Tobacco Documents Online

~ M ~ ~ M M ~ M M no M No M M so M W aft 101011 I CIIAPTER I Trends in Cigarette Consumption and the Sociodemographic Structure of the Smoking Population in Developed Industrial Countries Information on the development of cigarette consumption and the changes in thc sociodemographic structure of smoking populations constitute a major element in the description and analysis of smoking behaviour. tiowever, the collection of reli- able, sufficiently detailed data on this subject is beset with difficulties. Not only do we find differences in survey methods and definitions between the various coum tries, but there is also a lack of agreement on these issues even within individual countries. For example, there are several definitions of the category "regular smok- , er and many different methods of determining actual cigarette consumrtion.' Thc problems increase if we attempt to identify international trends. Indeed, closer ex- amination of apparently identical facts concerning the "typical smoker" reveals that the data are in reality highly heterogeneous and that it is in many cases impossible to make international comparisons.2 After all, "the different traditions, economic conditions and sociocultural backgrounds, together with the system of social values which varies from country to country ... (make) a direct comparison of findings" ~. even more problematic (Merzdorf et al. 1982 p.9). In view of these problems - which are dealt with more fully in Chapter 1.2 - it seemed sensible to confine our international comparison of the development of cig- arette consumption to cases where the figures were compatible. In the first instance this meant restricting our study to the highly developed industrial nations because only in these countries is a meaningful comparison of economic and sociocultural factors possible. We further limited the scope of our survey to Ihe USA and coum tries within the European Community because the data available are both method- ologically reliable and form a good basis for comparison. Finally, in order to attain a differentiated breakdown of the sociodemographic features of the smoking popu- lation we decided to concentrate our attention on one country - West Germany. This was necessary as the various international differences in the definition of such ~ sociodemographic factors as class structure or educational level simply exclude I comparison. The attempt to create a single common denominator would, in this ) case, have distorted the facts. GVV®QYMg

m2 ww~ so mim I renrmarcur aftntwo ws W 1Mnds t1~ttc ( ,1Mton Mao M M am ow 1.1 Trends in Cigarette Consumption in the USA and the Member Countries of the European Community t 1 1.1.1 1)erelopnteiu of Cignrelle Cnn.rwnplinn ~ C'igareuc consumption in the United States and the countries of the European Community' has increased considerably over the past 20ycars. Nonetheless, as Fig. I indicates, the rates of increase differ enormously. The increase in cigarette consumption for West Germany, which has nearly doubled during the period in question, is much higher than in other EC countries. At the same time the rate of in- crease in these countries is also much higher than in the USA. The rate of increase in consumption, which varies greatly in the EC and ranges from 11% in Great Britain to 84% in Italy, can he attributed to two factors - the ab- solute increase in the size of the smoking population and the increase in smoking in- tensity, i.e. (he amount smoked per smoker. Between 1960 and 1979 the potential smoking population or adult population (15 years and older) of.the European Community increased by 16%. Theoretically, this could account for one quarter of the increase in cigarette consumption. How- ever, Table 1.1 clearly shows that the increase in consumption must also be attrib- uted to the increased intensity of smoking. In addition, this table indicates why the rates of increase vary to such an extent from one country to the next. Countries which had a relatively low per capita consumption in 1960 - such as Italy an(I West Germany - subsequently experienced a high rate of increase, where- as the increase in countries which already had a high per capita consumption in 196(1(e. g. Great Britain) was lowest. The same explanation can be applied at least partially to the slight increase in consumption indicated for the USA by the data in Fig. l.l. In 1960 cigarette consumption in the USA among the potential smoking population (in this case everyone aged 18 years and older) was 11.4 cigarettes per capita/per day. FRGm +71% - ~ ~ / ~ \V~ . ~ ........................... ...: / EC - + 49% USA - + 24% FiR.1.1. Changes in cigarette consumption (ntanufaqured cig.) in West Germany, the Europcan Community and the USA from 1960 to 1979/1983/1984 (19fi0-100)- (Slatistisches 13undesamt 1985; Mcra(lorf ct al. 1982; Rccent Advanccs in Tobacco Rescarch 1984) t Table L1. 