Tobacco-free Economy: A SAM-based Multiplier Model to Quantify the Impact of Changes in Tobacco Demand in Bangladesh

Abstract

In Bangladesh, where tobacco use is pervasive, reducing tobacco use is economically beneficial. This article uses the latest Bangladesh social accounting matrix (SAM) multiplier model to quantify the economy-wide impact of demand-driven changes in tobacco cultivation, bidi industries and cigarette industries. First, we compute various income multiplier values (i.e., backward linkages) for all production activities in the economy to quantify the impact of changes in demand for the corresponding products on gross output for 86 activities, demand for 86 commodities, returns to 4 factors of production and income for 8 household groups. Next, we rank tobacco production activities by income multiplier values relative to other sectors. Finally, we present three hypothetical ‘tobacco-free economy’ scenarios by diverting demand from tobacco products into other sectors of the economy and by quantifying the economy-wide impact. The simulation exercises with three different tobacco-free scenarios show that, compared to the baseline values, total sectoral output increases by 0.92, 1.3 and 0.75 per cent. The corresponding increases in the total factor returns (i.e., gross domestic product, GDP) are 1.57, 1.75 and 1.75 per cent. Similarly, total household income increases by 1.40, 1.58 and 1.55 per cent.

JEL Classification: E1, C6, O2

Keywords

Social Accounting Matrix Bangladesh SAM Tobacco Consumption Tobacco-free Economy Multiplier Model Multiplier Values Input–Output Tables Demand-driven Interventions

1. INTRODUCTION

The tobacco epidemic in Bangladesh is pervasive. The Global Adult Tobacco Survey (GATS) estimates that 43.3 per cent of adults in Bangladesh (41.3 million persons) use tobacco in smoke and/or smokeless forms (World Health Organization [WHO], 2009). As the global call for the tobacco ‘endgame’ strategy is gaining momentum (Warner, 2013), it is relevant to quantitatively assess the significance of tobacco products and their place in the economy. This article offers a better understanding of the way tobacco-related production activities interact with various agents (i.e., other production activities, factors of production and socio-economic groups) in the Bangladesh economy.

Public health efforts to reduce tobacco consumption have strengthened over the last half century, often involving wide-ranging policy initiatives, such as smoke-free policies, mass media campaigns, restriction on youth access to tobacco products, price increases driven by fiscal measures, tobacco smuggling deterrence and cessation assistance programmes. The World Health Organization Framework Convention on Tobacco Control (WHO FCTC) is the first global health treaty that is providing the foundation for countries to implement and manage tobacco control (WHO, 2003). In addition, for effective implementation of some specific WHO FCTC provisions, WHO introduced the MPOWER measures in 2008, particularly to assist in reducing the demand for tobacco products at the country level (WHO, 2008). In response to the increasing tobacco control outreach, the tobacco industry frequently highlights the employment and income implications of reduced tobacco consumption, asserting that tobacco farming and the manufacturing, distribution and sale of tobacco products constitute a vital part of the economy. However, if tobacco use is reduced, resources previously spent on tobacco production would not disappear from the economy, but could be put to alternative uses, and the redistribution of resources from tobacco consumption to other goods and services could create jobs and generate income in other sectors of the economy.

Several independent studies have estimated the net impact on economic activity resulting from eliminating or reducing expenditure on tobacco, using assumptions about how the alternative expenditure would be redistributed in the economy (Jacobs et al., 2000). The results generally suggest that economic losses in the tobacco and associated sectors are outweighed by increasing employment in other industries, generated from the shifting of expenditure from tobacco to other sectors. For instance, a study on Scotland by McNicoll and Boyle (1992) assumed the elimination of domestic tobacco consumption expenditures and the redistribution of the amount to other sectors, according to average expenditure patterns. Similarly, van der Merwe (1998a, 1998b), Buck et al. (1995) and Irvine and Sims (1997) assumed full elimination, a 40 per cent decline and a 20 per cent decline in domestic tobacco consumption, respectively. All studies projected net economic gains. However, a study on Zimbabwe by van der Merwe (1998c) reported net losses under the scenario of elimination of domestic consumption expenditures and all tobacco production in 1980, and redistribution according to ‘average’ input–output (IO) patterns, with all production shifted to alternatives in agriculture. In terms of the assumptions regarding the redistribution of the marginal increase in smokers’ income, other scenarios are also suggested in the literature, for example, an increase of expenditure on recreational goods and services rather than on essential items (Buck et al., 1995).

In this study, we use the latest Bangladesh social accounting matrix (SAM) for the year 2006/07, which was constructed incorporating 86 sectors and 86 commodities, 4 factors of production and 8 household groups as well as a SAM-based multiplier model, to track how demand-driven reductions of the products of tobacco cultivation, bidi industries and cigarette industries may affect the economy. First, we compute various income multiplier values (i.e., backward linkages) for all the production activities in the economy to quantify the impact of changes in demand for the corresponding products on gross output for 86 activities, demand for 86 commodities, returns to 4 factors of production and income for 8 household groups. Next, we rank tobacco production activities by income multiplier values relative to other sectors. Finally, we present three hypothetical ‘tobacco-free economy’ scenarios by diverting exogenous demand amounts from tobacco-related products into other sectors of the economy proportionately and by quantifying its impact on sector-wise output, on the consumption of different commodities, on income generation for the various factors of production and on the income of different socio-economic groups. The simulation scenarios entail elimination of tobacco consumption (i.e., products originating directly from tobacco cultivation, bidis and cigarettes) in monetary terms, and redistribution of the same monetary amount to (i) all other commodity items, (ii) food items only and (iii) recreational or entertainment items only, according to the average consumption pattern.

2. METHODS

2.1 Bangladesh SAM 2006/07

A social accounting matrix is a data system in the form of a square matrix, which records the monetary transactions taking place in an economy during a specific period of time, generally 1 year. As a data framework, it may be described as a natural extension of the IO accounting systems that not only bring together disaggregated data on the inputs and outputs of the productive branches in the economy but also data concerning the distribution of the various kinds of factor incomes across institutional groups, the redistribution of income among these groups, the expenditure made by these groups on different types of commodities and savings and investments made by them. The data blocks in the SAM follow, in disaggregated terms, the main consecutive stages which can be distinguished in the circular flow that characterises the full economic process (Alarcon et al., 1991, p. 2; Pyatt and Round, 1979).

