 period. Balassa’s RCA index (Balassa 1977), which compares the export share of a given
sector in a country with the export share of that sector in the world and emerging market, is used for
analyzing the degree of specialization (comparative advantage) quantitatively.
1
111
ij ij
i
ij ij ij
jij
XX
RCA XX
=
===
=
∑
∑
∑∑
The numerator represents the share of a given sector in national exports, where ij
X
is the exports of
sector i from country j;
1
ij
iX
=
∑is the total exports of country j. The denominator represents the share
of a given sector in the world/emerging market exports, where
1
ij
jX
=
∑
is the world/emerging market
exports of sector i, and
11
ij
ij
X
==
∑
∑are the world/emerging market exports. Thus, when RCA is above 1
then the country is said to have a relative comparative advantage, in that sector, put differently,
country is specialized in that sector. When RCA is below 1 then (ranging from 0 to 1) the country is
said to have a relative disadvantage in that sector.
Table 3.23. Revealed Comparative Advantage Index and Percentage Change Between
1994–2000 and 2001–2004 (World)
Type of Exports RCA index
1994-2000 average
RCA index
2001-2004 average
Index change
(%)
High-technology (R&D) intensive Exp. 0.25 0.40 59.1
Resource-intensive Exp. 0.32 0.29 -11.3
Labor-intensive Exp. 4.97 4.99 0.4
Capital-intensive Exp. 1.78 2.08 17.0
Agriculture-intensive Exp. 2.20 1.60 -27.1
Source: Own compilations and calculations on the basis of UNCTAD data.
Tables 3.23 and 3.24 contain the RCA indices for Turkey in the 1994–2000 and 2001-2004
periods, as well as percentage changes in those indices over those sub-periods. They also indicate
Turkey’s relative competitive position in the world and the emerging markets, respectively. For the
2001-2004 period, the RCA indices for labor, capital and agriculture intensive exports were greater
than 1. In other words Turkey had strong revealed comparative advantage with respect to these
sectors both in world and emerging markets.

40
On the other hand, structural shift in the specialization of product groups in terms of factor intensity
of Turkey was observed for the 2001-2004 period compared to the 1994-2000 period. Even though
the RCA index for high-technology intensive exports is less than unity, there has been a remarkable
upward trend towards specialization in those commodities. The average RCA index of Turkey high-
technology intensive exports was 0.4 for the 2001-2004 period in the world market, which implies
59.1 percent increase from its 1994-2000 period value. In the same vein, the RCA index for
agriculture-intensive exports decreased by 27.1 percent in the said period. In other words, Turkey
succeeded to decrease its competitive disadvantages in high-technology-intensive products after
2001, while lost its comparative advantage in the agriculture-intensive products. Concerning the
capital-intensive products, Turkey raised its competitiveness significantly. We reached to similar
results when we compute RCA index for emerging markets too.
Table 3.24. Revealed Comparative Advantage Index and Percentage Change Between
1994–2000 and 2001–2004 (Emerging Markets)
Type of Exports RCA index
1994-2000 average
RCA index
2001-2004 average
Index change
(%)
High-technology (R&D) intensive Exp. 0.30 0.46 56.4
Resource-intensive Exp. 0.30 0.30 0.2
Labor-intensive Exp. 2.87 2.97 3.5
Capital-intensive Exp. 2.19 2.52 15.2
Agriculture-intensive Exp. 2.29 1.85 -19.2
Source: Own compilations and calculations on the basis of UNCTAD data.
Table 3.25. Revealed Comparative Advantage Index for Selected Emerging Economies
(2001-2004, average)
High-technology
(R&D) intensive Exp.
Resource-intensive
Exp.
Labor-
intensive Exp.
Capital-
intensive Exp.
Agriculture-
intensive Exp.
Turkey 0.40 0.29 4.99 2.08 1.60
Czech Rep. 0.91 0.45 1.37 1.93 0.46
Bulgaria 0.30 0.99 3.09 1.57 1.24
Romania 0.34 1.01 3.48 1.27 0.44
Malaysia 1.41 0.93 0.57 0.28 0.36
Argentina 0.25 2.04 0.23 1.05 5.60
Chile 0.07 2.07 0.14 3.90 4.12
Brazil 0.45 1.59 0.43 1.70 3.44
Source: Own compilations and calculations on the basis of UNCTAD data.
Table 3.25 displays RCA index calculations for major emerging market trading partners in
terms of factor intensity. As far as the high-technology products concerned, Turkey has a noticeable
comparative advantage over new EU member countries such as Bulgaria and Romania and Latin
American countries such as Chile and Argentina. As it is expected, Turkey has comparative
disadvantage in high technology products compared to East Asian emerging economies.