1)aily per capita consumption of cigarclics anwng tile po- tential sntoking poprdauion of EC countrict' (19(d)- I979). (IS1crzdorf c(aL 1982) Number of cigarencs 1960 1970 1975 1979 West Germany 4.9 7.4 7.7 7.8 France 5.3 6.0 6.5 6.3 Italy 4.5 5.7 6.9 7.2 Thc Netherlands 7.3 8.1 9.0 9.6 13clgium/Luxcmhourg 7.1 8.1 8.6 7.9 Great Britain 8.2 8.9 9.2 t,4 Ireland 7.3 8.1 9.7 9.3 I)cnmark 5.5 6.11 6-9 6.7 L:C total 5.9 7.1 7.6 7.6 ' Unlcs.c stated otherwise, the txricntial smoking Ixipulation (i.e. adult twtiulation) refers to the total population aged I5ycars and older While we can, perhaps, assume that the sharp increase in consunttxion which oc- curred in several countries was due to tile existence of some kind of vacuunt, it can- not be concluded that there is a "saturation level" beyond which consumhtion can- not rise. Although it does appear that 1973 to 1975 saw the end of a period during which cigarette consumption in the EC countries increased (consumption subse- quently stagnated or actually declined), the general trend in consuntrtion during the seventies was marked to such an extent by sudden increases, declines, Iluctua- tions and contradictory developments that it is impossible to say whether in the long term we should expect a renewed increase in cigarette consumption, or whether per capita consumption will remain stable or even decline. Whereas the annual fluctuations in consumption in the EC countries can be traced back to short-lived reactions to increased tobacco duties or retail hrice,;° Fig. 1.2 (see p.4) reveals that the developments in per capita consumption in the USA are closely linked to the publication of statements about the health risks posed by smoking. Publication of the first Surgeon General's Report in 1964 had similar effects to those "following the release of the first study linking cigarette smoking and lung cancer" (Grise 1983, p.20) in 1954. These reports, and the numcrous publi- cations on the risks of smoking to health which followed them in tile late sixties, heralded a decline in per capita cigarette consumption.5 As a result of this trend per capita consumption in the USA in 1983 was 20% below that of 1963, the "record year" (Recent Advances in Tobacco Research 1984, p. 11). For the reasons which we have outlined above it is not possible to say whether a similar development - at a cor- respondingly laterstage - will also occurin the countries of tile European Community. 1.1.2 Changec in tile Size o% Ihe Smoking Poptdnlion It is self-evident that findings on changes in cigarette consumption can only be meaningfully interpreted if changes in the potential smoking population are also taken into account. Furthermore, in order to gain an insight into changes in smok- i' OY "J0g' vJVVtt!

a'" "~ S- - '-" ~ .llcncl. Ili (ar.ti~• (iin,umpw11' ` ~~rend. ~~~ ~ ig:trcuc ~ptiot.~ ~ ~ ~ ~ S ~ 60 50-i 40-i 30- 20-I t0- t ..... 4 ? _ 3 70 - E:1 58 20 5] 22 f~rtu~le Snx~kr~5 E] Mnle s-tic- 29 29 9 1960 1905 1970 1975 1979 3800 -1 1963 '65 '67 '69 '71 '73 '75 '77 FiK.1.2. ('hanges in annual consumption of cigarettes per person (agcd 18 years or oldcr)in thc t1SA hctwecn I%3 and 1977. 1. based on Ihlx. of Agriculture total USA consumption sericc; 2. based on Federal Trade Commission domestic consumption scrics; 3. based on 1)ept. of Agricul- turc domestic consumption scries; 4, based on Maxwcll Report's domestic consumption series (Surgeon Gcncral's Report 1979 Appendix) '1'abte 1.2. Number of smokers as a percentage of the adult population in 13C member countries, 1960-1979. (Mcr~.dorf ct :d.1982) 1960 1970 1979 AII smokers 43.8% 417•/. 39.2% Male smokers 62.0% 58.0% 49.0% Female smokers 27.0% 29.0% 30.0% ing behaviour it is necessary not only to examine the potential smoking population but also thc proportion of the population who actually smoke. As can be seen from Table 1.2, during the same period when cigarette consumption in the EC countries rose by 49% (i.e. 19(>n-1979) the proportion of smokers within the potential smok- ing population fell from just under 44% to a little over 39%. The two main features of this development can be seen in Fig. 1.3, which illus- trates trends in West Germany. There is a clear decline in the proportion of male smokers, and this has been accompanied by a continual, although less substantial, incrcase in the proportion of women smokers. This phenomenon is not peculiar to the EC countries but also typifies trends in the USA.' The possible reasons for the different trends in the smoking behaviour of men and women will be discussed at a later stage. At this point attention should be given to the fact that during the period in which total cigarette consumption increased, the ~ number of smokers - as a proportion of the total population - actually declined. This confirms the suspicion that an increase in the intensity of smoking, i.e. per ca- pita consumption, was responsible for the increase in overall cigarette consump- 0 Fig. 1.3. Number of smokcrs as a percentage of the Ixitcntial smoking popwlation in West Germany from I1X,(1 to 1979 (Mcadorf ct al. 1982) Table 1.3. 