PHA-Exchange> Hungry for tobacco: an analysis of the economic impact of tobacco consumption on the poor in Bangladesh

[FCA Secretariat]

Introduction 

UNICEF estimates that over 200 million children aged under 5 years in low income countries are malnourished, and that malnutrition contributes to about six million deaths among these children each year.1 Bangladesh, with a population of about 130 million people, is one of the poorest countries in the world. While life expectancy has increased over the past decade, in 1998 it was still only 60.5 for women and 60.7 for men.2 Nearly half the population of Bangladesh in 1995-96 lived below the poverty line, with about half of these living below the "hard core" poverty line.2 In 1998, most households spent less than $82 each month.3 In 1995, 30% of families were classified as very poor, 22% as poor, and less than 1% as rich.4 

Coexisting with this extreme poverty is a thriving tobacco industry. The use of chewing tobacco, bidis, and cigarettes is widespread. About 15 local companies compete for the lower end of the cigarette market. British American Tobacco (BAT), which owns the controlling share of Bangladesh's former tobacco monopoly, is a ubiquitous presence through its glossy media advertising, cigarette display cases, and storefront signs. In 1998, BAT Bangladesh reported pre-tax profits of approximately $15.9 million, while it spent $3.4 million on brand promotions and development.5 

Twenty years ago, it was suggested that "the nutrition-mediated effects of smoking, in terms of chronic undernutrition as well as survival, are likely to be far more important than the direct consequences of smoking on health".6 Although some statistics exist on tobacco use and household expenditures as part of national household expenditure data collection, we are unaware of any research specific to this area in Bangladesh. Tobacco has remained an issue of health and drug addiction, rather than being investigated in its relation to poverty. 

Methods 

Data on tobacco use and on household expenditures including tobacco was collected from the Bangladesh Bureau of Statistics (BBS). The following surveys on poverty, tobacco use, and sales figures for BAT Bangladesh, were also collected. Bangladesh has fairly complete data, much of it based on national surveys. Information in Analysis of basic needs dimension of poverty on household expenditures on basic goods and on tobacco are based on a stratified sample of households throughout Bangladesh, for a total of 32 000 households. Age specific smoking rates and average expenditure for tobacco utilises data from the national Health and Demographic Survey. Statistics on smoking by income group, from Prevalence of smoking in Bangladesh, utilises information from 1299 respondents. The Household Expenditure Survey surveyed 3840 households in 1991-92, and 7420 households in 1995-96. When repeated visits to the Bangladesh Bureau of Statistics failed to yield the primary sources, we used the summaries published in BBS's Statistical pocketbook, which are based on BBS surveys. Limitations include differing definitions of tobacco use (some surveys only inquire about smoking, which thus disregards all smokeless tobacco use); difficulty of respondents in calculating daily food consumption and tobacco expenditures; and the unavailability of some of the primary sources for data from BBS. 

The data were analysed using Microsoft Excel to make comparisons of different expenditures and to calculate potential food purchases from tobacco expenditures. Calculations of potential food purchases from money spent on tobacco and their caloric value were made using food prices and calorie counts listed in BBS statistics for poor to medium grade foods. Advertised cigarette prices for 2000 were used where available. Where prices were not listed by the companies, the common market price of the cigarettes was used. Food prices were taken from a large market in Dhaka. (Food prices vary by type of foodfor example, types of leafy greenand by the quality, as well as by the location. Some items are more expensive in rural areas, some less, but a comparable diet could be calculated with substitute foods for other parts of the country.) 

All figures are adjusted for inflation to 1999/2000 values and, where originally in taka, converted to US dollars. 

The calculations for year 2000 are based on the cigarette prices during that period. BAT has since decreased the price of Gold Leaf light from $.90 to $.73, and Benson & Hedges from $1.43 to $.99, but raised the price of Star from $.38 to $.52. 

The exclusion of oral tobacco use from prevalence data for women represents a major flaw in the data. However, this exclusion was necessary as data were not available and this makes the results more conservative, as the estimation of poor women tobacco users is thus a gross underestimate. 

