Jump to:

Philip Morris

An Economic Assessment of the Role of Advertising in Cigarette Consumption

Date: Jan 1992
Length: 23 pages
2060476160-2060476182
Jump To Images
spider_pm 2060476160_6182

Fields

Author
Alchin, T.M.
Area
LEGAL DEPT
Type
REPT, REPORT, OTHER
ABST, ABSTRACT
BIBL, BIBLIOGRAPHY
CHAR, CHART, GRAPH, TABLE, MAPS
Litigation
Bcnj/Produced
Named Person
Alchin
Alemson
Baltagi
Bass
Bishop
Boddewyn
Chetwynd
Clarke
Clements
Cowling
Cox
Doroodian
Eckard
Ekelund
Foster
Grabowski
Hamilton
Harrison
Herington
Jackson
Johnson
Johnston
Laugesen
Leeflang
Levin
Lewit
Mcguinness
Mcleod
Meads
Metwally
Miller
Parsons
Peles
Raftery
Rao
Reuijl
Roberts
Samuelson
Schmalensee
Schnabel
Schneider
Schoenberg
Seldon
Smith
Stewart
Telser
Walker
Waterson
Yoo
Named Organization
Australian Tobacco Marketing Advisory Co
Ceasa
Metra Consulting Group
Nsw Government
Toxic Substance Board New Zealand
Uk Tobacco Advisory Council
Australian Government
Site
N28
Author (Organization)
Univ of Western Sydney
Master ID
2060476000/6364

Related Documents:
Date Loaded
16 May 2000
Brand
Benson & Hedges
UCSF Legacy ID
mtd42d00

Document Images

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size:

Page 1: mtd42d00
AN EODNOlIIC JLSSPS,4II+STP OF THE ROLE OP ADVffitTISING 321 CIC,7,RETTS CDl7SUKMOei by Terry Maxwell Alehin Department of Economics Faculty of Commerce Working Papers in Economics No. WP92l01, January 1992 TSSN 1034-0238 ISBN 1 86341 073 2 2060476i60
Page 2: mtd42d00
XCAb6i4l An Economic Assessment of the Role of Advertising in Cigarette Consumption in Australia by Terry Maxwell Alchin Faculty of Commerce University of Western Sydney, Nepean PO Box 10 Kingswood NSW 2747 ABSTRACT This study fooks at the determinants of cigarette smoking in Australia via the use of Ordinary Least Squares regression. Advertising expenditure, both in an absolute and per adult form, is analysed as a determinant as well as cigarette price, income. price of substitutes and a habit variable. Advertising expenditure is also tested in a disaggregated (into electronic and print media) form as well as in aggregate. The• results indicate that advertising expenditure does not statistically significantly affect the demand for cigarettes while both cigarette price and the habit variable are significant.. This has important policy implications in that cigarette advertising bans may not reduce cigarette consumption while increased cigarette price via taxation would be expected to reduce consumption. 2060476161
Page 3: mtd42d00
IECA%1., Introduction From September 1973 to September 1976 the Australian government introduced a ban on the advertising of cigarettes on television and radio as a matter of health policy. Also from January 1991 a ban in print media across Australia applied. These were intended to reduce the level of cigarette consumption in Australia in response to the accumulation of health data relating morbidity and high mortality amongst cigarette smokers. !n late 1991, the NSW Government followed the lead of Victoria and South Australia and banned cigarette advertising In areas such as on billboards and taxis. Much debate continues in Australia as to whether such bans work in reducing cigarette consumption or whether, since cigarettes are a legal commodity, such advertising bans are an infringement of individuai rights or such. The debate in Australia is rife at the moment and some answer to the question whether cigarettb advertising affects cigarette consumption is required to conclude this debate. Many empirical studies on this issue have been conducted around the world and have some up with varying results. This study differs from previous studies in a few respecta. This paper ueas both disagflregated (into both electronic and print media) and aggregated cigarette advertising expenditure data, uses tagged cigarette prices and advertising data as well as lagged consumption (the habit) variable and incorporates a substitute for cigarettes which is not pipe or roR-your-own tobacco (in this case food). As well this study examines cigarette advertising in an absolute dimension and also as a per capita variable. In this papef some of the previous evidence is examined from Australia, USA, United Kingdom and New Zealand. This does not imply that no evidence' is available for other economies but Instead the jiim of this short look at previous studies is to illustrate the types of evidence and modelsz used. There are various studies that have examined the effects of advertising on cigarette sales or the role of the electronic advertising bans imposed in 1965 in Great Britain, 1971 in the USA and 1976 in Australia that serve as representative of such evidence. : __2060476162
Page 4: mtd42d00
tECA55ul Australian Evidence 3 The earliest Australian study of cigarette consumption is that provided by Metwally (1980) who estimated that the sales elasticity of advertising for the then leading brand of cigarettes was 0.2. This meant that as advertising expenditure increased by 100 per cent, cigarette consumption increased 20 per cent. Clements et al. (1985) provided estimates of several models relating per adult annual tobacco consumption to tobacco price, per capita income and a set of dummy variables designed to capture the phasing-in of the electronic media advertising ban from 1973 to 1976 and the possible differential effects of the ban in the years after 1976. However, they did not directly consider the effects of actual advertising expenditure. they found that the ban in 1976 in Australia of electronic media advertising, while a significant influence, reduced consumption by 2 to 3 per cent per annum. McLeod (1986) furthered this work and examined the role, of the 1976 electronic advertising ban in Australia. He concluded that the ban produced an immediate but short-lived effect of 5-6 per cent reduction in cigarette consumption. An investigation of advertising expenditure and aggregate demand for cigarettes in Australia by Johnson (1986) found no relation between total expenditure on press, radio and television advertising and total cigarette sales. Also, no effect on sales was found In 1976-77 from the ban of cigarette advertising on television but it is contended by Walker (1984) that some of the ban's intended effect had probably been achieved before this by the warnings that were required on television from 1972. Chetwynd et al. (1988) criticised the results of this study when they argued that analyses using annual data such as Johnson (1986) may mask significant relationships which are present with quarterly or bi-monthty series. This it is argued may explain the inconsistent'results of studies refatingyadvertising expenditure to sales of cigarettes. The studies which used annual data which included Leeflang and Reuijl (1885) and Johnson (1986) found little significant advertising effect while those using shorter time periods showed more significant results. 2060476163
Page 5: mtd42d00