1):uly cigarette consumption per smoker in LC countries from 1960 to 1979 (Merzdorf ct al. 1982) Number of cigarcttcs I4X,II 1965 1970 1975 1979 West Germany 11.3 14.6 17.7 19.6 21.0 France 11.2 11.5 12.2 14.0 16-5 Italy 7.4 - - - 17.2 lhc Netherlands 11.9 11.8 14.2 18.5 21.3 ftelgium/Luxemhourg 18.1 - 20.2 22.7 21.3 Great Britain 15.4 15.9 18.2 20.8 19.8 Ireland - - 20.4 24.6 26.5 I)cnmark 15.6 14.9 14.2 17.6 18.2 GC total 13.5 - 16.3 - 19.4 ~ tion. Table 1.3 demonstrates this increase in cigarette consumption per smokcn ce- 1 twccn 1900 and 1979. At the same time as per capita cigarette consuntnlion antong the potential smok- ing (i.e. adult) population increased in the countries of the European Cautntwtity by almost 29%, it is apparent from Table 1.3 that the daily consumption of smokers rose by approximately 44°/m.7 The trend towards a continually decreasing (trohor- tion of smokers who smoke an increasing number of cigarettes per head is also evi- dent in the USA. Table 1.4 illustrates that the proportion of smokers who smoke t~re than 25 cigarettes a day grew steadily between 1965 and 1976. 1.1.3 Ti-endc in Con.runtprion n•idr Regard lo rlte Mo.cr Firroured Ti(re.r q/'Cigarelte Tltroughout this present discussion we have simply referred to smokers without dis- tinguishing between cigarette smokers and those who consume tobacco in any other fonn. The fact that it seems unnecessary to make such a distinction shows that ciga- ~oVooVM9

in Cige cttcZ:onsumpl i 1 ~ Trends in C igarcttc Consumption M l'able 1.4. Pcrccntagc of regular smokers in thc USA who smoke more than 25 ciga- 100-, rcltes daily (1965-1976). (Surgeon Gcncr,d's Rcport 1979 Appendix) 90-I 1%5 1970 1974 1976' 80 -~ All smokcrs 19.9"/. 23.3% 24.7% 25.3% Male smokcrs 24.5% 27.6% 311.3'Y. 30.8% 70-I Female stnokcrs 133% 18.1% 18.4% 19.4% 60 -~ ° Smoking population aged 20years and ovcr, all other figures based on smoking populnlion aged 17 years and over 50 1.0 Table 1.5. Breakdown of lhe smoking population in EC countries according 30-I to forms of smpking (1979)- (Mcrzdorf et al. 1982) 20 -1 Pipc Cigars/ Cigarillos Cigarettes West Germany 11.4% 9.1% 79.5% France 8.0% 16.0% 76.0Y. Italy - - 100.0% lhc Netherlands 5.9% 13.7'Y. 80.4% Iklgium/Luxenthourg - - - Great Britain 15.2% 19.7% . 65.2% Ireland 7.5% 2.5X 90.0% Denmark 17.1% 10'/.' 20% 52.9"/, " 111A'Y. cig:vs/cigarillos 20.0% chcroots rette smoking has become the dominant form of tobacco consumption. This is rein- forced by the figures for EC countries shown in Table 1.5. As we can see in the case of Denmark, tradition continues to be a major factor in maintaining specific forms of tobacco consumption. Nonetheless, it is quite clear that the trend in the countries of the European Community is similar to that in the USA, where "cigarettes account for about 85% of the tobacco used" (Grise 1983, 1,p.20). Similarities between Europe and the USA can also be seen with regard to the preference for certain types of cigarette. Filter cigarettes, in particular, have rapidly gained in popularity. This trend is demonstrated in Fig. 1.4 - between 1950 and 1980 the market share of filter cigarcttes in the USA rose from 0.5% to 92.5%. In 1979 filter cigarettes accounted for between 61% (in France) and 92% (in Great Britain) of the market in the European Community (Merzdorf et al. 1982, p.63, Table 2.5.2). Filter cigarettes increased in popularity and achieved a market share of approximately 60% even in countries where non-tipped cigarettes were traditionally favoured. Although this is far below lhe level reached in the USA it still represents a considerable increase in sales between 1960 and 1979. A second major trend in (he type of cigarette smoked can be seen in the increas- ing popularity of the so-called "low tar" cigarette, which yields a lower level of par- Iiculate matter. As shown in Fig. 1.5, these "milder" cigarettes have enjoyed an in- creasingly greater share of the market in Great Britain since the early seventies. 10 0 12 3 75 995 813 [9 1 356 199 05 187 509 64 L 80 1 877 92 5 1950 1955 1960 1965 1970 1975 1980 El Qgarelles wdhoul hlter a Fdter ugaretles Fig. 1.4. Filter tip share of total manufacturcd cigarettes in the USA, 195(1- 198(1(Grise 1983) Table 1.6. Filter lip share of total manufactured cigarettes in thc EC countrics. 1960-1979. (MeriAorf ct al. 1982) 1960 1979 I ltcrcasc Wcst Germany 68%. 88% +20'Y. France 12% 61% +49",•. Italy 13% 89% +76% The Netherlands 12% 62% +50".4 ftclgium/Luxenltwurg 30"/0 8[l"G +50"/. Great Britain 16% 92% +76"/. Ireland 10% 87% +779L I)cnnlark 29% 62% +33% A similar trend can be observed in the USA. In 1967 only 2% of manufactured cigarettes yielded 15 mg (or less) tar content. By 1933 this applied to some 65% of all manufactured cigarettes. Low tar cigarettes are particularly popular in West Ger- many, where they account for 88% of sales, however in Italy their market share is only 20% (Advertising and Cigarette Consumption 1983, p.8). Research has thus far been unable to explain these considerable national differences. Similarly, there ts still no answer to the question whether the reduction in tar and nicotine levels is re- sponsible for the increase in consumption per smoker. Although it seems an obvi- ous conclusion to draw, research has not yet established whether the reduced nico- tine content has led to an increase in the number of cigarettes smoked per head. "Studies of changes in cigarette consumption among those who voluntarily switched to lower tar and nicotine cigarettes have yielded equivocal results, with 90VOaV9V99