Results 

TOBACCO VERSUS BASIC NEEDS
The poor in Bangladesh spend most of their money on food and other essential goods, yet are still unable to provide even the basics for their families. But despite the high level of poverty in Bangladesh, smoking rates are relatively high. Men aged 35-49 years have the highest smoking prevalence at 70.3%2 and smoking prevalence is highest among the poorest. The highest rate, 58.2%, is among men with a household income of less than $24/month (table 1). Smoking prevalence declines proportionally as income increases, with the lowest rate, 32.3%, being for men with a monthly household income of $118/month or more.7 

While smoking rates by women are much lower than for men, rates of oral tobacco use by women, for which no reliable figures are available, are thought to be quite high. Where women do smoke, they tend to smoke tobacco in forms that are cheaper than cigarettes, such as bidis, which women are six times more likely to smoke as compared to cigarettes.7 

Calculations of the number of poor smokers in Bangladesh are shown in table 2. Statistics from the Household Expenditure Survey show poverty, in terms of the population unable to consume 2122 calories/day, as being over 47% for the three periods measured between 1988 and 1996.8 

The cutoff point we utilised for poverty in table 2 is based on the Bangladesh Bureau of Statistics criteria3 4 is a monthly household income of less than $70the category into which 40% of Bangladeshi households, and presumably most poor families and malnourished children, fall. The figures yield a total of 9.87 million poor male smokers and 612 000 poor female smokers aged 15 years and over, for a total of over 10.48 million poor smokers. (While figures for smoking by poverty rate are available for men, the calculation for women involves the smoking rate across income groups, by age.) 

Cigarettes are the most expensive form of tobacco consumed, followed by bidis, with hukka, pipes, and other forms the cheapest. In 1997, average tobacco expenditure for those who use it ranged from a low of $1.29 a month for women who smoke hukkas, to a high of $7.24 a month for men who smoke cigarettes.2 By comparison, per capita expenditure on clothing, housing, health, and education totals only $2.92 per month, or 40% of the average male monthly expenditure on cigarettes.9 Cigarettes are by far the most widely advertised tobacco product. 

The typical male cigarette smoker spends over five times as much on cigarettes as the per capita expenditure on house rent, 18 times as much as for health, and 20 times as much as for education (fig 1). For women, the figures are only slightly less striking; women who smoke hukkas spend almost as much on tobacco as the per capita expenditure for housing, and over three times as much as the per capita expenditures for health and education. As there is no government support for the poor for clothing or housing, and virtually none for education or health care (per capita government spending on health in 1999 was $3.5610) these numbers represent actual total expenditures on these items. Women's expenditure on average for cigarettes is more than half the per capita expenditure on food. Men spend more than 86% as much on their cigarettes as the average per capita expenditure on food. Spending per household on tobacco represents 1.3% of total household expenditure for rural areas, 3.3% for urban areas, or 1.4% for the country as a whole.9 

As shown in table 3, the poorest households spend half as much on tobacco as on health, and almost 10 times as much on tobacco as on education. While the lowest income groups spend the lowest percentage of their income on tobacco, the numbers do not increase steadily; for instance, both those with household expenditure between $35-$47 and $213-$236 spend 2.6% of their income on tobacco, with the average for all groups being 2.8% (not shown). The poorest spend much more on tobacco than on education; this levels out for middle income groups, with the upper half of income groups spending a smaller percentage of their income on tobacco than on education. For most groups, a similar amount is spent on tobacco as on health. 

In fig 2 we reallocated average male smokers' monthly cigarette expenditure of $7.24 to basic needs, using the proportions given as national averages (58.8% for food, 5.6% for clothing, 9.5% for housing, 2.8% for health, 2.5% for education, and 20.8% for other).9 This represents the way a typical non-tobacco using person might be expected to spend the money that would otherwise go to tobacco use. (The additional money was added to the average expenditures rather than shifting it from "other", since the average smoker spends more on cigarettes than the average household for the category "other".) 

Across income groups, people would be likely to spend an additional $1.70-$4.40 per month for food, with similar but smaller increases across other categories. For cigarette smokers, this would mean an average increase of over 50% in their monthly food expenditure. Rural cigarette smokers would have over 50% more money available for health care, and an urban cigarette smoker an additional 34% for education.9 

TOBACCO VERSUS FOOD
As spending on food increases, malnutrition (underweight for age and stunting) decreases.11 The poorest households spend the highest proportion of their income on food, but are also the most likely to have malnourished children in their household. Thus, they would benefit the most by shifting their tobacco expenditures to food. Most Bangladeshis live in the countryside, but the rural poor are mostly landless or virtually landless. Use of cigarettes and bidis in rural areas is highest for those who own less than half an acre of land,8 so growing extra food to compensate for tobacco expenditures is not an option. 

