Article Text

Counting 15 million more poor in India, thanks to tobacco
1. Rijo M John1,
2. Hai-Yen Sung2,
3. Wendy B Max2,
4. Hana Ross1
1. 1International Tobacco Control, American Cancer Society, Atlanta, Georgia, USA
2. 2Institute for Health and Aging, University of California San Francisco, San Francisco, California, USA
1. Correspondence to Dr Rijo M John, Director, International Tobacco Control, American Cancer Society, 250 Williams Street, Atlanta, GA 30044, USA; rijo.john{at}cancer.org

## Footnotes

• Competing interests None.

• Provenance and peer review Not commissioned; externally peer reviewed.

View Full Text

## Introduction

Tobacco use is primarily concentrated among the poor, and this disparity holds both within and across countries. Globally, the majority of smokers, 84%, live in low and middle-income countries (LMICs).1 In high-income countries and certain low-income countries, including India, tobacco use is primarily concentrated among the lowest socioeconomic groups.2 By 2030, over 80% of these all tobacco-related deaths are likely to occur in LMICs.2 Therefore, the most vulnerable segments of the population carry most of the burden of tobacco-related mortality and morbidity, as well as the adverse economic consequences of tobacco use. Tobacco thus indirectly contributes to widening the gap between the rich and the poor.

Studying the impact of tobacco use on poverty is important because the economic performance of many LMICs is measured by their progress towards poverty alleviation. In addition, poverty reduction is one of the Millennium Development Goals (MDG) set out by the United Nations.3 Research has demonstrated the link between poverty and tobacco use, but most of the studies do not quantify the impact.4–10 We attempt to do so using data from India, a low-income country where 10% of the world's tobacco smokers live, the second largest group of smokers in the world after China.2 This is the first study that enumerates how many people are de facto below the poverty line in India as a result of tobacco consumption. It is of significant political and social interest11 to count correctly the actual number of people living in poverty as even small changes in poverty rates capture the headlines and the attention of many policymakers.

Tobacco use contributes to poverty in several ways. First, since poor households typically devote a large share of their income to food, spending on tobacco often takes away resources that would otherwise be available for food, making the difference between an adequate diet and malnutrition. A study in Bangladesh showed that a typical poor smoker could purchase food worth at least 500 additional calories every day if he/she did not purchase tobacco. If the money spent on tobacco were directed towards food, about 10.5 million people in Bangladesh would have an adequate diet instead of being malnourished.4 Apart from food, tobacco expenditure reduce consumption of other important goods and services. Studies from China,12 13 and India,14 have shown, for example, that spending on tobacco reduces investment in education, thus further perpetuating the cycle of poverty.

Second, tobacco consumption causes various diseases15 resulting in increased spending on healthcare. The direct medical costs of treating only four major tobacco-related diseases in India in 2004 reached $1.2 billion, more than 4.7% of India's national healthcare expenditure.16 These costs present a huge burden for any household's budget, let alone for the budget of a poor family. Third, tobacco use has negative effects on earnings owing to the decreased productivity17 when a person is ill or absent from work because of a tobacco-related disease. Fourth, a death of a family breadwinner can have severe consequences for the welfare of the entire family. Approximately one million deaths a year are attributed to smoking in India.18 The majority of these deaths are among men, the primary economic providers for Indian families. The loss of a household head can increase the chances of the entire family falling below the poverty line. Thus, spending on tobacco exacerbates poverty both in the short and long run. In the short run, it results in a direct drain on disposable income, while in the long run the use of tobacco increases the risk of morbidity and mortality leading not only to higher healthcare expenditure, but also to reduced income for the entire family. Over 300 million people in India live in poverty and the poor represent 28.3% and 25.7% of the rural and the urban population, respectively.19 A typical Indian household with tobacco expenditure uses approximately 3% of the household budget for tobacco purchases.14 Many tobacco-consuming families live below or just above the poverty line, and owing to their tobacco consumption the resources available to them for other goods and services are on a par with families that have lower total expenditure but do not consume tobacco. Therefore, in order to make proper classification of who is and who is not below the poverty line, tobacco expenditure and expenditure for treating tobacco-related diseases need to be subtracted from the total expenditure of tobacco-consuming households. ## Data and methods The Indian government estimates the level of poverty using data from the National Sample Survey (NSS), a nationally representative survey conducted every 5 years since 1950. The survey contains detailed household expenditure data in categories such as food, education, clothing and tobacco products. The poverty line in India is defined by an income that can purchase at least 2400 calories per person per day in rural areas, or 2100 calories per person per day in urban areas.20 The average monthly per capita consumption expenditure (MPCE) is used as a proxy for the per capita monthly household income and the poverty level is expressed as the proportion of the population living below the poverty line. We employed data from NSS 2004–200521 on consumer expenditure, employment and unemployment for 124 644 households (79 298 living in rural areas and 45 346 living in urban areas). Since the household is the unit of observation, we calculated per capita monthly expenditure by dividing the household monthly expenditure based on a 30-day recall by the household size. In 2004–2005 the poverty line was set to 356.30 (US$8) and 538.60 (US\$12) per capita monthly expenditure in rural and urban areas, respectively.19 We used sampling weights to derive population level poverty estimates.

