or ene rde urk le c h a oss rt. T equal. While an overall deterioration in the OIVI has been observed during periods of increasing oil prices, better diversification of oil import sources has lead to significant improvements. We suggest Turkish policy-makers implement sound policies, emphasizing diversification of crude oil import sources and reduction of the share of crude oil in primary energy imports to increase energy supply source world od of lje´ and then, t licies, limited energy resources compared to demand, highlighted the ajor rices aded drop irect macroeconomic activity in most Asian countries during the Asian Contents lists available at ScienceDirect .el Energy Energy Policy 39 (2011) 2132–2142 of the U.S. DOE are energy import dependent, i.e. their consumptionE-mail address:
[email protected] (V.S- . Ediger). issue with a marked increase in the total amount of imported oil financial crisis in 1997. As Kilian (2009) has clearly noted, oil prices tend to respond to some aggregate economic forces that also drive global demand for crude oil. The problem is a global one; 115 of 167 countries recorded by EIA 0301-4215/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.enpol.2011.01.058 n Corresponding author. Tel.: +90232 488 98 67. prices during last decade which is reminiscent of the 1970s crises. Depleting crude oil reserves in such countries, which may have effect on worldwide trade because rather than a supply shock, it was primarily a demand shock caused by significant decline in disruptions associated with over-dependence on oil imports. At present, however, new concerns about energy security have emerged (Umbach, 2010). The issue of oil import dependency has regained importance, especially in developing countries, due to abrupt increases in oil the international oil trade. Following the crisis period, a m trend reversal occurred in 1987 as a result of decreasing oil p to $13.10 per barrel in 1986. The share of internationally tr production increased from 41% in 1987 to 68.2% in 2007. The in prices to $12.72 per barrel in 1998 had, however, no d dependent countries. As noted correctly by Lo¨schel et al. (2010, p. 1666–1668), the energy export/import relations are used as a ‘‘political weapon’’ and have an impact on the vulnerability of the entire economic system. The international concern was in the beginning directed mostly towards potential costs of supply One of the main factors affecting the international oil trade was the price. The sharp increase in oil prices from $3.29 per barrel in 1973 to $11.58 in 1974, and then from $14.02 in 1978 to $31.61 in 1979 caused a notable decrease in consumption, especially in oil-importing countries, resulting in a decrease in 1. Introduction Dependence on imported energy become the most important issue in beginning with the oil crises peri Toman, 1996; Toman, 2002; Corre Yergin, 2006; Hughes, 2009). Since oil’’ has deeply affected energy po security. This study has also demonstrated that it is possible to construct an index representing crude oil vulnerability caused by import dependency. & 2011 Elsevier Ltd. All rights reserved. s, especially on oil, has energy supply security 1970s (e.g., Bohi and van der Linde, 2006; he concept of ‘‘foreign specifically of import all over the world. While major oil-exporting countries have benefited from this situation, the additional overburden on the budget of import-dependent countries has contributed to internal political and economic instabilities. The volume of globally traded crude oil has been increasing for decades. In 2008, the world’s oil production rose to 81.82 mbbls (million barrels) daily, of which 54.63 mbbls (66.8%) was traded internationally, and 27.19 mbbls (33.2%) was consumed domestically. Crude oil import policy of Turkey: Hist implications since 1968 Volkan S- . Ediger n, Istemi Berk Izmir University of Economics, Sakarya Caddesi, No. 156, 35330 Izmir, Turkey a r t i c l e i n f o Article history: Received 7 July 2010 Accepted 26 January 2011 Available online 19 February 2011 Keywords: Oil import policy Vulnerability index Turkey a b s t r a c t Turkey is one of the most economic and strategic bu crude oil import policy of T We implemented a princip based on four factors, whic import bill as a share of Gr of oil in total energy impo journal homepage: www ical analysis of determinants and rgy import dependent countries in the world, suffering deeply from the ns of oil importation. Our purpose is to determine the factors behind the ey and to measure their contribution to a well-organized import strategy. omponent analysis to construct an Oil Import Vulnerability Index (OIVI) re crude oil import dependency of primary energy consumption, crude oil Domestic Product (GDP), non-diversification of import sources, and share he contribution of these factors to the OIVI is found to be approximately sevier.com/locate/enpol Policy to construct a crude oil import vulnerability index in order to V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–2142 2133 is greater than domestic production. Since renewables are all domestically produced and consumed, the energy import depen- dency is only directly related with fossil fuels, including coal, oil, and gas. The dependence of the top 20 countries and the European Union on imported fossil fuel sources is shown as Production–Consumption (P–C) in Table 1. The EU has the highest dependency, with a P–C of 996.2 mtoe, nearly 60% of which is comprised only from oil. USA, Japan, and Germany are the three countries most dependent on imported fossil fuels with P–C values of 615.3, 434.2, and 213.6 mtoe, respectively. Crude oil also constitutes the greatest shares of their dependency. As is seen in the Table, Turkey is one of the most energy import dependent countries; ranking 12th in fossil fuels overall, 17th in oil, 10th in natural gas, and 8th in coal dependencies. In 2008, the share of net import in consumption of natural gas, crude oil, coal, and total primary energy was 99.4%, 94.4%, 90.5%, and 73.5%, respectively. The importance and risks of Turkey’s foreign dependence on Table 1 Rankings of top 20 countries based on P–C of fossil fuels for the year 2008. Source: BP (2009) Country Fossil fuels Oil Rank mtoe Rank US 1 �615.3 1 Japan 2 �434.2 2 Germany 3 �213.6 4 South Korea 4 �203.8 5 China 5 �181.9 3 Italy 6 �155.0 9 India 7 �145.8 6 France 8 �143.8 7 Spain 9 �121.3 8 Taiwan 11 �101.8 13 Turkey 12 �77.3 17 Belgium & Luxembourg 15 �61.2 16 Singapore 16 �58.2 14 United Kingdom 17 �52.9 42 Ukraine 18 �51.5 26 Thailand 19 �41.4 20 Brazil 20 �33.8 33 European