Discussion

Coupons or other price-related discounts from PM, RJR and Lorillard were used by 14–25% of their consumers and have provided price reductions for the smokers who used specific brands. Coupons or other price-related discounts from other companies did not result in statistically significant price per pack reductions for their consumers. Therefore, after controlling for the use of other price minimisation strategies and respondents’ smoking intensity and nicotine addiction, the three leading cigarette companies provided price reductions for their products through coupons or other price-related discounts.

To put the range of price reductions associated with coupons or other price-related discounts into context, the cigarette federal tax was $1.01 per pack starting 1 April 2009, and the weighted average state cigarette excise tax rate was $1.17 per pack in 2010. These numbers imply that the coupon discounts from the three leading cigarette companies (about 24–29 cents) offset 23.8–28.7% of the price impacts from the federal tax or 11–13.3% of the price impacts from the federal and state excise taxes combined to the smokers of cigarettes produced by these manufacturers. As a result, these offers brought actual average prices down for users of specific brands, while the price reductions associated with these discounts were likely to increase the sales in these companies. For example, with a price elasticity of overall demand for cigarettes among adults at somewhere between −0.3 and −0.7,22 ceteris paribus, these price discounts can be translated into 10.2–23.9 million packs of cigarette sales in 2010 for Marlboro, 3.0–6.6 million packs for Newport and 1.8–4.1 million packs for Camel.21

Although it is true that statistically significant or marginally statistically significant price reductions were observed for the three leading US cigarette companies, significant reductions are brand specific. The significant reductions were observed only for Marlboro (US leading brand), Camel (second US leading brand), Kool, Doral and Salem, while marginally for Newport (third US leading brand). Thus, PM, RJR and Lorillard have concentrated their efforts to provide price discounts mainly to their best selling brands. This might be one of the reasons that a previous analysis failed to identify significant price reductions associated with promotional offers.12

This study has some limitations. First, the study design is cross-sectional. Therefore, the study findings may be specific only to the period October 2009–June 2010. However, the study covers the entire USA, and there is a variety of prices as a function of brand smoked and other factors. Second, most variables in the analysis are collected from recent purchases (ie, price paid, coupon and other price-related discounts and carton purchase are for the latest purchase; premium or generic brands are for the last 30 days), but others are collected with a time frame of 1 year (ie, purchase on Indian reservations, purchase through internet). However, when we excluded Indian reservation and internet purchases from the analysis, the adjusted average prices and price discounts associated with coupons did not change much. Third, because of an approval delay, only approximately 20% of respondents interviewed during the first 2 months of the survey were asked if purchases had been made on an Indian reservation anytime during the past year. In subsequent months, this question was asked for more than 90% of respondents. In the full sample, total missing responses for this question were 18.4% (3503). However, sensitivity analysis has shown that dropping these observations does not significantly affect the results.16 As noted in the Methods section, we also excluded respondents who failed to report price paid for their latest purchase, the brand name they used most often in the past 30 days or some of their demographic characteristics. We compared smoking and social-demographic characteristics between individuals with incomplete information and individuals with complete information and have found little differences at the mean level. Fourth, as noted above, the 2009–2010 NATS does not collect information for all price minimisation strategies, including cigarette purchases from states with lower price. Although the cross-border purchase is an issue in tobacco control, the prevalence of this behaviour was quite low in the USA compared with other forms of price minimisation strategies. For example, data from the 2003 and 2006–2007 Tobacco Use Supplement to the Current Population Survey (TUS-CPS) suggest that about 5% of smokers made purchase across a state border,23 while in the 2010–2011 TUS-CPS, approximately 3% of smokers purchased cigarettes from non-tribal land in lower-taxed and non-residential states (estimates not shown). In order to account for unmeasured price minimisation strategies, we controlled for smoking intensity and level of nicotine addiction in the analysis since the literature suggests that these are important risk factors of using any price minimisation strategies. Additionally, the paid prices were determined by using self-reported information from the smoker, which may be subject to recall bias. However, existing evidence indicates that the average of self-reported prices per pack in the 2009–2010 NATS was very consistent with the corresponding 2009 national average price reported in the Tax Burden on Tobacco (TBOT).16 Another benefit of using self-reported price in this analysis is that we are able to control for the corresponding smoking intensity of each smoker, which is not available in market-scanned data but closely related to potential use of unmeasured price minimisation strategies. Finally, smokers’ self-reported use of coupon or other types of price-related discounts in the 2009–2010 NATS only reflects direct-to-consumer discounts from the industry. As a result, the discount measure in the analysis does not include tobacco industry's promotional allowances directly paid to cigarette retailers or wholesalers, as these disaggregated promotional allowances by cigarette manufacturers or by brands have not been disclosed to consumers or to the public.

In addition to cigarette companies directly influencing cigarette retail prices by providing coupons or other price-related discounts, companies may also indirectly affect cigarette prices by offering discounts to retailers and by promoting cigars or pipe tobacco that can be used in roll-your-own cigarettes.24 Although these indirect influences are critical in tobacco control and certainly warrant additional studies, the NATS survey did not collect such information thus they are not within the scope of this analysis.

Our results show that the three leading cigarette companies in the USA continue to offer price discounts to smokers of their brands, although these promotions appear to be concentrated among their top-selling cigarette brands. Cigarette companies can be strategic when offering price discounts. For example, existing literature suggests that young adults, women and heavy smokers are more frequently targeted for these promotions.19 Other studies have shown that cigarette brands with low market share target young adults with the goal of encouraging brand switching, while major brands target older smokers to facilitate brand loyalty.25 As pointed out earlier, smokers who use price-related discounts are less likely to make quit attempts or to successfully quit in the future. The length (duration) of smoking and the amount of cigarettes smoked per day on days the person smoked is strongly associated with a higher likelihood of developing and dying from a smoking-related disease, such as lung cancer, chronic obstructive pulmonary disease or heart attacks.1 Therefore, even though price minimisation strategies may increase sales and profits for cigarette companies, these price discounts are likely preventing or delaying some smokers from permanent cessation. Policies that decrease price-minimisation strategies will benefit public health.