Addictive Behaviors 45 (2015) 294–300

Contents lists available at ScienceDirect

Addictive Behaviors

Women and smoking — Prices and health warning messages: Evidence from Spain Ana Isabel Gil-Lacruz a,⁎, Marta Gil-Lacruz b, Stephen Leeder c a b c

Department of Business Management, University of Zaragoza, Zaragoza, Spain Department of Psychology and Sociology, University of Zaragoza, Zaragoza, Spain Public Health, School of Public Health, Menzies Centre for Health Policy, University of Sydney, Sydney, Australia

H I G H L I G H T S • Prices and pictorial health warning labels are effective anti-tobacco policies • Smoking determinants vary across generation cohorts • Women from younger cohorts are more likely to smoke, even highly educated

a r t i c l e

i n f o

Available online 25 February 2015 Keywords: Cigarette Prices Pictorial images Women Generations

a b s t r a c t Objective: In Spain, fewer men are smoking every year yet the number of women smokers remains relatively high. This paper examines the impact of two anti-smoking policies (increased prices and obligatory pictorial health warning labels) on womens smoking decisions; generation cohorts are used to elucidate the determinants of those decisions. Data source: We have drawn 48,755 observations of women living in Spain from the Spanish National Health Surveys of 2001, 2003, 2006 and 2011. Data synthesis: Among the main results, we highlight that belonging to a particular generation modulates the manner in which individual characteristics and tobacco policies determine smoking decisions. For example, women's smoking was not considered as socially acceptable until the 1960s and therefore older women have lower smoking rates. However, for the younger female cohorts (generations X and Y) smoking was seen as an act of rebellion and modernity, so women belonging to these groups, irrespective of educational level, are more likely to smoke. Conclusions: The price of cigarettes and pictorial health warning labels on cigarette packets also influence the smoking behaviour of Spanish women. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction In Spain, smoking rates are currently falling among men but they are still increasing among women: the gender gap is narrowing (Franco, Perez-Hoyos, & Plaza, 2002; Jiménez Rodrigo, 2010). In many highincome countries (e.g. Australia, Canada, the United States of America and most countries of Western Europe), women smoke at nearly the same rate as men. In many low and middle-income countries, women smoke much less than men, for example, in China, 61% of men are smokers, compared with only 4% of women (WHO, 2008). Even if

⁎ Corresponding author at: Universidad de Zaragoza, Esc. Univ. Ingeniería y Arquitectura, C/ María de Luna, 50018 Zaragoza, Spain. Tel.: +34 976761000; fax: +34 976842189. E-mail address: [email protected] (A.I. Gil-Lacruz).

http://dx.doi.org/10.1016/j.addbeh.2015.01.016 0306-4603/© 2015 Elsevier Ltd. All rights reserved.

the feminisation of smoking is not a truly global phenomenon, it is geographically widespread. The reasons for gender differences on smoking decisions are many and complex. Increasing rates of female smoking have been attributed, among other things, to socioeconomic factors such as social acceptance and increasing economic resources. Due to the burden of smoking related illness on public health expenditures, a wide range of policies to reduce smoking rates (smoke-free public spaces, increased tobacco taxes, restrictions on purchasing tobacco products, health warnings etc.) have been implemented around the world. However, tobacco companies overcome legal barriers by targeting specific population groups that are more likely to provide regular smokers: women and young people. Whilst tobacco companies devise strategies to attract consumers in specific target groups, anti-smoking public policies are designed to cover the broadest possible segment of the population.

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

Scientific studies that produce evidence on women's smoking are especially important in a context of increasing information and awareness about the adverse health effects of smoking. This paper examines the impact of two anti-tobacco policies on the smoking decisions of women living in Spain: prices and pictorial health warning messages. Data has been taken from the Spanish National Health Surveys of 2001, 2003, 2006 and 2011. The results of this research may be of international interest given that specific indirect taxes on tobacco products and pictorial labels have been introduced by many governments around the world. The main contribution of this paper is the consideration of generation cohorts as a control for the different time periods or stages that women experience during their life cycle. People make decisions based on their past choices, experiences and messages they have received during their lives. This article examines the way that women from different generational groups have responded to two recent public policies on smoking: indirect tobacco taxes and pictorial health warning labels. The article is organised as follows: after this Introduction, Section 2 comprises a review of the published literature on gender and age differences and the individual response to cigarette health warning labels and the price of tobacco products; Section 3 describes the data and the econometric model; Section 4 presents the empirical results; Section 5 discusses the most relevant conclusions and possible implications for public policies. 2. Literature review The concept of feminisation is frequently employed in contemporary social science research to describe smoking patterns among women. In this context, feminisation refers to smoking as a behaviour that was exclusively related to men but then adopted by women (Jiménez Rodrigo, 2010). Given the importance of the time perspective, studies on tobacco use should consider how time scales and birth cohorts modulate gender differences — female birth cohorts have revealed smoking patterns that have persisted through the years (Keyes et al., 2013). People from the same generation usually feel that they belong to a population group, but more importantly, they share behaviours and beliefs and they have lived through the same historical events (Howe & Strauss, 2000). This present study is based on four different generational groups in Spain (Garcia, Stein, & Pin, 2008): a) Traditional (born before 1950): These are practical people, they work for a living and accept that decisions are taken by the upper levels of the hierarchy and they are loyal to it. They were born during the Franco dictatorship and spent their early adulthood under the regime. These women were educated to be housewives and to take care of their families. b) Baby boomers (1951–1964): Optimistic and idealistic, they live for work and the hope of being rewarded for their efforts. They believe that decisions should be taken by consensus, so they are not subservient to the hierarchy. They were born during the dictatorship but they show rebellious attitudes towards the establishment. This group of women had access to higher and further education and were able to enter the labour market. c) Generation X (1965–1983): Sceptical and independent, they believe decisions should be taken by those who are competent. Work is seen as enjoyable but they can differentiate between private and professional life. They were born at the end of the dictatorship or during the early years of democratic Spain. These women achieved educational levels that were as high as men but there were important gender differences in the labour market that limited their salaries and possibilities for promotion. d) Generation Y also known as Millennials (1985–1999): Realistic and determined, they are pragmatic, creative and oriented to obtaining results. They understand the value of collectivism but they are courteous to the hierarchy. They were born when Spain was already a member of the European Union. Spain's membership of the