Union # �996.2 energy, especially on oil, has been previously debated by several authors, such as Guven (1994); Balat (2004, 2010), Og˘ulata (2004), Altınay (2007), Yılmaz and Uslu (2007), Ediger (2008, 2009), So¨zen (2009), and Uslu (2008, 2010). However, most of these studies are descriptive, with only the exceptions of Altınay (2007), and So¨zen (2009), who constructs an artificial neural network to forecast future energy import dependency of Turkey and predicts that dependency will rise to 82% by 2020. Moreover, Altınay (2007) examined short-run and long-run price elasticities of demand for imported crude oil in Turkey for the period 1980–2005. His conclusion was that ‘‘the own price elasticity of oil clearly indicates that a 1% increase in the price of oil decreases the demand for foreign oil by 0.10% in the short-run, and by 0.18% in the long-run, holding all else constant. This makes Turkey, which is heavily dependent on foreign crude oil, vulnerable to shocks in oil prices. An increase in oil price will certainly have a strong negative impact on the trade balance.’’ (p. 5835). There seem to be two basic solutions for the problem of energy import dependency: (1) to change the country’s energy mix by increasing the share of domestic sources on the demand side, and (2) to apply a well organized energy import policy on the supply side. This study focuses on the second option and analyzes Turkey’s crude oil import policies. The main purpose of this paper is to investigate historically the important factors behind the crude oil import structure of Turkey, capture the relative importance of these effects. Section 2 consists of a historical analysis of the energy import structure of Turkey, with special reference to crude oil. Crude oil import dependency of primary energy consumption, share of crude oil in primary energy import, the crude oil import bill, and non- diversification of crude oil import sources are the topics discussed in this section. Section 3 constructs an oil import vulnerability index to determine the effects of each import related factors. Section 4 discusses the quantitative results provided in the previous section and Section 5 will conclude with policy suggestions. 2. Energy import structure of Turkey with a special reference to oil and to measure, if possible, the effect of each factor in imple- menting an effective import strategy. In this regard, we attempted Natural gas Coal mtoe Rank mtoe Rank mtoe �579.4 2 �67.7 31.9 �221.8 1 �84.4 1 �128.0 �118.3 4 �62.1 5 �33.2 �103.3 7 �35.7 2 �64.8 �186.0 30 �4.2 8.2 �75.7 3 �62.4 7 �17.0 �98.9 19 �9.7 4 �37.1 �92.2 5 �39.8 10 �11.9 �77.1 8 �35.1 14 �9.1 �50.1 15 �11.5 3 �40.2 �32.3 10 �32.4 8 �12.5 �41.3 14 �15.3 21 �4.6 �49.9 23 �8.3 – �6.5 11 �21.9 6 �24.5 �15.5 6 �36.9 0.9 �23.4 25 �7.7 11 �10.3 �11.4 18 �10.2 9 �12.2 �596.7 �269.8 �129.7 As a result of rapid increase in demand in comparison with production since 1950, Turkey has been an energy import dependent country (Ediger et al., 2007). Its primary energy consumption increased from 6.92 to 106.34 mtoe between 1950 and 2008 with an average rate of increase of 4.9%. On the other hand, its domestic production rose from 6.43 mtoe in 1950 to its peak level of 29.32 mtoe in 1998, before decreasing to 23.78 mtoe in 2003. Even though production has gradually increased since 2003, reaching a level of 29.26 mtoe in 2008, this level has not yet proven to be sustainable. As a result, its primary energy import increased continuously from 0.55 mtoe in 1950 to 85.36 mtoe in 2008 with an average rate of increase of 9.7%. Fig. 1 represents the historical development of net energy import dependency, i.e. difference between import and export per consumption, of each fossil fuel and of primary energy consump- tion. As clearly seen from the figure, since 1987 Turkey has been critically dependent on imported fossil fuels as well as on primary energy. We have considered 50% as the critical limit for import dependency, because when this level is exceeded, imports over- take domestic supply. Two structural break points exist, first in 1969 when crude oil slightly crossed the limit, and second in 1987 when hard coal, natural gas, and primary energy all passed the limit. With the introduction of natural gas in 1987 the country became more net import dependent on natural gas than on oil 50ep en Oil 980 V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–21422134 -10 0 10 20 30 40 1950 N et Im po rt D Gas Hard Coal 20.0% 1955 1960 1965 1970 1975 1 Fig. 1. Net import per consumption of fuels, 1950–2007. Source: ETKB (2010) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% sh ar e 1977 3.86 mtoe 3.97 mtoe 18.09 mtoe 6.42 mtoe 60 70 80 90 100 de nc y, % Primary Energy 1969 51.7% and coal. At present, the dependency of the Turkish energy system on net energy imports is overwhelmingly critical, reaching 99.4% in natural gas, 94.4% in oil, 90.5% in hard coal, and 73.5% in primary energy. However, oil has always been the dominant energy source in Turkish primary energy demand, which increased steadily from 13.90 mtoe in 1965 to 106.33 mtoe in 2008 (Fig. 2). In 1965, the shares of crude oil and coal in primary energy mix were 27.8% (3.97 mtoe) and 28.6% (3.86 mtoe), respectively. Oil then increased reach- ing a peak of 55.8% (18.09 mtoe) in 1977 when coal has declined to the minimum of 19.8% (6.42 mtoe). Afterwards, the share of coal ranged between 20% and 32%, currently at 29.5% (31.39 mtoe) while the proportion of oil has continuously decreased, reaching a current low of 29.9% (31.78 mtoe). The share of natural gas, on the other hand, has continuously increased, reaching 31.8% in 2008 after overtaking coal and oil in 2005 and in 2007, respectively. It is, therefore, obvious that crude oil import has always played the most significant role in the energy mix as well as the import structure of Turkey. Four parameters are considered to be impor- tant for the crude oil import policies of Turkey, (1) crude oil import dependency of primary energy consumption, (2) share of crude oil in primary energy import, (3) the crude oil import bill, and (4) non-diversification of crude oil import sources. 