~ M M M M M M M M 8 l rcnds in ('ig.rrcttc C onsumpluun 55- 45- 4 15-i 10 14 10-1 0 t...___r_......r._.__ 71 22 19 4 28 r.9 53 1972 '73 '74 '75 '76 '77 '78 "79 80 81 '82 ,'83 Fig. 1.5. MarkG sh:ue of low tar' cigarettes in Grcal Britain, 1972-1983. (Advertising and Cigarcue ('onsumption 1983) " Not cxcccding 16 mg tar contcnl sonic smokers reporting increased consumption, many smokers reporting no change, and still others reporting a decrease" (Surgeon General's Report 1979, Ap- ndix, p.A-20). L 1.1.4 Srnnrnarv TI1e trends in cigarette consumption in the USA and in the member countries of the European Community exhibit a number of similarities: - Since I960 cigarette consumption has increased (in absolute terms) in every country. - At the same time the number of smokers (as a proportion of the potential smok- ing population) has declined - despite an increase in the number of female smok- crs. - The daily consumption of cigarettes per smoker is considerably greater than it was in 1960. - These changes in cigarette consumption coincided with a rapid increase in the popularity of filter and low tar cigarettes. Although these developments tend to confirm a certain overall trend, it would be incorrect to make sweeping generalizations or to apply these findings universally. The trends outlined here for the European Community countries are by no means identical. National differences do exist which should at least modify any general conclusions that might be made. There are also a number of other factors - such as :tn increase in tobacco dutv or retail nricc. as well as the role nlaved hv the "anti- ~ M as M M IM M so ~ WE A('ritical Look au Available Data on Ilte Sociodcmographic Slnicturc 9 snloking lobby" - which account for national differences. Finally, there are clear ill- dications that since the nlid-seventies the increase in consunlption per smoker - one of the ntost significant trends of thc last few decades - may actually be levelling off or even declining in sonic countries. ~ These reservations alone make it clear that tile postulation of international trends on the basis of evidence collected in various countries should be viewed cautiously - even where, as in this case, the countries considered have a similar socioccononlic and cultural character. International comparisons are valid, but national trends should not be underestimated, particularly when the intention is to present an ex- ` r,lanation of behaviour rather than a descriptive study. 1.2 A Critical Look at Available Data on the Sociodemographic Structure of the Smoking Population in West Germany The difficulties associated with making international comparisons of trends in ciga- rette consumption have already been discussed. Similar Ixoblenls are posed by (lie comparative analysis of data collected within individual countries. Our present con- cern is to illustrate why such data cannot be used uncritically. This is rerhaps best achieved by taking West Germany as an example in order to point out sonic of Ihe typical problems associated with the evaluation of availablc d:ua. The analyses of the West German tobacco product market that we have consid- ered are based on several rcpresentalive surveys: however, a conlliarison reveals that these findings are often highly inconsistent. Table 1.7 lists sonic of the findings from surveys conducted during 19R4 and 1985. It is noteworthy that even tile total number of smokers recorded differs by nlore than onc and a half nlillion. I)iscrepancics of this nature can probably be attributed to the different nlethods used to extrapolate from of licial government statistics. The various studies thus assume a substantially different population base. Inconsistencies are also to be found in the breakdown of the snloker population (in terms of percentages). Figure 1.6 (see p. 10) illustrates (he variations in the find- ings of Burda and the Allensbacher Werbetriigeranalyse (AWA) with respect to tlte age distribution of nlale and female snlokers." We can only speculate on the reasons for such dilTcrences. The procedure of ran- dom sampling may, in sonic cases, lead to an overrepresentation of certain scclions ldble 1.7. I)iffercnces hctween the findings of four rcpresentaticc sun•eys of srnokcrs, 1984/ 19ft5 Source S:tnlplc Pnpulation base Cigarctle Malcs Fent:llcs in millions smokers (14 years and older) Burda 1984 8(147 4g.22 AWA 1984 g(ln4 49.00 AWA 1985 9918 49.(10 VA 1985 10767 48.22 Y4 ntillions ,6 millions millions 32 15.24 41 9.28 23 5.96 34 16.65 44) 9.17 26 7.48 34 16.73 41 9.38 28 7.35 35 16.94 42 9.511 29 7.44 LoPOoVsvog

