The economics of tobacco in Turkey : new evidence and demand estimates

:t a o . D mm Cu~C a. z Gi ,o I:- p- S0 z E 0 0 _______ Co u 0~~ E 4)~~~~~~~~~~~~~~~~ C ~~~~~~~~~~~~~~~~0   THE ECONOMICS OF TOBACCO IN TURKEY New evidence and demand estimates Zeynep Onder November 2002   Health, Nutrition and Population (HNP) Discussion Paper This series is produced by the Health, Nutrition, and Population Family (HNP) of the World Bank's Human Development Network (HNP Discussion Paper). The papers in this series aim to provide a vehicle for publishing preliminary and unpolished results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual authors whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Editor in Chief. Submissions should have been previously reviewed and cleared by the sponsoring department which will bear the cost of publication. No additional reviews will be undertaken after submission. The sponsoring department and authors bear full responsibility for the quality of the technical contents and presentation of material in the series. Since the material will be published as presented, authors should submit an electronic copy in a predefined format as well as three camera-ready hard copies (copied front to back exactly as the author would like the final publication to appear). Rough drafts that do not meet minimum presentational standards may be returned to authors for more work before being accepted. The Editor in Chief of the series is Alexander S. Preker (aprekeraiworldbank.org; For information regarding this and other World Bank publications, please contact the HNP Advisory Services (healthpop(i).worldbank.org) at: Tel (202) 473-2256; and Fax (202) 522-3234. The Economics of Tobacco Control sub-series is produced jointly with the Tobacco Free Initiative of the Worl( Health Organization. The findings, interpretations, and conclusions expressed in this paper are entirely those o the author/s and should not be attributed in any manner to the World Health Organization or to the World Bank their affiliated organizations or to members of their Executive Boards or the countries they represent. The editors for the Economics of Tobacco Control papers are: Joy de Beyer (idebeyerAworldbank.org), Emmanuel Guindon (guindone@who.int) and Ayda Yurekli (ayurekli@worldbank.org). For free copies of papers in this series please contact the individual author whose name appears on the paper, or one of the editors. Papers are posted on the publications pages of these websites: www.worldbank.org/hnn and www.worldbank.org/tobacco ISBN 1-93212649-X C 2002 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved.   Health, Nutrition and Population (HNP) Discussion Paper ECONOMICS OF TOBACCO CONTROL PAPER NO. 2 The Economics of Tobacco in Turkey: New Evidence and Demand Estimates Zeynep Ondera aFaculty of Business Administration, Bilkent University, Ankara, Turkey Paper prepared for the World Bank, with funding from the US Centers for Disease Control/Office on Smoking and Health Abstract: This report reviews the key issues in the economics of tobacco control in Turkey, including taxation, tobacco control policies, income, price and expenditure elasticity of cigarettes and the link between poverty and tobacco use. In the first part of the report, information is presented on tobacco and cigarette production, cigarette consumption, and government policies related to tobacco, especially taxation and tobacco control. The second part analyses aggregate cigarette consumption. Using aggregate level data from 1960 to 2000, price, tax and income elasticities for cigarettes are estimated. In addition, substitution elasticities among three types of cigarettes, filter, non-filter and foreign cigarettes, are calculated. A simulation analysis estimates the impact that an increase in excise tax would have on cigarette consumption and government tax revenues. The third section examines demand for cigarettes in Turkey at household level using the 1994 Household Expenditure and Consumption Survey. Using these data, some key economic policy issues related to cigarettes are discussed including the impact that changes in the cigarette tax rate and price of cigarettes would have on the smoking prevalence rate and number of cigarettes smoked, controlling for social and demographic factors affecting the decision to smoke. A simulation analysis is conducted to examine the impact of changes of cigarette taxes on government tax revenues and cigarette consumption. Keywords: cigarette consumption Turkey; tobacco tax; tobacco control in Turkey; demand for cigarettes in Turkey; price elasticity of demand for cigarettes in Turkey. Disclaimer: The findings, interpretations and conclusions expressed in the paper are entirely those of the author, and do not represent the views of the World Bank, its Executive Directors, or the countries they represent. Correspondence Details: Zeynep Onder, Bilkent University, Faculty of Business Administration, Bilkent, Ankara 06533, Turkey; phone: (90-312) 290-2038; Fax: (90-312) 2664958; E-mail: zonder@bilkent.edu.tr  Al  Table of Contents ACKNOWLEDGMENTS .................................................................... 1X 1. INTRODUCTION ....................................................................l1 2. THE ROLE OF TOBACCO AND CIGARETTES IN THE TURKISH ECONOMY ...................2 2.1 SMOKING PREVALENCE ....................................................................2 2.2 CIGARETr1E CONSUMPTION ...........3.........................................................3 2.3 CIGARE1TE PRODUCERS ...................................................................4 2.4 TOBACCO GROWING ...................................................................9 2.5 TOBACCO TRADE (EXPORTS AND IMPORTS) .................................................................. 14 2.6 TAXES ON CIGARErTES .................................................................... 17 2.7 GOVERNMENT REVENUES .................................................................. 19 2.8 TOBACCO CONTROL POLICIES AND THEIR ImPLICATIONS ................................................................ 1 9 3. AGGREGATE DATA ANALYSIS .................................................................... 20 3.1. DATA .................................................................. 20 3.2. MODEL .................................................................. 23 3.3. RESULTS .................................................................. 24 4. HOUSEHOLD LEVEL ANALYSIS .................................................................... 34 4.1. DATA .................................................................. 34 4.2. METHODOLOGY .................................................................. 40 4.3. RESULTS .................................................................. 42 4.4. SIMULATION ANALYSIS .................................................................. 49 4.5. SUMMARY AND CONCLUSION .................................................................. 51 REFERENCES .................................................................... 53 STATISTICAL APPENDIX .................................................................... 54 List of Figures Figure 2.1 Price of Different Brands of Cigarettes (Packs of 20 sticks) ................. ....................... 6 Figure 2.2 Share of Tobacco in Cultivated Areas .............................................. ................ 13 Figure 2.3 Tobacco Exports and Imports (Thousand Tons) ......................................................... 16 Figure 2.4 Tobacco Exports and Imports (Thousand US $) ........................................................ 17 Figure 3.1 Cigarette Consumption in Turkey, Period: 1960-2000 ............................................... 21 Figure 3.2 Average Price and Tax on One Cigarette Pack (US $) ............................................... 22 Figure 3.3 Consumption of Cigarettes per Adult per Year by Type ..................... ....................... 31 Figure 3.4 Prices of Cigarettes by Type (TL in terms of 2000 prices) ................. ....................... 31 Figure 3.5 Prices of Cigarettes by Type in US $ .............................................................. 32 Figure 4.1 Smoking Prevalence Rate by Education Level ........................................................... 37 Figure 4.2 Share of Cigarette Expenditures by Education Level ...................................... 37 v  List of Tables Table 1.1 - Domestic Cigarette Consumption in 1990 and 1999 in Five Countries with Fastest Increase, and in the World ...................................................................1l Table 2.1 - Smoking Prevalence Rates in Turkey ................................................................... 