Article Text

Estimating the price elasticity for cigarette and chewed tobacco in Pakistan: evidence from microlevel data
Free
1. Durre Nayab,
3. Junaid Alam Memon,
4. Mahmood Khalid,
5. Anwar Hussain
1. Pakistan Institute of Development Economics, Islamabad, Pakistan
1. Correspondence to Dr Muhammad Nasir, Pakistan Institute of Development Economics, Islamabad 44000, Pakistan; nasir84{at}pide.org.pk

## Abstract

Background While there is a large literature on the magnitude of price elasticity of cigarette demand, less is known about the same for chewed tobacco. Moreover, the studies on cigarette demand in Pakistan tend to ignore the heterogeneity in the smoking behaviour. This study estimates price elasticity for cigarette and chewed tobacco demand across different income groups, provinces and regions; and use these coefficients for simulating Pakistan’s tax policy and its impact on revenue and health outcomes.

Method Deaton model was applied on the 2015–2016 Household Integrated Income and Consumption Survey dataset to estimate price elasticities of different tobacco products.

Results The demand for cigarettes is unitary elastic (−1.06), suggesting that a 1% increase in cigarette price would reduce its consumption by 1.06%. On the other hand, the demand for chewed tobacco is relatively inelastic (−0.55). Provincially, the own-price elasticity of cigarettes is negative and significant for all but Kyber Pakhtunkhwa (KP) province; whereas that of the chewed tobacco is negative and significant only for KP and Balochistan. Besides, the price elasticity of demand for both tobacco products is negative and significant only for lower income group and the rural region. The tax simulations favour a two-tiered tax system over the existing three-tiered system as the former will bring significantly better tax revenue and health outcomes.

Conclusions While confirming the effectiveness of tax policies in curbing tobacco use, this study concludes that higher tobacco taxes could increase tax revenue and improve public health in Pakistan.

• tobacco consumption
• cigarette demand
• price elasticity
• Pakistan

## Introduction

With around 24 million active adult tobacco users, Pakistan ranks among the largest tobacco consuming countries in the world.1 The prevalence rates among men and women are 32.4% and 5.7%, respectively.2 Among in-school youth (aged 13–15 years) in different cities, tobacco prevalence ranges between 6.1% and 14.1%. According to Household Integrated Income and Consumption Survey (HIICS 2015–2016), 45% of the Pakistani households reported some sort of tobacco consumption. Considering that tobacco consumption takes heavy toll on human health,3 4 Pakistan needs strict tobacco control measures. To this end, the country has already signed the WHO’s Framework Convention on Tobacco Control in 20055 and since then has introduced various non-taxation policies including pictorial warnings on cigarette packs; smoking bans; ‘No Smoking’ signs in public places and vehicles; bans on cigarettes sale to minors, active advertisements campaigns and distribution of free samples.4 6 However, the policies’ implementation is generally weak7 and raising concerns that contrary to its commitment of 30% reduction in the tobacco prevalence by 2025, the trend may actually rise in Pakistan.8

The weakest action area in the country’s combat against tobacco is its taxation policy despite it being an effective tool to reduce tobacco consumption. Globally, ample evidences exist for tobacco taxation policies that simultaneously result in curbing tobacco consumption and raising revenues.9 Assuming a complete pass-through of tax to price, the extent of reduction in tobacco consumption in response to a tax increase depends on price elasticity of the tobacco demand. Hence, price elasticity is an important factor for tobacco prevalence and may vary between −0.25 to −0.50 and −0.5 to −1.00 in high-income and low-income countries, respectively.10

## Results

### Tobacco prevalence

Tobacco consumption is widely prevalent with 45% of the HIICS households reporting using it in some form, of these 26% had cigarettes smoking and 13% had tobacco chewing members (table 1 Panel A). Prevalence of all tobacco products is reportedly higher in rural than urban areas. Furthermore, households allocate ~3% of their budget to tobacco products. In both regions, this share is higher for cigarette consumption than for users of chewed tobacco owing to the latter’s lower prices. While budget shares for chewed tobacco do not show regional differences, urban households do allocate higher budget share than their rural counterparts for cigarette consumption.

Table 1

Household-level tobacco prevalence across regions, provinces and income groups (%)

The provincial layer reveals interesting regional variations in tobacco prevalence (Panel B). Two-thirds of the households in Balochistan reported tobacco consumption having the highest and second highest percentages of cigarette and chewed tobacco users, respectively. Punjab has the lowest percentage of tobacco consuming households (40%) among all provinces. Khyber Pakhtunkhwa (KP) presents the lowest prevalence of smoking but has the highest percentage of chewed tobacco. The highest chewed tobacco consumption in KP followed by Balochistan has to do with the tradition of chewing naswar by Pashtuns—the ethnic group dominating the population in both provinces. Sindh maintains its third position in consuming cigarettes and chewed tobacco though its overall tobacco prevalence stands as the second highest in the country. Budgetary allocations for tobacco expenses are the highest in Sindh and lowest in KP.

