<<

Chapter 4 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

Introduction

The widespread implementation of to go smoke-free should be left to in more detail, after a discussion of smoke-free policies in many countries the businesses themselves and that the strengths and limitations of the has been slowed by fears that if these policies are good for their methods used in these studies. This restrictions on may have an establishment, owners will voluntarily is followed by a brief review of the adverse impact on businesses. It is adopt them. The limited evidence on other incidental clear, however, that allowing smoking has supported these arguments with and/or unanticipated effects of in the workplace adds considerable claims that smoke-free policies result smoke-free policies not covered in costs for businesses. Lost productivity in lost revenue for , bars, other chapters of this Handbook. results from disease and premature and other hospitality establishments; caused by smoking and fewer jobs in the hospitality sector; Potential costs and benefits exposure to in the and business closings (KPMG, to businesses of smoke-free workplace. Establishments which 1998; Deloitte & Touche LLP, 2003; policies allow smoking face higher health U.S.Departement of Health and Hu- and hazard premiums, man Services, 2006). This strategy, smoking among employees and cleaning and maintenance of making claims about the harmful and customers imposes a variety of costs. Those that restrict smoking to economic effects of costs on businesses, ranging from designated areas assume the costs of policies, is not unique to the industry’s lost productivity among employees and maintaining them. attack on smoke-free policies. It has to higher insurance, cleaning, and Particularly prominent in the also been used in opposition to other other costs (Hallamore, 2006; Javitz debate over smoke-free policies have policies, including higher tobacco et al., 2006). Businesses can incur been concerns about the economic taxes and comprehensive bans on costs, however, from policies limiting impact on restaurants, bars, and other tobacco product advertising and or banning smoking in the workplace. hospitality sector establishments. promotion (Jha & Chaloupka, 1999; For instance, there are the costs of Some and bar owners, Chaloupka & Warner, 2000). creating and maintaining smoking for example, express concerns that This chapter reviews the evidence areas, potential lost productivity due smoke-free policies will drive their on the costs to businesses of allowing to increased smoking breaks for smoking patrons to other venues smoking in the workplace, and of smoking employees, and the loss where smoking is allowed (particularly the potential costs and benefits of business from customers who those in nearby jurisdictions without to businesses that restrict or ban smoke. Policy implementation and smoke-free policies), or lead them workplace smoking. The extensive enforcement costs will be shared to cut their visit short or even stay and growing literature on the econ- by businesses and governments. home, reducing their establishments’ omic impact of smoke-free policies Nonetheless, these policies can business. Others argue that decisions on the hospitality sector is reviewed significantly reduce the expenses

75 IARC Handbooks of Cancer Prevention to businesses that more than offset productivity on the job, resulting from While the subject of considerable any costs that result from their the diseases caused by smoking discussion, limited empirical evidence implementation. The potential costs and exposure to tobacco smoke; exists on the magnitude of these costs and benefits for businesses of time spent on smoking breaks by to businesses, particularly those in smoke-free policies are summarised smoking employees; increased developing countries. Briefly, existing in Table 4.1. This section briefly health and costs for evidence includes: reviews the limited evidence on these employees; increased fire and hazard costs and benefits; those related to insurance costs; higher cleaning and • Lost productivity from health gains and losses to businesses in maintenance costs; and the potential consequences of smoking: In the hospitality sector, that result from for significant legal costs resulting from a recent study from Sweden, changes in patronage by smoking claims filed by employees seeking using nationally representative and nonsmoking customers, are compensation for damages caused data from 1988 through 1991, described in more detail in the by exposure to tobacco smoke in the the Swedish Survey of Living following section. workplace, or by customers seeking Conditions was linked to register- protection from tobacco smoke. The based data on the number of days Costs of smoking to businesses relative magnitude of costs will vary missed from due to sickness, by type of businesses that have from the National Board of Social A growing body of research clearly many smoking patrons (e.g. bars, Insurance (Lundborg, 2007). illustrates the costs to businesses restaurants), com-pared to those It was estimated that smokers that allow smoking by employees where the costs are primarily from a were absent between 7.7 and and customers. These costs include: limited number of smoking employees 10.7 days more each year than increased absenteeism and reduced (e.g. white collar offices). were nonsmokers. Based on a

Table 4.1 Potential costs and benefits to businesses of smoke-free policies

Costs Benefits

Lost business due to smokers visiting less frequently or cutting visits Increased business from nonsmokers visiting more frequently or staying short longer

Costs of establishing and maintaining smoking lounges for smoking Reduced cleaning and maintenance costs employees

Implementation and enforcement costs Reduced fire, accident, and life insurance premiums

Lost productivity due to increased or longer smoking breaks for smoking Increased productivity as smoking employees quit or cut back and employees require fewer smoking breaks

Increased productivity due to reduced absenteeism and improved health Costs of establishing and maintaining smoking areas for patrons among smoking employees

Increased productivity due to reduced absenteeism and improved health among nonsmoking employees

Reduced health care costs from reductions in smoking among smoking employees

Reduced health care costs from reductions in exposure to secondhand smoke among nonsmoking employees

Avoidance of potential litigation costs from nonsmoking and smoking employees and/or customers

