3.1 METHODOLOGICAL FRAMEWORK

3.1.1 The methodological approach

The methodology essentially includes three stages. In the first stage production, demand, export and import equations were estimated for the major producing countries (China, India, Brazil, the United States, the EU, Turkey, Zimbabwe and Malawi) and also for groups of countries (other Europe, area of the former USSR, Oceania, other developed, other Africa, other Latin America, other Near East and other Far East). The specification and the estimation results using data for the period 1970 to 1998 are provided in Annex B.

In the second stage the model consisting of five equations (demand, area, yield, exports and imports) was calibrated to the base year (1998, i.e. the three-year average 1997-1999) by adjusting the intercept. The year chosen as the base year, 1998, is in fact the three-year average 1997-1999.

The model was validated by examining out-of-sample predictive ability (1992 to 1997). The validation results show that variability at the country level was high and predictive ability at the country level was rather poor but adequate. This implies that projections at the country level may not be good enough to capture all country details and particularities. However, at the regional level (such as Africa, Latin America and the Caribbean, etc.) or at the world level the predictive ability is very good because projection inaccuracies at the country level are averaged out. This does not invalidate country level projections, but it implies that such projections cannot be used for country level work. They serve a completely different purpose and as such have shortcomings which in order to make them useful at the country level need additional work to provide greater detail at the modelling stage and incorporation of qualitative information.

In the third stage, projections were obtained for each country of the standard country list, using the model. The projection results were examined at the country level to conform with a priori knowledge about developments in each country concerning production policies, consumption trends and trade flow developments. Some adjustments in certain parameters were necessary to achieve consistency of projection results. The consistency of the projection results was examined mainly by comparing the net trade position projected by supply and demand for main producing and trading countries with the results of the independently projected exports and imports.

In addition, results were examined for consistency with a priori knowledge about the various countries and their production, consumption and trade conditions. Such country knowledge was incorporated into the projection results for countries for which country level studies were available, such as China, Turkey, India, Zimbabwe, Malawi, and also for the EU and the United States for which some general information is available from other sources.

3.1.2 Data and definitions

The data used for the projections in this study have been derived from various sources, including FAO, USDA, the World Bank and various other sources.

The study covers all major countries of the world, although some minor ones are left out. The standard country list, given in Annex A, used for the study includes 164 countries grouped into Developed and Developing countries, grouped further into regional groups, the same as those used in the analysis of past trends in Chapter 2. China is in some cases singled out because of its size.

Production data include tobacco leaf production in farm weight, as well as the area under tobacco in hectares and yield in farm weight for all countries of the world for the period 1970 to 2000. There are also detailed data by type of tobacco, such as Virginia, Burley, etc. Consumption data are in dry weight for all countries of the standard country list. Similarly, there are export and import data (quantity and value) in dry weight for each country in the country list for the period 1970 to 1999. Consumption and trade data are aggregate data, and thus no distinction is possible of the various tobacco types. Tobacco leaf consumption data at the country level are derived using a supply utilization account and data on stocks, production, imports and exports. Consumption data do not include 'on farm' consumption of home produced tobacco, which is believed to account for a substantial part of consumption in many developing countries.

Cigarette production and consumption data exist for a number of countries, but they are incomplete. Therefore the analysis did not include cigarette demand and supply, but consumption of all tobacco products is translated into leaf equivalent. Thus, the analysis includes only one level (tobacco leaf).

Price data are very scarce. There are several reasons for this deficiency. First, tobacco is a differentiated commodity with various tobacco types produced and traded. Countries and manufacturers buy particular types that they need to produce the desired blend for the types of cigarettes they produce and therefore they may appear as both importers and exporters. Therefore, a price corresponds to a particular tobacco type, and prices of different tobacco types differ widely. Thus, there is no homogenous commodity called tobacco but various tobacco types traded internationally at prices that differ substantially. Furthermore, there are substantial quality differences within each tobacco type reflecting substantial differences in quality. In addition, some leaf producing countries such as in Brazil, manufacturers purchase by direct contract so price data are not available. Producers in countries such as Zimbabwe and Malawi sell in auction markets and there is good price data for these countries. For Turkey also there are good price data maintained by the national tobacco monopoly. For the United States and the EU price data were obtained from other sources. For other countries, however, no price data are available for either production or consumption.

3.1.3 Factors affecting demand

The main factors that determine the growth in consumption over the 1970 to 1998 period are population and income growth, urbanization and changing tastes. Among them population growth is perhaps the most important determinant of growth in tobacco consumption.

Population

World population for the period 1970 to 2010, actual and projected, is shown in Table 3.1. Population has increased during the period 1970 to 2000 at about 2 percent annually, with developed country population increasing at about 0.6 percent per year and developing country population increasing at about 2.2 percent per year. This divergence in population growth between developed and developing countries is projected to continue to the year 2010 increasing further the share of developing country population from 79 percent in 1998 to 81.5 percent in 2010.