~ M M I(1 60 50- c0-I 30_~ 20- 10-I 0- M no --- fernole srnu*ers - Mole smokers \ ~ \\ ~\ \ \ \ \ \ -t------ _T---.~--r-_T----- f----F--- 16-19 20-29 30-39 40-49 50-59 60-69 70 orolder FiR.1.6. Age distrihutnon of male and fcmaie smokers according to A WA (1984) and l3urda (1984) 1 Male smokers according to Burda 2 Male smokers according to AWA 3 Female smokers according to AWA 4 Fcmale smokcrs according to Ilurda of the population. Whercas Allensbach (AWA 1984, 1985) employed a random sample of individuals selected by a quota procedure,' the 13urda sample was based on addresses. Ttrc definition of the various categories is also probably an important factor. VA defines an average daily consumption of 1-10 cigarettes as "low", 11 and above as "medium", and 21 or more as "high". Burda adopts a scale of 1-9, 10-19, 20-29 and 30 or more per day, while AWA uses a scale of less than 10, 10-20 and more than 20 cigarettes daily. As 10 or 20 cigarettes appear to represent sonie kind of sub- jective barrier or "milestone" when smokers assess their own smoking intensity,10 it is by no means irrelevant whether a person who claims to smoke 10 cigarettes a day is classified as a light or medium smoker. The very fact that the various studies use different smoker categories emphasizes the point that a comparison of several studies can be highly problematic. There are also differences in the household income and professional status categories used to determine sociodemographic classification.rr In fact, even the number of categories employed differs. For example, the Allensbach study which we examined did not contain information on academic achievement or personal income. The absence of the category "personal income" brings its to another problem which is illustrated by other studies that will be considered at a later stage. Virtually all of them indicate major differences in the smoking behaviour of men and wom- Trcnds in (-ig.rrcttc Consumplion ~ M ow M M W M ~ M so The Sociodcmogr:rphic 'mructurc or ruc ~mokinr . .•r..Jation ... .. - .1 Gcrn , I I en.l-' Consequently, it is important to allow for differences between the seXcs whcn analysing the composition of the smoker population. Unfortunately, in those sur- veys with questions based on bou.seholrd income it is difficult to distinguish between lhe man's and the woman's respective incomes unless there is a question about per- .conal income. This weakness is aggravated by lhe fact that in order to establish pro- fession or occupation the subject is usually asked about the profession of the head of the household. As this is nearly always the husband, the only means of determin- ing the sociodernographic status of the woman smoker is according to educational stalus. 1)uc to the problems posed by comparing a variety of studies we declded to use just one representative survey in order to determinc the sociodemogralrhic structurc of the smoker population. After careful consideration we chose the study commis- sioned by Burda called Tipologie der 1Viiir.cche, volume 6 (tobacco products), which was published in 1984. Like the AWA 1984 and 1985 reports this study contains all the latest data, but it is preferable for our purposes because it includes a greater numbcr of categories which are further broken down, thus pennitting a better in- sight into the differences between the sexes. The tables and figures which follow arc therefore based on data presented by Burda. Undoubtedly the customary Ixoce- clure of rounding off figures will have Icd to certain inaccuracies. As a result - and especially in view of the othei problems involved by such surveys - none of the findings can be regarded as definitive, but they are accurate enough at least to point to a few general trends or conclusions. 1.3 The Sociodemographic Structure of the Smoking Population of West Germany 1.3.1 A verage Catsunrplinn According to the Burda survey findings, 32% of all West Germans aged ovcr 14 years are smokers, 61% of the smoking population are men and 39% are women. Table 1.8 indicates the average daily cigarette consumption. Most women are classl- lied as light or medium smokers and only a minority smoke more than 30cigarettes a day.t' Approximately half of all male smokers belong to the category "heavy smoker". Table 1.8. Average daily consumption of male and female smokers Number of cigarettes All smokers Male smokers Female smokcrs 1-9 19% 15 % 24 /. 10-19 38% 36% 39%. 2(1-29 33%. 37% 26% 30 and more 9% 1 I Y. 6% Itx1H.' 1aM/u' 100% a Total does not always equal IOOY% as ligures arc rounded off 80V00M09