76 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes telephone survey of 200 randomly • Lost productivity from exposure Penner, 1990). A thorough review selected Scottish businesses to secondhand smoke: As estimated that fire insurance with 50 or more employees, described in Chapter 2, there is costs were US$11-21 higher per linked to evidence on the costs strong evidence that exposure smoker in the USA (Javitz et al., of smoking drawn from a review to secondhand tobacco smoke 2006), while fire insurance costs of the literature, it was estimated (SHS) causes a variety of health attributable to smoking for Scottish that absenteeism among Scottish consequences in nonsmokers. workplaces were estimated to be smokers reduced productivity Among nonsmoking workers, the approximately £4 million annually by £40 million, while productivity death and disease caused by this (Parrott et al., 2000). Similarly, losses due to the premature exposure in the workplace leads smoking increased life insurance death caused by smoking totaled to additional lost productivity premiums by CA$75 per smoking approximately £450 million in and increased health care costs employee (Conference Board of 1997 (Parrott et al., 2000). More for businesses. To date, only Canada, 1997), while the cost comprehensive estimates of the one study has estimated these to a business of providing US$ lost productivity costs resulting costs. In 2005, using the same 75 000 in life insurance was an from premature deaths caused well developed methods used to approximately additional US$90 by smoking, based on well estimate the lost productivity costs per year for a smoking employee developed methods for estimating resulting from premature death (Javitz et al., 2006). economic expenditures, have caused by smoking, the Society been produced for many other of Actuaries (Behan et al., 2005) • Increased cleaning and developed countries, including estimated that SHS exposure maintenance costs: The US Australia (Collins & Lapsley, 1996, increased health care costs in the Environmental Protection Agency 2002, 2008), Canada (Kaiserman, USA by nearly US$5 billion, and (Mudarri, 1994) estimated that 1997), Ireland (Madden, 2003), led to an additional almost US$5 the adoption of a comprehensive the USA (Centers for Disease billion in lost productivity, due to smoke-free policy in 1994 would Control and Prevention, 2005a), lost wages, fringe benefits, and have reduced building operation and a growing number of others. value of services. This clearly and maintenance costs for US underestimates lost productivity businesses by US$4.8 billion per • Lost productivity from smoking costs to businesses, as it does not year, based on detailed estimates breaks: Based on a comprehensive account for the lost productivity of the costs of cleaning and review of existing literature on the due to work days missed from maintaining different types of costs to employers resulting from diseases caused by smoking. workspaces (office, assembly, smoking in the workplace, it was and /industrial space). estimated that smoking employees • Higher insurance premiums: These figures, updated to account take an additional four to thirty Similarly, studies have doc- for inflation, estimated that in minutes in break time each day umented the higher costs of 2005 the additional smoking- for on-the-job smoking (Javitz et insurance coverage for smoking related costs per 1000 square al., 2006). Using similar estimates, employees and/or workplaces that feet of workspace ranged from the Conference Board of Canada allow smoking. For example, in the US$305 for warehouse space to (Hallamore, 2006) estimated that USA, using data on paid health US$728 for office space (Javitz et unsanctioned smoking breaks cost care claims for a large group al., 2006). Canadian employers an average indemnity plan, it was estimated of CA$3053 per year in 2005. that average health care insurance • Potential litigation costs: There premiums for smoking employees are a variety of potential legal chal- were about 50% higher than lenges businesses may face as a those for nonsmokers (Penner & result of allowing smoking in the

77 IARC Handbooks of Cancer Prevention

workplace (Sweda, 2004).These Costs of smoke-free policies of persons using it. The costs range from claims for workers to businesses and governments of establishing seven smoking compensation and disability rooms for employees, patients, benefits, resulting from exposure While allowing smoking in the work- and visitors in a new building to smoking in the workplace, to place results in potentially significant being constructed by the Royal lawsuits from customers arguing expenses to businesses, policies that Victoria in Belfast, were that persons sensitive to smoke limit or ban workplace smoking are estimated to be approximately are being discriminated against not without cost. This is particularly £500 000 (McKee et al., 2003). by being denied the ability to true for policies that allow for smoking Similarly, it was reported that the enjoy a smoke-free environment. in restricted or designated areas. cost of creating seven smoking In the USA, hundreds of legal These costs include: rooms for employees and travelers challenges document key suc- in St. Louis’ Lambert Airport in cesses in litigation brought • Costs for smoking areas: 1997 was US$450 000 (Manor, on by those exposed to SHS. As described in Chapter 3, 1997). In response to comprehen- Anecdotal evidence indicates smoke-free policies vary in how sive smoke-free policies, some that similar legal disputes have restrictive or permissive they businesses have turned to build- been successful in a variety of are with respect to allowing for ing “smoking huts” or “smoking other countries. While the award smoking areas. Some workplace shacks” (partially enclosed amounts in these cases vary policies permit the creation of shelters to accommodate their widely, it is clear that the potential designated smoking rooms smoking employees and patrons). costs of unsuccessfully defending for smoking employees, while These can range in cost from less against these litigations can be others completely ban smoking than US$2000 for small, no-frills significant. in all enclosed areas, but allow shelters, to tens of thousands of smoking in non-enclosed areas; dollars for larger shelters with In summary, cigarette smoking others designate smoking and more amenities (Ford, 2008). imposes significant costs on nonsmoking sections for their businesses, which can be customers. If employers opt to • Implementation and enforce- considerably reduced if policies provide a smoking area for their ment costs: In general, within a that restrict or ban smoking in the employees and/or patrons, there few months of implementation, workplace are enacted. Some costs will be expenses associated compliance with smoke-free can be entirely avoided by complete with building and maintaining policies is high, and the policies bans on smoking in the workplace, these areas. Costs will vary become self-enforcing in most but only somewhat reduced by considerably depending on the places that have adopted them (see more limited restrictions (i.e. the size of the area, whether or not Chapter 5 for a thorough discussion need would still exist for cleaning it is enclosed, how it is ventilated, of support for and compliance and maintenance due to smoking). local costs, and other with smoke-free policies). Nearly Other costs will be reduced (e.g. lost factors. The Ontario Campaign all of the costs of implementation productivity and higher insurance for Action on Smoking (Wong, and enforcement will be taken costs), as employee smoking de- 2002), estimates that the cost of on by governments rather than clines (as discussed in Chapter 7) establishing a designated smoking by businesses. In addition to the and nonsmoking workers’ exposure room for use by employees who costs associated with creating and to tobacco smoke in the workplace smoke can range from about maintaining designated smoking falls (as discussed in Chapter 6) in CA$55 000 to over CA$268 000, areas, as described above, response to smoke-free policies. with monthly maintenance fees expenses to businesses are largely adding CA$200-600, based on the limited to signage and minimal size of the room and the number enforcement costs.

78 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

• Lost productivity: Some suggest (those that respond to complaints Summary that smoke-free policies will result from non-smokers) are relatively in lost productivity, as employees less costly, while more “proactive” Cigarette smoking in the workplace who smoke will take more enforcement efforts (those involving imposes a variety of costs on smoking breaks in response to active compliance checks) will be businesses, including the lost the policies. Others argue that more expensive. Greater proactive productivity resulting from smoking employees who smoke may be enforcement efforts, however, seem breaks, absenteeism, and premature less able to concentrate and less likely to raise compliance more deaths; higher health care and productive if their opportunities to quickly, suggesting that they will be other insurance costs; increased smoke during working hours are needed for a shorter period, making maintenance and cleaning costs; and limited. To date, there is no good it unclear which approach is more potential costs of litigation. Smoke- empirical evidence to support cost-effective in the intermediate free policies will reduce the costs to either issue. These policies could to long-term. For example, WHO businesses associated with work- raise the costs of smoking breaks recommends proactive enforcement place smoking. to businesses, by forcing smokers efforts in the first weeks and months To date, little solid evidence outdoors, and thus, increasing after the implementation of a smoke- exists about the costs of smoke- their time away from work. How- free policy (WHO, 2007b). free policies to businesses and/or ever, this increase in costs is likely Limited data from hospitality governments. While there is much to be offset by the reductions in sector employees in Norway indicate speculation about these costs, most time lost for smoking breaks by that greater enforcement may be appear minimal, short-term, and/or some smokers who quit or cut needed for smoke-free policies that likely to be offset by reductions in back in response to the policy (as restrict, rather than comprehensively related costs. It does appear, how- described in Chapter 7). Similarly, ban, smoking given greater non- ever, that the costs of a complete reductions in productivity among compliance with the partial policies ban on smoking will be lower than smoking employees, when they (Hetland & Aarø, 2005). the costs of policies that allow for are denied the opportunity to As discussed in Chapter 5, some smoking in designated areas, given smoke while working, are liable evidence suggests that compliance the costs of maintaining these areas, to be offset by the productivity with smoke-free policies is enhanced the remaining exposure that results, gains that accrue from reductions by investments in media advocacy and the greater need for enforcement in absenteeism and premature and public campaigns of these partial restrictions. More deaths caused by smoking. that strengthen social norms against research is needed to fully under- smoking, before and/or during the stand the costs to businesses Costs to governments implementation of these policies and governments of adopting, of implementing and enforcing (Ross, 2006; U.S. Department implementing, and enforcing smoke- smoke-free policies of Health and Human Services, free policies. 2006). This implies that more active The relatively quick compliance with enforcement of these policies, and/or Impact of smoke-free policies on smoke-free policies in most countries greater investments in the hospitality sector suggests that the implementation and campaigns, may be needed in some enforcement costs to governments developing countries where anti- The most vigorous debate over the will likely be short-term. Little smoking norms are weaker (Ross, economic impact of smoke-free empirical evidence is available on the 2006). policies has been with respect to the costs or cost-effectiveness of efforts business activities of restaurants, to enforce smoke-free policies. bars, gaming establishments, and Anecdotal evidence suggests that other firms in the hospitality sector. “reactive” enforcement efforts The debates center on the potential

79 IARC Handbooks of Cancer Prevention for lost revenues resulting from in this sector (Scollo et al., 2003; • Inclusion of data from smokers visiting these establishments Scollo & Lal, 2005, 2008). comparable jurisdictions where less frequently (or forgoing visits Studies of the impact of smoke-free no policy changes occurred altogether), cutting their visits shorter policies on the hospitality sector vary that can act as controls for the and spending less money than they considerably in their methodological jurisdiction(s) where the policy would have if smoking were permitted, quality, with the best of these studies change(s) being assessed took and/or taking their business to sharing most or all of the following place. establishments in jurisdictions that characteristics: do not have similar policies. Many of While many of the studies to these arguments have been voiced • Use of valid, reliable measures date share these characteristics, by the tobacco industry or various of business activity (e.g. official others do not. The findings from groups supported by the industry (e.g. reports of tax or business studies that use less reliable data, smokers’ rights associations, local revenues, employment, and/or fail to control for overall economic restaurant and/or bar associations) the number of licensed estab- activity, or otherwise deviate from (U.S.Departement of Health and lishments; population level, these guidelines, are mixed in their Human Services, 2006). representative survey data) that conclusions about the economic In nearly all countries, however, can be used to detect the real impact of smoke-free policies. In the number of nonsmokers in the impact of a change resulting from contrast, as described below, the population exceeds the number of the adoption of a smoke-free findings from studies with these smokers. This raises the likelihood policy; characteristics consistently find that that any revenues lost from smoke-free policies have no negative changes in smokers’ patronage • Use of data for several years economic impact on restaurants, bars, will be offset by greater revenues covering the period before and and other segments of the hospitality from nonsmokers increasing their after the implementation of a industry, with the possible exception patronage of businesses who enact smoke-free policy, in order to of gaming establishments. Indeed, smoke-free policies. As described separate out the impact of the many studies provide evidence of a in Chapter 3, smoke-free policies policy from underlying trends in small positive effect of smoke-free have been widely adopted in recent business activity, and to allow policies on business activity. years, generating a series of natural sufficient time for businesses, experiments that allow researchers smokers, and nonsmokers to Studies based on official reports to assess the impact of smoke- adapt their behaviour to the of business activity free policies on business activity, policy; attitudes towards these policies (as Studies based on sales data described in Chapter 5), on exposure • Use of appropriate statistical – restaurants and bars to SHS (as described in Chapter methods that include controls 6), and on smoking behaviour for underlying trends in the data, A large and rapidly expanding (as described in Chapter 7). With and other factors that lead to literature uses measures of taxable respect to business activity, over 160 fluctuations in business activity sales, sales tax revenues, or other studies have examined these issues, (most notably, overall economic official reports of sales data, to applying diverse analytic methods conditions), and that apply assess the economic impact of to a variety of data from hospitality appropriate tests for the statistical smoke-free policies on restaurants sector businesses in numerous juris- significance of the relationship and bars. Many of these studies dictions, and they have been compiled between the policy and measure include appropriate controls for (Scollo & Lal, 2008) in an update of of business activity; underlying economic conditions the previous comprehensive review and/or other jurisdictions where no on the impact of smoke-free policies policy changes occurred, and most

80 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes apply appropriate statistical methods that adopted them. Similarly it was (Blecher, 2006). Specifically, under to data for several years before and concluded that smoke-free bar policies the new policy, restaurateurs were after the policy change of interest. had no economic impact on bars. given the option of going entirely The first of these studies Comparable studies, based on smoke-free or creating separately examined the impact of local smoke- sales data from restaurants and/or ventilated smoking (in up to 25% free restaurant policies adopted in 15 bars, have been done for different of the restaurant) and nonsmoking California and Colorado communities jurisdictions in the USA. These sections. Using annual provincial between 1985 and 1992 (Glantz studies reflect the diversity of the USA value-added tax (VAT) revenues & Smith, 1994). Fifteen nearby geographically, demographically, for restaurants from 1995 through communities without a smoke-free socioeconomically, and with respect 2003, alternative models were restaurant policy were included as to the strength and estimated that controlled for overall controls, with selection of the control control efforts, from North Carolina economic conditions, and, in one, communities matched to communities counties in the heart of the US the efficiency of VAT tax collection. It where policy changes occurred based tobacco growing region (Goldstein & was concluded that “the restrictions on population, urbanicity, median Sobel, 1998), to large cities or states placed on smoking in restaurants in household income, and smoking like New York City (Hyland et al., 2001 have had at worst no significant prevalence. Using linear regression 1999) and New York State (Engelen effect on restaurant revenues, and at methods applied to a measure of et al., 2006). Likewise, a growing best a positive effect on revenues.” taxable restaurant sales revenues number of studies have used bar as a share of total revenues before and/or restaurant sales data from Studies based on employment data and after the implementation of the developed countries, including – restaurants and bars local policies, the authors concluded Australia (Wakefield et al., 2002; Lal that businesses were not adversely et al., 2003, 2004), Canada (Luk et Several studies use measures of affected in the communities that al., 2006), Norway (Lund et al., 2005; employment to assess the economic adopted and implemented policies Lund, 2006; Lund & Lund, 2006), and impact of smoke-free policies banning smoking in restaurants. New Zealand (New Zealand Ministry on restaurants and bars. These A few years later, the 1994 of Health, 2005; Thomson & Wilson, measures include direct counts of analysis was extended to include 2006; Edwards et al., 2008). Nearly employed persons, and more indirect three additional years of data for the all of these studies reached the same measures, such as official reports 30 communities originally analysed, general conclusion: that smoke-free of unemployment, insurance claims, and to add five cities and two counties policies do not adversely affect the and payroll tax collections. As with that had adopted smoke-free bar business activity of restaurants and the studies that use measures of policies between 1990 and 1994, with bars, with several providing evidence sales, the best of these studies will comparable control communities/ of a small positive impact of the policy control for underlying economic counties for all but one of these (Glantz on sales. conditions, include several years of & Smith, 1997). A few minor errors In contrast, given the slower pre- and post-policy change data, in the coding for the implementation diffusion of these policies to include similar control jurisdictions dates of the policies included in the developing countries described in where no policy changes occurred, earlier study were corrected. Using Chapter 3, almost no studies exist on and employ appropriate statistical similar outcome measures, linear the economic impact of smoke-free methods. and non-linear regression methods policies in these countries. The one To date, all studies using were applied to both the matched exception is an analysis of the impact employment-based outcomes have data and pooled data, confirming an of the 1999 amendments to South been conducted for jurisdictions earlier finding that the smoke-free Africa’s tobacco control policies that in developed countries. Given the restaurant policies did not adversely introduced restrictions on smoking relatively early diffusion of smoke- affect restaurants in the communities in restaurants beginning in 2001 free policies at the local level in the

81 IARC Handbooks of Cancer Prevention

USA, the majority of the studies and restaurants rose by 9%, and value based on the sales price of have focused on this locale. Findings by 24% in drinking establishments, establishments that were sold during from studies that meet the standards while employment in clubs fell by 8% the periods examined. For the 608 described above are quite consistent following the implementation of New restaurants sampled, 118 were in with the results from studies based Zealand’s comprehensive smoke- smoke-free jurisdictions, and sold on measures of sales. They generally free policy in late 2004 (Thomson & between 1991 and 2002; for the 197 conclude that smoke-free policies Wilson, 2006). bars sampled, 17 were in smoke-free have had either no significant jurisdictions, and sold between 1993 impact or a small positive impact on Studies based on the number of and 2005. Controlling for underlying employment. businesses – restaurants and bars economic conditions (using measures For example, the impact of the of gross state product and state level comprehensive smoke-free policy Other studies of the economic impact unemployment rates for the state in implemented in April 2004 by of smoke-free policies have used which each establishment was lo- Lexington-Fayette County , various measures of the numbers of cated), type of establishment (e.g. fast in the middle of one of the largest restaurants and bars, such as counts versus full- restaurant), tobacco growing and of businesses, business openings and general trends, it was concluded states in the USA, was studied (Pyles and closings, and the number of that the value of restaurants was 16% et al., 2007). It was concluded that bankruptcies, with the findings higher in smoke-free jurisdictions than restaurant employment rose after from studies that meet the criteria in those that allowed smoking, while the policy was put in place, while described above consistent with the value of bars was unaffected by bar employment was unchanged. those based on measures of sales policies banning smoking. In addition, restaurant and bar and employment. For example, in the employment in neighboring counties study of the impact of the Lexington- Studies based on revenue data was unaffected, a finding inconsistent Fayette County Kentucky smoke-free - gaming establishments with opponents’ arguments that policy, it was concluded that there smokers would take their business was no significant impact of the policy In contrast to the relatively large to nearby jurisdictions that allowed on business openings and closings, literature using objective measures smoking following the Lexington- both for those that served alcoholic to assess the economic impact of Fayette County . beverages and for those that did smoke-free policies on restaurants Though relatively few non-US not (Pyles et al., 2007). Similarly, in and bars, there are relatively few studies have looked at the impact of a study of the Ottawa smoke-free studies that have examined the smoke-free policies on employment policy, bankruptcy and insolvency impact on gaming establishments. outcomes, the methodologically indicators were found to be lower This is largely the result of the sound studies have come to the in the period immediately after the exclusion of these establishments same basic conclusions reached by ban was implemented than in the from the venues covered by most the ones from the USA. For example, two years prior to the ban (Bourns & smoke-free policies, as described in analysis of Ottawa, Canada’s August Malcomson, 2001). Chapter 3. Nevertheless, there are a 2001 ban on smoking in restaurants, few studies that have looked at the bars, and pubs found that employment Studies based on business value impact of smoke-free policies on in affected businesses rose in the – restaurants and bars various gaming activities. period immediately following the The first study to examine the im- ban, while unemployment insurance Two innovative studies looked at pact on gaming venues, investigated claims fell, despite an overall the impact of smoke-free policies the effects of local smoke-free policies decline in employment (Bourns & on the value of restaurants (Alamar in Massachusetts that limited smoking Malcomson, 2001). Similarly, it was & Glantz, 2004) and bars (Alamar in bingo halls and at gambling events concluded that employment in cafes & Glantz, 2007) using a measure of sponsored by local charities (Glantz

82 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

& Wilson-Loots, 2003). It was found (Mandel et al. used relatively simple to SHS, it was found that Victoria’s that increases in other gambling weighted regression methods based smoke-free policy was effective in opportunities led to reductions in on the number of video lottery reducing problem gambling, and that bingo and charitable gambling event machines, while Pakko used a more the money gamblers did not spend profits, but that the magnitude of the general approach to accounting gambling would likely be spent in drop was not related to the presence for the heteroskedasticity that also other sectors of the economy. or absence of a smoke-free policy. corrected for the serial correlation Clearly, more research is needed More recently, a similar approach in the data). The approach used by to sort out the economic impact was used to examine the impact of Pakko appears more appropriate of smoke-free policies on gaming Massachusetts’ state-wide smoking than that used by Mandel and establishments, both on the gaming ban implemented in mid-2004 on colleagues, and is robust to other sector directly, as well as the broader Keno sales. This report concluded specifications including those that economic impact. As described in that there was no impact of the ban replace the quarterly indicators for Chapter 3, the spread of increasingly on this type of gaming (Connolly et seasonality with monthly indicators, comprehensive smoke-free policies al., 2005). and that drop 1996 (the year the that ban smoking in casinos and in Two recent studies considered the three racinos opened, which appears other gaming establishments will effects of Delaware’s comprehensive to account for the heteroskedasticity provide new, natural experiments smoke-free policy that covered three in the data). allowing researchers to assess the horse racing tracks that allowed Most recently, the effects of economic impact of these policies on video lottery gambling (“racinos”). Victoria, Australia’s policy banning the gaming sector. The first of these studies (Mandel smoking in most gaming venues, et al., 2005) concluded that the implemented in September 2002, Studies based on revenue and/or state smoking ban had no impact on was studied (Lal & Siahpush, 2008). employment data – other hospitality either total revenues from the video Interrupted time series methods were sector establishments lottery machines, or on the average applied to monthly expenditures on revenues per machine; a subsequent electronic gaming machines (EGM), Finally, several studies have used re-analysis that corrected for a data from July 1998 through December measures of economic activity in entry error and for heteroskedasticity 2005, for both Victoria and South other parts of the hospitality sector (unequal variance in the error term) Australia (their control jurisdiction). It to evaluate the financial impact of in the data (Glantz & Alamar, 2005) was concluded that the Victoria policy smoke-free policies. These studies confirmed the findings from the led to “an abrupt, long-term decrease have generally focused on the impact original analysis. In contrast, the in EGM expenditures” of about 14%, of the policies on , using same data was used and reached the comparable to the 13% decline measures of revenues generated by opposite finding: that the Delaware estimated for the Delaware racinos and motels and/or employment smoking ban reduced gaming (Pakko, 2006). The report goes in these establishments. The most revenues by nearly 13% in the year on to note that the decline in EGM methodologically sound of these following the implementation of the expenditures was much larger than studies share the characteristics of ban (Pakko, 2006). The differences observed at Victoria’s casino, which the best studies described above, and in findings are accounted for by was also covered by the smoking ban, are consistent in their conclusions alternative approaches to modeling but subject to a number of exemptions that smoke-free policies do not have the seasonality in the data (Mandel that allowed the proprietor to cater an adverse economic impact on et al. included an indicator for winter to high-spending patrons in private these segments of the hospitality months only, while Pakko included rooms. Also, employment in Victoria’s industry. indicators for winter, spring, and gaming sector was at historically A good example of research on the summer), and by differences in high levels three years after the tourism sector is a study which looked the statistical methods employed ban. In addition to curbing exposure at measures of revenues (both

