Article Text
Abstract
Objective To measure the impact of a 15-fold licence fee increase on tobacco retailer licence renewals.
Methods The regulatory change increasing tobacco licence fees (from $A12.90 to $A200 per annum) took effect on 1 January 2007. Government Tobacco Licence records (n=7093) were audited for 1 year prior to, and 2 years after the change. An interrupted time series analysis using ARIMA modelling was conducted to examine the impact of fee increases on the number of active licences.
Results The total number of tobacco licences decreased by 23.7% from December 2007 to December 2009. The increased tobacco licence fee implemented on 1 January 2007, was associated with a significant reduction in the number of tobacco licences purchased or renewed in subsequent years. Of the 1144 entertainment licensees holding valid licences in December 2007, 30.9% no longer held a licence by December 2009, and 19.9% had reduced the number of points of sale within the same venue.
Conclusions Licensing of tobacco retailers has received little attention in tobacco control in Australia and internationally. Our data add to the growing body of evidence supporting further regulation of retail sale of tobacco. The results demonstrate that a tobacco licence price increase off a low base is a potentially effective method of reducing tobacco points of sale when consumer demand for cigarette products is low. However, further research is needed to identify additional measures that may be necessary to reduce the availability of tobacco products in areas where consumer demand for cigarettes is high.
- Public policy
- Environment
- Price
Statistics from Altmetric.com
Introduction
Despite being an exceptionally harmful consumer product, tobacco is a widely available consumer product, sold in virtually every supermarket, petrol station and corner store. This contradiction has prompted calls for tobacco sales reform1–3 and highlights the need for research into the potential of policy measures, such as retailer licensing.
‘Availability Theory’ has a long history within the alcohol consumption literature, and has influenced regulatory strategies aimed at reducing alcohol-related harm (eg, restrictions on the number and type of outlets, restricting trading hours).4 ,5 More recently, researchers in tobacco control have explored the relationship between the availability of tobacco and tobacco consumption. Several cross-sectional studies have found an association between the distributions of retail outlets and smoking behaviour, although the relationship may be modified by individual characteristics (ie, level of cigarette consumption and socioeconomic disadvantage).6–8 A longitudinal study examining the effect of tobacco outlet density and proximity on residents’ smoking abstinence 6 months following a quit attempt found an inverse relationship between proximity to a tobacco outlet and abstinence.9 The relationship between tobacco availability and smoking behaviour may also be influenced by outlet type. For example, light smokers have been found to be more susceptible to impulse purchases from outlets, such as convenience stores and pubs and clubs.7 ,10 Reasons offered for these findings include cueing effects, where the sight of a tobacco display or other smokers creates a desire to smoke,11 ,12 and the increased likelihood of smoking when consuming alcohol.13 ,14
Retail tobacco licence regulations vary across Australian states and territories. In South Australia, retailers selling tobacco products must have a Retail Tobacco Merchant's Licence. One licence must be purchased for each unique point of sale (POS). Only one POS is allowed per retail outlet. Outlets that are also licensed to sell liquor have an exemption, allowing five licences (for five unique POSs) per location, one of which may be a vending machine. Non-compliance has a maximum penalty of $A10 000.
On 1 January 2007, the SA Government increased the fee of a retail tobacco licence from $A12.90 to $A200 per annum, with annual indexation from 1 July 2007 onwards, making the fee comparable with jurisdictions, such as Western Australia ($A204), New York (US$300), Indiana (US$200), and Singapore (SGD 360). This study investigates the impact of the licence fee increases on the number of licences.
Data and methods
The SA Government Tobacco Licencing System database records between December 2005 and 2009 were audited. Each record contained a unique licence number as well as the start and end date, trading name, premises name, business type and address of the licensee. Business type was grouped as follows: ‘Entertainment’ (liquor-licenced; eg, pub, club, hotel, cafe, restaurant), ‘Retail-Food’ (eg, supermarket, service station, corner/convenience store, snack bar, deli, takeaway), and ‘Retail-Other’ (eg, tobacconist, hairdresser, newsagent, caravan park, outdoor festival). Tobacco licences are valid for 12 months, so each licence was assigned ‘active’ or ‘inactive’ status on the last day of each month, starting on 31 December 2005 and ending on 31 December 2009.
An interrupted time-series analysis using autoregressive integrated moving average (ARIMA) modelling was conducted to examine the impact of fee increases on the number of active licences. ARIMA modelling was necessary because the data had significant autocorrelation. Further details regarding ARIMA modelling and notation are available from Wakefield et al.15 Following the identification of the series, intervention models of the form developed by Box and Taio16 were tested for each price increase. The ARIMA procedure in SAS/ETS17 was used to conduct the analysis, and parameter estimates were computed using the maximum likelihood method.