.r .. ..r WMrrr rr r r r.rr +r s• rMMMM im 12 Trcnds in (ig:ucllc (bnsuropuun Iltc Suaoocnlograpi".....-aclurc... ...- "nwkir_ ..Ialion I(icri 13 Nnn- Srrx*Ms 1-9 cg 10-19 c9 20-29 ug 30 ug or more Women Men - 60 . . . 40 52 48 58 H 68 20 72 1 Fig. 1.7. Breakdown of smokcrs ncl n n-sm k li crs accor( a o o n f; to sex 10 20 30 40 50 60 70 80 90 100 and daily consumption Forty-eighl percent of male smokers smoke more than 20cigarettes daily. Simi- larly, as shown in Fig. 1.7, thc proportion of women within each category (i.e. the same level of consumption) decreases as the numbcr of cigarettes consumed rises. Not only are there more male smokers than women smokers, but men also slnoke morc. l.3_2 Age Uislribulion Howevcr, if we look at the age distribution of smokers as depicted in Fig.1.8 it be- comes clear that the differences in smoking behaviour and consumption are less pronounced among the younger generations. Although there are fewer female smokers than male smokers, the discrepancy is much smaller than in older age groups. The curves in the upper part of Fig. 1.8 - which are plotted according to the age distribution of the entire population - serve to emphasize the size of the gap be- tween the proportion of older male smokers and their female counterparts. Whereas the distribution of older male smokers roughly corresponds with their representation within the population as a whole, in the age groups where women are predominant there is, in fact, a disproportionately small percentage of smokers. Table 1.9 reinforces this point. In the age group 14-19years there is virtually the same proportion of male and female smokers. The age group 20-29 year olds con- tains the highest proportion of smokers in both sexes. The proportion of male smokers declines steadily from one age group to the next, whereas the proportion of smokers drops dramatically among women aged 40 years and older- As we know from other research Gndings that women generally find it more diffi- cult to cease smoking than men (see Siegrist 1981, p. 16), we cannot attribute the high proportion of older non-smokers among women to (he presence of a large 18 16 14- 12-I 10-i B-i 6-I 4- 2-I 0- 22 20 i Yo 20 14-19 20-29 30-39 40-r.9 SO-59 60 69 70«olc}er Fig.1.8. Age distrihutiun of lotal population and snwkinF population (urdcs and fcm:dcs). la. lh. age distribution of lotal pupul.uion; 1rL 1h, agc distrihnuon of smoking popuLdion (- m:des: - - - fcm:des) Tabte 1.9. Male and femute sm+ikers as a perccntaFc of their age groups All snlokera Malc srnokers Fcm:dc simikers 14 - 19 ycars 27.1": 29.IX 29 r)"'^ 211 29 ycars 4X.4"a S 1.R"/« 42 .Y".: 30 - 39 years 41.3% 46.4% M..l° h •71/ 49ycars 35.1% 45.7% 24.5% 5(1-59ycars 31.R".: 47. IR.7".u 60 69 years 18.7°r: 10.2% 70 years or older 11.3N, 21.5 ; S.-PP6 number of ex-snlokers. It is probable that among the older generations fewer wonr en adopted the habit. The fact that there are almost as many younger femalc smok- ers as men indicates that smoking behaviour not only differs between tht sexes hut also between different generations. 1.3.3 Educational .Slalus "The distribution of smokers according to educational qualifications (Fig. 1.9) is sini- ilar to that of the population as a whole.t{ Tllis applies - with certain rescrvations - to both nlale and female smokers. The percentage of smokers who attended elenlen- tary school and have no subsequent training or allrrenticeship is considerably high- er among females than males. This is partly accounted for by the larger number of women who leave elementary school and do not receive any training.t` The correla- V0Y'JOY a7Y aJ(!

14 100 ~ e0- ro 6o-i 50-I 40 30-I 20 -{ 10-I 0 % ® Au smokers Mote smokers M M M M M ~ M M ~ M M M Trcnds in ( igarcllc ( on.umplion Tltc Sociodcmographic Slructurc of Ihc Smoking Population ot wc>t Gcrmany I S Fernale smolrers Fkrcientary school ~ ~a~ry school witrwut sc.laol-leavng cerhircate Elementary school with [] School-Ieavag cerlihcate vocatrorxil training college. unversrly Fig.1.9. Ilrcakdown of sntoker population according to educational slalus 55- 50- 46 45-{ 41 40- 35- 30 - . 25-I 20-I 16 15-I 10-I 5-I 0 24 [ 1 h7ate smakers Female smokers 35 26 Elernenlary school Elernentary school Secorxkxy school wilh vocat.onal withord school- tra-ng leaving cerhhcale School leavrng certdscote,college, uruversdy Fig. 1.10. Male and fcmalc smokcrs as a proixtrtion of trxal population accorciing to cducation:d clalus I lion between smoking and education is shown more precisely in Fig. 1.1(1, which in- dicates ihe percentage of male and female smokers according to educational status. It is clear fronl thcse data that women with a higher level of education are ntore like- ly to smoke Ihan less-educated women. Among men the converse is true; men with fewer educational qualificartions are nlore likely to smoke. An examination of daily cigarette consumption according to educational catego- ries docs not reveal any significant differences between nun and women with thc same educational status. Figures 1.1 1 a and 1.11 6 confirm (lnce morc that daily cig- arette consumption is higher antong men than women. 1-9 c:.g laay 10-19 ug ioay 20cg•more/doy Flernentary school 22 39 withrwt vocatronal traYUng ElennxNntary school 25 GL : w,lh vocatiorx7l ? trairung SecorWa'y school 25 G3 ~ ~` ` 32 w tho~ul school-leavn}g cerlrlrcote --- Secorviary school wrth 32 32 sr.had-leavuig cerlilrcate• 1 cdMge. urwversdy Fig. 1.11 a. Avcragc consumption of femalc smokers according Io cclucalionai status 10 20 30 40 50 60 70 80 90 100 Elemer,tary school tt 33 without vaC:atwrxll traUtMlg Elernentary school 12 37 4 51 1~ with vacatrooul 'k Iravwx3 ~~ Secondary school 19 -39 [2 wiltxwl school-leov~rg cerhfrcale Secondary school w~th 19 32 IMM ~9 school-6eovrng cert~ticote, college. rnwersrty Fi1t.1.11 h. Average consuntption of ntalc smokers according to educational sl:uus OY1,00191`99

W M M M M M M M M M 16 I rcnd, in ('igarcllc ('onsunq-iIion 1.3.4 Ocr'upaliun The qucstion whether the snloker population (lil7ers in sociodenlographic composi- tion from the rest of the population must also be examined in the light of informa- tion on occupational status. Figure 1.12, which depicts the proportion of trainces/ apprentices, employed persons, uneulhloyed and retired persons, etc. within the smoker population, shows that smokers do not constitute a specific sociodcnlo- graphic group and are identical in these terms to the rest of the population. The one excehtion - the relativcly low number of retired persons in Ihe smoker population - can he interpreted as a result of the age distribution of smokers and does not indi- cate a"specif ic smoker characteristic". The large number of women withoul a stat- ed occupation - some 31°/n of female smokers - compared with only 2% of male smokers undoubtedly reflects the fact that houscwives still represent the largest group within this category. Another feature is Ihat employed women smokers tend to smoke more than wom- en who are not cmhloycd. Whereas 14n/~ of thc total female population are enlploy- ed and 39% arc not employed, the proportion of female smokers who are employed (50%) is higher than those without employment (31°/m)-16 The data available to us did not permit a more detailed analysis of the smoker hopulation according to specific occupation or.profession. As previously nlen- tionecl, inquiries ahout employment were expres.sed as a question about the occupa- tion or profession of the head of the household. This has meant that data which could, perhaps, have provided a better insight into the differences between male and female smokers was not collected. It would, for example, have been interesting MOK/1 All srllOkers Mole srnolrers Femole smokers ApprentKes.troinees a Unemployetl persons ~ FmPloyerl pe, snns a Unerrydoyed persons relned perscuis Fij!. 1. 17- Itrcakdnwn of snmkcr nnnulntiun accnrdine In ixrun: tinn ~ ~ M M ~ M M M go Ihe Suclortcnlograpm" .xructur, ... ..... Snlok..., ., ulatiu. sl Gcl - 17 to discover whether the divergent trends in smoking behaviour among nlen and women with the same educational status is also reflected in men and womcn wilh elevated professional status. In short, whether women with inlportant jobs arc more likely to smoke than men in similar positions who tend to snlokc less. The distribu- tion of smokers according to various occupational categories confirnls our original contention that the composition of the smoker population is virtually identical to that of the population as a whole. The differenccs are so minor that they cannot be regarded as typically characteristic of the smoking population." 1.3.5 Inconre Figure 1.I:S shows the proportion of male and fenlale smokers lvithin the total hop- ul:uion according to household income categories. Figure 1.14 shows a hreakelown of the smoker population according to personal income. Both Ggures reveal certain clifferenees between male and female smokers. Especially in the case of householet income, which is regarded as an indication of living standard, it is eviclent that thc proportion of women who snloke is higher in the urrer income categories than in the lower income categories. Among men the opposite trend can be observed. At present we do not have suf ficient data to explain why this fineling is not repeated in the hersonal income categories. 13.6 Smninan, Statistical findings on trends in cigarette consumption, the composition of the smoker population, the differences between men and wonlen smokers - or indeed the differences hetween smokers and non-smokers - provide little nlore than a gen- 55 50-I 45 40 35 30- 25- 20 15 10-a, 5- © Noles Q ferrKNes 41 6 42 21 6 !. 1 25 less Oan 1500 1500-2000 2000-2500 <.>00-3000 4 38 28 3O00-4000 r.(><tt7 a~d nx.e F'ig.1.13. Male and fcnlale snlokers as a proportion of the total pclpulation according lo nlonlhly household incomc (in (icrnlan marks) TTV00V91i09

M MMM M~ M MMM MMMM ~MMMMW , 18 Trcnd, in ( igarcnc ( onsumptiun Note. 55 -{ 5ll ll, 45 40-1 35 .10 25- 20- 15 10 5 ° Moles Fenxde5 I ( 45 23 43 24 48 !.1 23 lesslhon 1000 t000-1500 1500-2000 200p orui more h'ig.1.14. Male and female smokers as a proportion of the lotal population according to monthly personal income (in Gcrntan marks) eral description of the respective groups. On their own these ffndings explain noth- ing, but they do rcveal in which arcas an explanation is required. On the basis of the data discussed above it can be said that there are few signifi- cant sociodemographic differences between smokers and the population as a whole, i. e. between smokers and non-smokers. But this impression arises only if we consider the smoker population in its entirety. A comparison of male and female smokers classified according to the proportion of male and female smokers within each defined sociodemographic category reveals more about the character of the smoking population. Two aspects are worth emphasizing. Although there are still more male smokers in proportion to women smokers, the gap narrows among the younger generation. Almost as many young women as men now take up the habit. Secondly, more attention should be devoted to the differences between ntale and fe- male smokers in terms of individual cigarette consumption and sociodemographic characteristics, particularly as most data has hitherto been collected on the basis of (he total smoker population. Notes I hc division into dilTcrcnt categories (rcgular smokcrs, light and heavy smokers, etc.) is based on different working definitions and could in theory be standardized. The problems involved in try- ing to establish the actual nunlbcrs of cig;rrcucs smokcd are still nlorc conlplicalcd. llrc only prc- cise figures available are for m.r -1+nid factory cig;rrcltes sold per year. This is, however, not identi- cal with actual cigarette consumption, since this also includes hand-rolled cig;rrenes and so-called "hrxder sales'. Tilc statistics for these last two groups arc estimated figures. 'lhe following figure (cigarette consumption in (he Federal Republic of Germany) shows how a drop in the consumption of tax-paid factory cigarettes due to an increase in tobacco lax is compensated for by hand-rolled cigarettes anc("bordersales°. ,9 lhc figures for total cigarette consumption should bc approachcd with caution Tllc nuthod~ used in calculating the actual cig;rrcttc consuntption V:ny considcrablv. Figurc 1.2, which .vill he rliscusscd lalcr, is a good cxamplc. Thc four different studies show basically sinlilar curvcs: how- cvcr, the ligures for cigarette consunl)rtion vary quite considcrahly. % C: 2 6, 5-i 4-i ( [ iolol c.goretle consurrpt,cx, Tox po+d morxAoctureo aqorelles I I 3 29 23 21 20 1 0 B 2 t 7 I ° .._-- T -1" ~0 ~ ~~~~~'~L8 ...-~ -- --- r--_T` 1 . _. _ . -1 2 -08 1` {__--_u I f -1 9 -1 4 y' Y -10 1 22 1972 1973 1971. 1975 1976 1977 1978 1979 Fig. i. Fluctuations in Cigarette Cottsulnptlon 1972-1979: proportional dccrcase/incrcasc in manufactured cigarettes and in totad Cigarette Consumption as comparcd with prcrious year (I)TZ-Dokunlcntation 198(1) Meadorf et al. (1982) in their analysis of cigarette consumption in the countries of tlte I:uro1lcan ('onlmunity provide an illuminating breakdown of the methodological problems which make in- tenlalional comparisons so difficult. Sec here particularly p.Rff. and p.318ff. The statistical data on trends in cigarette consumption in countries of the Guropcan Community arc taken from a survey provided by Mertdorf at al. (1982), who have reviewed all the available information and related them to comparable data. Since it refers to the period 19fif)-1979, Greece has not yet been taken into considcration, as it was not a nlcmhcr of the European Community at that tinlc. When short-tcnn flucluations are ixing analyscd, another problem has to bc taken into accomlt which makes the calculation of actual cigarette consumption difficult. ilnnounccnlcnls of an in- crease in price or tobacco tax lead to hoard purchasing before Ihc hcginning of a new year, tvhcn the increases take effect. C'onsequently, cigarette turnorer and cigarette roavcmption are by no means idcntical in a particular year when such increases havc bccn annorulcccl. In this case, too, statistics are based on estimations only. In footnote I we have already indicated the reasons why the four studies shown in Fig. 1.2 arrive at different results for the per capita consunlplion. The Surgeon (iener:d's Report (1979) conl- rnents as follows on the different methods of calculating total cigarette consumption: Series "I- reported by the 1)epartntent of Agriculture is based upon feclcral taxable rentovals, plus domestic tax-cxcmpt deliveries, plus shipments to U. S- overscae forces, plus intlxtrts. 13c- cause the federal excise tax is applied to cigarettes transferred from manufacturers' factories to regional warehouses where they await distribution to wholesalers, these data may differ frnnl ac- lual cigarette consumption. Since 1970, the 1)epartment of Agriculture has adjusted this scries for estimated changes in warchoucc inventory. Series "2" (...) represents total per capita consumplion reported by the Fedcral -Radc Conl- ntission, based upon reports of cigarcttc sales filed by individual nlanufacturers pursuant to the Public tiealth ('igarcttc Smoking Act. Series "3" represents domestic per capita consumption, calculated from I)epartnlnu of Agri- cullure data, in which shipments to U.S. ovcrscas forccs are excluded from total consunlption, and in which overseas forces arc excluded from the population base. Finally, Series "4" is calcuhucd from total domestic consumption, gross of inventory adjust- mcnt, as published in various Maxwell Rcports:' (Appendix, p. A-6) 4 c 9r-VOotM9