83 IARC Handbooks of Cancer Prevention absolute revenues and revenues as A number of patron or consumer business owners/managers to either a share of total sales revenues) surveys collected information on fear that the ban will have a negative and number of tourists (Glantz & actual dining/drinking out behaviour impact on their business, or to attribute Charlesworth, 1999). Data were before and after a policy change, while any declines in business after the examined before and after the some pre-implementation surveys policy implementation to the policy adoption of comprehensive smoke- inquired about anticipated changes change, rather than to other factors free policies in three states (California, in behaviour in response to the that may account for the decline Utah and Vermont) and six cities policy. In some post-implementation (Glantz, 2007). Similarly, researchers (Boulder, Colorado; Flagstaff and surveys, individuals were asked observed that “it seems likely that Mesa, Arizona; Los Angeles and San about actual changes in behaviour owners of businesses that are faring Francisco, California; and New York resulting from the policy change, poorly in a highly volatile market may City). It was found that hotel revenues while others gathered information on be more likely to blame external forces grew faster following the adoption of respondent’s preferences for smoke- (such as the adoption of a smoke-free the smoke-free policy in four of these free dining/drinking areas, and other policy) rather than their own business jurisdictions, the rate of growth was related attitudes and perceptions. decisions for their problems” (Eriksen unchanged in four others, and that Similarly, surveys of business owners & Chaloupka, 2007). revenues grew more slowly (but did or managers collected self-reported Given the potential for biased not fall) in the last. In these analyses, information on business revenues responses, particularly in surveys which pooled the data from the that, in general, were not validated, as of business owners or managers, it communities, no significant impact of well as owner/manager perceptions is not surprising that these studies the policy adoption on either measure of the impact of the policy on their are much more likely to conclude of revenue was detected. Finally, business (either anticipated or that the economic impact of smoke- in analyses that used measures of realised), in addition to their attitudes free policies is negative. In a the number of international tourists about the policies. comprehensive review of studies visiting the three states studied, there In addition to meeting the other published through August 2002 was either no impact of the policy criteria described above, the best (Scollo et al., 2003), it was noted or the number of visitors increased of the studies based on survey data that studies based on this type of following the implementation of the used appropriate sampling and sur- survey data are four times more policy. Given these findings, the re- vey methods to collect validated likely to conclude that these policies searchers concluded that “smoke- measures of relevant outcomes. For have a negative economic impact, free ordinances do not appear to example, a convenience survey of bar than are studies based on official adversely affect, and may increase, patrons, prior to the implementation reports of sales, employment, and tourist business.” of a ban on smoking in bars, related data. Despite the potential for that asks about their anticipated bias, the majority of studies based Studies based on survey data response to the policy, is much more on survey data, particularly those likely to provide biased estimates based on patron/consumer survey A second set of studies has relied on of the impact of the policy than data, conclude that there is either no measures drawn from survey data would randomly selected consumer impact or a small, positive economic to assess the economic impact of surveys, representative of the local impact from smoke-free policies. smoke-free policies on hospitality population, that collect actual data on Studies that employ survey data sector businesses. These studies bar patronage conducted before and come from a wider variety of countries include data from surveys of patrons after the policy implementation. The than the studies based on sales, or more representative, population vociferous debate over potentially employment, and other related data level consumer surveys, and surveys adverse economic effects of smoke- described above, including Australia, of owners/managers of businesses free policies can create a “negative Canada, Hong Kong, Ireland, Italy, affected by the smoke-free policy. placebo effect” that leads some New Zealand, Norway, Scotland,

84 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

South Africa, Spain, the UK, and smokers were much more likely to visiting by the smoky environment, the USA. However, as seen by this report a decrease in their frequency and, as a result, do not pick up the list, these studies add relatively little of visiting affected establishments, increases in their patronage after information on the economic impact with 42% of daily smokers and 10% policy implementation that offsets of smoke-free policies in developing of occasional smokers reporting any anticipated reductions in patron- countries, as few of these countries reduced frequency, as compared to age among existing customers. have adopted such policies. Given 2% of nonsmokers. Given the much Moreover, the anticipated reduction the potential misuse of studies based greater prevalence of nonsmokers in patronage that smokers describe on flawed survey data, this section among Norwegian adults, 12% of the may not end up happening after the reviews methodologically sound overall sample reported an increase implementation of a smoking ban, studies and highlights the potential in their frequency of visiting hospitality as opportunities for smokers to go biases that result from the use of venues, while 12% reported a to alternative venues are limited, unrepresentative survey data and/or decrease. The majority of the resulting in few smokers actually unreliable measures. population (76%) reported no changes altering their establishment patron- in their patronage in response to the age in response to the ban (in contrast Studies based on consumer/patron smoke-free policy. to nonsmokers who are increasingly surveys These findings are consistent attracted to now smoke-free venues) with those from other studies that use (Cowling & Bond, 2005). Analyses of survey data from Nor- population-based consumer surveys way, collected before and after the to assess the economic impact of Studies based on owner/manager June 1, 2004 implementation of the smoke-free policies in a variety of surveys country’s ban on smoking in bars other jurisdictions. In general, these and restaurants, helps explain the studies find that the implementation The studies most likely to conclude consistent finding from studies based of the policy has no significant impact that smoke-free policies have a on sales and employment data that on dining/drinking out practices, and negative economic impact on the smoke-free policies do not have an that any reductions in the frequency businesses targeted by the policies adverse economic impact (Lund of such practices among smokers are those based on surveys of et al., 2005; Lund, 2006; Lund & are made up for (often more than business owners and managers. Lund, 2006). Annual representative made up for) by increases in the Many of these studies are based on surveys of Norwegian adults showed frequency of dining/drinking among proprietor expectations, rather than no significant changes in the nonsmokers. on the actual impact on business frequency of pub/bar and restaurant These studies also illustrate the after the smoke-free policy has been visits following the implementation bias that results from convenience implemented. Studies based on pre- of the ban. Responses to the post- samples of hospitality sector cus- implementation surveys of business ban survey question “has the ban tomers prior to a change in policies. owners/managers appear most likely on smoking in hospitality venues For example, a survey of current to be subject to the “negative placebo changed your patronage habits?” patrons’ anticipated responses to a affect” (Glantz, 2007). In contrast, illustrate the differences in the impact proposed ban on smoking in Hong well-designed, post-implementation of the ban on patronage by smokers Kong restaurants, bars, and cafes surveys of business owners/man- and nonsmokers. Nonsmokers were was administered, and concluded agers, which collected more valid significantly more likely to report that that the ban would reduce revenues measures of business activity, often the ban increased their frequency of for these businesses by more than concluded that their businesses visiting hospitality venues, with 18% of 10% (KPMG, 2001). In general, were not negatively affected by the nonsmokers reporting an increase, as these types of convenience surveys policy. The differences between the compared to 1% of daily smokers and of current patrons do not include perceived and actual business impact 3% of non-daily smokers. In contrast, the nonsmokers deterred from of smoke-free policies is clearly