Table 3.1 Population growth

Source: World Bank

Tobacco, however, is consumed largely by the adult population (usually 15 years and over). Thus, it is important to take into account not only the total population, but also the adult population.^[2]

Adult population (15 years and over) as a percentage of total population is increasing rapidly in developing countries, thus increasing the number of potential tobacco consumers in the population. (see Table 3.2).

The adult population increased from 4.3 billion in 1998 to 5.2 billion in 2010. With 1.1 million smokers (one in four) in 1998 (World Bank), if smoking rates continue the number of smokers may reach 1.32 billion in 2010, about 20 percent more than in 1998, or an increase of about 1.5 percent annually. Hence, an increase in tobacco demand may be expected simply as a result of the increase in the number of smokers world-wide with present smoking rates, as a result of population growth.

Table 3.2 Adult Population (15 years and over) as percentage of total population

Source: World Bank

Income

The second most important factor determining tobacco consumption is income growth. World income (GDP) increased during the period 1970 to 2000 at slightly over 3 percent. In developed countries incomes increased by about 2.5 percent annually while in developing countries income growth was about 4 percent. In China it was about 8 percent annually. Similar trends are expected to continue in the period 1998 to 2010 with a total GDP increase of 41 percent in developed countries, 127 percent in China, and 70 percent for the other developing countries, corresponding to average annual GDP growth rates of 3.1 percent, 7.7 percent, and 4.9 percent respectively.

Estimates of income elasticities are available in the literature for several countries. Evidence shows that estimated income elasticities are much higher in developing countries than in developed countries. The literature suggests that income elasticities for tobacco products are about 0.5 for developing countries and about 0.2 for developed countries (see Zhang, 2000). Thus, income growth may have, other things being equal, a potentially strong positive impact on tobacco demand that amounts to an average annual growth rate of 0.6 percent in developed countries, 3.8 percent in China and 2.4 percent in other developing countries.

The combined effect of population and income growth on tobacco demand may reach 1.4 percent in developed countries, 4.6 percent in China and 4.2 percent in other developing countries. However, due to other factors such as prices, taxation, and health awareness, such high rates will not probably materialize as actual consumption, as it was the case in the period 1970 to 1998. During this period the actual observed growth rate of tobacco consumption was much lower than the combined population and income effect.^[3]

Table 3.3 GDP growth

Source: World Bank

Prices and other factors

Price is obviously a very important factor determining demand for tobacco products. Cigarette demand is generally price inelastic. According to evidence available in the literature (see Zhang, 2000), in developed countries price elasticities of cigarette demand range from -0.2 to -0.4, while in developing countries price elasticities range from -0.6 to -0.9.

The price of tobacco at retail level includes a substantial amount of excise taxes, and thus the leaf value content of the final consumption product is quite low. The evidence available on demand elasticities for cigarettes implies that demand for cigarettes in developed countries is less responsive to taxes than in developing countries. As a result, any future changes in tobacco tax policies in developing countries that would result in significant increase of the retail price level may have a strong negative impact on cigarette demand.

Another factor affecting cigarette demand is urbanization. Other things being equal, consumption of tobacco products is higher among urban dwellers than among rural dwellers. With increasing urbanization expected in the period to 2010 in developing countries, consumption of tobacco products is expected to increase further, in addition to the effects of population and income growth.

Growing awareness of the health risks that are associated with tobacco use induces many consumers to quit tobacco consumption, in particular in developed countries. This changes consumer preferences and introduces a negative time trend in tobacco consumption.

Available evidence shows that demand for cigarettes in developed countries responds strongly to direct measures such as counter-advertising, banning of advertising and smoking restrictions. In developing countries, such measures may have a rather lower impact on cigarette demand.

Overall, it seems much harder to curtail tobacco demand in developing countries because of the strong impact of population and income growth. From surveying the available literature it appears that taxation may have a strong impact on cigarette demand and thus it becomes an effective measure to use in developing countries in parallel with other direct measures such as counter-advertising, banning of cigarette advertising and smoking restrictions.

Finally, as the objective of this effort is to estimate demand for tobacco leaf, trends in the use of tobacco leaf in cigarette manufacturing must be considered. Past trends show that the use of tobacco leaf per cigarette stick in cigarette manufacturing has been reduced by as much as 15 percent over the past 30 years, thus reducing the actual use of tobacco leaf in the final product. This reduction is due to the use of filter tips and new technologies which reduce waste and pack less tobacco into cigarette sticks. This is more pronounced in countries catching up in technology in cigarette manufacturing, such as China.