To estimate medical expenditure attributable to tobacco use, we adopted an epidemiological approach. First we calculate a smoking-attributable fraction (SAF) and then multiply it by total medical expenditure in tobacco-consuming households to obtain tobacco-attributable medical costs. The SAF can be estimated using either an inclusive approach or a disease-specific approach.22 The inclusive approach considers the impact of tobacco use on all types of medical expenditure and goes beyond the expenses on those diseases that are tobacco-related. The disease-specific approach considers only particular tobacco-related diseases. This study uses the inclusive approach and calculates the SAF as follows:

SAF=PE(RR1)PN+PE(RR)(1)

where PN and PE denote the percentage of people who have never used tobacco and the percentage who have ever used tobacco, respectively. RR represents the all-cause RR of mortality for ever tobacco users compared to those who have never used tobacco. RR of value x means the risk of mortality from any medical cause is x-times higher for ever tobacco users compared to those who never used tobacco. Tobacco use prevalence came from the National Family Health Survey,23 a nationally representative survey conducted in 2005. The RR estimates for India have been reported for the age group 30–69 separately for males and females, and for rural and urban areas.18 We computed a combined male and female RR (RRC) in urban areas using the following formula:

RRc=DFRRF+DMRRMDF+DM(2)

where D and RR refer to all deaths and the RR, respectively, and the subscripts F and M denotes females and males, respectively. A similar calculation was done for rural areas.

The impoverishing effects of tobacco use are expressed as the difference between the officially reported number of people below the poverty line and the actual number of people below this line when the expenditure on tobacco products and on tobacco-attributable medical care are subtracted from the average MPCE. This is an established method in the developmental literature24 and has been applied to tobacco use in China.6 25

It is possible that within a household only some of its members spend money on tobacco. In that case, one can argue, tobacco-related costs should not be attributed to the whole household. We assumed that the income brought into the house as well as the expenses are shared among all family members, because this practice is common in India. Moreover, even if not all members of a household are using tobacco, the non-smoking members are exposed to secondhand smoke and can have medical expenditure associated with this exposure. Therefore we apply the SAF to all medical expenditure by a family with a smoker to calculate the portion of those expenditure attributable to tobacco use.

All estimates are presented separately for rural and urban areas. It is a standard practice to treat rural and urban areas separately in the development literature from India owing to their vast socioeconomic differences.

## Results

### Household expenditure on tobacco

In 2004–2005, 55% of Indian households spent money on tobacco products. This share was 61% and 43% in the rural and the urban areas, respectively. Table 1 shows that, on average, tobacco consuming households spend 2.3% of their monthly expenditure on tobacco, a slightly smaller proportion compared to 5 years ago when the share of tobacco in total expenditure was approximately 3%.14 On average, tobacco-consuming households have lower MPCE than non-tobacco consuming households. This difference is statistically significant and is larger in urban areas (30% gap) compared to rural areas (19% gap). It can be partly explained by the larger average household size of tobacco-consuming households (6.43 and 6.07 in rural and urban India, respectively) compared to non-tobacco consuming households (5.38 and 5.24 in rural and urban India, respectively). The discrepancies in MPCEs and household size have serious implications for the poverty level. The lower the MPCE of a tobacco-consuming household, the higher the probability that a larger size household falls below the poverty line.

Table 1

Monthly per capita consumption expenditure (MPCE) in India in , 2004–2005

### Impoverishing effects of tobacco and misclassification of the poor

In 2005 the population of India was projected at 1.1 billion with the majority (71%) living in the rural areas (table 2).26 Officially, 28.3% of rural and 25.6% of the urban population lived below the poverty line that year,19 meaning that approximately 301.6 million people in India were classified as poor.

Table 2

Changes in poverty level after accounting for tobacco-related expenditure in India (2004–2005)

Accounting for tobacco purchases increases the poverty level to 29.8% and 26.3% for rural and urban India, respectively, and raises the total number of people classified as poor by 14 million (table 2). In a separate analysis, we found that smoked tobacco, non-smoked tobacco and dual use are responsible for 58%, 23%, and 19% of this increase in poverty as a result of tobacco expenditure, respectively.

The gender-combined RR for all-cause mortality for ever tobacco users was 1.76 in the rural areas and 1.90 in the urban areas (equation 2). Inserting the 2005 tobacco use prevalence (29.6% and 23.3% for rural and urban areas, respectively23) to equation (1), we estimated that 18% of medical expenditure by rural households and 17% of medical expenditure by urban households can be attributed to tobacco use. Accounting for these tobacco-related medical expenditure increased the poverty level to 28.4% in rural India and to 25.7% in urban India, meaning that an additional 0.9 million people should have been classified as poor (table 2).