A centerpiece of Bangladesh’s Sixth Five-Year Plan (SFYP, 2011–15) was the delineation of the country’s macroeconomic outlook, and the SAM 2006/07 provides an important data framework for economic model construction (GoB Planning Commission, 2011; Khondker and Selim, 2011). Construction of the 2006/07 SAM is based on several data sets drawn from diverse sources, including the IO Table 2007 for Bangladesh prepared as a background document for the technical framework for the SFYP, the SAM for Bangladesh for 2000 by Bangladesh Planning Commission, Bangladesh Bureau of Statistics, Household Income and Expenditure Survey (HIES) 2005, Economic Survey of Bangladesh 2008 by the Ministry of Finance, Export Promotion Bureau and Bangladesh Bank (i.e., the Central Bank of Bangladesh) and the National Board of Revenue (GoB Planning Commission, 2011).¹

The Bangladesh SAM 2006/07 identifies the economic relations through four types of accounts: (i) production activity and commodity accounts for 86 sectors; (ii) four factors of productions with two types of labour and two types of capital; (iii) current account transactions between four main institutional agents, namely, household-members and unincorporated capital, corporation, government and the rest of the world and (iv) two consolidated capital accounts distinguished by public and private origins to capture the flows of savings and investment. The disaggregation of activities, commodities, factors and institutions in SAM 2006/07 is given in Table A1. There are three tobacco-related production activities (and correspondingly three commodities) in the SAM: (i) tobacco cultivation, which entails tobacco farming activities and the final product directly originating from that sector; (ii) the bidi industry (bidis and bidi-related manufactured tobacco products); and the (iii) cigarette industry (manufactured cigarettes).

Table 1 presents the Bangladesh SAM in its aggregate form. The SAM follows the fundamental accounting principle that for every income or receipt there is a corresponding expenditure or outlay, which underlies the double-entry accounting procedures embedded in the macroeconomic accounts of any country. However, instead of the double-entry conventions of national accounts used to depict the correspondence between income and expenditure, SAM uses single-entry accounting to show the income and expenditure correspondence. Thus, SAMs embody this principle but record the transactions between accounts in a square matrix (Alarcon et al., 1991; Pyatt and Round, 1979). The transactions or accounts constitute the dimension of the square matrix. Table 1 shows that the SAM 2006/07 ensures equality between supply and demand of production activities and commodities, between factor receipts and outlays, income and expenditures of institutions and between the savings and investment identities. This consistency is maintained not only at the macro level but also for each of the meso-level disaggregated accounts of the SAM.

2.2 The Tobacco Economy in the SAM 2006/07

According to the Bangladesh SAM, total household consumption of tobacco products amounts to 90,743 million Bangladeshi taka (BDT), of which 71,724 million BDT is spent on cigarette consumption, 17,931 million BDT on bidi and 1,088 million BDT on other tobacco farming products (mainly smokeless and non-manufacturing tobacco products). Households spend about 2.54 per cent of their total consumption expenditure on tobacco products.² The share of tobacco consumption to total GDP is 1.99 per cent (i.e., 1.57 per cent on cigarettes, 0.393 per cent on bidis and 0.024 per cent on other tobacco products). Rural and urban households contribute 65 and 35 per cent of the cigarette expenditure, respectively. The contributions of rural and urban households in bidi consumption are 91 and 9 per cent, respectively. Figure 1 reveals the tobacco consumption pattern in monetary terms by different household groups. Expenditure on cigarettes constitutes the bulk of total tobacco expenditure. Expenditure on bidis is higher in rural household groups than urban households and poorer households than their richer counterparts. Rural landless, small farmers and non-farm household groups spends about 30 per cent of their total tobacco expenditure on bidi products. Urban high-educated households mainly consume cigarettes.

Table 1
Aggregate Social Accounting Matrix (SAM) for Bangladesh, 2006/07

(1)(Act) (2)(Comm) (3)(Fact) (4)(HH) (5)(Crptn) (6)(Indtax) (7)(Imp Dty) (8)(Prd Sub) (9)(Exp Sub) (10)(Govt) (11)(ROW) (12)(Pvt Cap) (13)(Pub Cap) (14)(Pvt Inv) (15)(Pub Inv) Total Income

(1) Activities(Act) 9,375,170(M:86x86) 9,375,170(V:86x1)

(2) Commodities(Comm) 4,807,024(M:86x86) 3,576,006(M:86x8) 261,056(V:86x1) 934,403(V:86x1) 898,618(V:86x1) 257,258(V:86x1) 41,058(V:86x1) 18,002(V:86x1) 10,793,424(V:86x1)

(3) Factors(Fact) 4,468,549(M:4x86) 4,468,549(V:4x1)

(4) Households(HH) 4,200,436(M:8x4) 110,660(V:8x1) 413,040(V:8x1) 4,724,136(V:8x1)

(5) Corporation(Crptn) 268,113(V:1x4) 6,000(S:1x1) 274,113(S:1x1)

(6) Indirect Taxes(Ind Tax) 125,067(V:1x86) 125,067(S:1x1)

(7) Import Duty(Imp. Duty) 156,626(V:1x86) 156,626(S:1x1)

(8) Prod. Subsidy(Prd Sub) –17,473(V:1x86) –17,473(S:1x1)

(9) Export Subsidy(Exp Sub) –7,998(V:1x86) –7,998(S:1x1)

(10) Government(Govt) 25,860(V:1x8) 60,350(S:1x1) 125,067(S:1x1) 156,626(S:1x1) –17,473(S:1x1) –7,998(S:1x1) 342,433(S:1x1)

(11) Rest of World(ROW) 1,261,628(V:1x86) 1,261,628(S:1x1)

(12) Pvt Capital(Pvt Cap) 1,122,271(V:1x8) 213,763(S:1x1) –396,357(S:1x1) 939,676(S:1x1)

(13) Public Capital(Pub Cap) –53,284(S:1x1) –85,815 396,357(S:1x1) 257,258(S:1x1)

(14) Private Investment(Pvt Inv) 41,058(S:1x1) 41,058(S:1x1)

(15) Public Investment(Pub Inv) 18,002(S:1x1) 18,002(S:1x1)

Total Outlays 9,375,170(V:1x86) 10,793,424(V:1x86) 4,468,549(V:1x4) 4,724,136(V:1x8) 274,113(S:1x1) 125,067(S:1x1) 156,626(S:1x1) –17,473(S:1x1) –7,998(S:1x1) 342,433(S:1x1) 1,261,628(S:1x1) 939,676(S:1x1) 257,258(S:1x1) 41,058(S:1x1) 18,002(S:1x1)

Source: Authors’ calculation using SAM for Bangladesh 2006/07.