2.1. Crude oil import dependency of primary energy consumption The first factor which we have suggested to be considered as a parameter in crude oil import policy of Turkey is crude oil import dependency of primary energy consumption. The underlying reason for including this factor in our analyses is the rising 1965 1970 1975 1980 198 Fig. 2. Share of fossil fuels in Turkey’s energy mix, 1965–2008. Source: ETKB (2010) 2005 91.0% 52.3% 49.3% 59.6% 1987 99.4% 94.4% 90.5% 73.5% 2008 1985 1990 1995 2000 2010 2015 oil gas coal 31.78 mtoe 31.39 mtoe 33.81 mtoe importance of diversification of fuels in countries energy mix, or balanced distribution of all energy options, as clearly noted by Fanchi (2000). Since Turkey’s primary energy mix is highly dependent on imported sources, as we pointed out in Section 2, diversification in fuel mix for Turkey means a balanced portfolio of import dependencies on all sources. Fig. 3 shows the historical trends of crude oil, coal, and gas import dependencies of primary energy consumption in Turkey. In 1968, 23.9% of primary energy demand was met by imports, 20.3% of which was crude oil, while the remainder was petroleum products such as gasoline, diesel oil, fuel-oil, etc. Since most of the primary energy imports were crude oil, a good correlation only exists between the primary energy and crude oil imports in total consumption with a correlation coefficient (R2) of 91.5 until 1988. At this time, crude oil peaked with a value of 45.2%, whereas coal and natural gas contributed only 6.6% and 2.2%, respectively. Within this period, global oil crises have affected crude oil with a sharp decrease in 1979. However, after 1988 primary energy and crude oil imports in total con- sumption diverge and the correlation decreases as a result of increasing natural gas and coal imports. Another turning point occurs in 2005, when increasing natural gas overtook falling crude oil at about 26%. Finally, in 2007 81.4% of primary energy demand in Turkey was met by imports, consisting of natural gas (30.8%), crude oil (21.8%), coal (15.1%), and petroleum products (13.7%). 2.2. Share of crude oil in primary energy import Another factor which appears to be important for Turkey’s crude oil import policies is the share of crude oil in primary energy import, 5 1990 1995 2000 2005 1 .2% .6% .2% .4% 5 ene V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–2142 2135 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1965 1968 20.3% 45 6 2 23.9% 59 1970 1975 1980 198 Fig. 3. Relative shares of import of each fossil fuel and primary energy in primary Source: PIGM (2010) 80 90 100 99.9% which has always been higher than other fuel imports, increasing vulnerability. Turkey’s energy import structure has changed signifi- cantly over time. While total energy import consisted of only crude oil and petroleum products until 1973, coal began to be imported after this time because of increasing concerns on the diversification of fuel mix (Fig. 4). Since then, a major trade-off between coal and oil has been witnessed, resulting in shares of 43% and 17.1% for oil and coal, respectively, in 2007. Imported coal is almost entirely bituminous coal, although between 1991 and 2006 lignite was imported in amounts varying between 2 and 56 thousand tons per annum. Another break in the structure of Turkish energy imports is the introduction of natural gas imports in 1987, with a total of 0.40 mtoe. Since then, natural gas imports have increased rapidly, and have recently reached 34.01 mtoe, slightly less than the current 36.68 mtoe for oil. The year 1994 was an important turning point, when the level of natural gas imports reached a level equal to those of coal, at approximately 13%. Gas imports are likely to reach a level equal to those of oil in the near future. 2.3. The crude oil import bill Bohi and Toman (1996) have identified ‘‘volume of energy imports’’, ‘‘energy price variability’’, and ‘‘national security and military expenditures’’ as three potential energy security 0 10 20 30 40 50 60 70 1960 Sh ar e, % Oil Gas Coal 0.1% 1973 1965 1970 1975 1980 19 Fig. 4. Share of fuels in Turkey’s energy import, 1950–2007. Source: PIGM (2010) primary_energy oil coal gas 988 2005 25.7% 26.5% 14.7% 21.8% 30.8% 15.1% 80.2% 81.4% 1990 1995 2000 2005 2010 rgy consumption, 1968–2007. 72.7% 2007 externalities which caused the loss of economic welfare. In this respect, we suggest that the crude oil import bill, which con- tributes the greatest proportion of total energy imports, creating an overburden on the Turkish economy, is one of the major determinants of crude oil import policy of Turkey. The crude oil import bill has always been an important factor in Turkish macroeconomy, constituting one of the biggest por- tions of the import bill of Turkey. Fig. 5 represents the share of crude oil import in total imports and in Gross Domestic Product (GDP) of Turkey. The share of oil in total imports represents firstly a sharp increasing and then a decreasing trend, separated by a peak with a value of 40% in 1982, as a result of 1973 and 1979 oil crises. From 1998 until the present, however, the trend has been relatively stable. The share of oil import in GDP of Turkey also follows a very similar pattern to the share of oil in total imports. In 2007, shares of value of oil import in total import bill and GDP of Turkey were 6.9% and 1.8%, respectively. The crude oil import bill is expected to be highly dependent upon benchmark prices determined in the crude oil market, one of the most globally liberalized markets. The international crude oil prices and Turkey’s unit import prices are compared in Fig. 6. As can be easily observed, with a correlation coefficient (R2) equal to 0.99, both prices correlate well with each other, indicating the management of a good pricing policy. Fig. 6 shows two occurrences (from 1974 to 1980 and during the post-2000 period) where there were significant 43.0% 39.9% 17.1% 13.8% 13.5% 19941986 85 1990 1995 2000 2005 2010 4) V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–21422136 10% 15% 20% 25% 30% 35% 40% 45% Sh ar e i n to ta l i m po rt s 40.0% (1982) 5.6% (198 21.3% (1974) differences between international prices and Turkish import prices. A possible explanation for these differences is Turkey’s use of long- term trade contracts. However, the high correlation between international prices and unit import prices does not necessarily mean that the crude oil import bill loses significance in Turkey’s energy security agenda. As an emerging economy, Turkey is a highly import dependent country in terms of both energy and other commod- ities and services (Table 2). The monetary policy of Turkey has promoted imports by implementing a higher interest rate–lower foreign exchange rate regime during last two decades as a result of financial liberalization in 1989 (Kepenek and Yentu¨rk, 2008). 