(EG39+I) a Johnson (1988) in a later paper confirmed his earlier conclusions regarding the lack of evidence linking aggregate cigarette advertisements to the demand for cigarettes in Australia by reestimating McLeod's (1986) equations with the inclusion of aggregate advertising expenses as an additional variable. United States Evidence Schnabel (1972), Bass and Parsons (1969), Bishop and Yoo (1985), Lewit et al. (198t) and Schneider et al. (1981) found a statistically significant cigarette advertising effect on cigarette consumption. Schoenberg (1933) found no evidenoe suggesting that total cigarette advertising has an effect on total consumption of cigarettes. Telser (1962), Peles (1971), Baigati and Levin (1986), Hamikon (1972), Hamilton (1977) and Schmalensee (1972) also found no significant effects. A recent study by Roberts and Samuelson (1988) analysed annual advertising and sales data. Their results indicated that firms' advertising primarily affects the level of market demand, (that is total cigarette sales) with a firm's affect on rival firms' market shares arfsing through variations in the number of brands sold. Advertising by any firm may increase total market demand. Low-tar advertising has a positive and statistically significant effect on market demand in both the low-tar and high-tar markets. This suggests that advertis- ing for low-tar products, possibly because of the audience toward which it is directed, increases the demand for cigarettes in general. Advertising of high- tar brands was found not to affect total consumption significantly. Eckard (19911, studied the 1976 American cigarette advertising ban and found, first, that shares of the market held by different fim s were more stable after the advert"ising ban, net of trend effects. Second, leading-brand shares and brand and firm concentration were declining before the ban, and are stable or increasing after the ban. Third, the ban appe4s to have increased profit margins, and these margins are negatively related to advertising intensity during the study period. Fourth, new brand entry virtually ceased during the first four years after the ban, resuming only in conjunation with a substantial increase in advertising Intensity. Finally, brands introduced by leading firms have been relatively more successful after the ban. These illustrate that 2060476164
Page 6: mtd42d00
S advertising bans affect cigarette consumption In a variety of ways but did not reduce aggregate consumption. United Kinadom Evidence McGuinness and Cowling (1975) and Cowling et al. (1975) examined the relationship between the aggregate demand for cigarettes in the United Kingdom and price, income and a variable designed to measure advertising messages. They concluded that advertising has had a statistically significant effect on the expansion of sales. However, these studies were strongly criticised by Johnston (1980), akhough McGuinness and Cowling (1980) dismissed much of this because Johnston's comments were apparently solicited by the U.K. Tobacco Advisory Council. Furthermore, re-estimation of the MoGuinness and Cowling models by the Metra ConsuRirig Group (1979), using more recent data led to the conciusionn that the significant relationship found using the 1960s data between cigarette demand and advertising did not hold in the 1970s. Further analysis by Metra of this relationship, using a variety of econometric and statistical techniques, failed to turn up any statistically significant relationships between cigarette consumption and advertising. Waterson (1984) concluded that cigarette advertising bans have not reduced consumption. He showed that such bans reduced the replacement of regular cigarettes by filter or low-tar brands and as such had a negative effect on nicotine consumption in the countries such as the UK with partial advertising bans. New Zealand Evidence Chetwynd et at. (1988) studied the relationship between cigarette consumption and cigarette advertising over the period 1973 to 1985. The data on cigarette consumption was drawn from government statistics. Expenditure on cigarette advertising included only that on print media advertising. Both quarterly and annual time series were examined and two 2060476165
Page 7: mtd42d00
0 different models were estimated. Results from the quarterly analysis suggested that advertising did affect overall consumption, with an elasticity of +.07. with the carry-over effect of advertising lasted for approximately 4 quarters. The results of Chetwynd et al. were challenged by Jackson and Ekelund (1989), Boddewyn (1989) and Raftery (1989) on a number of technical grounds, some of which were specific to this study and some of which were generic to all such attempts. After further analysis along the lines suggested, they reported in Harrison et al. (1989) and Chetwynd at al. (1989) that the original model and conctusfons appear to be very robust, and the additional long run analyses indicate that the impact of . cigarette advertising on consumption may, in fact, be longer than was indicated in the original work. Meads (1990) studied real price, cigarette advertising, and an Income measure, in relation to total advertising sales, using monthly data from 1973 to 1986. Meads-conoluded that cigarette price and advertising has significant: short run and cumulative effects on cigarette consumption as the same~ percentage reduction in advertising expenditure. The Chetwynd at al. (1988) and Meads (1990) studies both showed that with a 10 per cent rise In advertising expenditure, eonsumption would rise at least 0.7 per cent. Meads' study and the latest revision of Chetwynd's study show a net effect of advertising on consumption, over and above any reverse effect of sales on advertising budgets. Harrisort..and Chetwynd (1990) further extended prior econometric research by updating the database and including an additional variable to examine the influence of the recent New Zealand anti-smoking advertising campaigns. The results support the conclusions of their earlier work, that the industry's advertising does stimulate aggregate consumption, In addition, this study demonstrated that recent anti-smoking advertising:'6mpaigns in New Zealand had been effective in reducing cigarette consumption with sales levels after the campaigns significantly lower than would otherwise have been expected. Harrison and Chetwynd (1990) concluded that the proposed New Zealand ban on cigarette advertising would not devastate sales, but could be expected to 2060476166
Page 8: mtd42d00