85 IARC Handbooks of Cancer Prevention illustrated by analysis of the Québec reported a decrease in business Industry-sponsored studies policy limiting smoking in restaurants following its implementation; however, (Crémieux & Ouellette, 2001). Based illustrating the importance of including The tobacco industry, and groups that on a survey of restaurateurs, that appropriate controls, they also found support it, have been vocal opponents included both those in compliance that 34% of small restaurants and 36% of smoke-free policies, arguing among and those not in compliance with the of restaurants with significant income other things that these policies will ad- policy, the researchers concluded from alcoholic beverage sales also versely affect the businesses covered that “impacts on… restaurant reported a decline in business. Given by the policies (U.S.Departement of patronage were widely anticipated this, they concluded that “there is no Health and Human Services, 2006). but not observed.” evidence to suggest that the smoke- Many of the studies researchers The most methodologically sound free law has had a detrimental effect reviewed (Scollo & Lal, 2008) have studies in this group are those based on the city’s restaurant business.” been funded by the tobacco industry on representative surveys of business As with the studies based on (e.g. through the Accommodation owners/managers conducted long sales, employment, or related data, Grant Program) or by groups supported enough after the implementation the only evidence from developing by the tobacco industry (e.g. various of the policy for its impact on their countries based on survey data comes bar and/or restaurant associations that business activity to be clear. Ideally, from South Africa’s experiences received funding from the industry). such studies would also include similar following its 1999 Tobacco Products In a comprehensive review of the surveys in comparable jurisdictions Control Amendment Act that limited existing literature through August where no policy changes occurred, smoking in restaurants and other 2002, an assessment was made of the and/or other approaches to control public places. Between November association between funding source for general trends and underlying 2004 and January 2005, a survey and study findings (Scollo et al., 2003). economic conditions, in order to was conducted of 1431 restaurant It was concluded that all studies that isolate the effects of the policy from owners/managers (1011 surveys were found smoke-free policies to have a those of other factors. Few of these successfully completed) identified by negative economic impact had been studies, however, used this approach. searching online, publicly accessible funded by the tobacco industry, an Analysis of the economic impact databases (van Walbeek et al., 2007). organisation that had received tobacco of New York City’s 1995 smoke-free This survey gathered data both on the industry funding, or an industry ally. In policy is one of the small numbers of costs of complying with the policy and addition, the vast majority (94%) of the studies that uses a representative on restaurant revenues. Based on the industry-supported studies concluded sample of restaurants and includes retrospective reports of restaurant that smoke-free policies had a negative appropriate controls (Hyland & owners/managers, it was found that economic impact. Cummings, 1999). Since this policy revenues in most restaurants (59%) An updated review available only applied to restaurants with more were unchanged following the policy, through January 2008, includes more than 35 seats, and did not cover while 22% of restaurants reported an recent non-industry funded studies that establishments that generated at increase in revenues and 19% reported find that there is a negative economic least 40% of their revenues from the a decrease. Some differences were impact of smoke-free policies (most sale of alcoholic beverages, small observed across restaurants, with notably for gaming establishments), restaurants and restaurants with franchised restaurants more likely but there continues to be a strong bars were included as control groups to report an increase in revenues association between industry funding for the larger restaurants that were and independent restaurants more (either direct or through affiliated affected by the policy, as all of these likely to report a drop. Given this, it organisations) and the conclusion would have been subject to the same was concluded that there was no that the policies negatively affect the underlying economic conditions. net negative impact of South Africa’s businesses they cover (Scollo & Lal, The researchers found that 35% of smoking restrictions on restaurant 2008). the restaurants subject to the policy business.

86 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

Summary of research on the underlying economic conditions. can impact a variety of other economic impact of smoke-free These studies were more mixed in behaviours and related outcomes. policies on the hospitality sector their findings, with 18 concluding that Some of these are covered in other the policies had either no economic chapters, including the impact of the As of January 2008, 165 studies impact or a small positive effect, and policies on attitudes towards tobacco were identified which examined the 19 concluding that they had a negative and related social norms (Chapter 5), economic impact of smoke-free impact. Given their limitations, it is on exposure to tobacco smoke and policies on the hospitality sector not surprising that only three of these its health consequences (Chapter 6), (Scollo & Lal, 2008). Eighty-six of 37 studies were published in peer- and on smoking behaviour (Chapter these studies employed official reviewed outlets. 7). This section reviews the evidence, reports of sales, employment, or other Seventy-nine of the studies used often anecdotal, about other incidental related measures in their analyses, survey data to assess the economic and/or unanticipated effects of smoke- while 79 of them were based on impact of smoke-free policies, with 34 free policies. These include effects survey data. Other characteristics of these based on consumer/patron on other problem behaviours (e.g. of these studies including whether surveys and the remaining 45 based drinking, gambling, and violence), or not they were peer reviewed and on owner/manager surveys. Given litter, and street noise. Finally, the their findings are summarised in the limitations of these surveys overall, broader economic impact Table 4.2. described above (particularly those of smoke-free policies and tobacco Forty-nine of the identified studies based on convenience samples control efforts is briefly discussed. This based on official reports of business and/or that collected information section is not a comprehensive review activity met the criteria described on anticipated rather than realised of all possible incidental/unanticipated above for a methodologically sound impact), only about one in five of effects, but rather a short discussion evaluation of the economic impact of these studies were published in a of those that have received some smoke-free policies; specifically, these peer-reviewed outlet. Nearly all of attention and a selected review of the studies used data covering a period the peer-reviewed studies (17 of 19) existing evidence on each. including several years before and concluded that there are no negative after policy implementation, controlled economic effects of smoke-free Smoke-free policies and other for underlying economic conditions policies. The majority of studies problem behaviours and other relevant factors, and used based on consumer/patron surveys appropriate statistical methods. These that are published in other outlets Drinking and its consequences studies use data on an assortment also found that there is no negative of economic indicators, including: economic impact of these policies. Concerns about the potential taxable sales, sales tax revenues, Of the studies based on survey data, economic impact of smoke-free or other sales data; employment; the those that relied on owner/manager policies on bars are often driven by number of establishments; measures survey data and that are not published the observation that smoking and of bankruptcy; and the value of in peer-reviewed outlets (the most drinking are frequently done together. businesses. Several of the studies methodologically flawed studies) are Given this observation, it is plausible examine more than one outcome. Of the only group more likely to conclude that by reducing smoking, smoke- the 49 identified studies, 47 concluded that there is a negative economic free policies might also reduce that smoke-free policies have either impact of smoke-free policies. drinking among smokers. Several no economic impact or a positive studies by economists have explored economic impact on the businesses Other incidental effects the potential relationships between affected by them. of smoke-free policies smoking and drinking by examining In addition, 37 other studies met the impact of tax or price changes for some, but not all, of these criteria; In addition to their economic effects on drinking behaviour and/ most often they failed to control for (or lack thereof), smoke-free policies or vice-versa (Dee, 1999; Cameron &