Note: M is matrix, V is vector and S is scalar, and the corresponding data dimensions are mentioned in the parentheses. All values are in million BDT and represent total amounts of the respective matrices, vectors or scalars. When read column-wise, the values indicate expenditures of respective accounts, and when read row-wise, the values indicate income for the corresponding accounts. For example, the total transaction amount of 4,807,024 million BDT between activity (Act) and commodity (Comm) accounts is derived from an 86 × 86 data matrix in the disaggregated SAM reporting how each of the 86 activities makes transactions with each of the 86 commodities in the production process. Similarly, in column 11, rest of world (ROW) account, the total amount of 934,403 million BDT is derived from the 86 × 1 vector in the disaggregated SAM reporting export amount of each of 86 commodities.

Figure 1

Source: Authors’ calculation using Bangladesh SAM 2006/07.

2.3 SAM Multiplier Model

As SAM inherits the feature of a modular analytical framework, it has frequently been used to examine the consequences of real shocks, using a multiplier model that treats the circular flow of income endogenously. More specifically, the SAM framework, under certain assumptions, can be used to estimate the effects of exogenous changes and injections, such as increases or decreases in the demand for specific products on the whole socio-economic system (Defourny and Thorbecke, 1984; Pyatt and Round, 1979; Robinson, 2006; Round, 2003; Thorbecke, 2000). The move from the SAM structure to a model structure requires that the accounts of this matrix be segregated into endogenous and exogenous. The need for this arises from the fact that there must be an entry in the system, that is, some variables must be manipulated exogenously via injections (such as a change in demand) in order to evaluate the consequences on the endogenous accounts.

As a general guideline, accounts a priori specified as objectives or targets when the SAM was built should be made endogenous. However, accounts intended to be used as policy instruments, or those that are not endogenously determined from direct interactions among domestic economic agents and institutions, should be made exogenous (Alarcon, 2000; Round, 2003). Following these criteria, the production account (sectors and commodities), the factors account and the households account are selected as endogenous accounts. This helps to focus on the interaction between two sets of agents (production activities and households) interacting through two sets of markets (factors and commodities). Government, corporations, the rest of the world (i.e., foreign countries) and the capital accounts are made exogenous, because government outlays are usually policy determined, the external sector is outside domestic control and investment is exogenously determined in the model (Alarcon, 2000; Round, 2003).

The impact of any given injection into the exogenous accounts of the SAM is transmitted through the interdependent SAM system among the endogenous accounts. The interwoven nature of the system implies that incomes of factors, households and the production sectors are all derived from exogenous injections into the economy via a multiplier process. Accounting multipliers are calculated according to the standard Leontief inverse formula:

Y = A * Y + X = (I - A) ​^{- 1} * X = M_{a} * X

(1)

where Y is a vector of endogenous variables (accounts), X is a vector of exogenous variables (accounts), A is the matrix of average propensities of expenditures for endogenous accounts, I is the identity matrix and M_a or ( I – A )^–1 is the matrix of aggregate accounting multipliers.

The multiplier process is developed here on the assumption that when an endogenous account receives an exogenous injection, it spends it exactly in the same proportions as shown in the matrix of average propensities to spend (APS). The elements of the APS matrix is calculated by dividing each cell by its corresponding column sum totals. The dimension of the M_a matrix is 184 × 184 with broadly categorised four endogenous accounts (i.e., 86 sectors and 86 commodities, 4 factors and 8 households). The dimensions of both the Y (endogenous) and X (exogenous) vectors are 184 × 1.

The interpretation of the values in M_a is straightforward. When read column-wise, the values show the increase of income in each of the 184 endogenous elements due to one unit of external injection into the column element via the exogenous accounts. The term ‘injection’ refers to the income increase via exogenous accounts due to the increased demand for sectoral outputs, or investment demand, or exogenous income transfers to households, and is expressed in monetary units. The sum of all the values in a particular column would then show the total backward linkage that is generated due to the one-unit injection in the corresponding column account. Modular partial backward linkages can be identified for each of the broad endogenous accounts in the SAM, that is, activities, commodities, factors and households. Table 2, in which the M_a matrix is partitioned and presented as a collection of sub-matrices, illustrates this further.

Table 2

Impact Sub-matrices of the Multiplier Matrix ( M_a )

	Activities	Commodities	Factors	Households
Activities	M₁₁(86 × 86)	M ₁₂(86 × 86)(output multiplier)	M₁₃(86 × 4)	M₁₄(86 × 8)
Commodities	M₂₁(86 × 86)	M ₂₂(86 × 86)(demand multiplier)	M₂₃(86 × 4)	M₂₄(86 × 8)
Factors	M₃₁(4 × 86)	M ₃₂(4 × 86)(GDP multiplier)	M₃₃(4 × 4)	M₃₄(4 × 8)
Households	M₄₁(8 × 86)	M ₄₂(8 × 86)(income multiplier)	M₄₃(9 × 4)	M₄₄(8 × 8)

Source: Authors’ own.

Notes: The dimension of each matrix is shown in the parentheses; ‘M’ in bold are the sub-matrices that are used in the analysis.

When demand-driven interventions occur through the commodity accounts (i.e., an exogenous increase or decrease in demand), the relevant blocks for the impact analysis refer to M₁₂ (gross output impact for 86 sectors), M₂₂ (commodity demand impact for 86 commodities), M₃₂ (GDP impact for four factors of production) and M₄₂(household income impact for eight household groups). As the present multiplier framework has four endogenous accounts, four types of multiplier measures can be calculated: the output multiplier, demand multiplier, GDP multiplier and household income multiplier.