lECGSSU) ~ reduce them by about 30 per cent ff the industry did not respond with other promotional activities, This New Zealand study demonstrated that the industry's advertising has increased cigarette sales in New Zealand over recent years. The proposed legislation to ban all tobacco advertising and other promotions could be expected to reduce overall sales but• not to devastate them. Seldon and Doroodian (1989) had suggested that tobacco companies may increase advertising In response to anti-smoking campaigns. Harrison and Chetwynd (1990) showed this causality not to be supported and also that the reverse is not supported. In the light of all the various studies, models, specifications and analyses there remains some uncertainty whether increased cigarette advertising increases cigarette consumption or not. This study aims to test this possible relationship using regression analysis with relevant statistics for Australia for the period 1962/63 to 1989/90 which includes the 1973-1976 electronic media ban. From these results policy conclusions may then be drawn regarding the role of cigarette a'dvertising. Empirical Data Economic theory suggests that the quantity demanded of any commodity will depend upon a series of variables including the price of the commodity, the prices of related goods and consumer income. In addition, where time series data are used to estimate the demand function, time-related variables such as lagged variables are often included in the demand model specification. Since the aim is to investigate the effects of cigarette advertising on the demand for cigarettes, advertising variables are to be included in the demand function specification. It is important to note that all dollar values are in real terms with the base year 1985. Annual observations on all variables necessary to istimate specific forms3 of a demand equation were available from the financial years 1962-63 to 1989- 90, so that the total sample size was 26 when one uses lagged variables. The first economic variable which might be expected to be of some importance is the level of real disposable income of the population. The a_ 2060476167
Page 9: mtd42d00
[ECd.r9t/) 0 riori expectation was that, as for all normal goods, the income elasticity would be positive and be expected to be less than unity. The measure of income considered here is Gross Domestic Product from Foster and Stewart .(1991), Table 5.3. This income measure (GOP) was deflated by both the CPI and adult population (POPN15A0) to produce GOPRC. Many demand studies incorporate the prices of substitutes and complements In the demand funotion. With regard to cigarettes, while other tobacco products could be used as a direct substitute to meet the 'demand to smoke', as Johnson (1986) shows, observation of cigarette smokers' behaviour shows that in general there is a strong preference specifioally for manufactured cigarettes and that consumers do not keep the respective price of cigarettes and smoking tobacco under surveillance to decide which is the better 'buy'. As a result it is considered sufficient to have the price of some other goods (such as food) accounted for in that they are possible alternative forms of expenditure. For the prioe of other goods we use the price index for food (FOODPRR) frora Foster and Stewart (1991), Table 5.7 since it is considered that cigarettes and food (at least some types of foods) are substitutes. Similarly, as with Johnson (1986), the explanatory influence of the price of complements (chiefly matches and cigarette lighters) on the demand for cigarettes Is considered, on priori grounds, to be not significant to warrant separate inclusion of such a variable in the demand equation. Quantity of cigarettes demanded was measured as cigarettes produced (CIGS) in million kg in Australia per adult from the Australian Tobacco Marketing- Advisory Committee (1991), Appendix L, while the per adult deflator (POPNISAO) was obtained from Foster and Stewart (1991), Table 4.1. Data on actuat cigarette safes in Austrafia are not available but imported brands hold a very small market share as Alemson (1970) shows and cigarette exports amount to approximately i per cent of total production. Hence production would seern to be a reasonable proxy foi actual sales. The data are for the period 1962J63 to 1989190 are presented in Table 1. Consumption' of tobacco by weight is measured on a per adult basis, where adult is defined as 15 years or over. This definition is similar to that used in previous studies and is the accepted approach to defining the smoking population. This is CIGCCAP. 2060476168
Page 10: mtd42d00
[ECJ.Mq 0 The real price of cigarettes (CIGPR) was measured using the cigarettes and tobacco -subgroup index (tobacco and alcohol group component) of the Consumer Price Index provided by Foster and Stewart (1991), Table 5.7a. This price was deflated by the CPI from Foster and Stewart (1991), Table 6.23, to provide a real estimate of price (C1GRLPR). Advertising expenditures by tobacco product manufacturers in the press media and on television and metropolitan radio have been estimated by C.E.A.S.A. (1991). These expenditures have been deflated by the CPI to obtain a measure of real advertising expenditures, since a price index for media advertising is not readily available in Australia. Both aggregated total advertising expenditure (CIGADVR) and disaggregated data into electronic (CIGELRE) and print media (CIGPRRE) were used. Real advertising expenditure was expressed in both absolute and per adult terms (CIGADVRC, CIGELREC and CIGPRREC) so that, in the empirical work, four separate advertising measures were tried. Both Clarke (1976) and Grabowski (1976) showed that there is ample evidence that an advertising carry-over effect is present at the brand level but there is much less evidence that such an effect is present for industry advertising. Clarke estimated that sales effects lasted for months rather than years and showed that demand for a frequently purchased, low priced product may increase up to 9 months of it being advertised. This ailows us to include the four different advertising expenditures to be lagged at least one period and preferable by more in some cases. Another variable whose influence on the demand for cigarettes was considered important on theoretical grounds was the addictive nature of tobacco smoking. In an endeavour to capture the effect of 'habit' nature, the per capita consumption lagged one period, CIGCCAP(-1), was included as an explanatory variable on the grounds that consumption in any given period is partly determined by consumption in the preced'ng period(s). It would be expected that the habit coefficient be posttive. 2060476169

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size: