Tobacco Documents Online

I 54 Excerpt from Proceedings/3rd World Conference' Smoking and Health, Volume II, Health Consequences,, Education, Cessation Activities, and Social Action; PRICING OUT TOBACCO: PRICE AS A FACTOR IN CIGARETT CONSUMPTlOI Robert H. Miller . :a For many centuries governments have levied various consumption taxes or'" tariffs (duties) which are paid by consumers in the price of the commodity they,' purchase. Whether conceived as a means of raising revenue or discouraging con-# sumption, these taxes occupy an important part in revenue systems all over the world. A Today all countries have some form of taxation on tobacco. There is usually an import duty, and in addition, an excise tax on the manufactured article. In some countries the government operates the industry as a monopoly with the profits turned over to the government. Measurement of factors that affect ciga- rette consumption is necessary for determining the long-run economic ouNook for the industry as well as for public and private officials who make decisions af- fecting the industry. This paper gives some time series analyses of cigarette demand and considers the implication of statistical findings to the assigned, topic of "pricing out to- bacco." The paper starts with a simple analysis of the demand for cigarettes. Some numerical results are given. The goal of the discussion is the determination of the consumption of cigarettes in response to changes in prices and in con. sumer incomes. THE CONCEPT OF DEMAND Laymen as well as economists have long observed that when the price of a commodity is high the sales of it are low, and when the price declines people buy more of it.' How responsive is the change in the amount of a good pur- chase&to a change in its price, in incomes, or in other factors? The degree of this sensitiveness economists call elasticity of demand. t The concept that sales of an article depends on the price of such article was stated by the Frenchmen Cournot and Dupuit in the 19th century. They stated that the price-quantity relationships could be tabulated and a formula developed to represent the relationship. Marshall extended and popularized these concepts. Henry L. Moore began publishing a series of books that have become classics in statistical dem;:nd analysis. Since the 1920s scores of reports of demand studies have been written in the U.S. Department of Agriculture and else- where, both in this country and abroad. In USDA '. - studies are used to estimate or fore- cast prices and consumption or trade as a par' ic:dtural',outrook service. Also, the, results are used'to estimate the probable eff"a. rnativeagriculturariprognmssuch as those to adjust production or support prices (1).. (or Thi+• mar ifE of I pric'pric• eIaS unii of vat tin p r, oIl tu ch ch to pr ca pe ag si; at ce ti, c< 0 r n ir S f d, C t t r 0

Elasticity is merely the ratio of the relative change in quantity demanded (or dependent variable) to the relative change in price (or independent variable). This ratio can fall into 3 categories: If E is greater than 1, then elasticity of de• mand is said to be elastic; if E = 1, themelasticity of demand is sai&to be unitary; if E is less than 1, then elasticity of demand is said to be inelastic. Whether elasticity is elastic, unitary, or inelastic will affect the consequences of price changes on consumption. If the total revenue of the industry after a price increase is smaller than the total revenue obtained from the sale before the price increase, the demand is elastic;'if total revenue is larger, the deman&is inr elastic; and if the total revenue stays unchanged, the demand has an elasticity of unity. But our interest goes further. We are interested in a measure of the amount of elasticity or inelasticity. Evem for a single commodity, elasticity study results vary from market to market, from grade to grade, from use to use, from time to time, etc. ` CONSUMPTION RELATED TO PRICE AND INCOME Statisticians typically measure the per capita use of tobacco and tobacco products in terms of per adult (18 years and older) or per person 15 years and older. (Use can be measured in terms of cigarettes, tobacco weight, or expendi- tures. The number of cigarettes is a key standard measure though product changes over time must, be recognized. Manufacturers have made numerous changes such as filters, lengthiand diameterof cigarettes, and composition of the tobacco blend. Since expenditures represent price times quantity, the multiplied product is not as useful for analysis as the two component parts.) Since asignifi- eant share of the totali population is below the smoking age, the consumption per adult is a more meaningful measure, particularly when countries of diverse age distributions are compared and age distribution changes over time. Consumer deman& for cigarettes is related to several variables. Besides the size of the adult population, the most significantt factors are per capita dispos- able income, retail price, and consumer tastes and preferences (including its per- ceived effect on health). Over the years economists have held that the price elas- ticity of demand for cigarettes is inelastic, i.e., that E is less than 1. Even though cigarettes are produced in relatively large-scale plants variable costs are dominant, so we can assume that production occurs under conditions of constant cost, i.e., the cost of production per unit is constant over a wide range of volume. Hence, relative elasticity of demand is the exclusive determi+ nant of the long-term reaction of price and output to a price change such as the imposition of a tax increase. The relationships between cigarette consumption, cigarette price, and con- sumer income can be estimated rather precisely by mathematical analysis. I have followed the standard practice of deflating both price an&consumer income by dividing,each by a consumer price index (CPI). This assumes that a doubling of cigarette prices and consumer incomes would not sign?"rantly change the quan- tity of cigarettes consumed. This is reasonable if all other prices were doubled, too:

56 Thus, Q is per capita cigarette consumption; P is deflated cigarette price cr 8. OBSERVATIONS ~~ 5! computed the means are shown in the Statistical Appendix. gression equations, squared multiple correlations, standard errors and elasticiti Data from several countries are analyzed by ordinary least squares regressi equations. This analysis assumes linear relationships between variables. The population CPI). ,T, price/CPI);, Y is per capita deflated consumer income (i.e., consumer incom (income taxes) and less on tariffs and consumption taxes like those on tobacco inc (cigarettes). Also, a more efficient production and distribution industry would or reduce the real cost of cigarettes if competitive conditions apply. Increased car- thi ton sales in supermarkets and chain drug stores and abolition of resale price thi maintenance are evidences of this trend. ab Typically, as a country develops, it tends to rely more heavily on direct taxes sumption before 1946 showed a higher elasticity than the postwar years. (Per of capita consumption = f.(Index of cigarette prices; 1967 = 100 = Index of per er capita disposable income, 1967 = 100),1935-45: e=-2.05;1946-74: e=-0.73.) sul portant. When the cigarette industry is growing rapidly and new smokers are being attracted, price may be more of a consideration. For example, U.S. con- The stage of the development of industry and government may also be im-1i : countries, more sharply. The US. situation, with a 3-stage tax pattern (federal, state, and local), contrasts with a single taxing authority in overseas jurisdictions. tries. Cigarette prices have increased in recent'years, due in part to tax increases/ ta But prices of other goods and services have likewise increased, and, in most , m into the 1970s. This pattern occurs in both monopoly and non-monopoly coun• While this finding is not unexpected, the interesting trend to emerge from e, the statistics assembled for the analysis is the sharp decline in the relative price ra of cigarettes in many countries since World War II, with the declines continuing tf recently from theUnite& States by Sutton (2) of-0.45. Coefficients have expected signs in these equations and are significantly dif m ferent from zero. The price elasticities at the means are close to those reporte 51 GOVERNMENT IMPACT ON PRICES In many countries the government controls the tobacco industry in atom- plete manner through a government monopoly. While the degree of control varies, this form is believed by many to insure maximum state revenue and to be the best way to combat tax evasion. China, Russia, Japan, Bulgaria, and France are the largest of the "monopoly" countries, which include countries aroun&the Mediterranean and Balkan area, as well as in Asia. The different~ patterns of tobacco duties and taxes can be illustrated with the United States and United Kingdom. The expenditures for tobacco products consist of charges for rnant+`actur- ing; wholesaling, retailing, and tobacco value, as welli as taxes and ol-ties. Since an

e Ptin yu cr~ tnt (fed dicna o ~ 0 ~. kers~ IS.c ars. (P ex of~ =-0.]3 rect ta~ n tobati try wou reased p esale prii A y in a coili? of controt~ eandtobe; and Frani+e around the, strated witk, manu `actUF 4ties. Since 0 ® W.r Il, about 40 to 50% of US. cigarette spending at retail has gone for t,axes. The federal tax rate has been unchanged since 1951, so with persist- tn+nt expenditures for tobacco products, the federal tax share has declined. sr,d focat tax rates have increased rapidly, but these too have leveled this rey. Unitcd Kingdom's excise duty for unmanufactured tobacco has been in- *Veyd periodically and now stands at the equivafentof $18.18 per pound, plus A% od rvlorem for nonpreference tobacco. Last year the duty represented #1Ai of consumer expenditures on tobacco and amounted to 6%of total revenues. The United Kingdom and the United States illustrate the variation in tax #wDtodt, as the former applies duties on leaf entering the country. The United *tte+tus nominal import duties but specific taxes apply to finished products. M contrast to a fixed rate, taxes may be ad valorem. Ad vcrlorem requires *ynrnire tobacco to carry a larger duty than less expensive tobaecol Specific tiMrt piice a relatively heavier burden on the cheaper grades of goods than on Ma better grades because the rate bears no relation to tobacco value. The US. dl,aL or; triple system of tax collection means manufacturers file am on the basis of finished product shipments while states and local govern- wmts re+ceive revenues from wholesalers who operate nearby. 5f11FTIhG INCIDENCE, AND EFFECTS ON CONSUMPTION PATTERNS In contrast, taxes may be multiple-stage, that is, imposed at every transfer irl tstle in a channel of production or distribution. But whether specific or gen- Wai, and'regardless of their form, commodity and sales taxes are shifted to con- wAws, the form of the tax may influence how the tax is shifted (3). Tobacco taxes are shifted to consumers largely by price increases. The price Mrrrux is rarely exactly the amount of the tax. Whether the price rises by more et kss than the tax depends on the elasticity of demand for the commodity and Mr drwmstances of _its supply. According to relative elasticity of demand for O+t tixed commodity or service, and the rate of the tax, sales might be consider- +bty modified, with appreciable effects on the pattern of personal consumption &d on the structure of production and distribution. Tobacco products excises or sales taxes obviously'fall on smokers and may o0t bear any relation to income levels. Typically, taxes on mass-consumed com- "odiues impose disproportionate burdens on the lowest income classes (4). MAnufacturers tailor their product changes to the prevailing tax policy. Thus b the United States the appearance of successively longer cigarettes (most re- C'"y 120'mm) is due to a specific tax per package or per thousand cigarettes. In Contrast, U.K. manufacturers have gone rapidly, to filters, smaller cigarettes, sr+d now are planning to introduce synthetic tobacco to reduce requirements Mkh bear duties on a poundage basis. Mr~thin the United States and across international boundaries, the movement of eiprettes from low-tax areas to high-tax areas is significant. Cigarettes from 0* low•tax states of Kentucky and North Carolina move to states that levy 11

u: from e price inuing cou n- reases. most ederal, ctions. be im- rs are con, (Per of per 0.73.) : taxes bacco would d car- price 57 World War 11, about 40 to 50% of US. cigarette spending at retail has gone for excise taxes. The federal tax rate has been unchanged since 1951, so with persist- endy rising expenditures for tobacco products, the federal tax sharehasdeclined: State and local tax rates have increased rapidly, but these too have leveled this past year. United Kingdom's excise duty for unmanufactured tobacco has been in- creased periodically and now stands at the equivalent of #18:18 per pound, plus 8.4% ad volorem for nonpreference tobacco. Last year the duty represented 55% of consumer expenditures on tobacco and amounted to 6%'0 of total revenues.. The United Kingdom and the United States illustrate the variation in tax methods, as the former applies duties on leaf entering the country. The United States has nominal import duties but specific taxes apply to finished products. In contrast to a fixed rate, taxes may be ad valorem. Ad valorem requires expensive tobacco to carry a larger duty than less expensive tobacco. Specific rates place a relatively heavier burden on the cheaper grades of goods than on the better grades because the rate bears no relation to tobacco value. The U!S. dual or triple system of tax collection means manufacturers file taxes on the basis of finished product shipments while states and local govern- ments receive revenues from wholesalers who operate nearby. SHIFTING INCIDENCE, AND EFFECTS ON CONSUMPTION PATTERNS In contrast, taxes may be multiple-stage, that is, imposed at every transfer of tide in a channel of production or distribution. But whether specific or gen- eral~ and regardless of their form, commodity and sales taxes are shifted to con- sumers, the form of the tax may influence how the tax is shifted (3). Tobacco taxes are shifted to consumers largely by price increases. The price increase is rarely exactly the amount of the tax. Whether the price rises by more or less than the tax depends on the elasticity of demand for the commodity and the circumstances of.its supply. According to relative elasticity of demand for the taxed commodity or service, and the rate of the tax, sales might be consider- ably modified, with appreciable effects on the pattern of personal consumption and on the structure of production and distribution. Tobacco products excises or sales taxes obviously' fall on smokers and may not bear any relation to income levels. Typically, taxes on mass-consumed com- modities impose disproportionate burdens on the lowest income classes (4). Manufacturers tailor their product changes to the prevailing tax policy. Thus in the United States the appearance of successively longer cigarettes (most re- cently 120 mm) is due to a specific tax per package or per thousand cigarettes. In contrast, U.K. manufacturers have gone rapidly to filters, smaller cigarettes, and now are planning to introduce synthetic tobacco to reduce requirements which bear duties on a poundage basis. Within the United States and across international boundaries, the movement of cigarettes from low-tax areas to high-tax areas is significant. Cigarettes from the low-tax states of Kentucky and North Carolina move to states that levy .

58 higher excise taxes. In Europe, considerable quantities of cigarettes are taken b~ tourists and by people living near frontiers where prices are higher. Movement o~ cigarettes from Switzerland to Italy, for example, is an important activity. In r~ cent years about one-seventh of Italian consumption is apparently accounted b~ Swiss consignments on which the Italian duty was not paid. The smoking rate the southern states is probably higher than the United States average. An ele- vated level 10% above average would mean that 29% of the cigarettes for whit~ Kentucky and North Carolina excise taxes were paid went to other states. , • SOME EXAMPLES OF PRICE CHANGES What happens when excise taxes change? Studies of marketing margins o individual foods suggest that price spreads increase somewhat with reduc quantities, that they tend to be somewhat between percentage and absolut~ amounts, and that they are probably closer to absolute amounts than percent ages. Theory may suggest reasons for expecting price spreads to come down with increases in quantity and also expecting them to go up in other cases (1, 3). Table I illustrates some possibilities of price, quantity, and revenue relation ships under various margin changes. This illustration assumes a 20% excise tax„ rise from a base level where the consumer price and revenue is divided equally' between the tax authority and the industry (firms). An elasticity of demand with respect to price of -0.5 is assumed. Thus, a 10% increase in price means a' 50,'o decrease in consumption. ' The large-scale production in the cigarette industry and consumer prefer- ences for certain brands means that a relatively few firms exercise some control over, industry output and selling price. Theoretically, a firm operates at an out- put level where marginal cost equals marginal revenue. Firms can, choose to in- crease their margin by the same amount as the tax increase, make no change, or add some intermediate amount. If a firm makes no change in selling price, its gross margin willi decline with reduced consumption. A margin increase that off- sets the quantity decline will leave gross margin unchanged. But if firms raise their margins in line with the tax increase, then firms can maintain their share of the sates price and maximize revenue short of further price increases. The above assumptions consider an essentially static market. If the market- ing firms consider that consumption is expanding they may prefer to obtain in- creased revenue from expanded volume instead of additional price increases that reduce the potential market and allow substitutes to enter. COMPATIBILITY OF FISCAL, SUMPTUARY, AND PROFIT OBJECTIVES The sizable tax yield from tobacco means a government has a strong interest in retaining a tobacco industry. To collect maximum tax revenue, a government would desire the largest output possible even though sales might be made at a loss. Clearly, with firms making the output decisions, production in this range (sales below cost)~is very unlikely. - - .Z_.