3 Table 2.2 - Cigarette Consumption by Type, 1987 to 2000 ..........................................................4 Table 2.3 - Tekel Domestic Sales Volumes, 1992-2000 (million pieces or thousand kilograms) .5 Table 2.4 - Sales of Some Private Companies in Turkey, 1998-2000 (million TL) ......................6 Table 2.5 - Exports of Companies in Cigarette/Tobacco Industry .................................................7 Table 2.6 - Before Tax Profits, Cigarette/Tobacco Industry Companies (million TL) ..................8 Table 2.7 - Tobacco Purchasing and Cigarette Production at Tekel in 1992-2000 ........................8 Table 2.8 - Employment at Tekel ....................................................................9 Table 2.9 - Employment in Private Tobacco/Cigarette Companies ................................................9 Table 2.10 - World Tobacco Production, and Twelve Biggest Producers, 1995-1998 ................ 10 Table 2.11 - World Oriental Tobacco Production, and Twelve Largest Producers, 1995-1998.. 10 Table 2.12 - Tobacco Growers in Turkey, 1990-1998 11 Table 2.13 - Tobacco Area per Grower by Year and Region (square meters) 11 Table 2.14 - Tobacco Production, Tobacco Area and Yield, 1970-1999 ..................................... 12 Table 2.15 - Average Tobacco Prices, 1995-1998 ................................................................... 14 Table 2.16 - Tobacco Exports and Imports of Turkey ................................................................. 15 Table 2.17 - Major Tobacco Trading Partners of Turkey in 2000 ............................................... 16 Table 2.18 - Taxes on Cigarettes in 2000 ................................................................... 17 Table 2.19 - Separation of Different Components of Cigarette Prices ......................................... 18 Table 2.20 - Taxes Obtained from Tekel Cigarette Sales, 1996-2000 (billion TL) ...................... 19 Table 3.1 - Descriptive Statistics of Variables Used in the Estimations for 1960-2000 ............... 24 Table 3.2 - Results of Demand for Cigarettes ................................................................... 26 Table 3.3 - Impact of Tax Changes on Consumption and Government Revenues ....................... 27 Table 3.4 - Impact of Tax Changes on Government Revenues with Smuggling .......................... 28 Table 3.5 - Results of Myopic Addiction (N=40) ................................................................... 29 Table 3.6 - Results of Rational Addition Model (N=39) .............................................................. 29 Table 3.7 - Prices and Consumption of Cigarettes by Type, 1987-2000 ...................................... 30 Table 3.8 - Regression Results for Different Types of Cigarettes ................................................ 33 Table 4.1 - Brands and Shares of Cigarettes Used by Households ............................................... 35 Table 4.2 - Smoking Prevalence, Number of Smoker Households by Type of Cigarette and Income Group, 1994 ................................................................... 36 Table 4.3 - Household Preference by Income Group .................................................................... 36 Table 4.4 - Average Packages of Cigarettes Consumed per adult per household, Price and Tax Paid by Smoker Households by Type of Cigarette and Income Group ................................ 38 Table 4.5 - Taxes Paid on Cigarettes ................................................................... 39 Table 4.6 - Alternative Ways of Spending Cigarettes Money ...................................................... 40 Table 4.7 - Estimated Coefficients of Tax Equation .................................................................... 43 Table 4.8 - Estimated Coefficients on Price Equation, by Income Quintile ................................. 44 Table 4.9 - Estimated Logit Coefficients for Smoking Prevalence Rates ..................................... 45 Table 4.10 - Estimated Coefficients for Total Packs Smoked ...................................................... 47 Table 4.11 - Price Elasticity ................................................................... 48 vi  Table 4.12 - Income Elasticity .................................................................. 49 Table 4.13 - Impact of Price Increase on Household Smoking Prevalence Rate .......................... 49 Table 4.14 - Impact of Tax Increase on Government Revenues ............................ ....................... 50 Table 4.15 - Collection of Tax From Different Income Groups ............................ ....................... 51 Table 4.16 - Comparison of 1987 and 1994 Household Expenditures ................. ........................ 52 Table Al - Ratio of Tax Rates to the Retail Price of Cigarettes in 1994 ............... ....................... 54 Table A2 - Test for Endogeneity of Price in the Logit Estimation ...................... ......................... 55 Table A3 - Test for Endogeneity of Price in the Model for Number of Cigarettes Smoked ........ 56 vii  viii  ACKNOWLEDGMENTS The author is grateful to Ayda Yuirekli for her help in various stages of this paper. She thanks Joy de Beyer for her editing, valuable comments and suggestions. Many people from State Institute of Statistics, Tekel, Department of Finance and the Turkish Treasury helped provide the data used in the analysis; the authors thanks all of them. The author is also grateful to the World Bank for having published the report as an HNP Discussion Paper. ix  x  1. INTRODUCTION Turkey has become one of the largest consumers of cigarettes. While world cigarette consumption declined by 4 percent between 1990 and 1999, in Turkey, it rose by 52 percent; the third largest increase in the world after Pakistan and Bulgaria, according to U.S. Department of Agriculture data. Turkey's domestic cigarette consumption increased from 73,270 million tons in 1990 to 111,500 million tons in 1999 (Table 1.1). Per capita consumption of cigarettes was 136 packs per adult in 1999. Table 1.1 - Domestic Cigarette Consumption in 1990 and 1999 in Five Countries with Fastest Increase, and in the World Domestic Consumption Domestic Consumption Percentage Change in in 1990 (million tons) in 1999 (million tons) Domestic Consumption Countries 1990-99 Pakistan 31,934 54,500 71.66 % Bulgaria 12,200 19,000 55.74 % Turkey 73,270 111,500 52.18 % Indonesia 140,936 207,685 47.36 % The Netherlands 23,251 30,124 29.56 % World (Total) 4,538,890 4,351,770 -4.12 % Source: USDA World Cigarette Consumption in Selected Countries. The rapid increase in tobacco consumption in Turkey will cause a large increase in the burden of disease and premature death in years to come. Evidence shows that half of all long-term tobacco users are killed as a result of their addiction to tobacco products, and half of these deaths occur prematurely. A huge body of scientific research has clearly established that tobacco use raises the risk of developing many diseases, including cancer of the lung, bladder, kidney, larynx, mouth, pancreas and stomach, heart attacks, strokes and other circulatory diseases, and respiratory diseases including emphysema. Global evidence also shows clearly that there is a set of policies and interventions that can be highly effective in reducing tobacco use, encouraging smokers to quit, and deterring young people from starting to smoke. These are: using tax rates to raise prices of cigarettes and other tobacco products; complete bans on all advertising and promotion of tobacco products and associated logos and trademarks; bans on smoking in public places, especially in enclosed spaces; good information on the health risks caused by tobacco use and the benefits of quitting, including strong, large clear warnings on cigarette packages, and help for people who want to quit. People who oppose strong measures to reduce tobacco use usually comment that tobacco makes an important economic contribution, through creating jobs and incomes, and generating tax revenue. This study shows that even in Turkey, one of the largest tobacco growing and tobacco exporting countries in the world, tobacco contributes only a very limited amount to exports, jobs and production. Tobacco exports account for only about 1% of all Turkey's export value, and a much smaller net amount, since tobacco and cigarette imports have been growing considerably. Cigarette producers employ only 0.13% of the workforce in Turkey, and although there are about 600,000 farmers who grow tobacco, most grow relatively small amounts (less than half a hectare on average, and the total land under tobacco is only around 1% of all cultivated land in Turkey. 1  Cigarettes taxes provide an estimated 6% of total government revenues, but this share has been decreasing as Turkey's economy has developed, diversified and modernized. Moreover, the policy of supporting tobacco farmers by buying agreed quotas of tobacco leaf at fixed prices has been a large fiscal burden in some years. This study was undertaken to review the data on the economic contribution that tobacco makes to Turkey's economy, and the information on tobacco use. There are no studies (that we were able to find) that show the extent of the economic damage that tobacco use causes, in health care costs bome by the state and by individuals, in lost eamings and productivity because of sickness and premature death of smokers and others who inhale their smoke, and in the opportunity cost especially to families, especially of smokers from low-income households, who spend scarce family resources on cigarettes instead of on products that could improve the health or well-being of the family. The analysis of household data shows that although cigarette prices have only a limited impact on smoking prevalence (the decision whether or not to smoke), they have a much stronger impact on the decision as to how much to smoke. At higher prices, the percentage of smokers would drop a little, but smokers -especially smokers with lower incomes- would cut back significantly. The study's most important conclusions are that higher tobacco taxes would help reduce tobacco use and generate govermment revenue, but that in themselves they will not be enough. Additional strong measures need to be taken and enforced, to reduce tobacco use in Turkey, and protect the future health of Turkish adults and young people. 2. T'=2 -| ,E 1 1E T--1)vlllb 1]l'1 _0|.E LLl.: ON THSE YINJL -b' 1-. [-u- ! : L rlSl- This section provides data on the role of tobacco and cigarettes in the Turkish economy. It reviews data on cigarette consumption, and cigarette production. The role of tobacco in Turkish agriculture and foreign trade is examined. Data on cigarette taxes and their contribution to govemment revenues are presented. Finally, tobacco control policies in Turkey are described. 2.1 SMOKiN( G R E ' A LE- 6, ,E The adult smoking prevalence rate was 62.8 percent for men and 24 percent for women aged 15 or older in 1988 (Bilir, 1997). This is the only study that covers the whole country. Other (more recent) studies examine smoking prevalence rates for specific groups. For example, the most recent study by Bilir (2000) showed that the smoking prevalence rate among young people is high, despite some small apparent decline between 1998 and 1999. Among a sample of 6,715 (5,792) young people in 1999 (1998) living in 34 districts in Turkey, Bilir (2000) found that 1 (2) percent of 7th grade students and 15 (16) percent of 12'' grade students were smokers in 1999 (1998), and 30 percent of 12t graders had ever smoked. Table 2.1 shows smoking prevalence rates for different groups in Bilir's surveys. Among adults, drivers have the highest prevalence rate, 74 (70) percent and religious leaders have the lowest rate, 25 percent. 2  Table 2.1 - Smoking Prevalence Rates in Turkey Never Smoked (%) Quit (%) Smoker (%) Groups 1998 1999 1998 1999 1998 1999 Students (7th Grade) 87.1 93.1 10.9 6.1 2.0 0.8 Students (12th Grade) 69.1 70.1 14.6 15.1 16.3 14.8 Teachers 38.6 37.0 14.1 14.5 47.2 48.5 Physicians 41.7 39.2 17.2 17.7 41.1 43.1 Police Officers 26.6 22.2 12.8 13.2 60.6 64.6 Religious Leaders 53.9 49.8 21.6 25.1 24.5 25.1 Drivers 17.4 14.8 12.5 10.9 70.1 74.3 Source: Bilir and Onder, 2000 2.2 CIGARETTE CONSUMPTION There are three basic types of cigarettes sold in Turkey: a) domestic filter cigarettes, b) domestic non-filter cigarettes and c) foreign brand cigarettes. Table 2.2 shows the domestic sales of these cigarettes from 1987 to 2000. Foreign brand cigarettes include both imported and domestically produced foreign brand cigarettes, which have been produced in Turkey since 1991. The market share of foreign brands has increased dramatically and consumption of unfiltered cigarettes has declined. Total consumption of cigarettes increased by 80 percent over the last decade (Table 2.2), an average annual increase of 5 percent.' Some of the rise is due to increasing population size, but cigarette consumption per adult also increased. The largest increase in consumption occurred in foreign brand cigarettes, consumption of which increased 46 fold from 1990 to 2000, from a negligible market share to 32 percent of the market. Total consumption of filtered domestic cigarettes increased by 31 percent but the market share fell from 92 percent to 67 percent. Total consumption of non-filtered cigarettes fell by 78 percent over this period, and the market share fell from 6 percent to less than 1 percent. The increase in the consumption of foreign brands can be explained by the several factors. First, the restrictions on imports of these products were eliminated. Second, these cigarettes started to be produced in Turkey. Third, the government allowed these companies to promote their cigarettes. ' These data on domestic cigarette sales in Turkey from the Annual Reports of TEKEL show a larger increase than the USDA aggregate data reported in Section 1. The USDA data use aggregate leaf production, net of imports and exports, and convert leaf volume to cigarette equivalents. 3  Table 2.2 - Cigarette Consumption by Type, 1987 to 2000 Total Consumption Consumption per Adult (Million Pieces) (Packs with 20 pieces) Year Foreign Filter Non-filter Foreign Filter Non-filter Total 1987 859 66,184 7,083 1.27 97.90 10.48 109.65 1988 853 65,708 7,032 1.23 95.02 10.17 106.42 1989 546 42,068 4,502 0.77 59.47 6.36 66.61 1990 793 58,964 4,091 1.08 80.27 5.57 86.92 1991 690 67,619 3,810 0.92 90.59 5.10 96.62 1992 7,003 63,068 5,324 9.23 83.15 7.02 99.40 1993 10,180 67,479 4,853 13.21 87.55 6.30 107.05 1994 15,228 72,345 7,054 19.44 92.37 9.01 120.82 1995 17,812 73,627 5,061 22.38 92.51 6.36 121.25 1996 22,974 72,112 3,423 28.41 89.16 4.23 121.80 1997 28,634 68,613 2,783 35.02 83.91 3.40 122.33 1998 34,410 72,825 2,065 41.41 87.64 2.49 131.54 1999 35,108 78,952 1,440 41.58 93.50 1.71 136.79 2000 37,535 77,097 903 43.75 89.85 1.05 134.65 Source: State Institute of Statistics and TEKEL Annual Reports. 2.3 Cl't.[.ETE 2rIYCIiIIz:E%3 Cigarette production in Turkey started in the 19th century in small shops called atalye. The first cigarette factory was established in 1939, and produced hand-made non-filter cigarettes and packed tobaccos. In 1958 Tekel started to produce filter cigarettes, the demand for which increased steadily. The factory for the first machine made cigarettes started to operate in 1969. At the end of the 1970s, cigarette factories had problems importing some raw materials and spare parts because of a shortage of foreign exchange in Turkey. This caused a decline in cigarette production and factories operated below capacity. In order to satisfy domestic demand, tobacco was sent to factories in Yugoslavia and Bulgaria, and the cigarettes imported into Turkey. In 1983, Tekel started to import cigarettes with foreign brand names. As foreign brands became popular, Tekel started to produce new cigarettes that use blends of American tobaccos. Until the early 1980s, Tekel, a government owned company, had a monopoly over manufacturing and sale of tobacco products in Turkey. Its factories produced poor quality cigarettes, and Tekel's inadequate distribution system made it necessary for store owners to visit Tekel's warehouses to pick up their inventories. Tekel rarely advertised and never promoted its products to women or young people (Dagli, 1999). In 1984, the Turkish government allowed foreign tobacco companies to export their products to Turkey, while Tekel maintained the exclusive right to import, price and distribute domestic and foreign tobacco products. Later, the Turkish government eased the control on foreign tobacco products, and in 1991 other tobacco companies were allowed to price and distribute their own cigarettes in Turkey. One of the immediate effects of this new regulation was the establishment of Philsa in 1991, a joint venture of Philip Morris with one of the largest holding companies in 4  Turkey, Sabanci Holding. Philip Morris owns 75 percent and Sabanci Holding has a 25 percent ownership stake. Their factory started to operate in 1992. With the entrance of foreign companies into the Turkish tobacco market, Tekel's power and market share started to decline. Although Tekel's domestic sales volume and value increased (Table 2.3), Tekel's market share fell from 82 percent to 70 percent between 1995 and 1997. Philsa increased its market share from 15 percent to 23 percent in this period, to become the dominant foreign cigarette company operating in Turkey. The second foreign tobacco company, R.J. Reynolds, increased its share from 3 percent to 7 percent (IMF, 1999). (In 1999, Japan Tobacco International (JTI) bought the international operations of R.J. Reynolds.) All three major producers of cigarettes in Turkey -- Tekel, Philsa and JTI-are among the largest companies in Turkey. In 2000, Tekel was the 6th largest company, Philsa ranked 20th, and JTI was 53rd. Three other companies operating in the tobacco industry usually make the list of the largest 500 companies: Salatab Tobacco Company, Dimon Turkish Tobacco Company and Sunel Tobacco Trading Company. In 2000, a new tobacco company, TTL Tobacco was also included in the list. Table 2.3 - Tekel Domestic Sales Volumes, 1992-2000 (million pieces or thousand kilograms) 1992 1993 1994 1995 1996 1997 1998 1999 2000 Domestic Sales Volume (Million pieces, Thousand kg) Filter Cigarettes of 63,068 67,479 72,345 73,627 72,112 68,613 72,825 78,952 77,097 Tekel ilter Cigarettes 567 4,013 4,825 36 22 0 - - Outside of Tekel Nonfilter Cigarettes 5,324 4,853 7,054 5,061 3,423 2,783 2,065 1,440 903 Imported Tobacco 9,979 8,663 982 131 73 81 62 54 51 Products Other Tobacco 462 399 419 350 293 259 157 127 125 Products Domestic Sales Value (Billion TL) ilter Cigarettes of 12,936 23,081 51,386 80,538 143,620 288,481 480,200 820,223 1,621 Tekel Filter Cigarettes 194 3,068 7,268 82 45 - - - Outside of Tekel onfilter Cigarettes 171 251 989 1,018 1,488 2,853 3,806 5,157 9,13 ported Tobacco 4,845 5,566 1,190 333 303 286 169 101 218 Products Other Tobacco 25 30 66 92 173 368 513 911 1,422 Products Total 18,171 31,996 60,899 82,064 145,630 291,989 484,688 826,392 1,631,91 Source: Tekel, Annual Reports, 1994-2000.  Table 2.4 - Sales of Some Private Companies in Turkey, 1998-2000 (million TL) 1998 1999 2000 Philsa 96,880,998 426,070,651 233,500,223 RJ Reynolds/JTI Tobacco Products n.a. 66,433,985 116,714,923 Dimon Turkish Tobacco Corp. 16,417,426 26,764,370 35,751,246 unel Trading Company 7,683,631 10,054,006 n.a. TTL Tobacco Industry Corp n.a. n.a. 24,874,965 Salatab Tobacco Corp 36,445,178 n.a. n.a. n.a. means not available. Source: Capital, 500 Largest Companies in Turkey. To compete with Philip Morris brands, Tekel introduced a new American blend cigarette, Tekel 2000. About two-thirds of Tekel's production are oriental blend cigarettes, the rest are American blends. Tekel produces the following cigarettes: Tekel 2000, Tekel 2001, Samsun, Maltepe. Yeni Harman, Meltem, Ballica, Bafra, Bitlis, and Birinci. The second company, Philsa, has a factory in the western part of Turkey, Torbali, Izmir, with a total investment of more than $200 million and capacity to produce 35 billion cigarettes per year. It produces American blend products, including Malboro, Parliament, L&M and Chesterfield. The third cigarette producer, Japan Tobacco International, has a plant in the same part of Turkey that produces 13 million sticks of blended cigarettes per year including Camel, Winston, Salem, Magna, Monte Carlo and Aspen brands. Price trends Figure 2.1 shows the price of different cigarette brands sold in Turkey between 1975 and 2000. Prices are expressed in US dollars for comparisons over time. Prices of domestic cigarettes fluctuated less than prices of American blend cigarettes. The government adjusted Tekel's prices periodically in response to the high inflation rate, and private companies adjusted their prices to maintain more-or-less constant differentials. Figure 2.1 Price of Different Brands of Cigarettes (Packs of 20 sticks) ,s 1 6 1 4 Parham-,i 1 2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Cn1-~ .. r .. I 4,'' . ,' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Tekel2000 08 06 / .sa~Long 06 - > ,G - _ _. - . --.. _'- /' San,,ms8 t S _- -_ ,_ _ (B., Bm,Y-m ) 02 _ _ 0 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1909 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Ycar Source: State Institute of Statistics 6  Exports Table 2.5 shows cigarette and tobacco exports. Panel A shows Tekel's exports and Panel B shows exports of private companies that were among the largest 500 companies in Turkey in 1998, 1999 and 2000. The combined exports of all tobacco companies amount to only 1.4 percent of total exports in Turkey in 1998, and this declined to 1.1 percent in 2000. In recent years, Tekel has accounted for approximately one-third of all tobacco exports. Tekel's exports rose in the first half of the 1990s and declined after 1996. Table 2.5 - Exports of Companies in Cigarette/Tobacco Industry Panel A - Exports of Tekel , 1992 1993 1994 1995 1996 1997 1998 1999 2000 Exports Volume (million pieces, Thousand Kg) Supported Leaf Tobacco 34,115 26,989 88,791 92,205 140,474 82,485 51,665 35,858 45,443 Blended Leaf Tobacco 68 17 51 17 - - - - Cigarette 2,407 4,176 1,889 1,423 2,426 3,460 494 218 33 Exports Value (Thousand $) SupportedLeafTobacco 98,985 64,320 144,510 173,148 367,518 198,111 131,493 94,275 98,121 Blended Leaf Tobacco 303 76 226 76 igarette 19,568 40,046 4,487 16,355 31,351 44,809 3,638 1,483 329 otal 118,856 104,442 149,223 189,579 398,869 242,920 135,131 95,758 98,45 Source: TEKEL Annual Reports Panel B - Exports of Private Companies (Thousand $) 1998 1999 2000 Philsa 82,172 n.a. 19,785 RJ Reynolds / JTI n.a. 85,960 114,473 Salatab Tobacco Corp. 86,351 n.a. n.a. Dimon Turkish Tobacco Corp. 51,030 54,721 43,680 Sunel Trading Company 32,000 27,342 n.a. TTL Tobacco Corp. n.a. n.a. 35,789 Total 251,553 168,023 213,727 Source: Capital, Largest 500 Companies of Turkey Profits Table 2.6 shows total profits of companies in the tobacco industry. Profits fell in 1999 and rose dramatically in 2000. Tekel's profit decline in 1999 was caused primarily by purchases of tobacco to support growers, which accounted for three fourths of Tekel's purchases; only one- fourth of total tobacco purchases were for production (Table 2.7).  Table 2.6 - Before Tax Profits, Cigarette/Tobacco Industry Companies (million TL) Panel A - Private Tobacco Companies 1998 1999 200 Philsa 2,085,475 1,897,723 12,097,173 JTI Tobacco Products Corp. n.a. 11,904,125 32,376,385 Solotab Tobacco Corp. 672,213 n.a. n.a. Dimon Turkish Tobacco Corp. 248,996 5,002,142 8,149,265 Sunel Trading Company 678,363 398,145 n.a. TTL Tobacco Industry Corp. n.a. n.a. 7,013,440 Panel B - Tekel I I 19951 19961 19971 19981 19991 200( Net Profit (Billion TL) 7,310 4,0851 24,027 15,173 -58,720 68,93 Source: Panel A: Capital, 500 Largest Companies in Turkey; B: TEKEL Annual Report Table 2.7 - Tobacco Purchasing and Cigarette Production at Tekel in 1992-2000 1992 1993 1994 1995 1996 1997 1998 1999 200 Volume ('000 kg) Leaf Tobacco 151,801 232,872 271,664 110,930 107,933 124,104 199,915 185,583 177,55 Need of Tekel 50,000 35,000 41,000 48,442 61,996 60,952 39,330 48,601 28,88 Supported 101,801 197,872 230,664 62,488 45,937 63,152 160,585 136,982 148,66 Value (Billion TL) Leaf Tobacco 3,903 7,603 14,943 14,324 24,066 59,303 139,217 205,698 237,26: NeedofTekel 1,304 1,161 1,518 4,949 11,776 24,014 27,657 55,530 37,75: Supported 2,599 6,441 13,425 9,376 12,290 35,289 111,560 150,167 199,51 Source: TEKEL Annual Reports Tekel has many factories: in 2000, there were 108 establishments for leaf tobacco processing and trading and 7 cigarette manufacturing plants. The total number of Tekel employees in 2000 was 26,334, most (73 percent) employed in tobacco processing and trading establishments, about half of whom are seasonal workers. Table 2.8 shows employment in all Tekel establishments.  Table 2.8 - Employment at Tekel 1995 1996 1997 1998 1999 2000 Leaf Tobacco Processing and Trading Establishment Officials 104 107 132 125 124 122 Contractual Employees 1,090 1,230 1,353 1,987 1,953 1,956 Permanent Workers 10,336 9,252 8,935 8,724 8,253 7,875 Seasonal Workers 8,326 9,587 8,028 9,891 8,489 9,348 Total 19,856 20,176 18,448 20,727 18,819 19,301 Cigarette Industiy Establishments fficials 64 58 54 38 52 49 Contractual Employees 329 319 377 629 612 554 Permanent Workers 7,695 7,138 6,636 6,146 6,020 5,913 Seasonal Workers 120 332 465 451 517 Total 8,088 7,635 7,399 7,278 7,135 7,033 otal Number of Employees 27,944 27,811 25,847 28,005 25,954 26,33 Source: TEKEL Annual Reports Table 2.9 shows the number of workers employed in the private tobacco companies that were on the largest 500 companies list. Together they provided 0.13 percent of total employment in Turkey, and 0.74 percent of industrial sector employment in Turkey in 2000. Table 2.9 - Employment in Private Tobacco/Cigarette Companies 1998 1999 2000 Tekel 28,005 25,954 26,33 Philsa 715 710 713 JTI Tobacco Products Corp. na 306 31 Salatab Tobacco Corp 784 na n Dimon Turkish Tobacco Corp. 590 413 423 Sunel Trading Company 310 325 n TTL Tobacco Corp. na na 339 Total 30,404 27,708 28,123 Source: Capital, 500 Largest Companies in Turkey 2.4 TOBACCO GROWING Turkey is the fifth largest tobacco producer in the world (Table 2.10). In 1998, world production was 7,066,875 tons and Turkey produced 258,811 tons, 3.7 percent of the total. Turkey is the world's leader in oriental tobacco production (Table 2.11). In 1998 total oriental production was 702,733 tons, of which Turkey produced 35.7 percent  Table 2.10 - World Tobacco Production, and Twelve Biggest Producers, 1995-1998 1995 1996 1997 1998 Tons % tons % Tons % tons % China 2,317,700 36.47 3,076,000 42.08 3,390,000 42.10 2,524,500 35.72 India 587,100 9.24 562,750 7.70 623,700 7.75 635,000 8.99 .S.A. 575,380 9.05 688,258 9.42 810,154 10.06 696,116 9.85 razil 398,000 6.26 439,000 6.01 576,600 7.16 442,500 6.26 Turkey 204,440 3.22 230,949 3.16 302,008 3.75 258,811 3.66 Indonesia 171,400 2.70 177,000 2.42 184,300 2.29 175,631 2.49 Zimbabwe 209,042 3.29 207,767 2.84 192,107 2.39 212,050 3.00 Italy 124,492 1.96 130,590 1.79 131,410 1.63 132,000 1.87 reece 131,875 2.08 131,000 1.79 132,450 1.64 132,000 1.87 Malawi 130,686 2.06 142,262 1.95 158,615 1.97 142,300 2.01 Argentina 79,010 1.24 98,200 1.34 123,200 1.53 117,300 1.66 Pakistan 80,917 1.27 80,760 1.10 86,279 1.07 90,450 1.28 Others 1,344,945 21.16 1,345,269 18.40 1,342,005 16.67 1,508,217 21.34 otal 6,354,987 100.00 7,309,805 100.00 8,052,828 100.00 7,066,875 100.00 Source: State Planning Organization, The 8 Five-Year Planning Report, 1998 Table 2.11 - World Oriental Tobacco Production, and Twelve Largest Producers, 1995- 1998 1995 1996 1997 1998 Tons % tons % tons % Tons % Turkey 199,434 40.18 224,962 38.40 295,506 44.96 251,109 35.73 Greece 85,000 17.12 83,500 14.25 83,500 12.70 82,800 11.7 ulgaria 16,339 3.29 33,392 5.70 33,392 5.08 65,359 9.3 Moldova 24,366 4.91 23,913 4.08 23,913 3.64 32,608 4.64 Macedonia 0.00 30,000 5.12 30,000 4.56 30,000 4.2 Kyzgistan 19,000 3.83 30,000 5.12 30,000 4.56 30,000 4.2 hina 15,000 3.02 20,000 3.41 20,000 3.04 25,000 3.5 Uzbekistan 22,000 4.43 24,700 4.22 24,700 3.76 23,300 3.3 Pakistan 10,553 2.13 14,101 2.41 14,101 2.15 17,030 2.4 Thailand 9,000 1.81 12,500 2.13 12,500 1.90 15,500 2.21 Iran 12,500 2.52 12,500 2.13 12,500 1.90 12,500 1.78 Italy 11,083 2.23 11,271 1.92 11,271 1.71 12,343 1.7 Others 72,117 14.53 64,946 11.09 65,910 10.03 105,184 14.97 Total 496,392 100.00 585,785 100.00 657,293 100.00 702,733 100.0( Source: State Planning Organization, The 8' Five-Year Planning Report, 1998 Tobacco growing in Turkey is usually carried out as a family business. There were around half a million tobacco growing families during the last decade, gradually increasing to over 600,000 in 1998 (Table 2.12). Most of the tobacco growers are in the Aegean region in western Turkey. On average each tobacco growing family plants 0.45 hectare (4,500 square meters) with tobacco 10  (Table 2.13). The average area allocated for tobacco production increased in 1998 compared to the average area in 1995. Table 2.12 - Tobacco Growers in Turkey, 1990-1998 Year Number of Tobacco Growers 1990 521,952 1991 468,361 1992 526,385 1993 543,923 1994 494,298 1995 550,016 1996 546,671 1997 560,380 1998 622,063 Source: State Planning Organization, The 8 Five-Year Planning Report, 1998 Table 2.13 - Tobacco Area per Grower by Year and Region (square meters) Year Aegian Marmara Black sea Eastern Southeastern Average 1995 4,110 3,350 3,592 2,477 3,452 3,810 1996 4,871 4,208 4,000 3,461 3,472 4,355 1997 6,699 4,471 4,450 3,087 5,135 5,751 1998 4,731 4,510 4,197 2,417 4,474 4,475 Source: State Planning Organization, The 8' Five-Year Planning Report, 1998 The total area cultivated with tobacco fluctuates from 1970 to 2000 (Table 2.14, Figure 2.2). It declines until the mid 1980s, and then tobacco increases as a proportion of the total cultivated land area. The yield per hectare increased from a very low level in the early 1970s, to 915.45 kg tobacco per hectare in 1999.  Table 2.14 - Tobacco Production, Tobacco Area and Yield, 1970-1999 Year Area (ha) Production (ton) Yield (kg/ha) 1970 328,498 149,861 456.20 1971 335,627 173,861 518.02 1972 352,383 179,799 510.24 1973 322,840 149,120 461.90 1974 230,149 203,487 884.15 1975 241,508 199,935 827.86 1976 315,315 323,963 1027.43 1977 276,550 247,952 896.59 1978 299,299 292,563 977.49 1979 232,505 216,585 931.53 1980 222,997 228,349 1024.00 1981 177,166 168,024 948.40 1982 206,113 207,735 1007.87 1983 229,544 233,843 1018.73 1984 188,494 177,529 941.83 1985 176,848 170,491 964.05 1986 169,919 158,480 932.68 1987 206,247 184,712 895.59 1988 237,068 219,063 924.05 1989 284,768 269,888 947.75 1990 320,236 296,008 924.34 1991 281,701 240,881 855.09 1992 331,158 334,321 1009.55 1993 339,860 338,800 996.88 1994 226,928 187,733 827.28 1995 209,919 204,440 973.90 1996 237,992 230,949 970.41 1997 322,500 302,008 936.46 1998 278,350 258,811 929.80 1999 283,444 259,478 915.45 Source: State Institute of Statistics 12  Figure 2.2 - Share of Tobacco in Cultivated Areas 35 1 - 1.6 > ~~~~~~~~~~~~~~~~~~~~~~142 20~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 12 8 15 0~~~~~~~~~~~~~~~~~~~~~~~~~~~ .0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ lo 10 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 5 -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' 0 2 o,N ,o,> ,o 9< > 9S 9 9 9 9 Years Source: State Institute of Statistics The major tobacco purchaser is Tekel, which buys tobacco for producing cigarettes and other tobacco products and also to implement the government's policy of support for tobacco farmers. As Table 2.7 shows, Tekel purchased far more tobacco for support purposes than for production, with tobacco bought for support reaching nearly 84 percent of total tobacco purchases in 2000. In addition to Tekel, 15 private tobacco companies and merchants purchased tobacco in 1998. The total amount of tobacco produced in 1998 and purchased in 1999 was 252 thousand tons. Tekel purchased 186 thousand tons (74 percent), and private companies bought the other 26 percent. Until 1994, the government followed a support price policy for tobacco that was implemented by Tekel. Each year, Tekel experts estimated the amount of tobacco production based on the cultivated area declared by farmers. Then the government set prices for three different quality grades of tobacco (Grade A, Grade B, and Kappa, the lowest grade) by region. Tekel was required to purchase all tobacco not purchased by private companies at these prices in order to support farmers. In December 1993, the government announced a new support program for tobacco in which quotas were set, which took effect from the 1994 crop year. Tobacco production declined from 340 thousands tons to 227 thousands tons. As part of this new program, an extra premium and compensation were paid to farmers. Table 2.15 shows average prices paid by Tekel and private tobacco merchants from 1995 to 1998. Private companies paid higher price per kilogram than Tekel. Historically, government has increased average prices by more than the inflation rate, but in recent years, government increased prices by less than the inflation rate. 13  Table 2.15 - Average Tobacco Prices, 1995-1998 Tekel Private Total Crop Purchases Change Purchases Change Average Change Inflation Year (TL) % (TL) % (TL) % Rate 1995 189,171 71.2 234,746 91.5 210,930 82.9 93.63 1996 410,619 117.0 447,017 103.2 440,752 108.9 79.41 1997 696,387 69.6 798,928 67.4 729,197 65.6 84.99 1998 1,108,385 59.1 1,198,210 49.9 1,131,843 55.0 83.61 Source: Source: State Planning Organization, The 8' Five-Year Planning Report, 1998 2.5 TOBACCO TL-- ;.!E (EXP?OIRTS X 'T.012?7fRTS) Turkey is a net exporter of tobacco. According to the USDA, Turkey ranked the 4th among tobacco exporting countries in terms of volume in the 1990s and the 12*' among tobacco importing countries. During the last decade, an average of 139,000 tons of tobacco were exported each year. Imports averaged 24,000 tons during the first half of the decade, and around 53,000 tons during the second part of the last decade. Unprocessed tobacco constituted 92 percent of all tobacco exported and 75 percent of total tobacco imported. Table 2.16 shows tobacco imports and exports of unprocessed and processed tobacco over the last thirty years. Tobacco constituted 1.8 percent of exports and 0.6 percent of imports in 2000. Turkey's major trading partner was the U.S.A. both in exports and imports (Table 2.17). Figures 2.3 and 2.4 present the trends in tobacco imports and exports over the last 20 years. The quantity exported fluctuated a lot, increasing in recent years. The quantity imported grew fairly rapidly after ban on tobacco imports was removed. As with quantity, the monetary value of tobacco imports and exports fluctuated over the last 30 years. The composition of tobacco imports and exports has changed over the years. In the 1970s, Turkey was not exporting processed tobacco but only unprocessed tobacco. The share of processed tobacco has increased over time, although unprocessed tobacco is still the major part of tobacco exports. Processed tobacco constituted 25 percent of all tobacco imported in 1999. 14  Table 2.16 - Tobacco Exports and Imports of Turkey EXPORTS IMPORT Unprocessed Processed Total Unprocessed Processed Total Unprocessed Processed Total Unprocessed Processed Total (thousand (thousand (thousand (Thousand (Thousand (Thousand (thousand (thousand (thousand (Thousand (Thousand (Thousand Year tons) tons) tons) US$) US$) US$) tons) tons) tons) US$) US$) US$) 1970 74,027 71,819 0 71,819 0 0 0 0 0 1971 84,412 79,251 0 79,251 0 0 0 0 0 0 1972 124,485 126,797 0 126,797 0 0 0 0 0 0 1973 108,443 129,759 0 129,759 0 0 0 0 0 1974 112,374 205,671 0 205,671 0 0 0 0 0 1975 65,640 183,222 0 183,222 0 0 0 0 0 0 1976 75,174 251,299 0 251,299 0 0 0 0 0 0 1977 61,836 175,827 0 175,827 0 0 0 0 0 0 1978 77,335 225,261 0 225,261 0 0 0 0 0 0 1979 69,554 176,971 0 176,971 0 5 5 0 40 4 1980 83,727 0 83,727 233,742 1 233,743 6 0 6 46 0 46 1981 130,969 22 130,991 395,013 135 395,148 0 0 0 0 0 0 1982 104,906 8 104,914 348,320 47 348,367 0 0 0 0 0 0 1983 69,529 14 69,543 237,758 70 237,828 0 0 0 0 0 0 1984 69,718 144 69,862 216,357 611 216,968 0 2,768 2,768 0 26,585 26,585 1985 102,726 64 102,790 330,144 310 330,454 0 5,978 5,978 0 55,870 55,870 1986 81,951 229 82,180 270,226 1,595 271,821 2 11,056 11,058 59 116,225 116,28 1987 106,322 580 106,902 313,954 1,854 315,808 4 15,678 15,682 145 177,955 178,100 1988 77,683 777 78,460 266,001 3,246 269,247 610 13,693 14,303 3,602 169,680 173,28 1989 116,868 196 117,064 479,083 729 479,812 4,081 9,366 13,447 23,561 197,459 221,020 1990 94,770 47,233 142,003 418,491 23,868 442,359 3,279 14,021 17,300 21,429 315,766 337,195 1991 136,572 895 137,467 563,463 8,412 571,875 10,984 14,432 25,416 68,634 298,011 366,645 1992 76,452 18,734 95,186 309,425 23,192 332,617 20,936 9,079 30,015 134,255 184,719 318,97 