295

European Union accelerated the modernisation of many Spanish state responsibilities such as taxation and public health policies on tobacco products.

In Spain, the biggest growth of cigarette smoking among women took place during the transition from dictatorship to democracy (the 1970s). At that time, women experienced intense social changes: female liberation; migration from rural to urban areas; easy access to the job market and university education. Smoking was adopted by Spanish women as a symbol of emancipation and sexual equality (Jiménez Rodrigo, 2010). Differences in birth cohorts underline the fact that anti-smoking policies are not only necessary for young women, they are also important for middle-aged women (Birkett, 1997; Marugame et al., 2010; Wagenknecht et al., 1998). One of the most common measures aimed at reducing tobacco consumption is taxation. Although increasing tobacco taxes is a populationbased control policy, there are differences in effect when adjusted by population groups; men and women react to tobacco taxes in the same way, but the intensity of the effect varies. Some authors have found that men are more sensitive to tobacco taxes than women (Chaloupka, 1990; Hersch, 2000; Lewit & Coate, 1981), whilst others have concluded that women are more responsive than men (Farrelly, Brady, Pechacek, & Woollery, 2001; Nonnemaker & Farrelly, 2011; Stehr, 2007; Townsend, Roderick, & Cooper, 1994). This indicates a wide dispersion in cigarette price elasticities (Stehr, 2007). One possible explanation might be found in the proxies of socioeconomic status. In general, women have lower incomes than men and people with low incomes are usually more responsive to price changes (Stehr, 2007). Low socioeconomic status is often associated with an increased likelihood of smoking (Gilman, Abrams, & Buka, 2003; McGee & Williams, 2006), but the association is far from evident. If men and women exhibit similar tobacco elasticities with regard to educational levels it infers that gender differences on smoking rates could be explained by gender gaps in education. In that case, policies that improve educational achievement among women might be effective in reducing women's smoking rates. Although access to education in Spain is universal, there are still socioeconomic differences, especially among older cohorts that might explain gender differences on smoking decisions. However, if men and women have different tobacco elasticities with regard to education, then education policies may even increase gender differences; for example, if education plays no role in smoking decisions among women (if educated women smoke in order to demonstrate a modern, liberal image), then policies that focus on education will reduce smoking rates among men but will have a limited impact on women. Two very recent studies (with data from France, Germany, Norway and the United States) conclude that educational differences in smoking are strongly determined by birth cohorts. Educational differences in both daily smokers and those that never smoke increased among young cohorts and levelled off among older cohorts. This result emphasises the importance of birth cohorts when studying behaviour and social backgrounds (Pampel et al., 2015; Vedøy, 2014). In terms of public policy, it seems that anti-smoking policies may have heterogeneous effects on smoking behaviours among men and women who share socioeconomic characteristics (Bauer, Göhlmann, & Sinning, 2007; Escobedo & Peddicord, 1997). Only 2% of the gender difference in smoking is explained by socioeconomic gender inequalities, for example, in education or income; 98% can be attributed to gender disparities with regard to attitudes, beliefs and social pressure (Chung, Lim, & Lee, 2010). Female smoking is reinforced by social, economic and political processes that change women's images and roles in society. Tobacco companies take advantage of this fact and target women, using aggressive marketing strategies (Suárez, 2011). The published literature indicates that changing gender roles lead to more women smoking, for example, even in countries such as Italy, where smoking rates are falling for both