41
4. Empirical Analysis: Time Varying Parameter Estimates for Export Supply and
Demand Functions
This chapter conducts an empirical analysis to estimate export demand and supply functions
for Turkey. The primary purpose of the analysis is to investigate if the coefficients of the export
functions exhibited a noticeable fluctuation over time that can be interpreted as structural changes in
export elasticities. Our focus will be on long run export demand and supply functions of Turkey.
These functions will be estimated separately in order to examine sources of potential structural
changes. In the case of parameter instability Kalman Filter approach is better approach than classical
regression methods since it allows one to estimate time varying coefficients. However, before
proceeding with Kalman filter analysis standard long-run procedures must be performed to ensure
whether there is a well-defined relationship among the specified variables in the supply and demand
equations or not.
In the literature a standard export supply equation is defined as a function of competitiveness
indicators such as relative prices, unit labor cost, effective exchange rate and scale variables such as
domestic output and output gap, as well as some form of import constraint variable such as imported
raw materials. On the other hand, a standard export demand equation is specified as a function of
competitiveness indicators and foreign income.7 After examining the stationarity properties of each
variable, we proceed with co-integration tests over different vector of variables to find out a well-
defined long-run export supply and demand equations for Turkey, for the 1987q1-2006q4 period. In
both supply and demand equations, we used export quantity index, which is taken from the Central
Bank of the Republic of Turkey (CBRT) website, as a dependent variable. Unit labor cost based real
effective exchange rate (REER_ulc) is included as a measure of competitiveness in the supply
function due to the fact that it reflects the cost of production better than the consumer price index
based real effective exchange rate (REER_cpi). Indeed, our analysis suggests that among the price
competitiveness indictors while REER_ulc produced theoretically and statistically better results in
the export supply function, REER_cpi worked better in the export demand function. Both of these
indicators, which include currencies of the 34 countries, are taken from Eurostat. In addition, we
include import quantity index, which is taken from CBRT website, in the long run supply equation as
a measure of import dependency of exports and found that it is a significant variable in determining
the long-run supply function. However, even if the actual and potential outputs are alternative
7 Interested readers may refer to Nowak-Lehman (2004) and Muscatelli et al (1995) for alternative export
function specifications.

42
measures of physical productive capacity we found poor evidence for their inclusion in the long-run
supply function. We also avoid using import and output variables in the same equation due to
endogeneity problem. For the export demand equation, we preserve the standard set up and in
addition to REER_cpi, we include foreign income in current prices and current PPP for the OECD
countries, which is taken from OECD website.
4.a. The Model
As a result, the following equations were estimated to analyze export demand and supply
functions respectively for Turkey.
01 2 3142531
_
ttttt tt t t t
x
smreerulcsssu
β
ββ βββ
=+ + + + + + (4.1)
01 2 3142532
_
f
tttt t tt t t t
x
d y reer cpi a s a s a s u
αα α
=+ + + + + + (4.2)
Here, t
x
s and t
x
d are the log of export quantity index;
f
t
y
is the log of foreign income; t
m is the
log of import quantity index, _t
reer cpi and _t
reer ulc are the log of REER_cpi and REER_ulc
respectively,
j
s (j=1,2,3) are seasonal dummies, i
α
’s and i
β
’s, where i= 0, 1…5, are the coefficients
of the explanatory variables. Equation (4.1) represents an export supply function that is determined
by imports quantity index and real effective exchange rate. Equation (4.2) is a typical export demand
function that relates exports to foreign income and real effective exchange rate. Since quarterly data
is used in the analysis seasonal dummies are added to deal with seasonality.
4.b. The Method: Kalman Filter Approach
Kalman filter approach or state space models developed by Kalman (1960, 1963), has been
used extensively in economics. The Kalman filter is a recursive algorithm for expressing dynamic
systems that involve unobserved state variables, conditional on observed vector (Kim and Nelson,
2000). A detailed discussion of Kalman Filter approach is presented in the appendix. Theoretical
explanation of the model can be found in the appendix. Application of a state space model on our
question is explained in this section. The model consists of two equations. The first equation is called
transition equation that describes the dynamics of the state variables. The second equation is the
measurement equation which points out a relationship between observed variables and unobserved
state variables. Since this paper is interested in analyzing how model parameters change over time,
we assumed that all parameters of the equation (4.1) and (4.2) follow a random walk process. Then
the transition equations for the demand and supply functions are:

43
12it it t
β
βε
+=+ i=0,1…5, (4.3)
11it it t
α
αε
+=+ i=0,1...5, (4.4)
here t
ε
is normal white noise processes. Then, the measurement equations can be written as
1
*'
ttt t
x
sFBu=+, (4.5)
2
*'
tttt
x
dHAu=+, (4.6)
where,
[
]
123
1_
tt t
Fmreerulcsss=,123
1_
f
tt t
H y reer cpi s s s
⎡
⎤
=
⎣
⎦,
[
]
012345t tttttt
B
β
ββββ β
=,
[
]
012345t tttttt
A
α
αααα α
=, 1(0,1)
t
uiidN
and 2(0,1)
t
uiidN. We also include seasonal dummies (sj) in both equations. Kalman Filter
approach is a recursive process that updates the estimated coefficients over time as new information
arrives. However, this may reduce reliability of estimated coefficients especially at the beginning of
the period. Instead, we may apply a fixed-point Kalman smoother, which gives the estimated value of
the state variable at time t based on all available information up to time T, where Tt>. The idea is
that as new data become available, we can improve our estimation result from the Kalman filter by
taking into account the additional information.
4.c Results
As a first step, unit root properties of the each variable are analyzed and found that all are
I(1).8 Then, co-integration analysis is conducted and found that there is a well-defined long-run
relationship among the vector of variables that are defined in each equation. Smoothed Kalman filter
estimation results based on the long-run relationship are presented in Figure 4.1.
Estimated coefficients for the export supply function are presented in the first column of
Figure 4.1. 1994, 1997-8, and 2001 crises and 1996 customs union are marked on these graphs as
shaded areas. Smoothed Kalman filter estimates show that intercept term can be considered as
constant through out the period in both supply and demand equations. On the other hand, other
parameter values vary over time. Both import and income elasticity, as well as the real effective
exchange rates have an upward trend. The rate of increase in the parameters accelerated right after
the CU in 1996. The trend in import elasticity of export supply was disrupted temporarily three times
in 1994, 1998 and towards the end of 2000. While the crises in 1994 and 2000-2001 had an upward
8 Unit root tests were not presented in the paper in order to keep the paper as short as possible. However, these
test results could be provided to an interested reader.

44
jump affect, 1998 had a downward effect. However, neither of these shocks had path-breaking
impact, albeit, the shock in 2000 had a level shifting impact on import elasticity of export supply. As
a result, it is estimated that the import elasticity of exports increased about 50 percent from 1987 to
2006.
Meanwhile, the responsiveness of the export supply to the changes in REER_ulc steadily
decreased during the 1987-2006 period, indicating smaller sensitivity of export supply to the shocks
in REER_ulc. The decrease in REER_ulc elasticity is estimated to be more than 50 percent.
Similarly, we may clearly identify three different jumps in this trend in 1994, 1997 and 2001. The
shock in 1994 changes the direction in REER_ulc elasticity between 1994 and 1997, such that export
supply becomes more elastic. However, after the CU the elasticity starts to decrease again. As in the
case of import elasticity, these shocks, which are originated from the domestic economy, had
temporary impacts, thus the REER_ulc elasticity of exports continues to fall during the 1987-2006
period.
On the right hand side column, parameter estimates for the export demand equation are
presented. As in the case of export supply function, elasticity parameters of the demand equation are
time variant. Income elasticity of exports, which was roughly constant during the 1987-1997 period,
gained an increasing trend a year after the CU, notwithstanding to the 1998, 1999 and 2000-2001
crises. Thereby, by the end of the period, the income elasticity increased by more than 50 percent
compared to its initial value. Beside, oscillations of import elasticity of supply are deeper than supply
function; therefore we may say that supply is affected more from external shocks than the demand
function during the 1987-2006 period.
REER_cpi elasticity, which showed a slight decrease during the 1987- 1997 period, followed the path
of income elasticity closely, afterwards. The decrease in REER_cpi elasticity of export demand is 70
percent throughout the sample period. As a result, even though REER_ulc elasticity of export supply
depicted a more volatile path than the same coefficients of the export demand function, the decrease
in sensitivity of export demand to the changes in REER_cpi was much higher than export supply
function. Residuals of the estimated equation is presented in panel d) and they dont show any
systematic error. However, all crises periods jumped the residual upward indirectly showing
their effect on the estimated coefficients. It also worth to note that residual volatility of
export demand and supply functions get smaller after 2001, which may indicate decrease in
uncertainty in economic environment due to prudent fiscal and monetary policies.