1993 96,354 6,361 102,715 395,563 45,467 441,030 12,497 12,191 24,688 93,667 233,618 327,285 1994 103,711 11,443 115,154 395,165 28,469 423,634 16,692 6,307 22,999 86,442 53,102 139,54 1995 82,590 59,514 142,104 244,545 136,660 381,205 21,639 4,693 26,332 134,397 25,214 159,611 1996 162,843 10,957 173,800 542,479 95,111 637,590 44,245 7,550 51,795 243,388 34,093 277,481 1997 162,527 14,506 177,033 564,514 118,231 682,745 54,396 10,350 64,746 341,158 41,837 382,995 1998 155,284 7,494 162,778 521,472 68,389 589,861 42,173 14,943 57,116 255,498 51,882 307,38 1999 129,053 11,460 140,513 477,459 83,330 560,789 48,846 16,347 65,193 247,591 45,795 293,38 Source: State Institute of Statistics 15  Table 2.17 - Major Tobacco Trading Partners of Turkey in 2000 Country Exports (U.S. $) Imports (U.S. $) U.S.A. 102,488,230 225,125,133 Germany 70,282,948 5,502,239 United Arab Emirates 46,844,893 Lebanon 39,260,618 Netherlands 28,497,230 4,722 Russia 26,047,942 Japan 19,397,261 Switzerland 16,024,214 17,396,243 Tunisia 14,066,861 861,411 South Korea 11,232,005 22,381,403 Brazil 719,327 10,688,791 Zimbabwe 27,902,569 Malawi 10,364,331 Total 491,418,606 350,726,549 Source: State Institute of Statistics, Foreign Trade Statistics Figure 2.3 Tobacco Exports and Imports 200,000 180,000 Tobacco Exports 160,000 ------- Tobacco Inports 140,000 Qu 120,000 anti ty 100,000 80,000 60,000 40,000 20,000 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 Years Source: State Institute of Statistics 16  Figure 2.4 - Tobacco Exports and Imports (Thousand U.S.$) 800,000 - 700,000 - Tobacco Exports - - - - - -Tobacco hnports 600,000- 500,000l > 400,000- 300,000 200,000- 100,000 -,.'| 0~~~~~~~~~~~~~~~~~~~~~~~~0 Years Source: State Institute of Statistics Before 1984 only Tekel was allowed to import tobacco, and private companies needed special permission to export tobacco. Although private companies have been allowed to import since 1984, they did not do so until 1991 because of high import duties. Tekel's share in total tobacco exports has changed over time, from 15 percent in 1990 to 30 percent in 1999. In monetary terms, Tekel's share was much lower; 12 and 20 percent of the total value of tobacco exports in 1991 and 1999, respectively. 2.6 TAXES ON CIGARETTES There are different types of taxes on cigarettes in Turkey. Table 2.18 shows the average taxes in 2000. Taxes constituted approximately 77 percent of the retail cigarette price in 2000, a considerable increase from 44 percent in 1994. Table 2.18 - Taxes on Cigarettes in 2000 Type of Tax Amount of Tax Tobacco Fund: Domestic Cigarettes US$ 3 / kg Tobacco Fund: Imported Cigarettes US $ 0.4 / package Defense Industry Fund 10 % on Factory Price Additional Tax 120 % on Factory Price Education Fund 15 % on Retail Price Grazing Ground Fund 2 % on Retail Price Veterans Fund 2 % on Retail Price Value Added Tax 17 % on Retail Price Source: TEKEL 17  Table 2.19 - Separation of Different Components of Cigarette Prices 1994 2000 2001 Factory Retail Factory Retail Factory Retail Values Price Price Values Price Price Values Price Price TL % % TL % % TL % % Panel A - Samsun (85 mm) Factory Price 3,500 100.0 23.3 92,000 100.0 30.67 122,000 100.0 27.1 Additional Tax 2,846 81.3 19.0 113,174 123.0 37.72 150,078 123.0 33. Federation Tax 17 0.5 0.1 472 0.51 0.16 625 0.5 0.1 Veterans Fund 299 8.6 2.0 5,200 5.65 1.73 7,800 6.4 1. Defense Industry Fund 200 5.7 1.3 20,749 22.55 6.92 27,514 22.6 6.1 ducation Fund 40 1.1 0.3 40,000 43.48 13.33 60,000 49.2 13.3 Grazing Ground Fund 1,840 2.00 0.61, 2,440 2.0 0.5 Value Added Tax 1,607 45.9 10.7 43,590 47.38 14.53 68,644 56.3 15.3 Commissions 1,800 51.4 12.0 36,000 39.13 12.00 54,000 44.3 12. Marketing & Distribution 615 17.6 4.1 12,300 13.37 4.10 18,450 15.1 4.1 reasury Share 6,724 7.31 2.24 TotalCost 10,925 312.2 72.8 365,324 397.09 121.77 511,552 419.3 113. Profit Margin 4,075 116.4 27.2 -65,324 -71.00 -21.77 -61,552 -50.5 -13. Retailer Price 15,000 428.6 100.0 300,000 326.09 100.00 450,000 368.9 100. Total Tax 5,010 143.2 33.4 231,747 251.90 77.25 317,102 259.9 70.5 Panel B - TEKEL 2000 (100 mm) Tobacco processing cost 4,660 42.7 13.3 28,410 21.9 3.9 69,402 27.8 6.3 Imported Tobacco (CIF) 3,181 29.2 9.1 66,286 51.0 9.1 142,905 57.2 13. port Duty 795 7.3 2.3 16,572 12.7 2.3 35,726 14.3 3. Tobacco Fund 2,272 20.8 6.5 28,408 21.9 3.9 61,245 24.5 5. Factory Profit/Loss 9,676 7.4 1.3 -59,279 -23.7 -5. Factory Price 10,909 100.0 31.2 130,000 100.0 17.9 250,000 100.0 22. dditional Tax 8,072 74.0 23.1 159,920 123.0 22.1 307,538 123.0 28. ederation Tax 55 0.5 0.2 66 0.1 0.0 1,281 0.5 0.1 eterans Fund 700 6.4 2.0 12,600 9.7 1.7 19,200 7.7 1. efence Industry Fund 200 1.8 0.6 29,319 22.6 4.0 56,382 22.6 5.1 Education Fund 40 0.4 0.1 95,000 73.1 13.1 140,000 56.0 12. Grazing Ground Fund 2,600 2.0 0.4 5,000 2.0 0.5 Value Added Tax 3,750 34.4 10.7 105,342 81.0 14.5 167,797 67.1 15.3 ommission 2,450 22.5 7.0 50,750 39.0 7.0 77,000 30.8 7. Marketing & Distribution 1,435 13.2 4.1 29,725 22.9 4.1 45,100 18.0 4.1 Treasury Share 31,399 24.2 4.3 Total Cost 27,611 253.1 78.9 647,320 497.9 89.3 1,069,298 427.7 97. Profit Margin 7,389 67.7 21.1 77,680 59.8 10.7 30,702 12.3 2. etailer Price 35,000 320.8 100.0 725,000 557.7 100.0 1,100,000 440.0 100. otal Tax 15,885 145.6 45.4 481,825 370.6 66.5 794,169 317.7 72. Source: TEKEL 18  The tax structure differs slightly depending on whether cigarettes are produced using imported or domestic tobacco (Table 2.19). In Panel A, the pricing schedule for cigarettes with domestic tobacco (Samsun) is shown for years 1994, 2000, and 2001. Panel B summarizes the tax structure for a cigarette brand with imported tobacco (Tekel 2000) that Tekel introduced to compete with American brand cigarettes. Tax rates on cigarettes have increased over time, especially the value added components. Most of the specific tax rates did not change in 2000 and 2001 but the tobacco fund rate increased. The defense fund tax as a percent of the price of Samsun (Tekel 2000) increased from 5.7 (1.8) percent in 1994 to 22.6 percent in 2001. Similarly, the rate for the Education fund increased from 0.3 (0.1) to 13 percent. The tax rate on Samsun was highest in 2000 and declined to 70.5 percent of the retail price in 2001 because of the removal of the Treasury share in the taxation of cigarettes in that year. This was an earmarked tax to offset Tekel's debt with the Treasury, incurred as a result of tobacco support purchases from farmers, which Treasury had financed through external borrowing. 2.7 GOVERNMENT REVENUES Table 2.20 shows some of the tax revenues collected from Tekel's cigarette sales. Only revenues from Tekel cigarette sales for the grazing ground fund, federation fund, additional tax, defense industry fund, and education fund are available. These funds correspond to 3.14 percent of total government revenues. Since the taxes for the Veterans Fund and the Value Added Tax are 1.7 and 15.3 percent of the retail price of the cigarettes, it is possible to calculate total taxes collected from Tekel cigarette sales approximately. They amount to 4.04 percent of total government revenues in 2000. Since the market share of Tekel in 2000 was 67.5 percent, this suggests that approximately 6 percent of total government revenues were obtained from cigarette sales. This had been much higher fifty years earlier; Tekel contributed 1.13 percent of government revenues in 1948 (Dogruel and Dogruel, 2000). Table 2.20 - Taxes Obtained from Tekel Cigarette Sales, 1996-2000 (billion TL) 1996 1997 1998 1999 2000 Grazing Ground Fund 2,336 4,469 7,037 Federations Fund 236 370 677 1,145 3,861 Additional Tax 47,223 74,510 135,383 234,301 432,765 Defense Industry Fund 6,568 14,902 27,077 46,321 174,208 Education Fund 686 17,563 56,957 110,253 214,705 Total 54,712 107,344 222,430 396,488 832,575 Total Government Tax Revenues 2,244,094 4,745,484 9,228,596 14,802,280 26,514,127 Percent of Total Tax Revenues 2.44% 2.26% 2.41% 2.68% 3.14% Source: Department of Finance 2.8 TOBACCO CONTROL POLICIES AND THEIR IMPLICATIONS Before the 1980s, when new brands were introduced, there was limited advertisement, in print only. In September 1981, the warning: "Harmful to Your Health" was printed on cigarette packages of some brands. After cigarette imports were permitted in 1984, advertisement of cigarettes was allowed in print but not on radio or television. There was a short-lived anti- smoking campaign in 1988 initiated by the Ministry of Health, with some posters in public places. 19  A new law (Law No. 4207) on cigarette smoking was enacted in November 1996. The law regulates smoking in public places and the sale and advertising of tobacco products, in order to discourage smoking, reduce cigarette consumption among young people and reduce the damage that results from smoking. According to this law, smoking is banned in health and education related places, at in-door sports and cultural activities, in all types of transportation vehicles, waiting rooms, and in work places where at least five people work. There is a penalty for individuals and companies that do not obey this regulation. In addition, it is forbidden to advertise, or to make advertisements using tobacco or tobacco products, their brand names, and to campaign to promote and motivate the use of tobacco or tobacco products. There is an age restriction on buying cigarettes: the minimum age to buy tobacco or tobacco products is eighteen. Furthermore, according to this law, tobacco products that are produced in Turkey or that are imported have to carry the warning "Legal Precaution: Harmful to Health." Products without this warning may not be sold or imported. Lastly, both public and private radio and television channels have to broadcast information about the harmful effects of the use of tobacco and its products for at least 90 minutes every month. Even though this regulation is comprehensive, there are problems in its implementation. For example, the law does not state explicitly who will be responsible for punishment or collecting penalties. A study by Bilir and Onder (2000) shows that the 1996 Law has affected significantly the number of cigarettes smoked but not the smoking prevalence rate and that enforcement declined in 1999 relative to 1998. 3u 'i'a-S'i''!S^--)-i 1^A -11I-lh_ D Cigarette demand in Turkey is analyzed using annual data for 1960 to 2000, using the standard models in the literature to estimate price, tax and income elasticities. The substitution elasticity of intemational brands produced in Turkey is also estimated, to understand the extent to which smokers shift to a different type of cigarette as the price of one type of cigarette increases relative to other types. Furthermore, the impact of a recent regulation that bans smoking in public places is examined. 3.11. DATA The variables used in the estimation are based on the economic theory of demand for cigarettes and on previous empirical studies: the quantity of cigarettes smoked is a function of the real price, per capita income, regulations that ban smoking in public places, and a time trend variable. Quantity The quantity consumed is defined as the number of packages of cigarettes consumed per adult per year in Turkey. Total consumption in Turkey is taken from the database of the U.S. Department of Agriculture. Total packs consumed are divided by the population 15 years of age or older in order to get consumption of cigarettes per adult. The adult population is calculated by interpolating the population in the eight census estimates that span the 41-year period under analysis. 20  Figure 3.1 - Cigarette Consumption in Turkey, Period: 1960-2000 140000 160 120000 140 8 ' , _ _ _ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~120 v 100000 & 10 | 10-0[ m 80000 - 40 -~~~~~~~~~~ -- -- oluo pTAtprY - - C. C 9 s o0 9q 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~6 6 0000 0 Y4 20000AnulCiumo 20000 ~ ~ ~ ~ ~ ~ ~~... CdisunVitton perAdult perYee 20 0 0 Source: Author's estimates using data from US Department of Agriculture Figure 3.1 shows aggregate consumption and packages smoked per adult per year. Total consumption has increased over the sample period. Per adult consumption increased from 1960 to 1979, declined in 1980 and stayed stable until 1993, but then started to increase again after 1993, which is when private companies started to produce American brand cigarettes in Turkey. In 2000, 134.65 packs of cigarettes were smoked per adult per year, or 11.22 packs per adult per month. Given that 64 percent of adults smoke, this corresponds to an average smoking intensity of 12 cigarettes per day per adult smoker. Price The price variable used in the estimation is the weighted average price of a pack of cigarettes. Data for the period 1960-1988 are taken from Tansel (1993). Prices for the rest of the period are calculated using the prices of individual brands weighted by their market shares. Prices and market shares are obtained from the State Institute of Statistics and TEKEL records. All prices are presented in terms of prices in Turkish Lira in December 2000, using the Consumer Price Index (CPI) as the deflator. Figure 3.2 shows the weighted average price of a pack of cigarettes and the average taxes paid per pack. Inflation is very high in Turkey, and over the sample period, the price of cigarettes has increased as the government adjusted prices periodically in response to inflation. When private companies started to operate, they also adjusted their prices from time to time, usually just after the government increased the prices of cigarettes produced by Tekel. Tax rates have risen over the years from 39 percent in 1993 to 72 percent in 2000, and the average tax paid per pack of cigarettes has increased. Trends in cigarette prices and taxes are similar, whether prices and taxes are expressed in Turkish Lira or in US $ (Figure 3.3). 21  Figure 3.2 - Average Price and Tax on One Cigarette Pack (US $) 0 9 l . ._... 0 7 0 6 0 4 0 3 01 8. |- -. - -a 02 Source: State Institute of Statistics and Tekel The average price of cigarettes was $0.75 in 2000. This price is low compared to average cigarette prices in upper-middle-income countries ($0.81) and very low compared to average prices in high-income countries ($3.23), but higher than most of the lower-middle income countries ($0.34).2 Although Turkey's tax rate as a percent of price is similar to high-income countries, because the price is low in Turkey, the average amount of tax per pack of cigarettes was only $0.54 in 2000. Moreover, over the whole period analyzed in this study, the average price of cigarettes was only $0.35 with an average tax of $0.18 and average tax rate of 49 percent of retail cigarette prices. Income Income is calculated from the Gross Domestic Product (GDP) in Turkey, obtained from the State Institute of Statistics. Income per adult is calculated by dividing GDP by the population aged 15 years or older. Income per adult is adjusted for inflation using the CPI. Regulations In order to examine the impact of smoking regulations, a variable is created which takes different values depending on the smoking regulations that were in force in Turkey. Since 1992, the warning message on cigarettes packs: "Harmful to Your Health" has been required. In 1997, a new regulation was enacted to ban smoking in public places and end all advertising of cigarettes. The regulation variable is set to 0 before 1992, has a value of 0.25 from 1992 to 1996, and a value of 1 from 1997 onwards. Trend In the literature, a trend variable is included in cigarette demand models in order to control for other changes that could affect consumption over the time period analyzed. 2 The average cigarette price in lower-middle income countries is $0.34, but the range is huge: from $0.0004 in Indonesia to $0.67 in El Salvador (Chaloupka, et al., 2000). 22  3.2. MODEL Following the current literature, the following double-log model is estimated: In Qt = Po + PI In Pt + P2 In Yt + f33 Regulation, + 4 Trend + et where Qt is the quantity (in packs) of cigarettes consumed per adult in year t. Pt is the weighted average price of a pack of cigarettes in year t. Yt is per capita income in year t. Regulation, is the regulation variable that takes a value of 0, 0.25 or I depending on year t. Trend, is the trend variable and et is the error term. Is are the coefficients estimated. The model was also estimated without the trend variable. In addition to this basic model, myopic addiction and rational addiction models are also estimated, in which the quantity of cigarettes demanded in year t is also affected by consumption in the previous year (myopic addiction) and in the following year (rational additction). Myopic addiction model: In Qt = Po + PI In Pt + 02 In Y, + D3 Regulation, + P4 Trend + P5 In Qt-l + et Rational Addiction model: In Qt = PO + I In Pt + P2 In Yt + P3 Regulationt + P4 Trend + PS In Qtll + 16 In Qt+1 + et All of the models were first estimated using Ordinary Least Squares. However, since Durbin- Watson test statistics indicated autocorrelation, Generalized Least Squared (GLS) Estimations were obtained. Before estimating the models, endogeneity of cigarette prices was tested. The results indicate that prices are endogeneous, so two stage estimation was used. (This is the standard procedure used when a variable is endogenous.) In the first step, the price of cigarettes was predicted using tax rate data. The predicted cigarette prices are used in the demand estimation in the second step. In addition to the basic demand model, in order to examine substitution elasticities, cigarettes are grouped into three categories: domestic filter cigarettes, domestic non-filter cigarettes and cigarettes with foreign brand names like Malboro, Parliment, Camel, that are either imported or produced in Turkey. The availability of data restricts the analysis to the period between 1987 and 2000. Since the degrees of freedom is low, only the major model is estimated separately for each type of cigarette. Additionally, the price of the type of cigarettes that can be considered as a substitute is included in the model in order to calculate cross-price elasticity. Descriptive statistics for the variables in the analysis are shown in Table 3.1. On average, 62,533 million cigarettes were consumed per year over the period analyzed. This corresponds to 106.38 cigarettes per adult. The average price is $0.35 per pack and average tax paid per package is $0.18. Expressed in 2000 prices, these correspond to 414,849 TL and 206,215 TL respectively. The average price and tax in 2000 were 505,089 TL and 363,664 TL, much higher than prices and taxes in the earlier years of the analysis period. Consumption was lower in earlier years. 