296

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

men, women and teenagers, young women feel that they are under pressure to start smoking (Verlato et al., 2006). Young women might be more vulnerable to tobacco products than young men because they are more concerned with their weight, body image, or social acceptance. Women are still encouraged to smoke by images in magazines and movies which are, ultimately, little more than subtle marketing campaigns (Eriksen, Mackay, & Ross, 2012). Identifying the reasons why women are more susceptible to tobacco companies' strategies is essential for fighting against them (Maralani, 2013). Decisions on stopping smoking have also been the subject of research. The main determinants of smoking cessation are: high socioeconomic status (Gilman et al., 2003; McGee & Williams, 2006); low alcohol intake (Pomerlau, Zucker, & Stewart, 2003); and partner support for stopping (Roski, Schmid, & Lando, 1996). In general, fewer women stop smoking than men, a possible explanation for this is that women perceive smoking as less addictive (Jiménez Rodrigo, 2010; Lundborg & Andersson, 2008). Among women, the prevalence of cessation increases with age (Kendrick & Merritt, 1996; Wagenknecht et al., 1998). Young people who start experimenting with smoking ignore anti-smoking messages about the long‐term health risks because they think they will be able to stop before they do themselves too much harm (Gerking & Khaddaria, 2012). With the aim of reaching the largest population possible, health messages on smoking are generalised and unequivocal (Smoking Kills) and tobacco companies are obliged to include health warnings on tobacco products. It is argued that these health warnings are based on behaviourally-compatible incentives (Chandon, 2013) and they include health related messages on beliefs about the health risks of smoking, particularly less well-known health effects such as gangrene, impotence and strokes (Mutti, Hammond, Reid, & Thrasher, 2013). Health warnings can help ex-smokers because they remind them why they made the decision to stop (Partos, Borland, & Yong, 2013). Research on lifestyle changes motivated by negative health information has found that those with higher income levels are more likely to alter their behaviour (Zhao, Konishi, & Glewwe, 2013). Another area of interest is the kind of information that is offered to smokers. The first health warnings were simple messages such as “Smoking Can Kill” or “Smoking is Addictive”; these messages had greater impact on individuals with a lower socioeconomic status (Hitchman et al., 2011). The next step was to display more information about ingredients, on the basis that more comprehensive information would be understood by the smoker; it has been suggested that informative texts are more effective at encouraging people to stop smoking than the simple display of numerical facts (Gallopel-Morvan et al., 2010; Hammond & White, 2012). More recently, pictorial labels have been included on the packets and these seem to be especially effective at reducing demand for cigarettes among younger smokers (Rousu and Thrasher, 2013), women (Muñoz et al., 2013) and smokers with low literacy levels (Thrasher et al., 2012). Health warnings on tobacco products require government legislation but they are appealing because public health costs are reduced (Parkinson & Goodall, 2011). When tobacco control strategies are aimed at reaching all population groups they can influence the more vulnerable populations (Bittencourt, Sharina, Person, Cruz, & Scarinci, 2013). Australia has introduced plain cigarette packaging, a measure that is being considered by several other countries (e.g. Canada, New Zealand and the European Union). Recent research has shown that young people are responsive to the strategy (Bansal-Travers, Hammond, Smith, & Cummings, 2011; McCoola, Webb, Cameron, & Hoek, 2012) and this is especially true for young women (Moodie, Bauld, Ford, & Mackintosh, 2014). 3. Data and methodology Data has been taken from the Spanish National Health Surveys of 2001, 2003, 2006 and 2011. The surveys began in 1987, but for reasons of data comparability, we have only used estimation models with the

Fig. 1. Smoking rates. Source: Spanish Health Surveys (1987, 1993, 1995, 1997, 2001, 2003, 2006, 2011).

last four waves. The survey is cross-sectional and is representative of the current health situation in Spain. The main reason for using the survey is because it gives information on individuals' socio-demographic characteristics and smoking decisions. Our sample includes the responses of 48,755 women, of 16 years and older, living in Spain. Fig. 1 shows that Spanish men's smoking rates have been declining in the last decades, however, women's rates have still not reached the maximum level, considered as 50%. Given that anti-smoking programmes were introduced when Spain joined the European Union in 1986, it could be argued that a specific analysis of the low impact of the programmes among women during the last three decades is required. One of the main research contributions of this paper is that we have repeated the empirical analysis by generational samples. The classification of women by generational cohorts allowed us to control for different backgrounds. People from the same generation usually feel that they belong to a population group, but more importantly, they share behaviours and beliefs and they have lived through the same historical events (Howe & Strauss, 2000). Given our time horizon, we considered four different generational groups: Traditional (born before 1950); Baby boomers (1951–1964); Generation X (1965–1983); Generation Y/Millennials (1985–1999). The older women of the traditional cohort have the lowest smoking rates. They are people that grew up under Franco and were conditioned by a strong catholic culture in which women were seen as housewives and mothers. The heaviest smokers are women from the baby boom and generation X cohorts. The baby boomers experienced early adulthood at the end of the dictatorship and rebelled against the traditional hierarchy. Smoking is identified as a sign of dissent and modernity. Women who belong to generation X were born at the end of the dictatorship or during the early period of Spain's democracy. At that time, smoking became socially acceptable. Neither generation was exposed to anti-smoking messages until the beginning of the 1990s. Smoking rates for women who belong to the generation Y cohort are not conclusive; there are young women who do not smoke, or did not smoke, but might do so in the future. The Y cohort is the youngest group and many of them may be experimenting with tobacco products rather than thinking about stopping. However, this is the only group that will have been exposed to anti-tobacco campaigns throughout their lives. An analysis of socioeconomic characteristics by generations shows that educational levels have greatly increased, particularly from the traditional group to the baby boomers. In 1970, a new education law was passed1 that provided the foundations for the modern education 1 Law 14/1970 of the 4th of August, on General Education and Finance of Educational Reform.