3.1.4 Price and non-price factors in projecting supply

Tobacco supply is determined by area cultivated under tobacco times yield obtained per unit of area. Average world yields increased from about 1.3 to 1.7 tonnes/ha in farm weight during the period 1970 to 2000 or an average annual growth rate of about 0.9 percent. Area under tobacco increased steadily from 1970 to 1990 but it shows a strong variability over the past decade, mainly due to China.

Production conditions

Production conditions differ significantly across producing countries, reflecting the production of different tobacco types and differences in comparative advantage of tobacco production of a specific type. It should be stressed that it is difficult or sometimes impossible for a country to change from the production of one type of tobacco to another, either in the short term or in the long term. It is not only that natural conditions differ from country to country, but also fixed capital produced has no use when production is switched to a different type and production knowledge becomes obsolete.

Thus, the factors used in tobacco production are usually product or type specific and cannot be easily switched between types. Hence, although aggregate supplies may appear to be in balance, prices of particular types of tobacco may exhibit different trends because of shortages or surpluses of certain tobacco types or grades within a certain tobacco type. In particular, high-quality grades used in making high quality cigarettes are usually in short supply and are high priced, whereas other grades are plentiful and low priced.

In addition, in the manufacturing of cigarettes, individual brands contain a mixture of tobacco types producing certain taste and consumption characteristics. Any changes in the mixture are implemented over long periods of time. Thus, substitution between types is limited and it is possible only between certain types.

Producer prices

Producer prices are influenced by price and trade policies implemented at country level either for tobacco production or for commodities competing with tobacco for the same resources. Although price support schemes in the developed countries have an impact on world market supplies, the profitability of tobacco as a cash crop in most developing countries is several times higher than that of any other competing commodity. There are also some countries, such as Brazil, China, Zimbabwe and Malawi, where tobacco production does not receive any government support, implying that they have comparative advantage at world level.

Price is the most important factor in determining the supply of tobacco leaf, through its influence mainly on the area harvested. The impact of price levels on yields is generally much smaller and for all practical purposes can be ignored since yields are usually determined by technological and physical factors. Area harvested is determined mainly by past land use levels, price levels and trend factors (partial adjustment).

Application of this well-known partial adjustment methodological framework, however, is difficult for tobacco, in particular at world level, because of lack of price data. Tobacco is a differentiated product and there are many prices corresponding to different product types. Finally, quality is an important factor in determining prices, resulting in substantial price differences between different types and within each type. Quality varies considerably between types and within each type, as do buyers' requirements for quality and type.

Changing support policies

Recent trends in price support policies towards reducing support levels may decrease production in developed countries.

The pressure against policies supporting tobacco production in the developed countries indicates that sooner or later these support policies will weaken, if not be eliminated, thus increasing further the possibility of shifting tobacco production to the developing countries.

Tobacco production in some developing countries, such as India and Turkey, receives direct price support or indirect input subsidies. However, even if such subsidies were to be removed, tobacco profitability would not be challenged in comparison with alternative cash crops. Thus, even in countries where such subsidies are used, production might be expected to continue to increase, even if these subsidies are partly removed.

In China, tobacco may not be more profitable than other crops but it offers farmers lower risk and uncertainty, since production is based on contracts. Thus, given the institutional framework of tobacco production and use in China, production is expected to follow trends in demand. In other developing countries, such as Brazil, Zimbabwe and Malawi, where tobacco is produced without any government support and is competitive internationally, production is expected to take over the gap in world supplies that will occur from the decline in production in the developed countries.

3.1.5 Determinants of exports and imports

Trade flows are determined by trends in demand and supply in trading countries and regions of the world, and also by trade distortions introduced by exporters or importers, such as tariffs and non-tariff barriers, export promotion and domestic policy.

The best approach for projecting trade flows at world level is to use a world commodity model, where producing and consuming regions interact in the world market and determine export and import flows and a world price. As noted above, this first best option is not available in the present study for several reasons. First, tobacco is a differentiated commodity and various tobacco types have substantially varying prices in the international market. Second, for the same reason, many countries find themselves to be both importers and exporters of different tobacco types. Third, the lack of adequate price data does not allow the construction of a full commodity model that would allow an accurate forecast of various commodity flows.

Given the above limitations, there are two alternative options that can be utilized. First, since we have available independent projections of demand and production, we can derive from them projections for net trade flow requirements (net import requirements or net export availabilities) for each country. Alternatively, we can estimate export and import functions, as we did for production and demand, and project for each country in the standard country list independent export and import flows. A first test of the reliability of these estimates could be used to examine whether projected trade flows generally conform with the net trade position implied by the independent supply and demand projections.

In this study we implemented the second option and estimated export and import functions for major producing and consuming countries and regions. Export and import function specifications and estimation results are given in detail in Annex B. Export and import flows depend on self sufficiency ratios (production to consumption) and relative prices (domestic to world market prices), as well as the lagged dependant variables.