Shocks occurring in a particular production account (e.g., commodity demand) will impart their impact predominantly to the account’s linkage industries. The inter-industry transactions in the SAM 2006/07 reveal that tobacco cultivation mainly has backward linkages with the fertiliser industry, wholesale and retail trade, water and land transportation, banks, insurance and real estate and other services. The bidi and cigarette industries have their main linkages with tobacco cultivation, the paper industry, basic chemicals, wholesale and retail trade, water and land transport, bank insurance and other services, and rural and urban building infrastructure. All these sectors are also linked with other sectors of the economy. The ultimate economic impact of the exogenous reduction in tobacco consumption and the concomitant increase in demand for other commodities from the redistribution in our simulation scenarios is the net benefit, which depends on the values of the multipliers.

While the multipliers obtained using the SAM as a linear model allow us to capture the structural features of income distribution and interrelations among various economic agents, the model rests on some assumptions. It assumes the existence of excess capacity that would allow relative prices to remain constant in the face of demand shocks, that expenditure propensities of endogenous accounts remain constant and that production technology and resource endowments are given for a period. Therefore, the SAM-based multiplier model inherits the assumptions of the traditional IO analysis, particularly the following (Alarcon, 2000, p. 16): (i) APS are fixed, linear and considered constant or at least stable over the short to medium term; (ii) relative prices are constant over the time horizon of the model, usually the short term, implying that the components which make up any account bunch have substitution elasticities which are zero across accounts and infinite within accounts, that is, they are homogeneous within and heterogeneous across accounts; (iii) expenditure–income elasticities are constant and equal to unity; (iv) there is a perfect complementarity between capital and other factor inputs; (v) it offers a nominal analysis in current prices; (vi) the economy has idle capacity utilisation; and (vii) a number of accounts are exogenous.

The richness of the SAM multipliers comes from their tracing out chains of linkages from changes in demand to changes in production, factor incomes, household incomes and final demands (Thorbecke, 2000, pp. 21–22). Therefore, the SAM framework permits tracing and quantifying all the propagation channels in the economy, and in doing so, provides a very useful policy instrument for meso-level economy-wide impact analysis of demand-driven interventions.

2.4 Simulation Design

We assumed three scenarios within a hypothesised tobacco-free economy context, based on the premise that when resources are no longer devoted to a given economic activity, they do not simply disappear; rather, they are redirected to other economic activities (Warner, 2000). If people stop tobacco consumption, for instance, the money that would have been spent on tobacco products could now be spent on something else. The new spending could stimulate demand in other production activities. We designed the following three simulation scenarios to evaluate and compare the economy-wide impact of changes in demand for tobacco-related commodities:

A reduction of exogenous tobacco demand by 90,743 million BDT, equalling the total household consumption demand amount spent on products directly originating from tobacco cultivation (mainly smokeless non-manufactured items), bidi industries and cigarette industries, and redistributing the total 90,743 million BDT among all other commodity demands according to weighted household consumption shares.

A reduction of exogenous tobacco demand by 90,743 million BDT as above and redistributing that demand to only food items (i.e., sugarcane, potatoes, vegetables, pulses, oilseeds, fruit, cotton, tea, spices, other crops, livestock, poultry, shrimp, fishing, forestry, rice milling products, grain milling products, processed fish, oil, sweetener products, tea, refined salt and processed food) according to the weighted household consumption shares of these items.

A reduction of exogenous tobacco demand by 90,743 million BDT and redistributing that amount to the commodity category ‘entertainment’ only.

3. RESULTS

3.1 Ranking of Tobacco Commodities in Terms of Their Backward Linkages

The M₁₂, M₂₂, M₃₂ and M₄₂ sub-matrices of the M_a multiplier matrix show column-wise the increase in gross outputs in the sectors, demand for the commodities, incomes of the factors of production and incomes of the household groups, respectively, that results from one unit of injection into any particular column account. Therefore, row sums of the respective sub-matrices show the total increase in output (i.e., the gross output multiplier), commodity demand (i.e., demand multiplier), GDP (i.e., the GDP multiplier) and household income (i.e., income multiplier).

The values in Table 3 indicate how a one-unit increase in the demand for each of the commodities leads to a total increase in the income of four endogenous accounts as a whole. For instance, considering the gross output multipliers in panel 1 on the left of Table 3, a one-unit injection in ‘fish process’ leads to 5.08 units of output increase in the economy, compared to the 2.45 unit increase when the injection occurs in the cigarette industry. The top five sectors in terms of generating the highest gross output multipliers are fish processing, shrimp farming, rice milling, hotels and restaurant and poultry rearing, which indicates their high integration with other sectors. The bottom five sectors that generate the least gross output multiplier values are fertilisers, petroleum, yarn, basic chemicals and cotton, indicating their lower level of integration with other sectors and high leakages attributable to imports from the rest of the world.

Table 3
Ranking of Commodities in Terms of Their Backward Linkages

Panel 1Ranking of Commodities in Terms of Generating Total Output Multiplier Value
Panel 2Ranking of Commodities in Terms of Generating Total Demand Multiplier Value

Value Rank Value Rank

Top 5 Commodities Top 5 Commodities

Fish process 5.08 1 Fish process 5.46 1

Shrimp farming 4.86 2 Shrimp farming 5.22 2

Rice millings 4.62 3 Rice millings 5.02 3

Hotel and restaurant 4.55 4 Poultry rearing 4.97 4

Poultry rearing 4.55 5 Hotel and restaurant 4.96 5

Ranking of Tobacco-related Commodities Ranking of Tobacco-related Commodities

Bidi industry 3.67 36 Bidi industry 4.04 39

Tobacco cultivation 2.63 67 Tobacco cultivation 3.24 67

Cigarette industry 2.45 68 Cigarette industry 2.68 70

Bottom 5 Commodities Bottom 5 Commodities

Fertiliser industry 0.91 82 Toiletries 1.73 82

Petroleum 0.33 83 Petroleum 1.22 83

Yarn industry 0.23 84 Yarn industry 1.21 84

Basic chemical 0.07 85 Basic chemical 1.06 85

Cotton cultivation 0.07 86 Cotton cultivation 1.06 86

Panel 1Ranking of Commodities in Terms of Generating Total Output Multiplier Value	Panel 2Ranking of Commodities in Terms of Generating Total Demand Multiplier Value
Top 5 Commodities			Top 5 Commodities
Fish process	5.08	1	Fish process	5.46	1
Shrimp farming	4.86	2	Shrimp farming	5.22	2
Rice millings	4.62	3	Rice millings	5.02	3
Hotel and restaurant	4.55	4	Poultry rearing	4.97	4
Poultry rearing	4.55	5	Hotel and restaurant	4.96	5
Ranking of Tobacco-related Commodities	Ranking of Tobacco-related Commodities
Bidi industry	3.67	36	Bidi industry	4.04	39
Tobacco cultivation	2.63	67	Tobacco cultivation	3.24	67
Cigarette industry	2.45	68	Cigarette industry	2.68	70
Bottom 5 Commodities			Bottom 5 Commodities
Fertiliser industry	0.91	82	Toiletries	1.73	82
Petroleum	0.33	83	Petroleum	1.22	83
Yarn industry	0.23	84	Yarn industry	1.21	84
Basic chemical	0.07	85	Basic chemical	1.06	85
Cotton cultivation	0.07	86	Cotton cultivation	1.06	86

Source: Authors’ calculation.

Bidis, tobacco cultivation and cigarettes are ranked 36th, 67th and 68th among the 86 commodities in the SAM. Similarly, the demand multipliers in panel 2 on the right of Table 3 report that bidis, tobacco cultivation and cigarettes are ranked 39th, 67th and 70th, respectively. A one-unit increase in the exogenous demand for cigarettes will create only 2.68 units overall demand in the economy compared to, for example, 5.46 units when demand increases for processed fish.

3.2 GDP Multipliers

The GDP multipliers in Table 4 show that the sectors that produce high (low) gross output and demand multipliers do not automatically generate high (low) GDP multipliers, accordingly. Table 4 reports the ranking of the commodities in terms of total GDP multipliers as well as for each of the factors of production, separately. Tea cultivation produces the highest GDP multiplier value, that is, a one-unit increase in exogenous demand leads to 2.26 units increase in total factor returns. Bidis, tobacco cultivation and cigarettes are ranked 52th, 61th and 72nd among the 86 commodities, respectively. A unit injection into bidis generates 1.59 unit of GDP, of which 0.39, 0.41, 0.70 and 0.09 units are accrued to unskilled labour, skilled labour, capital returns (profits) and returns on land use, respectively. A one-unit increase in cigarettes produces only 0.83 unit of GDP. While increases in the demand for bidis and cigarettes generate relatively higher rewards for labour and capital factors, tobacco cultivation generates higher value for returns to land.

3.3 Household Income Multipliers

The multiplier values in Table 5 obtained from the income multipliers in the M₄₂ sub-matrix show the increase in incomes of respective household groups due to a one-unit increase in the corresponding exogenous demand for the commodity. For example, when read row-wise, a one-unit increase in the exogenous demand for bidi products increases a landless household’s income, as a group, by 0.1 units, the marginal farmer group’s income by 0.09 units and so on, resulting in a total increase in household income by 1.49 units.³ However, when read column-wise, the values show how a particular household group’s income increases due to a one-unit injection in different sectors. For example, a one-unit injection in tea cultivation would increase the landless group’s income by 0.11 units, whereas they accrue only 0.05 units when the injection occurs in the form of increased cigarette demand. The column-wise ranking of values in descending order for each of the household groups would then reveal the ranking of the sectors for corresponding households in terms of income generation, and therefore, poverty alleviation. Bidis, tobacco cultivation and cigarettes are ranked 51st, 61st and 72nd, respectively, among the 86 commodities in the Bangladesh SAM 2006/07.

Table 4
Ranking of Commodities in Terms of GDP Multiplier

Total GDP
Unskilled Labour
Skilled Labour
Capital
Land

Value Rank Value Rank Value Rank Value Rank Value Rank

Top 5 Commodities

Tea cultivation 2.26 1 0.63 5 0.40 39 0.57 57 0.65 2

Jute cultivation 2.23 2 0.68 2 0.47 20 0.66 42 0.42 10

Paddy cultivation 2.17 3 0.66 3 0.44 26 0.61 51 0.47 9

Shrimp farming 2.16 4 0.45 31 0.51 11 1.08 3 0.12 24

Baling 2.15 5 0.58 7 0.51 14 0.83 19 0.23 17

Tobacco Commodities

Bidi industry 1.59 52 0.39 46 0.41 35 0.70 36 0.09 55

Tobacco cultivation 1.32 61 0.36 51 0.30 65 0.43 68 0.23 16

Cigarette industry 0.83 72 0.21 71 0.22 69 0.36 71 0.05 73

Bottom 5 Commodities

Toiletries 0.32 82 0.08 82 0.09 82 0.14 82 0.02 82

Petroleum 0.08 83 0.02 84 0.02 83 0.04 83 0.00 85

Yarn industry 0.07 84 0.02 83 0.02 84 0.03 84 0.00 84

Basic chemical 0.03 85 0.01 85 0.01 85 0.01 85 0.00 86

Cotton cultivation 0.03 86 0.01 86 0.01 86 0.01 86 0.01 83

	Total GDP	Unskilled Labour	Skilled Labour	Capital	Land
Top 5 Commodities
Tea cultivation	2.26	1	0.63	5	0.40	39	0.57	57	0.65	2
Jute cultivation	2.23	2	0.68	2	0.47	20	0.66	42	0.42	10
Paddy cultivation	2.17	3	0.66	3	0.44	26	0.61	51	0.47	9
Shrimp farming	2.16	4	0.45	31	0.51	11	1.08	3	0.12	24
Baling	2.15	5	0.58	7	0.51	14	0.83	19	0.23	17
Tobacco Commodities
Bidi industry	1.59	52	0.39	46	0.41	35	0.70	36	0.09	55
Tobacco cultivation	1.32	61	0.36	51	0.30	65	0.43	68	0.23	16
Cigarette industry	0.83	72	0.21	71	0.22	69	0.36	71	0.05	73
Bottom 5 Commodities
Toiletries	0.32	82	0.08	82	0.09	82	0.14	82	0.02	82
Petroleum	0.08	83	0.02	84	0.02	83	0.04	83	0.00	85
Yarn industry	0.07	84	0.02	83	0.02	84	0.03	84	0.00	84
Basic chemical	0.03	85	0.01	85	0.01	85	0.01	85	0.00	86
Cotton cultivation	0.03	86	0.01	86	0.01	86	0.01	86	0.01	83

Source: Authors’ calculation.

Table 5

Ranking of Commodities in Terms of the Income Generation Effects to Households

	Total Household Income		Rural Landless		Rural Marginal Farmers		Rural Small Farmers		Rural Large Farmers		Rural Non-Farm Poor		Rural Non-Farm Non-Poor		Urban Low Education		Urban High Education
	Value	Rank	Value	Rank	Value	Rank	Value	Rank	Value	Rank	Value	Rank	Value	Rank	Value	Rank	Value	Rank
Top 5 Commodities
Tea cultivation	2.18	1	0.11	25	0.13	1	0.33	1	0.31	2	0.19	3	0.47	16	0.23	5	0.41	37
Jute cultivation	2.13	2	0.13	6	0.13	2	0.29	4	0.24	9	0.19	2	0.48	14	0.24	2	0.44	22
Paddy cultivation	2.09	3	0.12	12	0.13	4	0.29	3	0.25	6	0.18	4	0.46	25	0.23	3	0.42	33
Other services	2.05	4	0.15	1	0.13	3	0.25	13	0.16	17	0.20	1	0.47	17	0.29	1	0.40	38
Pulses cultivation	2.04	5	0.11	30	0.12	6	0.29	5	0.26	5	0.17	7	0.47	21	0.21	10	0.42	30
Tobacco Commodities
Bidi industry	1.49	51	0.10	47	0.09	51	0.17	53	0.10	52	0.13	50	0.38	50	0.15	48	0.38	44
Tobacco cultivation	1.26	61	0.07	65	0.08	62	0.17	52	0.14	22	0.11	61	0.29	65	0.13	59	0.28	66
Cigarette industry	0.78	72	0.05	71	0.05	72	0.09	72	0.05	72	0.07	71	0.20	71	0.08	71	0.20	70
Bottom 5 Commodities
Toiletries	0.30	82	0.02	82	0.02	82	0.03	82	0.02	82	0.03	82	0.08	82	0.03	82	0.08	82
Petroleum	0.08	83	0.00	83	0.00	83	0.01	83	0.01	84	0.01	83	0.02	83	0.01	84	0.02	83
Yarn industry	0.06	84	0.00	84	0.00	84	0.01	84	0.00	85	0.01	84	0.02	84	0.01	83	0.02	84
Cotton cultivation	0.03	85	0.00	86	0.00	85	0.01	85	0.01	83	0.00	85	0.01	86	0.00	86	0.01	86
Basic chemical	0.03	86	0.00	85	0.00	86	0.00	86	0.00	86	0.00	86	0.01	85	0.00	85	0.01	85

Source: Authors’ calculation.

3.4 Tobacco-free Economy: Simulation Outcomes

We present three ‘tobacco-free’ simulation scenarios of reduced tobacco consumption equalling the household tobacco consumption amount (90,743 million BDT) and concomitant increases in exogenous demands for other commodities under three different set-ups: (1) ‘scenario 1’ includes an increase in the exogenous demand for all other commodities according to weighted household expenditure shares, (2) ‘scenario 2’ includes an increase in the exogenous demand for ‘food’ commodities according to weighted household expenditure shares and (3) ‘scenario 3’ includes an increase in the exogenous demand for ‘entertainment’ only.

The four panels in Table 6 show the simulation outcomes in terms of total sectoral output, commodity demand, returns to factor and household income under the three scenarios and compare them with the baseline scenario that reproduces the original SAM data in the absence of any change in exogenous demand. Compared to the baseline values, total sectoral outputs increased by 86,200, 121,860 and 70,010 million BDT under scenarios 1, 2 and 3, respectively, resulting in 0.92, 1.30 and 0.75 per cent increases from the baseline values. The corresponding increases in total factor returns (i.e., GDP) are 1.57, 1.75 and 1.75 per cent, respectively. Similarly, total household income would increase by 66,030, 74,730 and 73,390 million BDT, respectively, resulting in increases of 1.4, 1.58 and 1.55 per cent, respectively.

Table 6
Simulation Outcome for the Four Endogenous Accounts

Total Amount(million BDT) Change from Baseline(million BDT) % Changefrom Baseline

Panel 1: Total Sector (Activity) Output

Baseline 9,375,170 — —

Scenario 1 9,461,370 86,200 0.92

Scenario 2 9,497,030 121,860 1.30

Scenario 3 9,445,180 70,010 0.75

Panel 2: Total Commodity Demand

Baseline 10,793,400 — —

Scenario 1 10,897,300 103,900 0.96

Scenario 2 10,932,200 138,800 1.29

Scenario 3 10,871,800 78,400 0.73

Panel 3: Total Factor Returns

Baseline 4,468,550 — —

Scenario 1 4,538,680 70,130 1.57

Scenario 2 4,546,930 78,380 1.75

Scenario 3 4,546,920 78,370 1.75

Panel 4: Total Household Income

Baseline 4,724,140 — —

Scenario 1 4,790,170 66,030 1.40

Scenario 2 4,798,870 74,730 1.58

Scenario 3 4,797,530 73,390 1.55

	Total Amount(million BDT)	Change from Baseline(million BDT)	% Changefrom Baseline
Panel 1: Total Sector (Activity) Output
Baseline	9,375,170	—	—
Scenario 1	9,461,370	86,200	0.92
Scenario 2	9,497,030	121,860	1.30
Scenario 3	9,445,180	70,010	0.75
Panel 2: Total Commodity Demand
Baseline	10,793,400	—	—
Scenario 1	10,897,300	103,900	0.96
Scenario 2	10,932,200	138,800	1.29
Scenario 3	10,871,800	78,400	0.73
Panel 3: Total Factor Returns
Baseline	4,468,550	—	—
Scenario 1	4,538,680	70,130	1.57
Scenario 2	4,546,930	78,380	1.75
Scenario 3	4,546,920	78,370	1.75
Panel 4: Total Household Income
Baseline	4,724,140	—	—
Scenario 1	4,790,170	66,030	1.40
Scenario 2	4,798,870	74,730	1.58
Scenario 3	4,797,530	73,390	1.55

Source: Authors’ calculation.