46
5. Conclusion
This study investigates the structural change in the Turkish exports. The analysis consists of
three broad sections. In the second section, performance of the Turkish economy is examined in
various perspectives, such as export and import performance, and developments in competitive
indicators compared to some of Turkeys’ trade partners. The focus of the third section is on both
changes in commodity and country composition of Turkish exports. Import dependency and
competitiveness of Turkish exports are also examined and compared to some of those emerging
markets. The final section conducts an empirical analysis and applies smoothed Kalman filter
approach to estimate demand and supply functions with time varying parameters to examine how
parameter values change during the 1987-2006 period. The outcomes of each section can be
summarized as follows:
Second Section:
1. With the liberalization of the Turkish economy at the beginning of 1980s, trade openness
rate had shown an increasing trend. Meanwhile, a similar trend is observed for other
countries, as well. Cross country comparison revealed that small economies tend to
increase their openness rate at relatively higher pace than large economies, and Turkey,
in this sense, is not an exception.
2. Increase in openness rate was followed by fast economic growth. However, the fast
growth raised the problem of trade and current account deficits. Cross-country
comparison showed that countries in the EU accession process also had and still have
large trade deficits. Indeed, for those countries, trade deficit in recent years is smaller
than their earlier years of their EU candidacy, implying that any potential negative effect
of EU accession process was temporary for few of those new members.
3. The average growth rate of GDP was about 4.2 percent between 1980 and 2006 in
Turkey. Financial and currency crises created instability in 1990s but growth rate got
back to its increasing trend afterwards.
4. Regarding competitiveness indicators, Turkish Lira appreciated in real terms during the
2002-2006 period, however, the same trend was also the case in most of the other
countries. The decline in the real wages in Turkey was much more evident compared
with the other trade partners. The positive effect of the decline in real wages
compensated the negative effect of real appreciation of Turkish Lira on competitiveness
on Turkish exports. As a result, external competitiveness indicators provided evidence in
favor of gaining competitiveness power of Turkish exports in the world export market.

47
5. Overall contribution of the rise of quantity of exports and imports to the value of exports
and imports increased during the 1994-2006 period, though after 2003 this trend slowed
down. Investigation of price and quantity indexes across the manufacturing sectors
shows that terms of trade improved after 2003 (particularly for food and beverages, metal
industry, machinery and equipment, electronics, motor vehicles together with furniture),
while real increase in exports was less than imports (except for chemicals, plastic and
rubber, basic metals and motor vehicles).
Third Section:
6. Post 2001 period witnessed a change in export commodity composition in favor of more
capital and technology intensive commodities. In general fast growing sectors are
relatively new commodities that are not considered as the traditional Turkish export
commodities. However, it is shown that even though there is a process of transformation
in the Turkish exports after the crises period, Turkey is still short of building any
comparative advantage in these new commodities.
7. Three measures of commodity concentration ratios, distribution of normalized exports by
commodity groups, weighted spread of exports by commodity groups and share of top 10
and 20 commodities in total exports, showed that commodity concentration of exports
increased after 2001, thanks to the expansion of exports of new non-traditional
commodities.
8. The rising industries of the post crises period are considered as relatively more capital
and high technology intensive commodities compared to the popular industries of 1980s
and 1990s. These new industries also have high intra-industry trade.
9. The classification of exports in terms of factor intensity reveals that concentration occurs
in the high technology products in the world. Although the share of R&D-intensive
product exports was below the world and emerging market averages, Turkey ranked the
first in the growth of R&D product exports among the emerging market economies in the
2001-2004 period.
10. Both measures of the country concentration ratios; weighted spread of Turkish exports
by countries and share of top 10 and 20 countries in total exports, indicate an increasing
trend in country concentration of Turkish exports. On the other hand, the share of
Turkish exports in the world market has been increasing since 1980s. Indeed, compared
to the emerging markets, performance of the Turkish export is striking.
11. Analysis of exports shows that the high import dependence of overall Turkish exports is
not exceptional. Indeed, import dependency of exports is higher in new EU members as
well. It is found that the rate of dependency which was almost the same in Turkey and
Poland during the 1998-2001 period, increased much faster in Poland in recent years.