3.1.6 Policy scenarios

The projections take into account changing consumer demand, production trends and likely developments in national and international policy relating to tobacco production, consumption and trade with particular attention given to the likely impact of the Framework Convention on Tobacco Control.

Two alternative policy scenarios were developed for the analysis of projection results, incorporating various price and non-price policy measures including raising taxes, reducing protection to tobacco production and the adoption of bans on advertising and promotion of cigarettes in various countries.

The baseline scenario assumes continuation of present policies. It assumes that real consumer and producer prices will remain at base year (i.e. 1998) levels throughout the projection horizon 1998 to 2010.

In projecting tobacco leaf demand, the baseline scenario assumes there would be no change in policies and a continuation of present price, tax and other smoking restriction policies. Countries have already introduced smoking restrictions and a slow-down or even decline in certain cases has been observed in cigarette consumption during the 1990s. However, actual data for 2000 that have recently become available indicate that in some countries, such as China, part of this decline in the 1990s may be a cyclical short term trend associated with changes in stocks and some growth in tobacco demand should be expected during the period 1998 to 2010.

In projecting tobacco leaf supply some assumptions were made to capture changes in the level of production support in the various countries. In the baseline scenario a continuation of present policies is assumed, thus maintaining the present level of production support. However, it is unlikely that such support policies will be fully maintained during the period to 2010 and some decline in production support should normally be expected. Several, mainly developed, countries have already introduced policies restricting production and a slow down or even decline in certain cases has been observed in tobacco leaf production during the 1990s.

The policy scenario assumes the adoption of strong policy measures against tobacco consumption and production that includes increasing taxation on consumption and reducing production support. As a result of such measures, it is assumed that consumer prices increase by 30 percent and production support is reduced by 40 percent in the developed countries and by 20 percent in the developing countries.

The policy scenario aims to capture the impact on demand of a strong anti-smoking policy, similar to the one introduced in most developed countries, mainly in the United States. This scenario assumes a strong increase in taxation in all developing countries, including China, resulting in an increase in tobacco price of 30 percent. In addition some stronger measures of smoking control are assumed, resulting in a relatively modest negative time trend. For the United States no change was introduced with the policy scenario and the results of the baseline and the policy scenarios are the same.

In projecting tobacco leaf supply the policy scenario aims to capture the possible impact of a reduction in the level of support in both developed and developing countries. We assume a strong anti-tobacco policy in the United States and the EU that would result in a 30 percent reduction in the tobacco price in 2005, remaining at that level to 2010. In addition, we assume a rather less strong anti-tobacco policy in developing countries that reduces the farm level price by 20 percent implemented in 2001 and remaining at that level for the entire period to 2010. No change, however, is assumed for Brazil, Zimbabwe and Malawi, where tobacco production does not receive any government support. In some countries, such as India, the decline in price due to a reduction in government support was assumed to be 10 percent.

The price changes assumed in the policy scenario are introduced in one step in 2001, with real producer and consumer prices remaining at the same level throughout the period 2001 to 2010. This does not change the results for 2010 and shows the immediate impact of policy on demand, production and trade. The policy scenario aims to evaluate the relative impact of certain policy measures, such as consumption tax and producer price support, that are considered to have a strong impact on demand, production and trade.

The assumed increase of 30 percent in consumer prices in developing countries with an increase in taxation rates is substantial and, perhaps, difficult for governments to implement. Further, a reduction in producer support resulting in a decline in producer price by 40 and 20 percent respectively in developed and developing countries may be considered quite realistic. It is realistic to expect an elimination of producer subsidies in the EU within 4 to 8 years and there is increasing pressure in the United States for elimination of the tobacco programme.

Projections of trade flows to 2010 for the baseline and policy scenarios were obtained by introducing projected production and consumption figures in the baseline and policy scenarios respectively and assuming continuation of present trade policies with respect to trade distortions by tariff and non-tariff barriers, and export promotion policies.

Finally, the projection results were compared with the net trade position implied by the supply and demand projections. Adjustments were made using judgement and discretion and some a priori knowledge about the countries so that the two approaches produce similar results.

3.1.7 Caveats and limitations

Modelling in such a detail at world level and for all countries of the world would have been an enormous task extending far beyond the limits of this exercise. Instead, the study followed a pragmatic approach, by constructing a commodity model, using several shortcuts to make up for limitations due to lack of data. Incorporation of qualitative information was implemented sometimes at the expense of modelling purity. The focus was to produce meaningful results useful for guiding policy makers.

A substantial amount of information results not from reading the projection numbers per se, but from the changes that result from changing assumptions and from introducing policy measures. Again, it is not the level of the projected variables that matters but the relative percentage change. Thus, it can be said that the projection numbers presented in this study reflect not only the outcome and the simulation results of the quantitative model, but also the results of substantial qualitative information available through country level studies.