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

297

Table 1 Descriptive statistics for Spanish women (16 years and over): 2001–2011. Percentage of current smokers

Percentage of ex-smokers

Population distribution

Traditional

Baby Boom

Generation X

Generation Y

Traditional

Baby Boom

Generation X

Generation Y

Traditional

Baby Boom

Age 16–19 20–29 30–39 40–49 50–59 60–69 70–79 80–89

– – – – 24.7 16.1 8.8 5.8

– – 49.2 40.1 31.6 25.2 – –

40.4 43.9 40.1 37.1 – – – –

26.8 41.2 – – – – – –

– – – – 19.7 22.2 21.1 18.5

– – 18.5 23.6 23.2 23.0 – –

4.7 9.1 17.3 19.0 – – – –

2.9 8.9 – – – – – –

– – – –

– –

Education level No formal education Primary Secondary Tertiary

9.5 12.5 18.2 20.1

34.7 37.6 39.7 31.9

38.8 50.0 43.3 29.7

29.8 37.9 30.3 32.9

17.7 19.3 25.5 33.8

15.9 19.6 23.4 29.7

7.1 14.2 14.7 15.7

3.5 3.1 5.0 6.9

31.1 42.5 19.8 6.6

5.3 25.4 51.6 17.7

2.2 12.7 62.0 23.1

1.5 14.0 79.8 4.7

Cigarette price b=2.55 € N = 2.56 €

14.3 12.2

42.5 33.6

44.9 37.3

27.4 34.4

20.7 21.1

21.4 24.4

11.4 17.6

2.4 6.5

46.1 53.9

43.8 56.2

47.4 52.6

42.0 58.0

Pictorial label No Yes Total

12.9 14.0 13.1

39.4 31.8 37.5

42.4 35.1 40.1

30.2 42.2 31.4

21.0 20.6 20.9

23.4 22.3 23.1

14.1 16.8 14.7

3.9 12.5 4.8

76.3 23.7 100.0

74.9 25.1 100.0

79.6 20.4 100.0

89.4 10.6 100.0

15.6 33.9 32.7 15.8

6.9 55.9 35.6 1.6 – –

Generation X 2.7 31.6 50.9 14.8 – – – –

Generation Y 67.7 32.3 – – – – – –

Number of observations by birth cohorts: Traditional — 21,519; Baby boomer — 11,066; Generation X — 14,372; Generation Y — 1798.

system in Spain. The law ensured that the state guarantees access to education to all. Women were able to study, enter the labour market and abandon their exclusive role of homemakers. As a consequence, women who belong to generation X are the most educated women in Spanish history (it is too early to fully analyse the educational achievements of generation Y). From Table 1, we can see that for women in the traditional cohort, education plays a positive role with regard to smoking. In generation X, women with tertiary studies are clearly the group with the lowest number of smokers. Education is positively correlated with stopping smoking for all age population groups. Two types of anti-smoking policies were contemplated: increased prices and pictorial warning labels. The prices of cigarettes were taken from the Spanish National Institute of Statistics website. In order to facilitate the temporal comparison of prices and remove the effect of inflation, we used constant price terms, taking 2011 as the base period. In addition, we employed squared price values to determine how price changes affect demand. The time period considered in this research is broad enough to analyse the impact of the last Spanish law on cigarette labelling which introduced images to cigarette packs in 2010.2 A dummy variable was introduced that takes the value of 1 if cigarette packs contained pictorial labels in that year, and 0 if not. Table 1 shows that the price of cigarettes is negatively correlated with smoking and positively correlated with stopping smoking. Nevertheless, there were no significant differences on smoking decisions and prices for women from the traditional cohort. Introducing pictorial labels on cigarette packs seems to have slowed smoking rates for women from the baby boom and generation X and encouraged more women of generations X and Y to stop. Given the dichotomous nature of the dependent variable Smoker (yes: 1: no: 0) and Ex-smoker (yes: 1: no: 0), we ran logit models. We have reported the data as the marginal effects of the probabilities (mfx) to provide information on the marginal effects and the strength

2 Royal Decree 639/2010, of the 14th of May: “Smoking May Kill” was removed from the list of phrases that tobacco companies could use to meet the labelling requirement. It became obligatory to state that it was illegal to sell tobacco to minors. It was recommended that images of the physical consequences of smoking should be included on the cigarette packs.