48
Sectoral analysis shows that import dependency rate increased much faster in the Turkish
manufacturing sectors, particularly in motor vehicles as well as electrical machinery and
apparatus sectors. Examination of the indicators across the manufacturing sectors for the
new EU member countries shows that the rate of increase is relatively high in Turkey
and Slovakia. Considering the cross sectoral development we observe relatively high
import dependency rate in motor vehicles and electrical machinery and apparatus sectors
.
12. Investigating the export performance and competitiveness of Turkey relative to the world
and emerging economies during the 2001-2004 period we find that Turkey is one of the
best performing countries in terms of both the number of competitive products and the
share of these products in total exports. The improvement of Turkey’s position in both
world and emerging markets together with the increase in Turkey’s exports of high-
technology intensive products signal steady integration to the world export market. With
regard to the continuity of the integration of Turkey to the world economy it is important
to introduce new products into the world markets` rising-star product spectrum. In this
sense, Turkey ranked the second in the number of competitive products introduced to the
world markets.
13. Finally export specialization index is calculated for Turkey to identify the Turkish
industries that have comparative advantages in world exports. The analysis suggests that
there is a remarkable upward trend towards the specialization in high-tech sectors.
Fourth Section:
14. Kalman filter approach is applied to examine how the value of each parameter values of
the export demand and supply function varies over time, without predetermining a
breaking point in time. The results suggest that none of the elasticity parameters in
respective equations are stable. There is a continuous increase in imports as well as
income elasticity as opposed to the persistent decrease in the real effective exchange rate
elasticity.
15. None of shocks in 1994, 1998 and 2001 had path-breaking impact, except; the shock in
2001 had a level shifting impact on import elasticity of export supply. As a result, it is
estimated that the import elasticity of export increased about 50 percent from 1987 to
2006.
16. We interpreted these findings in the following way: Some sectors were successful in
integrating to the world markets especially after the Turkey-EU CU and this helped them
to expand their export market share by producing for the external market during the
turbulent periods. High import dependence and low real effective exchange rate elasticity
shield them from the detrimental effects of real appreciation of Turkish lira. We believe

49
that there are two-way self-fulfilling dynamics between exchange rate sensitivity and
import dependence. One possible explanation is that, due to real appreciation of Turkish
lira, firms in these sectors were able to purchase inputs at lower price abroad.
17. Increase in the share of non-traditional commodities in total exports raised not only the
overall income elasticity of total Turkish exports but also its import elasticity, which
explains the recent surge in the import dependence of overall exports (Saygılı and
Saygılı, 2007). It is evident that exchange rate elasticity of non-traditional commodities
is smaller than that of the traditional goods, which also pulls down the overall exchange
rate elasticity of exports over time especially after 1996. The change in the composition
of Turkish exports in favor of low exchange rate elastic non-traditional commodities may
explain the seemingly puzzling coincidence of high growth of total exports and real
appreciation of the Turkish lira.