t ken b a 59 TABLE I n J. ent o .Inre-, ted by rate iti An elew ~whic J ;Ins O duced solute ;rcent= n with ~ation= se tax qually mand eans a refer- ~ ntrol out- to in- 1ge, or ce, its at off- s raise ~are of arket- in in- s that I VES iterest ment eata range I Tax and MargimChanges Under Assumed Price, Quantity Relationships, 3 Examples* Tax Increase Alone Tax and Morgin Tax and Margln lncrease Equally Increase Unequally Item Base Amount Change Amount Change Amount Change Number Percent Number Percent . Number Percent Price 100 110 +10 120 +20 113.6 +13.6 Quantity sold 100 95 - S 90 -10 93.2 - 6.8 Excise Tax 50 60 +20 60 +20 60:0 +20.0 Margin 50 50 0 60 +20 53.6 7.2 Total revenue 10,000 10;450 + 4.5 10,800 + 8 10,588 + 5.9 Tax revenue 5,000 5,700 +14.0 5,400 + 8 5,592 +11.8 Gross margin 5,000 4,750 - 5.0 5,400 + 8 4,996 - 0.1 • Assumes elasticity of demand with respect to price of-0.5. Like a revenue tariff, an excise tax may discourage consumption and at the same time yield sizable revenue. Raising an excise tax rate generally increases the tax yield but not proportionately, as an increase in consumer price reduces con- sumption. In a growth-oriented society, firms generally prefer larger output to moderate margins. With the sumptuary objective uppermost, a government could discourage use with increased taxes. In the United Kingdom, the 35% increase in tobacco taxes in April put the price of king-size filter-tip cigarettes up to nearly $1.10 per pack of 20 (5). This means taxes account for about 55-73 cents per pack ver- sus 12 to 29 cents per pack in the United~States. Before World War 11, per capita consumption of cigarettes was higher in the United Kingdom than in the United States, but in recent years the U.K. rate has been about one-fifth below the U.S. level (U.K.'s tobacco use per person was two-fifths below) (6). Keep in mind that the higher the tax the greater inducement to evasion, whether directly by consumers, as in purchase of bootlegged items, or by manu- facturers in the production of substitutes (mini-cigarettes, other tobacco prod- ucts) that avoid the highest rate or appear to give the consumer a more econom- ical buy. The history of international tobacco taxation is a varied one with some countries raising tobacco taxes relatively more than others. The prospects are probably best in countries thatdo not grow tobacco and import'all their require- ments. Prospects are less where tobacco is widely grown and manufactured, and least where the tax authority is diffused to different levels such as in the United States. These differences will probably never be completely resolved butcontinue to be governed by social, psychological, and political influences, as well as by the economic considerations of the different firms and'individuals. I.

60 REFERENCES 1. Waugh, F.V. Demand and Price Analysis, Some Examples From Agriculture. U.S' partment of Agriculture Tech. Bul. 1316, 1964. 2. Sutton, W. "An Econometric Analysis of the Structure of, the U.S. Tobacco Indu 0 Unpublished Ph.D. Dissertation, 192 pp., University of Kentucky, Lexington, tucky,1974. 3. Shuitz, W.J., and Harriss, C. Lowell. "American Public Finance," 8th ed., Prentice Englewood Cliffs, New Jeney, 1965. 4. Kalmus, T.S. The burden of federal excise taxes by income classes, Quarterly Rer of Economics ond Business 10:17-24,1970. 1 5. "U.K. Tobacco Taxes Jump," Foreign Agriculture, Wasliingtonj D.C., May 5, 1975.1~ 6. Beese, D.H. ed:, "Tobacco Consumption in Various Countries," Tobacco Resei Council Research Paper 6, 3rd ed., London, 1972. U. S. DEFA.HI'MENT OF HEAIbTH, EDUCATION, AND WELFARE Public Health Service National Institutes of Health~ .;~ Mr. to rr p p'"I l, 1 4 to trT rtiawi il t s+tit:a t.lktr p4mli avtl"6If' "I'h. 1(r Jf: \St runt: for t abr>u f ron that eron cil ? Cou t ice n T3cX, farr and S cati' loca thos belI C ~ 5pe clg: an cro If of an I