of the coefficients. The variable Age is a continuous variable that gives us the individual's age in the year of the interview. Educational levels are categorised into 4 dummy variables (No formal education, Primary, Secondary and Tertiary). The category No formal education is not included in the regression analysis and results are interpreted by taking this category as a reference point. If the estimated parameter for Primary in the regression analysis of Smoker is positive, it means that individuals with primary studies are more likely to be smokers than individuals with the same characteristics but with no formal education. The dummy variable Pictorial Label takes the value 1 if the individual was exposed to pictorial labels when interviewed and 0 otherwise. Cigarette Price is a continuous variable that gives the average price of cigarette packages in the year the individual was interviewed. Cigarette Price2 is the Cigarette Price squared, it informs us if the relationship between Cigarette Price and the dependent variable is roughly proportional to increases in prices. The robustness of the estimated coefficients was checked by repeating estimations of the total sample and sub-samples of generation cohorts. Two different models were also considered: Model 1, analyses socio-demographic characteristics (age and educational levels) and tobacco policies (cigarette prices and cigarette pictorial labels); Model 2, includes interactions of individual educational levels and anti-tobacco policies, with the aim of determining if education adjusts the impact of these policies on female smoking decisions. The interactions of educational levels and anti-tobacco policies are calculated by multiplying both variables. For example, the variable Label_noeducation is calculated by multiplying the dummy variables Pictorial Label and No formal education that identify individuals with no formal education who have been exposed to pictorial labels in the year of the interview. As there are four educational levels, there are four interactions for each anti-tobacco policy. One of the interactions is taken as a reference point. The results complete the information provided by the parameters of anti-tobacco policies, indicating if individuals are more sensitive to these policies by groups of educational levels. We have further reported estimated probabilities of the dependent variables (to be compared with the observed probabilities) and the pseudo-R2 to check the global fit of the models controlling for the number of explanatory variables. The pseudo-R2 ranges from 0 to 1 (from 0 to 100%) with higher values indicating a better fit.

298

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

Table 2 Women's smoking decisions (logit: mfx). Smoker Model 1 Age No formal educationa Primary Secondary Tertiary Cigarette Price Cigarette Price2 Price_noeducation Price_primary Price_secondary Price_terciary Pictorial label Label_noeducation Label_primary Label_secondary Label_terciary Estimated value (%) Pseudo-R2 (%)

Ex-smoker Model 2

0.010***

Model 1

0.011***

0.010***

–

–

–

0.045*** 0.078*** 0.029*** −1.562*** 0.294*** – – – – −0.180*** – – – – 24.6 14.0

0.190 0.749*** 0.823*** −0.884 0.192* – −0.029 −0.194*** −0.207*** −0.099 – −0.077** −0.001 0.002 24.6 14.6

0.044*** 0.118*** 0.189*** 0.762 −0.127 – – – – 0.094 – – – – 12.6 6.9

Model 2 0.010*** – −0.147* 0.009 0.083 0.826* −0.152* – 0.102** 0.053 0.041 0.617** – −0.087*** −0.100*** −0.091*** 12.6 7.7

***, ** and * indicate statistical significance at levels 0.001, 0.005 and 0.010. a Reference variable. Age2 was also included but results for this variable are not shown because they were 0 or very close to 0.

4. Results Table 2 presents a regression analysis for the total sample. The results for Model 1 show that older women are more likely to be smokers or ex-smokers. Education plays an important role — it is negatively correlated with the decision to smoke but positively correlated with the decision to stop. The correlation effects for both decisions are reinforced for higher educational levels. Higher prices and pictorial labels are negatively correlated with smoking rates. Although the estimated coefficients suggest the existence of monetary incentives for stopping smoking, the results were not statistically significant. When the interaction of anti-smoking policies and educational levels is considered, policies alone are not statistically significant; in any case, the coefficients remain constant for smoking rates. Estimations for Model 2 indicate that prices are a more powerful instrument for highly educated women, whilst pictorial labels might be a more effective strategy for women with no formal education. With regard to ex-smokers, anti-smoking policies not only retain their positive significance in

controlling cigarette demand but they also gain statistical significance. Once again, images are more relevant for the less educated. Both models reveal a similar goodness of fit in terms of the number of coefficients that are statistically significant, the estimated values of the dependent variables and the values of the pseudo-R2. Model 2 offers a broader picture; estimations by generation cohorts are shown in Table 3. Age is negatively correlated with smoking for older women (the traditional cohort), whereas it plays a positive role for the youngest women (generation Y). Educational levels do not appear to have any particular effect on the smoking decisions of older women, but education is positively correlated with smoking decisions for women who belong to the baby boom generation and generation X. The main difference between the groups is that the peak smoking rate is reached for women with tertiary education in the case of the baby boomers, and for women with secondary education in the case of generation X. For generation Y, having tertiary studies is negatively correlated with the probability of smoking, but this result is of limited validity as there are many women in this group who are still studying and it was not statistically significant. When the results are controlled by birth cohorts, price increases seem to be a good instrument for reducing smoking rates among older women, but not for women of other generations. The inclusion of pictorial labels may deter young women in generations X and Y from smoking. Education reinforces the impact of images in the case of the traditional cohort. Women from generation Y were not included in the results on exsmokers as many of them will still be at the stage of experimenting with tobacco, rather than thinking about stopping. Consequently their corresponding results involved more technical problems than explanatory information. Women from the generation X cohort are more likely to stop smoking. Cigarette prices might be a good instrument to encourage stopping smoking among women who belong to the baby boom cohort but not for women who belong to generation X. Education may reinforce the effect of cigarette prices, especially for baby boomers. The inclusion of anti-smoking images on cigarette packs may help smokers from the baby boom cohort to stop but this is not the case for generation X. Finally, there is empirical evidence that pictorial labels might be of most benefit for less educated women. 5. Conclusions and policy implications In Spain, whilst male smoking rates are falling, it seems that more and more women are taking up the habit. Measures are needed to