Panel C reveals the prevalence and budgetary shares of tobacco products across the two income groups, namely upper 40% and lower 60%. Clearly, the prevalence of smoking exceeds that of the chewed tobacco in both the groups. The lower income group, however, apportions higher share of their budget on tobacco consumption and may, therefore, face strong welfare toll compared to their higher-income counterparts.

Figure 1 demonstrates cigarette and chewed tobacco consumption across income quintiles by region and province. These tobacco products are mostly consumed by the lower/middle-income households while the prevalence rate for these products, barring a few exceptions, is lower among the highest income quintiles. The World Health Survey report for Pakistan corroborates the consumption pattern found in this study vis-à-vis the income levels.2

Figure 1

Tobacco consumption by income quintiles across regions and provinces. KP, Khyber Pakhtunkhwa.

As evident from figure 1 (panel A), regional cigarette consumption reveals different consumption patterns for rural and urban areas. Contrary to the trend in rural areas, the prevalence rate in urban areas is highest in the upper income quintiles. For chewed tobacco, higher consumption is observed in the lower income quintiles in the rural areas. The urban areas, on the other hand, have a mixed distribution of consumers; yet we observe that highest income group has the lowest proportion of tobacco chewers (figure 1, panel C).

The provincial graphs reveal interesting patterns of cigarette consumption across income quintiles (figure 1, panel B). For instance, both Punjab and Sindh follow a similar pattern of cigarette consumption as is observed for Pakistan. However, in case of KP, the pattern is reversed. Not only does the province have the lowest percentage of cigarette consuming households (see table 1, panel B), it also has the highest proportion of consumers in the top income quintiles. This can have important implications for price elasticity of cigarettes in the province. A similar trend, though a less pronounced one, can also be observed for Balochistan. Households in the fourth and fifth-income quintiles (the top 40%) in KP and Balochistan consume 54% and 47% of the total smoked cigarettes, respectively. Figure 1, panel D presents chewed tobacco consuming households across income quintiles in the four provinces and show trends somewhat similar to those of the cigarette, except in case of KP which has a bell-shaped distribution implying that chewed tobacco is consumed in significant proportions across all income groups. Interestingly, in all the four provinces, >50% of the consuming households lie in the lower income group (the bottom 60%).

### Unit values and price elasticity

If the unit values were to be treated as prices, then the price of cigarettes and chewed tobacco in Pakistan is PKR 1.91 (US$0.02) per stick and PKR 0.42 (US$0.003) per gram, respectively. These prices do not vary significantly across regions. Cigarettes are found to be expensive in Balochistan (PKR 2.28) and cheapest in KP (PKR 1.63). Similar pattern is observed for chewed tobacco. The unit value for cigarette is higher (PKR 1.87) for upper income group compared with the lower one (PKR 1.02). For chewed tobacco, however, these are broadly similar for both income groups.

Before discussing the price elasticities, one must look for evidence that spatial variation in unit values explains significant variation in prices. This can be obtained by regressing the log of unit values on village (cluster) dummies and observing the values of F-statistics and R squared. Due to relatively small intra-cluster variation over a short time period, a larger share of variation should be coming from inter-cluster disparities in prices. Empirically, this means a larger R squared or a significant F-statistics. The analysis of variance (ANOVA) analysis in Panel A of table 2 reports these values. These results suggest that the cluster effects explain 43% and 72% variation in prices of cigarette and chewed tobacco, respectively. The significance of both F-statistics at conventional significance level provides sufficient evidence for link between variation in unit values and spatial price variation.

Table 2

Variation in log unit values and price elasticity for tobacco products

The results of symmetry-constrained own-price elasticities for cigarettes and chewed tobacco as well as the cross-price elasticities are presented in table 2 (Panel B) with their respective bootstrapped SE estimates in parentheses. The own-price elasticities for the given tobacco products have negative signs and are statistically significant at 1% level. The own-price elasticity of cigarettes is −1.06 suggesting that a 10% increase in the price of cigarettes in Pakistan will decrease its demand by almost the same percentage (ie, 10.6%).