The combined effect of both direct tobacco purchases and tobacco-attributed medical expenditure is an increase in poverty rates by 1.5 percentage points in rural areas and by 0.8 percentage points in urban areas. It means 15 million people became poor after accounting for tobacco use in 2005. These 15 million people were misclassified as being above the poverty line even though their resources for obtaining goods and services (excluding tobacco) were similar to people being officially classified as poor.

## Discussion

We found that spending money on tobacco and tobacco-related medical expenditure in tobacco-consuming families increases the poverty level in India. Accounting for tobacco use increased the rate of poverty by 1.6 percentage points in rural and by 0.8 percentage points in urban India after we reclassified 15 million people living in tobacco-consuming households as poor. Direct spending on tobacco, the opportunity cost of tobacco use, has a greater impact on poverty than tobacco-related direct medical care expenditure. The indirect costs of smoking such as reduced labour productivity are partially captured by lower monthly per capita expenditure.

The study has several limitations. First, the official definition of the poverty line has been criticised for not taking into account some minimum expenditure on healthcare and education. A study that computed the level of poverty in India while considering a certain minimum expenditure on education and healthcare suggested the poverty rates in India to be 29% higher in rural areas and 32% higher in urban areas compared to the official estimates.27 To the extent that such basic services are not included in the calculation of the poverty level, our estimates for the number of people below the poverty line are conservative.

Second, the unit of our analyses is a household, because of the design of the NSS. Tobacco use, on the other hand, may be restricted to only a few household members. Our estimates of the poverty impact of tobacco use would be more accurate if we had individual level consumption expenditure data. Unfortunately, such data are currently not available.

Third, the RR for healthcare utilisation would be a better indicator for calculating SAF of medical expenditure attributable to tobacco use than the RR for all-cause mortality for ever tobacco users. Since the RR for healthcare utilisation in India is yet to be assessed, we used the RR for all-cause mortality as a proxy. Studies indicate that such substitution would lead to a conservative SAF of medical costs.16 28 Therefore, we are most probably underestimating the amount of medical expenditure related to tobacco use.

Based on our results, we can conclude that regulating tobacco use is not only a good public health policy, but also a good poverty alleviation strategy. In many LMICs, where poverty hinders development, reducing tobacco use could have an important role in achieving their development goals.

We found that the average age of a tobacco-consuming household (25.9) was approximately one and half year less than that of a non-tobacco-consuming household (27.3). This means that tobacco expenditure disproportionally affects younger generations by allocating fewer resources to their education14 and/or health, which could result in severe intergenerational consequences.

One of the most effective tobacco control measures is a price increase via higher tobacco taxes.29 Since the prevalence of tobacco use is higher among the poor than the rich in many LMICs, it is often argued that tobacco tax policy is regressive, meaning that it can disproportionally hurt the poor.30 However, the empirical evidence shows that the poor are relatively more responsive to tobacco price increases than the rich,14 therefore the poor as a group will reduce their expenditure on tobacco more relative to the rich. The merits of regulating tobacco use through taxation for a poor household are multifold. First, a tax increase will reduce tobacco use and free up money for the consumption of essential goods and services such as food, clothing, healthcare and education. If a rich family stops using tobacco, they may divert their expenditure to luxury items or save the money, since their essentials have been already covered even with tobacco consumption. Second, lower tobacco consumption improves the health of the entire family and reduces the likelihood of spending on healthcare. Third, improved health because of less tobacco use will improve labour productivity and increase earning potential for the family. Fourth, more investment into education made possible by not spending on tobacco can further increase the earning potential of the family.

Poverty eradication is the first of the eight MDGs and reducing tobacco use, along with other strategies, can be a very effective means of achieving this goal. A study by the World Health Organization demonstrates that including tobacco control within the MDG framework will help to make investments in health, poverty reduction and development more effective in the short term and, in particular, more sustainable in the medium to long term.31 As Efroymson4 puts it, if national economic improvements lead to greater expenditure on tobacco but not food, then the benefits of economic growth will be erased. Given India's 33% share of the world's poor,32 its success in poverty reduction plays an important part in achieving the MDGs from a global perspective. Hence, reduction of tobacco use should be given high priority by the government of India.

### What is already known on this subject

• Tobacco use is high among many poor nations. Often, tobacco consumption is higher among those who can least afford it. However, the efforts to quantify this relationship are limited. Despite the high level of poverty and tobacco use in India, no attempt has been made to quantify the impact of tobacco on poverty there.

• This study finds that approximately 15 million people dropped below the poverty line in India after accounting for their tobacco use. Of this number, direct spending on tobacco accounted for 14 million and tobacco-related medical care for one million. Hence this paper contends that tobacco control measures would not only improve public health, but would also reduce poverty in India.

## Acknowledgments

We would like to acknowledge Elizabeth Ward, Evan Blecher and Jojo Mathew George for their valuable comments and edits during the preparation of this manuscript.

View Abstract

## Footnotes

• Competing interests None.

• Provenance and peer review Not commissioned; externally peer reviewed.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.