Table A2 reports the simulation impacts on each of the 86 sectoral outputs. The drastic elimination of tobacco demand led to obvious net reductions in sectoral outputs for tobacco cultivation, bidis and cigarettes. We observe a net negative impact on a few other sectors, for example, paper, basic chemicals, wholesale trade, retail trade and water transport under scenario 1; paper, basic chemicals and the communication sector under scenario 2; and paper, basic chemicals, wholesale trade, retail trade, water transport, land transport and railway transport under scenario 3. However, we observe a net positive impact on the majority of the 86 production activities, and the impact magnitudes are much higher leading to an overall positive economic impact. Table A3 shows similar patterns, that is, the commodity demand in most sectors went up, leading to an overall net positive impact. Table A4 reports the impacts in terms of returns to each of the four factors of productions and incomes for the eight household groups. A net positive impact is observed throughout.

4. CONCLUSION

This article highlights the fact that reduced tobacco use does not lead to a loss in the economy, as money no longer spent on tobacco would be used to purchase other goods and services. This reallocation of spending creates demand stimuli in other sectors of the economy and generates larger multiplier effects and, in the process, the aggregate benefit to the economy outweighs the loss in tobacco-related sectors. This article does not take into account the negative health consequences and several associated types of societal costs of tobacco consumption, and merely presents an economic analysis by looking at the ways different agents in the economy interact in monetary terms. The use of the SAM offers a framework of analysis that brings together the growth and redistributive elements in a single framework, and also facilitates the conducting of simulation exercises to trace and quantify each stage of various demand shocks (stimuli) in the case of tobacco-related scenarios. The results show that tobacco farm products and bidis are ranked in the bottom third quartile and cigarettes are ranked in the bottom quartile in terms of income generation for various agents in the economy. The findings of this article support the core demand and supply reduction provisions in the WHO FCTC, particularly the full-scale implementation of WHO MPOWER measures to reduce the demand for tobacco. In Bangladesh, where a tobacco epidemic is pervasive, striving for the tobacco endgame strategy is economically beneficial.

Footnotes

Appendix

Table A4

Simulation Outcome in Terms of Factor Returns and Household Income

	Baseline	Scenario 1		Scenario 2		Scenario 3
	Value	Value	% Increase	Value	% Increase	Value	% Increase
Impact on Factor Returns
Labour unskilled	1,107,770	1,125,130	1.57	1,128,760	1.89	1,117,280	0.86
Labour skilled	1,130,940	1,143,290	1.09	1,144,940	1.24	1,159,970	2.57
Capital	1,941,430	1,971,080	1.53	1,967,830	1.36	1,977,460	1.86
Land	288,419	299,175	3.73	305,406	5.89	292,212	1.32
Impact on Household Income
Rural landless	300,255	304,102	1.26	304,397	1.36	304,678	1.45
Rural marginal farmers	283,097	287,117	1.40	287,679	1.59	287,170	1.42
Rural small farmers	549,960	558,396	1.51	560,352	1.85	557,416	1.34
Rural large farmers	341,538	348,071	1.88	350,468	2.55	345,990	1.29
Rural non-farm poor	433,474	438,990	1.26	439,853	1.45	439,225	1.31
Rural non-farm non-poor	1,156,860	1,173,960	1.46	1,174,560	1.51	1,175,550	1.59
Urban low educated	490,267	496,686	1.29	497,902	1.53	495,532	1.06
Urban high educated	1,168,680	1,182,850	1.20	1,183,660	1.27	1,191,970	1.95

Source: Authors’ calculation.

Note: All values except percentages are in million BDT.

Acknowledgements

The authors are thankful to Samira Asma, Deliana Kostova, Xin Xu, Krishna Palipudi, Rebecca Bunnell and Jing Xu (at the Centers for Disease Control and Prevention, Atlanta, USA), and Rubana Mahjabeen (University of Wisconsin-Superior, USA) for their comments and suggestions. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Notes

References

Alarcon

J.V.

, Heemst

J.V.

, Keuning

, de Ruijter

, & Vos

(1991). The social accounting framework for development: Concepts, construction and applications. UK: Aldershot-Avebury.

Alarcon

J.V.

(2000). Social accounting matrix-based modelling: Extension to wellbeing and environment and computable general equilibrium models (applications using the 1975 and 1980 Ecuador SAMs). The Hague, The Netherlands: Institute of Social Studies (ISS).

Buck

, Godfrey

, Raw

, & Sutton

(1995). Tobacco and jobs. York: Society for the Study of Addiction and Centre of Health Economics, University of York.

Defourny

, & Thorbecke

(1984). Structural path analysis and multiplier decomposition within a social accounting matrix framework. The Economic Journal, 94(373), 111–36.

Efroymson

, Ahmed

, Townsend

, Alam

S.M.

, Dey

A.R.

, Saha

, . (2001, September). Hungry for tobacco: An analysis of the economic impact of tobacco consumption on the poor in Bangladesh. Tobacco Control, 10(3), 212–17.

GoB Planning Commission (2011). Sixth five-year plan of Bangladesh, FY 2011–FY 2015, accelerating growth and reducing poverty: Part 3. Statistical annex and technical framework. General Economics Division, Planning Commission, Ministry of Planning, Government of the People’s Republic of Bangladesh, Dhaka.

Irvine

I.J.

, & Sims

W.A.

(1997). Tobacco control legislation and resource allocation effects. Canadian Public Policy, 23(3), 259–73.

Jacobs

, Gale

, Capehart

, Zhang

, & Jha

(2000). The supply-side effects of tobacco-control policies. In Jha

& Chaloupca

(Eds), Tobacco control in developing countries (pp. 311–41). Oxford, UK: Oxford University Press.

Khondker

B.H.

, & Selim

(2011, September). A social accounting matrix for Bangladesh 2006/07: Methodology and results. In Mustafa

K.M.