0 10 20 30 40 50 60 70 80 1960 In te rn at io na l O il Pr ic es , M on ey o f D ay , $/b b International Turkish 1974-1980 1965 1970 1975 1980 1985 Fig. 6. International prices vs. Turkish crude oil unit import prices. Source: PIGM (2010), BP (2009) Table 2 Total energy and foreign trade of Turkey, 2000–2008. Source: THM (2010) Years Total energy million US$ Import Export Net import 2000 9398 308 9090 2001 8175 423 7752 2002 9073 674 8399 2003 11,528 958 1057 2004 14,385 1366 13,019 2005 21,226 2534 18,692 2006 28,829 3404 25,425 2007 33,846 4928 28,918 2008 48,253 7407 40,846 0% 5% 1965 1970 1975 1980 1985 Fig. 5. Share of Oil import Cost in Turkey’s Total Import and GDP. Source: PIGM (2010), TUIK (2010), World Bank (2010) 1% 2% 3% 4% 5% 6% Sh ar e i n G D P oil import ($)/ total imports ($) oil import ($)/GDP ($) 6.9% 1.8% (2007) 0.8% (1998) From 2000 to 2008, imports increased from $40.67 bn to $201.96 bn, while exports rose from $27.78 bn to $132.03 bn, creating a gradually increasing foreign trade deficit. The net energy import has been the most significant overburden, forming the major part of foreign trade deficit in the last decade. In 2008, for instance, the net import bill for total energy reached $40.8 bn, constituting 58.4% of total trade deficit. 2.4. Non-diversification of oil import sources The final factor in crude oil import structure is assumed to be diversification, which Churchill referred to as ‘‘variety’’, as very 0 100 200 300 400 500 600 Tu rk ish B ill $/ ton 2000-Present 1990 1995 2000 2005 2010 Foreign trade million US$ Share of net energy import in deficit, % Import Export Deficit 40,671 27,775 12,896 70.5 41,399 31,334 10,065 77.0 51,554 36,059 15,495 54.2 69,34 47,253 22,087 47.9 97,54 63,167 34,373 37.9 116,774 73,476 43,298 43.2 139,576 85,535 54,041 47.0 170,063 107,272 62,791 46.1 201,964 132,027 69,937 58.4 0% (2007)4.5% (1998) 1990 1995 2000 2005 2010 Table 3 Source countries of Turkey’s imported crude oil. Source: PIGM (2010) V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–2142 2137 important for energy supply security in 1913 (Yergin, 2006). Diversification of sources of imported oil is closely related to energy security, especially in oil-importing countries (Vivoda, 2009). Li (2005) noted that promoting diversification is ‘‘sensible and feasi- ble’’ for sustainable development and also essential for energy security, stating that ‘‘y the dominance of a single energy source and system, no matter how ‘‘perfect’’ it might be at a time, would be unsustainable in the long-run’’ (p. 2240). Verrastro and Ladislaw (2007) pointed out that in 2006 the United States has imported Rank Country 1968–2009 2009 Tons % Tons % 1 Iran 155,964,310 22.7 3,262,648 23 2 Iraq 148,086,572 21.5 1,817,025 12.8 3 Saudi Arabia 143,142,710 20.8 2,096,277 14.8 4 Libya 93,882,851 13.6 139,286 1 5 Russian Federation 70,043,056 10.2 5,788,840 40.8 6 Syria 24,262,538 3.5 160,296 1.1 7 Algeria 12,914,852 1.9 – – 8 U.A.E 12,403,489 1.8 – – 9 Egypt 7,811,568 1.1 – – 10 Kuwait 3,206,858 0.5 – – 11 Kazakhstan 3,350,847 0.5 392,546 2.8 12 China 2,878,158 0.4 – – 13 Italy 2,150,707 0.3 232,909 1.6 14 Azerbaijan 439,854 0.1 76,739 0.5 15 Venezuela 381,118 0.1 – – 16 Qatar 367,630 0.1 – – 17 Switzerland 621,360 0.1 – – 18 Albania 122,221 a – – 19 Oman 169,002 a – – 20 Tunisia 292,974 a – – 21 Georgia 235,582 a 35,607 0.3 22 Others 9,162,742 1.3 190,351 1.3 Total 691,890,999 100.0 14,192,524 100.0 a Less than 0.1. crude oil and petroleum products from nearly 35 countries, the top five of which provide 58% of its total oil demand, in order to illustrate the importance of diversification in oil suppliers. These are also true for Turkey. From 1968 to 2009, Turkey imported oil from 21 different countries (Table 3)1 . Iran (22.7%), Iraq (21.5%), Saudi Arabia (20.8%), Libya (13.6%), Russian Federation (10.2%), and Syria (3.5%) have been the six major source countries, constituting more than 90% of total oil imports during this period. In 2000, when we believe diversification was at its all-time peak, Turkey imported nearly 21 million tons of crude oil, almost 91% of which was from seven different countries; Saudi Arabia (17.2%), Iraq (18.9%), Libya (16.6%), Iran (17.5%), Syria (10.2%), Russia (10.3%), and Kazakhstan (1.3%). This diversification pattern, however, has chan- ged recently with the Russian Federation replacing Iraq, increasing dependency on one single source. In 2009, the vast majority of oil (91.4%) was imported from only four countries, Russian Federation (40.8%), Iran (23.0%), Saudi Arabia (14.8%), and Iraq (12.8%). An examination of historical development of oil import struc- ture has shown that significant changes between 1968 and 2009 can best be explained by dividing this period into four separate periods, namely pre-1974, 1974–1990, 1990–2000, and post-2000 (Fig. 7). There were significant changes in each of these four periods, especially in the top 2 supplier countries. Russian Federation and Iran became the two major oil suppliers during the post-2000 period, whereas for the previous periods the two analysis to construct an energy index with four parameters, namely, 1 The details of the group ‘‘others’’ are not given by General Directorate of Petroleum Affairs, Turkey. primary energy consumption, electricity consumption per capita, energy intensity of GDP, and CO2 emission per capita. The study uses this energy index to modify the Human Development Index (HDI) to construct an energy-adjusted HDI. Percebois (2007) uses three major indicators of energy vulnerability, import concentration, energy bill, and black-out risk. Using decomposition analysis Bacon and Kojima (2008) find that total change in vulnerability of a country is the summation of oil price effect, import dependency effect, share of oil in fuel mix effect, energy intensity of GDP effect, and exchange rate effect. Gnansounou (2008) constructs a compo- site index using ‘‘Euclidian Distance’’ methodology, including such factors as energy intensity, oil and gas import dependency, CO2 content of primary energy supply, electricity supply weaknesses, and non-diversity in transportation fuels. Finally, Gupta (2008) forms an index of oil vulnerability by conducting ‘‘Principle Compo- nent Analysis’’ on five indicators; reserve to consumption ratio, geopolitical oil risk, cost of oil import to GDP ratio, share of oil in primary energy consumption, and oil intensity of economy. Lo¨schel et al. (2010) calculated the indicator aggregated by using energy import shares as a weight in the process of sensitivity analysis. In contrast to these studies, we used five indicators: oil import dependency of primary energy consumption (F1), the crude oil import bill as a share of GDP (F2), non-diversity