Those with a monthly expenditure of less than $45 a month spend 73% of that money on food, whereas the figure for those with a monthly expenditure of $45-$111 is 66%.9 This averages out to over 69% of household monthly expenditure going to food. Nearly 78% of calories in the Bangladesh diet are supplied by cereals.3 We can therefore assume that the poor would spend most of their increased food budget on rice, while smaller numbers of calories of less commonly consumed foods could have a huge impact on children's diet. 

Table 4 shows average daily expenditures for tobacco in 1997 for men and women who use it, by type of tobacco product, and the number of calories from rice that each sum could purchase, using the 1997 price of $0.08 for 1000 calories of rice. The average male tobacco user could purchase 1402 calories of rice per day with his tobacco money; the figure for women is 770 calories. The minimum daily calorie requirement varies by age and sex, from 1094 for children aged 3 years and under, to a peak of 2782 for males and 2544 for females aged 18-29 years.3 

Even cheap cigarettes are expensive when compared to foods. In year 2000 prices, a 20-pack of Navy (one of the most heavily advertised but mid cost brands) costs almost as much as a litre of milk; less than two packs would pay for a litre of soybean oil. A pack a day of Navy would consume 16% of the average income for a Bangladeshi.3 An egg costs little more than a stick of Gold leaf light. A pack of Gold Leaf light could pay for 3.4 dozen small bananas, 6.7 kg of spinach, 1.3 litres of soybean oil, or 3.4 kg of rice. With $0.73a few hours' wage for a rickshaw pullerone could buy either 0.5 kg of beef, or 5.5 kg spinach, or over 1 kg of lentils, or a dozen eggs, or one pack of John Player Gold Leaf regular cigarettes. A smoker of a pack a day of Star or Scissors spends $0.38 each day, or over $11.50 each month. If he spent 70% of that money on food instead, he could easily add 800 calories each day to his family's diet, in the form of lentils, potatoes, fish, beef, and spinach. 

BAT Bangladesh's gross turnover in cigarettes in 1998 was over $293 million.5 That figure could have purchased over 4.7 billion eggs, enough to feed almost 13 million children an egg a day. Meanwhile, egg consumption in 1996 (latest year in which statistics are available) averaged one egg per person per month.2 

Per capita monthly expenditure for tobacco is higher in both rural and urban areas than that for milk, and higher in urban areas than for leafy green vegetables. People spend nearly as much in cities for tobacco as for lentils (the main protein source in the Bangladeshi diet). Nationally, for each of the high nutrient foods shown in fig 3, tobacco expenditures represent more than half the expenditures for food.9 

TRENDS IN CONSUMPTION OF TOBACCO VERSUS FOOD
Poverty, as measured by daily caloric intake, increased in urban areas from 1991 to 1996 (the most recent years for which statistics are available) from 46.7% (6.82 million people) to 49.67% (9.56 million), though it improved slightly in rural areas from 47.64% (44.81 million) to 47.11% (45.73 million); for the country as a whole, poverty worsened from 47.52% (51.63 million) to 47.53% (55.28 million).8 Given the high rates of tobacco use among the poor, it is plausible that development gains over the past several years have been significantly offset by diversion of income to tobacco. 

Comparing the 1997 figures for those for 1995 shows that the general trend is towards increasing expenditures for tobacco. Daily expenditures for tobacco rose from $0.07 to $0.11 for men, and from $0.04 to $0.06 for women. Meanwhile, the price of rice fell, so that the opportunity costthe amount of rice that could have been purchased with tobacco moneyfurther increased. While the average male smoker could have purchased an additional 1837 calories of rice with his cigarette money in 1995, the figure rose to 2942 calories in 1997. For women smoking bidis, the figure tripled, from 302 calories in 1995 to 907 in 1997. The potential in calories of rice for the average tobacco user nearly doubled for both men and women, from 721 and 419 respectively in 1995, to 1402 and 770 calories in 1997.2