The own-price elasticity for chewed tobacco in Pakistan is −0.55, suggesting that a 10% increase in the price of chewed tobacco products will decrease its demand by 5.5%. There are no comparable studies for this product, perhaps due to the aggregated nature of this commodity. As far as the cross-price elasticities are concerned, they are economically small and statistically insignificant indicating that these products are independent of each other. Table 3 presents the disaggregate analysis of own-price elasticity estimates across regions, provinces and incomes groups. The own-price elasticities of cigarette and chewed tobacco products are negative and significant for the rural but insignificant for the urban regions. This could be due to the fact that, in contrast to urban consumers who mostly belong to higher income group, the rural consumers, constrained by their budgets, are forced to cut down the tobacco use in face of a price hike.

Table 3

Disaggregate analysis of own-price elasticities

On the provincial level, the own-price elasticity of cigarettes is insignificant for KP while it is negative and significant for the others. This outcome is substantiated with findings in table 1 and figure 1. The cigarette price elasticity estimates for the rest of the provinces are similar in sign, significance and magnitude. The own-price elasticity of chewed tobacco product is negative and significant for KP and Balochistan but insignificant for Punjab and Sindh. Although these elasticity estimates indicate provincial differences in demand for chewed tobacco products, one cannot accurately interpret these due to the aggregated nature of this commodity.

## Discussion and conclusions

This paper is the first attempt to estimate the price elasticity of cigarettes and chewed tobacco using household level data while also considering the spatial heterogeneity across regions and provinces as well as income groups. The unitary price elasticity for cigarette demand found in this paper is in contrasts with that of Burki and coauthors4 whereby the same elasticity was found to be –0.58 for Pakistan. This high price elasticity can be explained with the fact that majority of the consumers come from the lower income group who are relatively more responsive to price changes (see figure 1 and table 3). Existing literature from Pakistan,14 India,22 Bangladesh23 and Nepal24 too shows that getting a unitary elastic demand for cigarette is not uncommon in the South Asian region. From policymaking perspective, the introduction of a three-tiered tax structure by the government and the subsequent increase in cigarette sales confirm the validity of these findings of price responsiveness of tobacco demand. Thus, we tend to support the idea that tax increase would reduce cigarette demand without raising illicit trade as argued by tobacco manufacturers when opposing any increase in tax rates.

Moreover, the simulation exercise establishes superiority of the two-tiered over a three-tiered tax system. The two-tiered system would be easier to implement, disincentivise tobacco companies to manipulate prices and increase the overall tax collections. Although the total tax collections do not significantly differ between the two variants of the two-tiered tax scenarios, in one (scenario 3) we, however, see a significant improvement in public health outcomes. The decision of whether to opt for scenario 2 or scenario 3 is, thus, contingent on the importance assigned to the respective outcomes that is, the tax collection or public health outcomes. This finding accentuates the need for estimating the health cost attributable to tobacco consumption to get accurate estimates for the assessment of various tax regimes. An important limitation of the study is that we could not perform the scenario analysis for chewed tobacco. This was due to lack of data on different variables and parameters required for this exercise. Future studies may consider conducting this analysis if and when the data get available.

#### What is already known on this subject

• The price elasticity of cigarette demand in Pakistan has been estimated using time series macrolevel data. However, macro (particularly of the national scope) tend to aggregate the behaviour of different individuals with varying characteristics and thereby ignores the heterogeneity in behaviour. These studies do not control for potential covariates either.

#### What important gaps in knowledge exist on this topic

• Price elasticity of demand has not been estimated for chewed tobacco products in Pakistan. Moreover, price elasticity for tobacco products across income groups, regions and provinces are also not available.

• Using microlevel data, Household Integrated Income and Consumption Survey 2015–2016, the study estimated price elasticity of tobacco demand, and conducted a heterogeneous analysis with respect to income, province and region. These estimates advance the understanding of the impact of tobacco tax changes on different populations, and also provide inputs for the simulation modelling. The simulation exercise confirms the superiority of the two-tiered tax structure over the three-tiered system.

## Acknowledgments

These data were previously presented as part of 'Economics of Tobacco Taxation and Consumption in Pakistan', Nayab D, Nasir M, Memon JA et al, Pakistan Institute of Development Economics 2018.

## Footnotes

• Contributors DN conceived the idea, supervised all steps, obtained the funding and revised the draft versions. MN designed the study, conducted empirical analysis and drafted the manuscript. JAM provided support in drafting the manuscript and revised the draft versions. MK and AH obtained and cleaned all the data and critically reviewed the draft at various stages. All authors read and approved the final version.

• Funding This research is funded by the University of Illinois at Chicago’s Institute for Health Research and Policy through its partnership the Bloomberg Philanthropies.

• Competing interests None declared.

• Patient consent for publication Not required.

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

• Data availability statement Data are available in a public, open access repository.

## 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.