23  Table 3.1 - Descriptive Statistics of Variables Used in the Estimations for 1960-2000 Standard Mean Deviation Minimum Maximum Total Consumption (Million pieces) 62,533 25,295 26,284 115,535 Tax rate (%) 49.23 6.63 39 72 Consumption of Packs per Adult per Year 106.38 15.05 80.49 139.44 Price of Pack in terms of 2000 Prices (TL) 414,849 147,883 207,332 779,155 Price of Pack (in US$) 0.35 0.22 0.07 0.93 rax per Pack in terms of 2000 Prices (TL) 206,215 88,968 99,520 502,555 Tax per Pack (in US$) 0.18 0.13 0.04 0.60 Income per Adult in terms of 2000 Prices (million TL) 2,298 727 904 3,261 Regulation index 0.10 0.30 0 1 Source: Author's estimates 3.3. RESULTS Because of the endogeneity of cigarette prices, price is first estimated as a function of the tax per pack. The following estimate is obtained: ln P, = 2.4176w + 0.8601 ln Tax, with Adjusted R2=0.8993 and F = 358.369 (0.5526) (0.0454) The standard errors of the estimates are shown in parentheses. This model explains almost 90 percent of variation in prices. It shows that if taxes increase by 100 percent, prices will increase by 86 percent. This result is used to estimate the tax elasticity of demand for cigarettes (the extent to which tax revenues will change when cigarette demand changes as a result of a price change). The coefficient estimates for the basic model with and without a time trend variable and the regulation variable are shown in Table 3.2. Since a double log model is used, the coefficients on the price and income variables measure the elasticity, that is, the change in quantity demanded in response to changes in these variables. The OLS estimations indicate a substantial and significant effect of price on the demand for cigarettes. The results are robust across specifications with and without the trend variable. However, the Durbin-Watson test statistic indicates that there is autocorrelation in the error terms. Therefore, the discussion that follows focuses on the GLS results. The GLS estimations of the basic models indicate a surprisingly low effect of prices on the demand for cigarettes, and the coefficient is not statistically significant. The results show that if prices increase by 100 percent, cigarette consumption will decline by approximately 19 percent, controlling for income, the time trend and the effect of changing regulations. This is quite similar to the -0.21 price elasticity of demand estimated by Tansel (1993) for the period 1960 to 1988. 24  The trend variable is not statistically significant, and excluding it does not change the price elasticity estimate. Turkey's price elasticity of demand for cigarettes is low compared to other countries. For example, price elasticity of demand for cigarettes in high-income countries typically falls into a range of -0.25 to -0.50. For low- and middle-income countries, estimates usually range between - 0.50 and -1.00, suggesting much greater price sensitivity in lower income countrnes. For example, price elasticity was estimated to be -0.70 in Papua New Guinea (Chapman and Richardson, 1990) and -0.59 in South Africa (van der Merwe, 1998). Nor does the regulation variable appear to have a significant impact. Several hypotheses may explain this. First, the strong regulations were only in effect for the last three years of the sample period, perhaps not long enough to see a significant impact. Second, enforcement of this regulation is weak in many places, and Bilir and Onder (2000) show that the impact of this regulation depends on the degree of enforcement. Third, there are other factors that have an opposite effect on consumption. In particular, foreign brand cigarettes have been produced in Turkey since the mid 1990s, and their availability has 25  Table 3.2 - Results of Demand for Cigarettes OLS GLS Without Time Without Time With Time Trend Trend Without Regulation With Time Trend Trend Without Regulation Parameter Parameter Parameter Parameter Parameter Parameter t- Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate statistic Intercept -0.7920 -0.471 -0.8196 -1.195 -0.0731 -0.048 0.0911 0.431 0.1280 0.615 0.0747 0.335 Ln Pt -0.3129* -2.033 -0.3123* -2.094 -0.4110*** -3.518 -0.1924 -1.248 -0.1895 -1.229 -0.2836* -2.176 n Yt 0.2928*** 3.658 0.2940* 7.763 0.2704*** 3.526 0.2328* 9.452 0.2315 9.409 0.2451 11.029 rend 0.0001 0.018 0.0017 0.701 0.0017 1.013 0.0024 1.644 egulation 0.0678 0.979 0.0685 1.221 0.0686 1.025 0.0984 1.638 -statistic 25.3030 34.6740 33.454 56.805 75.345 69.866 rob>F 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 djusted R-squared 0.7085 0.7164 0.7088 0.8513 0.8512 0.8412 Number of Observations 41 41 41 40 40 40 Durbin-Watson Statistic 1.3470 1.3480 1.377 1.907 1.888 1.875 and show significance at 10, 5 and 1 percent levels respectively Source: Author's estimates  increased consumption in recent years, counteracting the impact of the regulations.3 The only statistically significant factor in the cigarette consumption equation is income. Cigarettes are a normal good; that is, people consume more cigarettes when income increases. The income elasticity of demand for cigarettes is 0.23, indicating that if income increases by 100 percent, cigarette consumption will increase by 23 percent. This is similar to income elasticity in the U.S. 30 years ago, but lower than many other countries (Townsend, 1998, Chapter 8, Page 88). The model was re-estimated without the regulation variable, since its coefficient was not significant. In this specification, the impact of price is significant and much greater: a 100 percent increase in the price of cigarettes would decrease consumption by 28 percent. The price and tax elasticities allow estimates to be made of the impact of changes in taxes on the demand for cigarettes and on government revenues (holding income constant). (See Table 3.3.) The base case is for the average cigarette price, average consumption per adult per year and average tax rate in 2000. The total tax rate comprises the value added tax (VAT) and excise tax, calculated as a percentage of the retail price of a pack of cigarettes. Government revenues are calculated as the sum of VAT and excise taxes on cigarettes. The simulation analysis shows that if the government were to increase the excise tax rate on cigarettes by 10 percent, the price of cigarettes would increase by 5.5 percent. This price increase would reduce cigarette consumption by 1.50 percent. The higher tax rate would increase government revenue by 7.08 percent. Table 3.3 - Impact of Tax Changes on Consumption and Government Revenues Percentage Changes in Excise Tax Rate Base Case -25 -10 10 25 50 Price (TL) 505,089 onsumption 134.65 139.90 136.75 132.55 129.40 124.15 Total Tax Rate (%) 72 58.25 66.5 77.5 85.75 99.5 VAT (%) 17 17 17 17 17 17 Excise Tax Rate (%) 55 41.25 49.5 60.5 68.75 82.5 hange in Price(%) -13.75 -5.50 5.50 13.75 27.50 New Price (TL) 505,089 435,639 477,309 532,869 574,539 643,988 Change in Consumption (%) 3.90 1.50 -1.50 -3.90 -7.80 Govemment Revenues (in Thousand TL) from VAT 11,562 10,334 11,097 12,008 12,639 13,592 from Excise Tax 31,048 24,194 28,377 33,620 37,296 42,939 Total 42,610 34,555 39,474 45,628 49,935 56,531 Change in Government Revenues(%) -18.90 -7.36 7.08 17.19 32.67 Source: Author's estimates 3 A dummy variable with the value =1 for years when foreign cigarettes were available was statistically insignificant, so the more standard time variable was used in the model. 27  In this estimation it is assumed that there is no smuggling. However, a tax increase that increases cigarette prices makes smuggling more profitable, and may induce an increase in smuggling, which will moderate the increase in tax revenue. Since smuggling is illegal, no data on smuggling in Turkey are available. In order to take into consideration the effect of smuggling, the potential supply of cigarettes through smuggling is calculated by assuming that any difference between the sum of domestic consumption and exports, and the sum of production and imports, is accounted for by smuggling. This difference (assumed to be the smuggled volume) is regressed against the tax rate on cigarettes for the sample period, making the (over) simplifying assumption that the volume of smuggled cigarettes is a function of the tax rate. Then, this estimated model is used in the simulation analysis to calculate smuggling at different tax rates. Incorporating an increase in smuggling in response to an increase in taxes and prices does not change the results much. For example, a 10 percent increase in the excise tax rate would increase government revenues by 5.5 percent with smuggling (Table 3.4), not much lower than the 7.1 percent revenue increase without smuggling. Table 3.4 - Impact of Tax Changes on Government Revenues with Smuggling Percentage Changes in Excise Tax Rate Base Case -25 -10 10 25 50 New Tax Rate (