of import source (F3), share of crude oil in primary energy import (F4), and oil import dependency of oil consumption (F5) to construct an oil import vulnerability index using statistical methods such as major suppliers had been Saudi Arabia and Iran (Period III), Iraq and Iran (Period II), and Saudi Arabia and Iraq (Period I). Iran has been listed in the first two major source countries in the Periods II, III, and IV, whereas Saudi Arabia and Iraq were the major suppliers in the Periods I and III, and I and II, respectively. The Russian Federation has only become the major oil source country in the most recent period (Period IV), Libya has consistently occupied the third place during all periods, and Syria mostly has only become a significant supplier in the two most recent periods. In order to better understand the diversification issue in the historical context, we used mean of shares of each source country together with their standard deviations following Eq. 1.3 as explained in the Section 4.1 (Fig. 8). The total numbers of source countries for each year are also compared with these values. The mean values fluctuate almost periodically with decreasing ampli- tudes and the standard deviations fluctuate in a similar pattern, implying an improvement in diversification. This result is also supported by a recent increase in the number of source countries. (Solid lines are standard deviations of shares of source coun- tries in each year, dots are mean values of shares, and the below columns represent the number of source countries each year). The following section describes the construction of a crude oil import vulnerability index designed to calculate the contribution of above mentioned indicators to the entire crude oil import policy of Turkey for the period between 1968 and 2007. 3. Crude oil import vulnerability index of Turkey In the literature, there exist a number of multivariate approaches to creating an index with different indicators, including descriptive multivariate methods, factor analysis, cluster analysis, MANOVA, discriminant analysis, canonical correlation analysis, multidimen- sional scaling, and principle component analysis (Abeyasekera, 2005). Our method is closest to those proposed by Ediger and Tatlıdil (2006), Percebois (2007), Bacon and Kojima (2008), Gnansounou (2008), and Gupta (2008), and Lo¨schel et al. (2010). The study of Ediger and Tatlıdil (2006) applies principle component Principle Component Analysis (PCA) and Euclidian Distance (ED). 86 1 9. V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–21422138 30 35 40 45 50 un tri es 0 10 20 30 40 50 60 70 80 90 100 19 68 Sh ar e, % Period I pre-1974 Period II 1974-1990 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 Fig. 7. Shares of six major source countries in Turkey’s crude oil import, 1968–200 Source: PIGM (2010) The SPSS software package was used for the data, which consisted of the following: crude oil consumption in tons, crude oil import from each source country in tons, primary energy consumption in tons of oil equivalent, primary energy import in tons of oil equivalent, value of total oil import in nominal dollars, and finally GDP, gross domestic product in nominal dollars. All data were obtained from the General Directorate of Petroleum Affairs except the last figure, from the World Bank. First, we have converted the original data into five factors (F1–F5) by using the following: F1,t ¼ oil import primary energy consumption ð1:1Þ F2,t ¼ value of oil import GDP ð1:2Þ F3,t ¼ s¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiP ðYi�YÞ2 n s ð1:3Þ where, Yi is the share of a country in Turkey’s total import, Y is the mean of shares, n is the number of countries, and s is the standard deviation. F4,t ¼ oil import total primary energy import ð1:4Þ F5,t ¼ oil import oil consumption ð1:5Þ 0 5 10 15 20 25 19 68 N um be r o f C o 19 70 19 72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 1 Fig. 8. Diversity of Turkey’s oil import and number of source countries, 1968–2009. Source: PIGM (2010) 30 40 50 60 70 ev ia tio n, % Iran Iraq S. Arabia Libya RF Syria Period III 1990-2000 Period IV post-2000 19 88 19 9099 2 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 These factors are then normalized as follows: Xn,t ¼ ðFn,t�minðFn,T ÞÞ maxðFn,T Þ�minðFn,T Þ � � for n¼ 1, . . .,5 ð1:6Þ where, t and T represents the corresponding year and whole period; 1968–2007, respectively. These normalized indicators are further used to implement PCA and ED methodology to form the index. Since both methodologies have very similar outputs, this section presents results for PCA in detail, whereas ED methodology is provided in the Appendix. The PCA methodology includes capturing correlation matrix, calculating eigenvalues, rotated component loadings using ortho- gonal rotation matrix, and finally, score coefficients (Pearson, 1901; Ho, 2006). Since PCA does not indicate the relevancy of indicators, we first calculated partial correlations of each variable to test for their significance in the means of controlling correla- tions between other factors in order to define all the indicators to be included in PCA properly, using the following: rðXiXj :XkÞ ¼ rXiXj�rXiXkrXjXkffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1�r2XiXk q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1�r2XjXk q ð1:7Þ where, r(Xi.Xj.Xk) is the partial correlation between Xi and Xj, provided that Xk is the controlling variable. Results of all partial correlations imply that X3 and X4 are mostly controlling, whereas X2 is secondarily controlling the correlation between all variables. Among the least significant controlling 20 00 -40 -30 -20 -10 0 10 20 Av er ag e an d St . D 99 0 19 92 19 94 19 96 19 98 20 02 20 04 20 06 20 08 variables X1 is included rather than X5 because X5 has already been represented in X1 as proved by partial correlation coefficients between these two variables. Therefore, X1, X2, X3, and X4 are considered as the controlling indicators in the following calculations. The correlation coefficient (R) matrix between the four indi- cators is given in Table 4, which is used to calculate the eigenvalues li (Table 5) by solving the following determinantal formula: R�lI �� ��¼ 0 for l ð1:8Þ where, I is the identical matrix. PCA has first reduced four components to two based on corresponding eigenvalues and then grouped all indicators in these two components since they explain nearly 84% of total variation in correlation, as seen in Table 4. The grouping has been done using rotated component loadings. Rotated component matrix (C) states that X1 and X2 are highly correlated with the Component 1 while X3 and X4 with Component 2 (Table 6). The rotated component matrix is used to calculate the coeffi- cient of each indicator as follows: bk ¼ CkiliP4 n ¼ 1 ln ð1:9Þ where k is from 1 to 4 for each indicator, i is the largest value of components 1 and 2 in each row. Each coefficient is divided by the summation of coefficients to calculate the weights of each indicator. The resulting index is as follows: OIVIt ¼ 0:284X1,tþ0:269X2,tþ0:227X3,tþ0:219X4,t ð1:10Þ where X1 is the crude oil import dependency of primary energy consumption, X2 is the crude oil import bill as a share of GDP, X3 is the non-diversification of crude oil import sources, and finally X4 shifted from OPEC to futures exchanges. While the major oil producers, especially those integrated into OPEC, were more effective Table 4 Correlation matrix (R) for normalized factorsa. X1 X2 X3 X4 X1 1.000 X2 0.591 1.000 X3 �0.156 �0.228 1.000 X4 0.464 0.169 0.525 1.000 a Result of Barlett’s test for sphericity (w2¼49.981) states that correlation matrix provides sufficient significance to implement the PCA methodology. Table 5 Eigenvalues for the correlation matrix of normalized factors. l1 l2 l3 l4 Eigenvalues 1.840 1.504 .449 .207 Variability (%) 45.995 37.596 11.230 5.179 Cumulative variance Explained (%) 45.995 83.592 94.821 100.000 Table 6 Rotated component loadings matrix (C). Component 1 Component 2 V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–2142 2139 X1 .904 .151 X2 .858 � .122 X3 � .290 .885 X4 .396 .854 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1965 O IV I 1968-1975 1970 1975 1980 1985 1975-1979 1979-1984 Fig. 9. Oil import vulnerability index of Turkey (OIVI) v in determining oil prices during the pre-1986 period, other market participants, notably consumers, gained more power after 1986. 1990 0 20 40 60 80 100 120 20 08 $/ bb OIVI Crude Oil Real Prices 2000-2007 1995 2000 2005 2010 1984-2000 is the share of crude oil in total primary energy import and t is each corresponding year from 1968 to 2007. 4. Discussions of results The historical development of the OIVI of Turkey between 1968 and 2007 is given in Fig. 9. In general, the OIVI consists of an increasing and fluctuating trend between 1968 and 1984 and a decreasing and relatively smooth trend between 1984 and 2000. These major trends are separated by a peak of 0.786. This peak represents the highest OIVI, indicating the least effective oil import management during the study period. During the period 1968–1984, the OIVI was generally affected by the price of crude oil. Two peaks occurred during this period. Firstly, an increase of 267.8% in oil prices, from $2.83 per bbl in 1973 to $10.41 per bbl in 1974 caused a peak in the OIVI in 1975, with a one year lag. The second price increase from $13.03 per bbl in 1978 to $29.75 per bbl in 1979, and then a further increase to $35.69 per bbl, representing a rise of 128.3%, created the second and greatest OIVI peak in 1984, after a four year lag. These two peaks are separated by a temporal improvement which occurred as a result of relatively stable crude oil prices, increasing only by about 25% from $10.41 per bbl in 1974 to $13.03 per bbl in 1979. A second period of improvement in the OIVI occurred between 1984 and 1986, with oil prices declining to $13.10 per bbl in 1986, the minimum level between 1979 and 1997. The year 1986 was an important milestone in the crude oil price mechanism since the dominancy in price determination process s. international crude oil real prices. (2008 $/bbls). The OIVI declined gradually between 1986 and 1998 when international oil prices were relatively low, varying between a minimum of $13.10 per bbl in 1986 and a maximum of $20.45 per bbl in 1990. Since 1998, however, the oil prices have been continu- ously increasing, whereas the OIVI has remained stable, indicating that other factors have a greater effect on the OIVI than the oil import bill. Fig. 10 shows the historical development of four components of the OIVI. As shown in Eq. 1.10 the overall contribution of X1, X2, X3, and X4 in the OIVI are very close to each other with values of 28.4%, 26.9%, 22.7%, and 21.9%, respectively. However, the contribution of each factor varies across different periods of time. For instance, in 1984, when the OIVI was at its highest, the contributions of X1 through X4 are 31.7%, 34.3%, 10.9%, and 23.1%, respectively. It is an expected result that the crude oil import bill as a share of GDP dominated during this period since oil prices rose dramatically as a result of the second oil shock, resulting an excess overburden in the budgets of oil-importing countries. As stated by a number of authors, international political relations. Since Turkey imported most of its oil from neighboring countries such as Iraq, Iran, Saudi Arabia, Russian Federation, Libya, and Syria, the short-term relations with these countries have been the major dynamic affecting the factor X3. 5. Conclusions and policy implications As a result of analyzing the important underlying factors of crude oil import policy of Turkey from 1968 to 2008, we have suggested that there exist four import related indicators, (1) crude oil import dependency of primary energy consumption, (2) the crude oil import bill as a share of GDP, (3) non-diversification of import sources, and (4) the share of oil in total energy import. These factors are then used to construct an Oil Import Vulner- ability Index (OIVI) using the PCA methodology. The resulting Eq. 1.10 shows that the four factors have very similar weights (28.4%, 26.9%, 22.7%, and 21.9%, respectively), indicating almost 86 1 V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–21422140 abrupt increase in oil prices has a dual-effect on macroeconomy, it increases total cost of oil imports while decreasing aggregate economic output and, therefore GDP (e.g. Hamilton, 1983; Mork, 1989; Kahn and Hampton, 1990; Huntington, 1998; Brown and Yucel, 1999, 2002; Gao and Madlener, 1999; Dickman and Holloway, 2004; Guo and Kliesen, 2005; Rogoff, 2006; Sill, 2007; Kilian, 2008; Oladosu, 2009). However, in 1975, when the OIVI reached its second highest value, the dominating factor is X3, non-diversification of oil import sources, with a contribution of 34.5%, indicating that it played a significant role in this case. Similarly, in 2000 when the OIVI was at its lowest point in the whole period X3 has made the largest contribution with the minimum value of all period. The historical data of each factor consisting OIVI show that a good correlation