Results
The total number of tobacco licences decreased by 23.7% from December 2007 (3734 licences) to December 2009 (2850 licences). Figure 1 demonstrates a distinct shift up in the numbers of active licences just prior to the initial fee increase (December 2006), and then down 12 months following the initial fee increase (December 2007), and a trailing down after that. An initial ARIMA model was identified using the number of active licences at the end of each month prior to the first intervention, from December 2005 to December 2007, and then fitted to the entire series. Specifically, this model indicates that there was an abrupt and permanent step increase of licences in December 2006 (ω1=240, p<0.001), an abrupt and permanent step decrease of licences in January 2008 (ω2=–224, p<0.001), and a gradual decrease in numbers of licences in July 2008 leading to a permanent reduction of around 350 licences a few months later (ω3=−184/ (1-d=0.48); p<0.001). There was no significant change in licences coincident with the price increase from indexation in July 2009. Ljung-Box Q statistics18 revealed that the model was statistically adequate.
Figure 1 also displays the trends in numbers of licences according to business type. Of the new licences purchased in December 2006, 52.0% were Entertainment licensees, and of these, 76.0% were increases in POS terminals rather than new outlets. Of the 1144 Entertainment licensees holding valid licences in December 2007, 30.9% of licensees ceased holding a licence by December 2009, and a further 19.9% of licensees reduced the number of POSs within the same venue. Of the 862 licensees holding valid licences in December 2009, 8.2% of licensees had bought new licences since December 2007, and a further 3.7% of licensees had increased the number of POSs within the same venue.
Discussion
The one-off 15-fold increase in the licence fee from the low base of $A12.90 to $A200, resulted in an apparent ‘buying up’ of licences before the fee increase, but a subsequent and corresponding decline in the overall number of licences a year later after those $A12.90 licenses expired. It was also possible for licence holders to renew early to delay paying the increased licence fee. The decrease in licences in January 2008 corresponds with the first point at which all new licences/annual renewals cost $A200. This point also corresponds with the introduction of the indoor smoking ban in liquor-licensed (ie, entertainment) outlets, introduced in November 2007. Smoking in outdoor areas remained unrestricted. The decreasing trend in numbers of licences between January 2008 and July 2008 was greatest for Entertainment outlets, suggesting that the combination of increased licence fees and reduced opportunity for smoking in Entertainment outlets resulted in a marked reduction in licences within these business types.
The tobacco licence fee increase had minimal impact on reducing licences in other retail outlet types, which may reflect the importance of tobacco sales in these outlets. For instance, a 2005 industry report19 showed that tobacco products accounted for 95% of sales for tobacconists, over 30% of sales for convenience stores, but only 1% of sales for hotels and clubs. In supermarkets, only 7% of sales were attributable to tobacco products, but the gross revenue exceeded that of other retail types due to the high volume of sales. Thus, it appears that the tobacco licence fee increase was not sufficient to impact on retailers that have a high volume of sales from tobacco products.
The decrease in licences could be the result of businesses ceasing to operate, or businesses not being able to afford a tobacco licence, however, published data suggests otherwise. The number of gaming and liquor licences increased each year from 2006 to 2009,20 ,21 and there was no significant change in business revenue before and after the implementation of smoke-free pubs and clubs.22
This observational study examined a possible mechanism for reducing tobacco availability by providing a disincentive for retailers to hold a tobacco licence (ie, a more expensive licence fee). However, this study had a number of limitations. It was assumed that business types classified as ‘Entertainment’ (eg, pub, club, hotel, cafe, restaurant) also held a liquor licence, as this data was not available from the database. A review of the database found a small number of duplicate records due to some licences not being cancelled when businesses ceased activities. The SA Government does conduct ongoing monitoring to ensure compliance, and the database is manually updated at a later date. The database was structured for licences rather than licensees, thereby limiting further analyses of change by location and density. Additionally, the database does not contain information on all businesses within SA, therefore, the completeness of coverage of the licensing system cannot be determined.
Licensing of tobacco retailers in relation to availability and consumption of cigarettes has received little attention in tobacco control in Australia and internationally. This study has important implications; the reduction in tobacco licensees in the ‘Entertainment’ category suggests that the substantial fee increase from a low base was a disincentive for renewing/purchasing new licences in outlets with low consumer demand. However, further research is needed to determine whether further fee increases would discourage retailers with high volumes of tobacco sales from renewing/purchasing new licences, or whether additional measures may be necessary to reduce availability of tobacco products where consumer demand is high.
What this paper adds
-
Licensing of tobacco retailers in relation to access and availability of tobacco products has received little attention in tobacco control.
-
To our knowledge, this is the first paper published examining the impact of an increase in tobacco retailer merchants licence fees.
-
This study demonstrated that if a licence fee is increased substantially (or potentially introduced in places where it does not exist) a reduction in tobacco points of sale is likely to occur as a result.
Acknowledgments
The authors would like to thank the Health Protection Operations Section for providing access to the data.
References
Footnotes
-
Contributors JAB and CLM conceptualised and designed the study; JAB and DLJ sourced the data; DLJ and JD cleaned data and conducted statistical analyses; JAB, JD, DLJ, and CLM drafted and revised the paper.
-
Funding This project was grant funded by SA Government.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.