50
References
Aydın, F., Çıplak, U. and Yücel, E., (2004), “Export Supply and Import Demand Models for the
Turkish Economy”, Working Paper No. 04/09, The Central Bank of Turkey.
Balassa, B., (1977), “‘Revealed’ Comparative Advantage Revisited: An Analysis of Relative Export
Shares of the Industrial Countries, 1953-71,” Manchester School of Economic and Social
Studies 45:327-344.
Bhagwati, Jagdish and Donald R. Davis (1994) “Intra-industry trade issues and theory”, Harvard
Institute of Economic Research, Cambridge.
Gönel, Feride Doğaner (2001) “How important is intra-industry trade between Turkey and her
trading partners?” Russian and East European Finance and Trade, 4, pp:61-76.
Grupp, H. (1995), “Science, High Technology and the Competitiveness of EU Countries”,
Cambridge Journal of Economics, 19:1 p.209.
Helpman, Ethan and Paul Krugman (1985), Market Structure and Foreign Trade, Cambridge,
MIT Press.
Kalman, R.E. (1960) “A New Approach to Linear Filtering and Prediction Problems. Transactions of
the ASMA.” Journal of Basic Engineering 82D, 35-45.
Kalman, R.E. (1963) “New Methods in Weiner Filtering Theory Problems.” In: Bogdanoff, J.L.,
Kozin, F. (Eds), Proceedings of the First Symposium of Engineering Applications of
Random Function Theory and Probability, New York, 270-388.
Kim, Chang-Jin and Nelson, C. R (2000), State Space Models with Regime Switching: Classical
and Gibbs Sampling Approaches with Applications. Cambridge, MA: MIT Press.
Muscatelli, V.A., Stevenson, A.A., and Montagna, C. (1995), “Modeling Aggregate Manufactured
Exports for some Asian Newly Industrialized Economies,” The Review of Economics and
Statistics, PP: 147-155.
Ng, Thiam Hee (2006). “Foreign direct investment and productivity: evidence from the East Asian
economies” UNIDO, Staff working paper No: 03/2006
Nowak-Lehmann, F.D. (2004), “Different Approached of Modeling Reaction Lags: How do Chilean
Manufacturing Exports React to the Movement of Real Exchange Rate,” Applied Economics,
36, pp: 1547-1560.
Sarıkaya, Ç. (2004), “Export Dynamics in Turkey”, Central Bank Review, No:2 p.41-64.
Saygılı, M. and Saygılı, H. (2007) “Investigating Structural Change in Turkish Exports: The role of
Turkey-EU Custom Union and Economic Crises,” unpublished manuscript.
Saygılı, M., Şahinbeyoğlu, G.; and Özbay, P.(1998), “Competitiveness Indicators and the
Equilibrium Real Exchange Rates Dynamics in Turkey,” Macroeconomic Analysis of
Turkey: Essays on Current Issues, ed. E.M. Üçer, CBRT.

51
Sönmez, Mustafa (2005), “Türkiye İhrcatının İthalata Bağımlılığı: 2000-2004”, unpublished
manuscript.
Şahinbeyoğlu, G. and Ulaşan, B. (1999), “ An Empirical Examination of the Structural Stability of
Export Function,” CBRT Discussion Paper No: 9907.
Thornhill, D. (1988), “The Revealed Comparative Advantage of Irish Exports of Manufactures”,
Journal of Statistical and Social Inquiry Socaaty of Ireland, Vol XXV, part V.
WTO (2006) International Trade Statistics,
http://www.wto.org/english/res_e/statis_e/its2006_e/its06_toc_e.htm.
Yükseler, Z. and Türkan, E. (2006) “Türkiye’nin Üretim ve Dış Ticaret Yapısında Yapısal Dönüşüm:
Küresel Yönelimler ve Yansımalar,” Ekonomik Araştırmalar Formu Çalışma Raporları
Serisi, TÜSİAD-Koç Üniversitesi.