exists between the general patterns X1, the share of crude oil in total primary energy import, and X4, the share of crude oil import in primary energy consumption. This is an expected result since Turkey’s energy mix has always been dominated by oil supplied mostly by foreign sources (Figs. 1 and 2). The reason for the overall declining trend of these factors following the oil shocks of the 1970s can be attributed to the decreasing share of crude oil in Turkey’s energy mix (Fig. 2). X2, the crude oil import bill as a share of GDP, also has a similar pattern with X1 and X4 until 1999. X2 declined during the period when oil prices were comparatively stable between 1986 and 1999 and increased during the period when oil prices begin increasing after 2000 (Fig. 5). X3, the non-diversification of crude oil import sources, appears to follow a fluctuating pattern highly independent from the other three factors of the OIVI. This can be best explained by the Turkey’s 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 19 68 19 78 19 70 19 72 19 74 19 76 19 80 19 82 19 84 19 Fig. 10. Relative shares of each indicator (X1, X2, 98 8 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 equal contribution to the oil import vulnerability index of Turkey. The OIVI reached its highest value in 1984 and the most important factors affecting the OIVI immediately before and after this date were the high oil prices and diversification, respectively. Since the index is mostly composed of the crude oil import bill as a share of GDP and oil import dependency of primary energy con- sumption in that year, an effective oil import policy for Turkey would be expected to minimize their effect as much as possible. On the other hand, this policy should maximize oil import source diversity, which is the major component of minimum OIVI in 2000. We, therefore, suggest the following policy implications to Turkish policy-makers to increase oil supply security, which is an important part of energy security of Turkey: the economic and strategic burden of oil import can be lightened by implementing sound policies which emphasize diversification of oil import sources while minimizing the cost of oil imports. Since Turkey unlikely to be in a position to influence the global crude oil prices directly, a reduction of the proportion of crude oil in primary energy imports is desirable to relieve the economic burden created by the cost of oil imports. This will also diminish the oil import dependency of primary energy consumption, which was the second major contributing factor in the most vulnerable year of 1984. Moreover, increasing effective use of spot market funda- mentals such as types of transactions, terms, and durations of contracts, and dealers will also reduce the cost of import. Further, with respect to strategic concerns of the country, the diversification should be accepted as the core of energy security and ‘‘sufficient and secure’’ crude oil supplies should be treated as X1 X2 X3 X4 X3, and X4) in oil import vulnerability index. the top priority. On the other hand, policy-makers should bear in mind that diversification in both fuel type and in source countries are two inseparable components of the overall diversification. For instance, diversifying of Turkish energy mix into natural gas does not, in itself, increase the energy security of Turkey, since majority of natural gas imports come from Russia. This study has demonstrated that it is possible to construct an index representing crude oil vulnerability caused by import dependency. Similar studies can also be conducted on other imported fossil fuels such as coal and natural gas, in Turkey and in other developing countries. Implementing the same methodol- ogy in other developing countries with other fossil fuels will certainly enhance the reliability of this analysis. Finally, the most crucial part of this study is to determine the underlying factors. We have, in that sense, first suggested five factors, which were then reduced to four by considering the partial correlations between them. However, in order to avoid subjective bias, further statistical diagnostic tests are needed to determine the significance of each of the factors included. Acknowledgements Most of the data used in this study was provided by the General Directorate of Petroleum Affairs of Turkey. We specially thank Mr. Erdal Gu¨rdelen, General Director and Mr. Selami _Incedalcı, Deputy General Director for their generous contribu- tions. Moreover, authors would also like to thank Dr. Sabri Erdem, 9 Eylu¨l University, Dicle Yurdakul, Benjamin Beranek, Cihangir Appendix 1. ED methodology Kepenek, Y., Yentu¨rk, N., 2008. Tu¨rkiye Ekonomisi (Turkish Economy), 21st edition V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–2142 2141 Euclidian Distance methodology to form an index simply con- siders the lowest geometric distance between two variable points. Table A1 Partial correlations of factors. Controlling variable X1 X2 X3 X4 X5 X2 1.000 X3 .142 1.000 X4 � .613n � .171 1.000 X5 � .713n � .148 .683n 1.000 Controlling variable X2 X1 1.000 X3 � .440n 1.000 X4 � .109 .587n 1.000 X5 .717n � .027 .459n 1.000 Controlling variable X3 X1 1.000 X2 .509n 1.000 X4 .343nn .349nn 1.000 X5 .846n .577n .650n 1.000 Controlling variable X4 X1 1.000 X2 .552n 1.000 X3 � .569n � .378nn 1.000 X5 .910n .587n � .530n 1.000 Controlling variable X5 X1 1.000 X2 .047 1.000 X3 .200 .601n 1.000 X4 .306 .453n .784n 1.000 n Significance at 1% confidence level. Kan, Ceki Franko and Itır Bag˘dadi, Izmir University of Economics, for their helpful comments and Simon Mumford, Izmir University of Economics for critically editing the manuscript. nn Significance at 5% confidence level. Remzi Kitapevi, Istanbul. Kilian, L., 2008. The economic effects of energy price shocks. Journal of Economic Literature, American Economic Association 46 (4), 871–909. Kilian, L., 2009. Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review 99 (3), 1053–1069. Li, X., 2005. Diversification and localization of energy systems for sustainable In our case, we have four variables so the equation is as follows: EDt ¼ P4 n ¼ 1 X 2 n,t 4 !0:5 Appendix 2. Partial correlations of factors Table A1. References Abeyasekera, S., 2005. Multivariate methods for index construction, in: Household Sample Surveys in Developing and Transition Countries, ST/ESA/STAT/SER.F/ 96, United Nations, New York. Altınay, G., 2007. Short-run and long-run elasticities of import demand for crude oil in Turkey. Energy Policy 35 (11), 5829–5835. Bacon, R., Kojima, M., 2008. Oil price risks: Measuring the vulnerability of oil importers. Viewpoint no. 320, Available at; /http://rru.worldbank.org/docu ments/publicpolicyjournal/320-OilPrices.pdfS. Balat, M., 2004. Oil and natural gas transport systems, trade and consumption trends in Turkey. Energy Exploration and Exploitation 22 (3), 207–216. Balat, M., 2010. Security of energy supply in Turkey: Challenges and solutions. Energy Conversion and Management 51 (10), 1998–2011. Bohi, D.R., Toman, M.A., 1996. The economics of energy security. Kluwer Academic Publishers, Boston–Dordrecht–London. BP, 2009. Statistical Review of World Energy, Available at /www.bp.com/ statisticalreviewS. Brown, S.P.A., Yucel, M.K., 1999. Oil prices and U.S. aggregate economic activity: A question of neutrality. Economic and Financial Review 2nd Quarter. Federal Reserve Bank of Dallas 16–23. Brown, S.P.A., Yucel, M.K., 2002. Energy prices and aggregate economic activity: An interpretative survey. Quarterly Review of Economics and Finance 42, 193–208. Correlje´, A., van der Linde, C., 2006. Energy supply security and geopolitics: A European perspective. Energy Policy 34 (5), 532–543. Dickman, A., Holloway J., 2004. Oil market developments and macroeconomic implications. Bulletin, October 2004, Reserve Bank of Australia. Ediger, V.S- ., Tatlıdil, H., 2006. Energy as an indicator of human development: A statistical approach. The Journal of Energy and Development 31 (2), 212–232. Ediger, V.S- ., Hos-go¨r, E., Su¨rmeli, A.N., Tatlıdil, H., 2007. Fossil fuel sustainability index: An application of resource management. Energy Policy 35 (5), 2967–2977. Ediger, V.S- ., 2008. National energy report of Turkey: Energy situation, challenges, and policies for sustainable development, in: AASA Beijing Workshop on Sustainable Energy Development in Asia, Beijing, China, InterAcademy council, pp 77–93. Ediger, V.S- ., 2009. Sustainable energy development of Turkey. TU¨BA-Gu¨nce 39, 18–25 in Turkish. Fanchi, J.R., 2000. Oil and gas in the energy mix of the 21st century. Journal of Petroleum Technology 52 (12), 40–46. Gao, W., Madlener, R., 1999. Oil price shocks and macroeconomic performance. Newsletter 1999/2000 (No.1) IHS, 3. Gnansounou, E., 2008. Assessing the energy vulnerability: Case of industrialised countries. Energy Policy 36 (10), 3734–3744. Guo, H., Kliesen, K.L., 2005. Oil price volatility and U.S. macroeconomic activity. Review, Federal Reserve Bank of St. Louis 84 (6), 669–683. Gupta, E., 2008. Oil vulnerability index of oil-importing countries. Energy Policy 36 (3), 1195–1211. Guven, C., 1994. Energy planning under import restrictions. European Journal of Operational Research 72 (3), 518–528. Hamilton, J.D., 1983. Oil and macroeconomy since World War II. The Journal of Political Economy 91 (2), 228–248. Ho, R., 2006. Handbook of Univariate and Multivariate Data Analysis and Inter- pretation with SPSS.. Chapman & Hall/CRC, London, New York. Hughes, L., 2009. The four ‘R’s of energy security. Energy Policy 37 (6), 2459–2461. Huntington, H.G., 1998. Crude oil prices and U.S. economic performance: Where does the asymmetry reside? Energy Journal 19 (4), 107–132. Kahn, G.A., Hampton, R., 1990. Possible monetary policy responses to the Iraqi oil shock. Economic Review, Federal Reserve Bank of Kansas City November/ December, 19–32. development and energy security. Energy Policy 33 (17), 2237–2243. Lo¨schel, A., Moslener, U., Ru¨bbelke, D.T.G., 2010. Indicators of energy security in industrialised countries. Energy Policy 38 (4), 1665–1671. Mork, K.A., 1989. Oil and the macroeconomy when prices go up and down: An extension of Hamilton’s results. The Journal of Political Economy 97 (3), 740–744. Og˘ulata, R.T., 2004. Energy planning for Turkey. Energy Sources 26 (3), 323–335. Oladosu, G., 2009. Identifying the oil price–macroeconomy relationship: An empirical mode decomposition analysis of US data. Energy Policy 37 (12), 5417–5426. Pearson, K., 1901. On lines and planes of closest fit to systems of points in space. Philosophical Magazine 2, 559–572. Percebois, J., 2007. Energy vulnerability and its management. International Journal of Energy Sector Management 1 (1), 51–62. PIGM (General Directorate of Petroleum Affairs of Turkey), 2010. Annual Reports of 1956–2008 on Petroleum Activities of Turkey from 1956 to 2008, Ankara PIGM Publications. Rogoff, K., 2006. Oil and the global economy. Available at /http://www.nes.ru/ public-presentations/Papers/Oil%20and%20the%20Global%20Economy_Ro goff__v2.pdfS. Sill, K., 2007. The macroeconomics of oil shocks. Business Review, Federal Reserve Bank of Philadelphia Q1 (issue no), 21–31. So¨zen, A., 2009. Future projection of the energy dependency of Turkey using artificial neural network. Energy Policy 37 (11), 4827–4833. THM (Undersecretariat of Treasury of Turkey), 2010. Available at /http://www. treasury.gov.tr/irj/portal/anonymous?guest_user=treasuryS. Toman, M.A., 2002. International oil security: Problems and policies. Resources for the Future, Issue Brief no., 02–04. TUIK (Turkish Statistical Institute), 2010. Available at /http://www.turkstat.gov.tr/ Start.doS. Umbach, F., 2010. Global energy security and the implications for the EU. Energy Policy 38 (3), 1229–1240. Uslu, T., 2008. Turkey’s Foreign Dependence on Energy. Journal of Environmental Engineering 134 (12), 113–120. Uslu, T., 2010. The necessity of nuclear-based energy production for Turkey. Energy Sources 5 (2), 155–164. Verrastro, F., Ladislaw, S., 2007. Providing energy security in an interdependent world. The Washington Quarterly 30 (4), 95–104. Vivoda, V., 2009. Diversification of oil import sources and energy security: A key strategy or an elusive objective? Energy Policy 37 (11), 4615–4623. World Bank, 2010, World Development Indicators. Available at /http://data. worldbank.org/data-catalog/world-development-indicatorsS. Yergin, D., 2006. Ensuring energy security. Foreign Affairs 85 (2), 69–82. Yılmaz, A.O., Uslu, T., 2007. Energy policies of Turkey during the period 1923–2003. Energy Policy 35 (1), 258–264. V.S- . Ediger, I. Berk / Energy Policy 39 (2011) 2132–21422142 Crude oil import policy of Turkey: Historical analysis of determinants and implications since 1968 Introduction Energy import structure of Turkey with a special reference to oil Crude oil import dependency of primary energy consumption Share of crude oil in primary energy import The crude oil import bill Non-diversification of oil import sources Crude oil import vulnerability index of Turkey Discussions of results Conclusions and policy implications Acknowledgements ED methodology Partial correlations of factors References