& Shamsul

(Eds), Technical framework papers: Sixth five-year plan. Background paper in volume 2. Dhaka: Bangladesh Planning Commission.

10.

McNicoll

, & Boyle

(1992). Regional economic impact of a reduction of resident expenditure on cigarettes: A case study of Glasgow. Applied Economics, 24(3), 291–96.

11.

Pyatt

, & Round

(1979). Accounting and fixed price multipliers in a social accounting matrix framework. The Economic Journal, 89(356), 850–73.

12.

Robinson

(2006). Macro models and multipliers: Leontief, Stone, Keynes, and CGE models. In de Janvry

& Kanbur

(Eds), Poverty, inequality and development: Essays in honor of Erik Thorbecke (pp. 205–32). New York: Springer Science.

13.

Round

(2003). Social accounting matrices and SAM-based multiplier analysis. In Bourguignon

& Pereira da Silva

L.A.

(Eds), The impact of economic policies on poverty and income distribution: Evaluation techniques and tools (pp. 301–24). New York: World Bank and Oxford University Press.

14.

SHDU (2000). SHD SAM 1993 for Bangladesh: A social accounting matrix for sustainable human development. In-house project report/output, Sustainable Human Development Unit, Planning Commission, Government of Bangladesh, Dhaka.

15.

SHDU (2002). Social accounting matrix 2000 for Bangladesh (SHD-SAM 2000): A social accounting matrix for sustainable human development. In-house project report, TP Series, number 7, Sustainable Human Development Unit, Planning Commission, Government of Bangladesh, Dhaka.

16.

SHDU (2003). Supply–use and input–output table 2000 for Bangladesh. Sustainable Human Development Unit, Planning Commission, Government of Bangladesh, Dhaka.

17.

Thorbecke

(2000). The use of social accounting matrices in modeling. Paper prepared for the 26th General Conference of The International Association for Research in Income and Wealth, Cracow, Poland, 27 August to 2 September 2000.

18.

van der Merwe

(1998a). The economics of tobacco control in South Africa. In Abedian

, van der Merwe

, Wilkins

& Jha

(Eds), The economics of tobacco control: Towards an optimal policy mix (pp. 251–71). Cape Town: Medical Association of South Africa Press.

19.

van der Merwe

(1998b). Employment and output effects for Bangladesh following a decline in tobacco consumption. Population, Health and Nutrition Department, World Bank, August 1998, Washington, DC.

20.

van der Merwe

(1998c). Employment and output effects for Zimbabwe with the elimination of tobacco consumption and production population. Health and Nutrition Department, World Bank, Washington, DC.

21.

Warner

K.E.

(2013, May). An endgame for tobacco? Tobacco Control, 22(Suppl 1): i3–5.

22.

Warner

K.E.

(2000). The economics of tobacco: Myths and realities. Tob Control, 9(1), 78–89.

23.

World Health Organization (WHO) (2009). Global adult tobacco survey: Bangladesh Report 2009. Dhaka, Bangladesh: Country Office for Bangladesh, 2009.

24.

World Health Organization (WHO) (2008). WHO report on the global tobacco epidemic, 2008: The MPOWER package. Geneva: World Health Organization.

25.

World Health Organization (WHO) (2003). WHO framework convention on tobacco control. Geneva: WHO. Retrieved 5 July 2015, from http://whqlibdoc.who.int/publications/2003/9241591013.pdf

	(1)(Act)	(2)(Comm)	(3)(Fact)	(4)(HH)	(5)(Crptn)	(6)(Indtax)	(7)(Imp Dty)	(8)(Prd Sub)	(9)(Exp Sub)	(10)(Govt)	(11)(ROW)	(12)(Pvt Cap)	(13)(Pub Cap)	(14)(Pvt Inv)	(15)(Pub Inv)	Total Income
(1) Activities(Act)		9,375,170(M:86x86)														9,375,170(V:86x1)
(2) Commodities(Comm)	4,807,024(M:86x86)			3,576,006(M:86x8)						261,056(V:86x1)	934,403(V:86x1)	898,618(V:86x1)	257,258(V:86x1)	41,058(V:86x1)	18,002(V:86x1)	10,793,424(V:86x1)
(3) Factors(Fact)	4,468,549(M:4x86)															4,468,549(V:4x1)
(4) Households(HH)			4,200,436(M:8x4)							110,660(V:8x1)	413,040(V:8x1)					4,724,136(V:8x1)
(5) Corporation(Crptn)			268,113(V:1x4)							6,000(S:1x1)						274,113(S:1x1)
(6) Indirect Taxes(Ind Tax)	125,067(V:1x86)															125,067(S:1x1)
(7) Import Duty(Imp. Duty)		156,626(V:1x86)														156,626(S:1x1)
(8) Prod. Subsidy(Prd Sub)	–17,473(V:1x86)															–17,473(S:1x1)
(9) Export Subsidy(Exp Sub)	–7,998(V:1x86)															–7,998(S:1x1)
(10) Government(Govt)				25,860(V:1x8)	60,350(S:1x1)	125,067(S:1x1)	156,626(S:1x1)	–17,473(S:1x1)	–7,998(S:1x1)							342,433(S:1x1)
(11) Rest of World(ROW)		1,261,628(V:1x86)														1,261,628(S:1x1)
(12) Pvt Capital(Pvt Cap)				1,122,271(V:1x8)	213,763(S:1x1)							–396,357(S:1x1)				939,676(S:1x1)
(13) Public Capital(Pub Cap)										–53,284(S:1x1)	–85,815	396,357(S:1x1)				257,258(S:1x1)
(14) Private Investment(Pvt Inv)												41,058(S:1x1)				41,058(S:1x1)
(15) Public Investment(Pub Inv)										18,002(S:1x1)						18,002(S:1x1)
Total Outlays	9,375,170(V:1x86)	10,793,424(V:1x86)	4,468,549(V:1x4)	4,724,136(V:1x8)	274,113(S:1x1)	125,067(S:1x1)	156,626(S:1x1)	–17,473(S:1x1)	–7,998(S:1x1)	342,433(S:1x1)	1,261,628(S:1x1)	939,676(S:1x1)	257,258(S:1x1)	41,058(S:1x1)	18,002(S:1x1)