52
Appendix
A. SITC 3 Classification of High-technology Intensive Sectors in Terms of Factor
Intensity
Product Group Product Group
266 Synthetic fibers 745 Other non-electrical machinery
277 Advanced industrial abrasives 749 Non-electrical machinery parts, accessories
515 Heterocyclic chemistry 751 Office machines
516 Advanced organic chemicals 752 Automatic data processing machines
522 Rare organic chemicals 759 Advance parts for computers
524 Other precious chemicals 761 Television equipment
531 Synthetic matter 762 Radio-broadcast receivers
533 Pigments, paints, varnishes 763 Sound recorders, phonographs
541 Pharmaceutical products 764 Telecommunication equipment
551 Essentials oils, perfume, flavor 772 Traditional electronics
591 Agricultural chemicals 773 Optical fibred and cables
598 Advanced chemical products 774 Medical electronics
663 Mineral manufacturers, fine ceramics 776 Semi conductor devices
689 Precious non-ferrous base metals 778 Advanced electrical machinery
714 Turbines and reaction engines 781 Motor vehicles for persons
718 Nuclear, water, wind power generators 782 Motor vehicles for goods transport
724 Textile and leather machinery 791 Railway vehicles
725 Paper and pulp machinery 792 Aircraft and spacecraft
726 Printing machinery 871 Advanced optical instruments
727 Industrial food-processing machines 872 Medical instruments
728 Advanced machines tools 873 Measuring equipment
736 Metal working machinery tools 874 Advanced measuring equipment
737 Other metal working machinery 881 Photogram apparatus, and equipment
741 Industrial heating, cooling equipment 882 Photographic and cinematographic supplies
744 Mechanical handling equipment 884 Optical fibres and contact lenses
Note: The classification of high technology sectors is based on Grupp (1995).
B. Kalman Filter
Kalman filter approach or state space model is developed by Kalman (1960, 1963). The
Kalman filter is a recursive algorithm for expressing dynamic systems that involve unobserved state
variables, conditional on observed vector (Kim and Nelson, 2000). A state space model consists of
two equations of which a general form of a linear state space system representation is written down
bellow:
1=
ttttt
zzGw
+Φ+ (B.1)
=
tttt
yHz
ζ
+ (B.2)
Here, n
t
zR∈ is a (1)n× state vector, m
t
y
R∈ is the vector of observation t
Φ
, t
H, t
G are known
matrices that are also allowed to vary over time, k
w, and t
ζ
are vectors of normally distributed i.i.d
shocks. The first equation is called a transition equation that describes the dynamics of the state

53
variables. The second equation is the measurement equation points out a relationship between
observed variables and unobserved state variables. The model satisfies the following assumptions:
[
]
0
t
E
ζ
= ,
[
]
0
t
Ew
=
(B.3)
tj ttj
ER
ζ
ζλ
′
⎡⎤
=
⎣⎦ , tj ttj
Eww Q
λ
′
⎡⎤
=
⎣⎦ (B.4)
0
tj
Ew
ζ
′
⎡⎤
=
⎣⎦
,
[
]
00
Ez z
=
(B.5)
[
]
0000 0
()()Ez z z z P
′
−−= (B.6)
[
]
00
t
Ezw
′=,
[
]
00
t
Ez
ζ
′= (B.7)
Under these assumptions, ˆt
zcan be determined by the Kalman filter:
()
11
ˆˆ ˆ
tttt
tt tt
zz KyHz
−−
=+ − (B.8)
0
ˆ(0)zz=
Here, t
K is the Kalman gain, which determines the weight assigned to new information about and
calculated by
()
1
11
ttttt
tt tt
KPHHPHR
−
−−
′′
=+
(B.9)
where t
P is the ()nn×covariance matrix of conditional on information up to )1( −t and calculated
as follows
11 1 1 1 1
1tt t t t t
tt
PPGQG
−− − − − −
−′′
=Φ Φ + (B.10)
()
1
ttt
tt
PIKHP
−
=− (B.11)
11
1
ˆˆ
tt
tt
zz
−−
−=Φ (B.12)
As it is clear from equation (B.12) the success of the estimation depends on the representation of the
dynamics of the system. If the best Kalman gain is used then,
1
ˆ
tt t
tt
gyHz
−
=− (B.13)
The residual vector satisfies all white noise properties and its covariance matrix can be calculated as
[
]
0, 1
tttttt
tt
CEggHPHR
−
′′
== +
. (B.14)
Kalman Filter approach is a recursive process that updates estimated coefficients over time as new
information arrives. However, this may reduce reliability of estimated coefficients especially at the
beginning of the period. Instead, we may apply a fixed-point Kalman smoother, which gives the

54
estimated value of the state variable at time t based on all the available information up to time T,
where Tt>. The idea is that as new data are made available, we can improve our estimation result
from the Kalman filter by taking into account the additional information. Then Kalman smoothing
gives
()
1/tt tt tt
PIKHP
+=− (B.15)
1
ˆˆ
tt
tt
zz
+=Φ (B.16)