Table 3 Women's smoking decisions by generation cohorts (logit: mfx): Model 2. Smoker

Age No formal educationa Primary Secondary Tertiary Cigarette Price Cigarette Price2 Price_noeducation Price_primary Price_secondary Price_terciary Pictorial Label Label_noeducation Label_primary Label_secondary Label_terciary Estimated value (%) Pseudo-R2 (%)

Ex-smoker

Traditional

Baby boom

Generation X

−0.014*** – −0.065 −0.020 0.209 −0.792* 0.155* – 0.035 0.031 −0.010 −0.067 – −0.037*** −0.043*** −0.028** 10.1 11.9

−0.005 – 0.483 0.543** 0.656*** −0.501 0.134 – −0.156 −0.182 −0.239 −0.189 – 0.140 0.114 0.100 33.6 1.5

−0.003 – 0.625*** 0.643*** 0.582** −2.203 0.431 – −0.248 −0.281 −0.261 −0.411*** – 0.267 0.307 0.325 38.4 1.6

Generation Y

Traditional

Baby boom

Generation X

0.314***

−0.006*** – −0.095 −0.048 0.019 0.231 −0.041 – 0.048 0.059* 0.033 0.269 – −0.040*** −0.049*** −0.042*** 17.4 10.4

−0.001 – −0.446** −0.507 −0.207 5.090*** −0.986*** – 0.293* 0.257* 0.204 0.953*** – −0.176*** −0.204*** −0.174*** 18.5 2.7

0.036*** – −0.052 −0.267 −0.216 −2.080* 0.386* – 0.066 0.147 0.178 −0.217** – −0.042 −0.098 −0.120* 14.3 2.9

– 0.704 0.781** −0.964 −5.785 1.204 – −0.225 −0.792 5.446 −0.522*** – 0.042 0.651 −0.409 30.8 2.9

***, ** and * indicate statistical significance at levels 0.001, 0.005 and 0.010. a Reference variable. Age2 was also included but results for this variable are not shown because they were 0 or very close to 0.

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

protect women from smoking and the negative health consequences associated with tobacco products. The results of this research are coherent with historical perspectives of epidemiological transitions by income. The generalised adoption of unhealthy behaviours (smoking, drinking, lack of physical exercise etc.) begins by causing more health problems for economically advantaged individuals, but in the long term, the greatest burden of chronic diseases is suffered by those on lower incomes. Women, especially women in low socioeconomic conditions, are a vulnerable group (Gersh, Sliwa, Mayosi, & Yusuf, 2010). The main contribution of this paper is the consideration of birth cohorts to study policy factors that affect smoking decisions among Spanish women. Characterising smoking patterns in birth cohorts is essential for evaluating the impact of tobacco control interventions. Women from different generations have experienced different policy interventions, marketing strategies or cultural changes that have determined their behaviour. More country-specific evidence on factors that influence tobacco consumption is required. Our results highlight that smoking determinants vary across generation cohorts. Smoking rates have always been low among older women but women from younger cohorts (generations X and Y) are more likely to smoke, even if they are highly educated. Smoking imposes health gender inequalities and health policies should address these disparities. If mandatory policies related to product labelling, prices and accessibility are effective, these policies should be included in national portfolios (Costa-Font, Hernández-Quevedo, Jayachandran, & Variyam, 2013). Our results indicate that cigarette prices and pictorial labels may be good instruments for reducing female smoking rates. Highly educated women are more sensitive to prices; lower educated women are more sensitive to pictorial labels. Although both policies are effective at reaching women, pictorial labels have a two-fold value because they target the most vulnerable (women with low educational levels). Differences in smoking across cohorts, genders and countries help identify the national and temporal circumstances that shape the relationship between education and health and the need for policies that target educational differences (Pampel et al., 2015). The effectiveness of anti-tobacco polices is modulated by generation cohorts, therefore, in terms of smoking rates, it is necessary to focus on vulnerable population groups and to design the strategies that best suit them. Policy proposals to reduce smoking among women and younger generation cohorts include plain packaging and uniform tax rates on tobacco products. Plain cigarette packaging reduces demand, particularly among young women (Moodie et al., 2014). The Spanish government has increased tobacco taxes, but mainly on cigarettes, the tobacco industry has responded by marketing cheaper alternative tobacco products, such as rolling tobacco, which are more popular among the young. Uniform tax rates for tobacco products could discourage people from substituting rolling tobacco for cigarettes (López-Nicolás, Cobacho, & Fernández, 2013). This paper is not without technical limitations. In future studies, it would be desirable to use panel data in order to follow individual lifecycle smoking patterns by birth cohort. Panel data allows the researcher to control for fixed effects and this could help explore the relationship between policies and individual smoking decisions. The use of fixed effects assumes that there may be a factor within the individual that may impact or bias the prediction of smoking decisions. As we have observed that women belonging to different generations react differently to anti-tobacco policies (independently of their educational levels), controlling for fixed effects might offer more refined results. Role of funding source No funding sources.

Contributors All authors have been contributing to the different sections of the article and all authors have approved the final manuscript.

299

Conflict of interest No conflict of interest.

References Bansal-Travers, M., Hammond, D., Smith, P., & Cummings, M. (2011). The impact of cigarette pack design, descriptors, and warning labels on risk perception in the U.S. American Journal of Preventive Medicine, 40, 674–682. Bauer, T., Göhlmann, S., & Sinning, M. (2007). Gender differences in smoking behaviour. Health Economics, 16, 895–909. Birkett, N. (1997). Trends in smoking by birth cohort for births between 1940 and 1975: A reconstructed cohort analysis of the 1990 Ontario Health Survey. Preventive Medicine, 26, 534–541. Bittencourt, L., Sharina, I., Person, D., Cruz, R. C., & Scarinci, I. C. (2013). Pictorial health warnings on cigarette packs and the impact on women. Revista de Saúde Pública, 47, 1123–1129. Chaloupka, F. (1990). Men, women, and addiction: The case of cigarette smoking. NBER Working Paper 3267. Chandon, P. (2013). How package design and packaged-based marketing claims lead to overeating. Applied Economic Perspectives and Policy, 35, 23–147. Chung, W., Lim, S., & Lee, S. (2010). Factors influencing gender differences in smoking and their separate contributions: Evidence from South Korea. Social Science & Medicine, 70, 1966–1973. Costa-Font, J., Hernández-Quevedo, C., Jayachandran, J. T., & Variyam, N. (2013). Understanding healthy lifestyles: The role of choice and the environment. Applied Economic Perspectives and Policy, 35, 1–6. Eriksen, M., Mackay, J., & Ross, O. (2012). Tobacco atlas (4th ed ). Atlanta: American Cancer Society, World Lung Foundation (Available at http://www.tobaccoatlas. org.). Escobedo, L. G., & Peddicord, J. P. (1997). Long-term trends in cigarette smoking among young U.S. adults. Addictive Behaviours, 22, 427–430. Farrelly, M., Brady, J., Pechacek, T., & Woollery, T. (2001). Response by adults to increases in cigarette prices by sociodemographic characteristics. Southern Economic Journal, 68, 156–165. Franco, J., Perez-Hoyos, S., & Plaza, P. (2002). Changes in lung-cancer mortality trends in Spain. International Journal of Cancer, 97, 102–105. Gallopel-Morvan, K., Moodie, C., Hammond, D., Eker, F., Beguinot, E., & Martinet, Y. (2010). Consumer understanding of cigarette emission labelling. European Journal of Public Health, 21, 373–375. Garcia, P., Stein, G., & Pin, J. (2008). Políticas para dirigir a los nuevos profesionales. Motivaciones y valores de la generación Y. Documento de Investigación 753 IESE Business School. Gerking, S., & Khaddaria, R. (2012). Perceptions of health risk and smoking decisions of young people. Health Economics, 21, 865–877. Gersh, B. J., Sliwa, K., Mayosi, B. M., & Yusuf, S. (2010). The epidemic of cardiovascular disease in the developing world: global implications. European Heart Journal, 31, 642–648. Gilman, S. E., Abrams, D. B., & Buka, S. L. (2003). Socioeconomic status over the life course and stages of cigarette use: Initiation, regular use, and cessation. Journal of Epidemiology and Community Health, 57, 802–808. Hammond, D., & White, C. M. (2012). Improper disclosure: Tobacco packaging and emission labelling regulations. Public Health, 126, 613–619. Hersch, J. (2000). Gender, income levels, and the demand for cigarettes. Journal of Risk and Uncertainty, 21, 263–282. Hitchman, S. C., Mons, U., Nagelhout, G. E., Guignard, R., Mcneill, A., Willemsen, M. C., et al. (2011). Effectiveness of the European Union text-only cigarette health warnings: Findings from four countries. European Journal of Public Health, 22, 693–699. Howe, R., & Strauss, W. (2000). Millennials rising: The next great generation. New York, USA: Vintage books. Jiménez Rodrigo, M. L. (2010). La feminización del consumo de tabaco: ¿convergencia o desigualdad? Revista Española de Drogodependencias, 35, 285–296. Kendrick, J. S., & Merritt, R. K. (1996). Women and smoking: An update for the 1990s. American Journal of Obstetrics and Gynecology, 3, 528–535. Keyes, K. M., March, D., Link, B. G., Chilcoat, H. D., & Susser, E. (2013). Do socio-economic gradients in smoking emerge differently across time by gender? Implications for the tobacco epidemic from a pregnancy cohort in California, USA. Social Science & Medicine, 76, 101–106. Lewit, E., & Coate, D. (1981). The potential for using excise taxes to reduce smoking. NBER Working Paper 764. López-Nicolás, A., Cobacho, M. B., & Fernández, E. (2013). The Spanish tobacco tax loopholes and their consequences. Tobacco Control, 22, 21–24. Lundborg, P., & Andersson, H. (2008). Gender, risk perceptions, and smoking behavior. Journal of Health Economics, 27, 1299–1311. Maralani, V. (2013). Educational inequalities in smoking: the role of initiation versus quitting. Social Science & Medicine, 84, 129–137. Marugame, T., Kamo, K., Sobue, T., Akiba, S., Mizuno, S., Satoh, H., et al. (2010). Trends in smoking by birth cohorts born between 1900 and 1977 in Japan. Preventive Medicine, 42, 120–127. McCoola, J., Webb, L., Cameron, L. D., & Hoek, J. (2012). Graphic warning labels on plain cigarette packs: Will they make a difference to adolescents? Social Science & Medicine, 74, 1269–1273. McGee, R., & Williams, S. (2006). Predictors of persistent smoking and quitting among women smokers. Addictive Behaviors, 31, 1711–1715. Moodie, C., Bauld, L., Ford, A., & Mackintosh, A. M. (2014). Young women smokers' response to using plain cigarette packaging: Qualitative findings from a naturalistic study. BMC Public Health, 14, 812.

300

A.I. Gil-Lacruz et al. / Addictive Behaviors 45 (2015) 294–300

Muñoz, M. A., Viedma-del-Jesus, M. I., Rossello, F., Sanchez-Nacher, N., Montoya, P., & Vila, J. (2013). The emotional impact of European tobacco-warning images. Tobacco Control, 22, 123–129. Mutti, S., Hammond, D., Reid, J. L., & Thrasher, J. F. (2013). The efficacy of cigarette warning labels on health beliefs in the United States and Mexico. Journal of Health Communication: International Perspectives, 18, 1180–1192. Nonnemaker, J. M., & Farrelly, M. C. (2011). Smoking initiation among youth: The role of cigarette excise taxes and prices by race/ethnicity and gender. Journal of Health Economics, 30, 560–567. Pampel, F., Legleye, S., Goffette, C., Piontek, D., Kraus, L., & Khlat, M. (2015). Cohort changes in educational disparities in smoking: France, Germany and the United States. Social Science & Medicine. Parkinson, B., & Goodall, S. (2011). Considering consumer choice in the economic evaluation of mandatory health programmes: A review. Health Policy, 101, 236–244. Partos, T. R., Borland, R., Yong, H. H., Thrasher, J., & Hammond, D. (2013). Cigarette packet warning labels can prevent relapse: findings from the International Tobacco Control 4—Country policy evaluation cohort study. Tobacco Control, 22, 43–50. Pomerlau, C. S., Zucker, A. N., & Stewart, A. J. (2003). Patterns of depressive symptomatology in women smokers, ex-smokers and never-smokers. Addictive Behaviors, 28, 575–582. Roski, J., Schmid, L. A., & Lando, H. A. (1996). Long-term associations of helpful and harmful spousal behaviors with smoking cessation. Addictive Behaviors, 21, 173–185. Rousu, M. C., & Thrasher, J. F. (2013). Demand reduction from plain and pictorial cigarette warning labels: Evidence from experimental auctions. Applied Economic Perspectives and Policy, 35, 171–187.

Stehr, M. (2007). The effect of cigarette taxes on smoking among men and women. Health Economics, 16, 1333–1343. Suárez, N. (2011). Women & smoking: Becoming part of the solution. MEDICC Review, 13, 56. Thrasher, J. F., Carpenter, M. J., Andrews, J. O., Gray, K., Alberg, A. J., Navarro, A., et al. (2012). Cigarette warning label policy alternatives and smoking-related health disparities. American Journal of Preventive Medicine, 43, 590–600. Townsend, J., Roderick, P., & Cooper, J. (1994). Cigarette smoking by socioeconomic group, sex, and age: effects of price, income, and health publicity. British Medical Journal, 309, 923–927. Vedøy, T. F. (2014). Tracing the cigarette epidemic: An age-period-cohort study of education, gender and smoking using a pseudo-panel approach. Social Science Research, 48, 35–47. Verlato, G., Melotti, R., Corsico, A. G., Bugiani, M., Carrozzi, L., Marinoni, A., et al. (2006). Time trends in smoking habits among Italian young adults. Respiratory Medicine, 100, 2197–2206. Wagenknecht, L. E., Craven, T. E., Preisser, J., Manolio, T. A., Winders, S., & Hulley, S. B. (1998). Ten-year trends in cigarette smoking among young adults, 1986–1996: The CARDIA Study. Annals of Epidemiology, 8, 301–307. WHO (2008). WHO report on the global tobacco epidemic, 2008: The MPOWER package. Geneva: World Health Organization. Zhao, M., Konishi, Y., & Glewwe, P. (2013). Does information on health status lead to a healthier lifestyle? Evidence from China on the effect of hypertension diagnosis on food consumption. Journal of Health Economics, 32, 367–385.