87 IARC Handbooks of Cancer Prevention 37 79 86 49 45 34 165 Total 1 1 2 0 1 6 7 1 19 19 21 25 26 33 54 Yes (through January 31 2008) (Number of studies) + 3 8 9 15 19 10 18 19 47 27 27 20 65 46 No 111 Reported a negative impact Total Subtotal Subtotal Yes (n=3) Yes Yes (n=9) Yes No (n=28) No (n=25) No (n=34) No (n=35) Yes (n=21) Yes Yes (n=10 ) (n=10 Yes criteria (n=37) criteria (n=49) Peer reviewed Total ConsumerTotal http://www.vctc.org.au/tc-res/Hospitalitysummary.pdf. Total Owner/ManagerTotal (n=45) Total forTotal studies meeting all four Total forTotal studies meeting some of Methodological quality* Meet criteria for methodologically sound studies (n=49) Met some of but not all criteria for methodologically sound studies (n=37) Patron/Consumer surveys (n=34) Owner/Manager surveys(n=45) Type ofType data Table 4.2Table Summary of studies on the economic impact of smoke-free policies on the hospitality sector Official reports of sales, employment or related measures (n=86) Survey data (n=79) + In+ preparation for this chapter, Scollo and Lal (2008) updated their previous, comprehensive review of the literature on the Januaryimpact of 2008.smoke-free These studies policies were identified on the hospitality searchingby numbera sector citationof to include indices, studies including available Medline, through the Science Citation HealthSTAR.Index, the SocialThe following Science search Citationterms wereused: “smok* Index, andrestaurants,”Current Contents“bars,” “hospitality,” Search, “economic,”PsycINFO, andOvid“regulation,” Unpublishedand“law.” studies, andthose foundnot through these searches,through otherwereidentified approaches, such solicitingas relevant studies from the members theof International Union Against Cancer’s International and tobaccoTobacco Control Network (GLOBALink), websitesandsearching by hospitality based industryin major English-speaking “economiccountries. impact” or “study.” Finally, additional searches studies were identifiedwere conducted through the monitoringwith of mediaGoogle, reports using and various the tobacco-related search For terms publications.their reviews,“smok* each bans”identified and “restaurants”study was assessed or by both“bars,” Scollolimited and Lalby and thetheand the followingsource, terms of if funding any, detailsfor the werestudy. Where tabulated:the funding author;source was yearunclear, published;Scollo and publisherLal searched nameindustry and type;document organisations.whetherarchives oraddition, to determine Innot thetheapproach if studyfunding findingsand waseach was studyofdetermined,provided peerwas reviewed;bytobacco including locationanddatecompanies, policyof implementation, or associated naturethepolicy of implemented, typeoutcome of measuremethod employed, analyticused, whether or not economic conditions were controlled for, and a brief descriptionpolicy at issue of onthe the outcome(s) key findings being analysed (including was negative). whether The detailed or not review the is available author(s) onlineconcluded at: that the actual or potential economic impact*These of studiesthe used data covering a period including several years before and after policymethods. implementation, controlled for underlying economic conditions and other relevant factors, and used appropriate statistical

88 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes

Williams, 2001), generally concluding bans may crossover to reductions this is likely to be accompanied by that there is some complementarity in social maladies tied to excessive a reduction in problem gambling between these behaviours; that is, drinking.” (Lal & Siahpush, 2008). Indeed, it increases in the price for one leads A different conclusion was was noted that “Gambling control to reductions in the consumption of reached in the assessment of the advocates expected the legislation both. A few of these studies have impact of selected smoke-free policies would be useful in curbing excessive considered similar relationships on drinking and driving between gambling among EGM users in that between smoking and illicit drug 2000 and 2005, based on the use enforcing a break in play would use, reaching similar conclusions of US county level data on fatal prompt many gamblers to reconsider (Chaloupka et al., 1999; Cameron & motor accidents attributable their gambling,” and suggests that Williams, 2001; Farrelly et al., 2001). to alcohol (Adams & Cotti, 2008). problem gamblers deterred by Few studies, however, have Jurisdictions selected for analyses smoke-free policies may pay off considered the impact of smoke-free are large US cities and counties existing debts, save more, spend policies on other substance use, with and counties in states that adopted more on housing, and increase those that have focused on drinking smoke-free bar policies between 2002 spending in retail establishments. and its consequences. For example, and 2005. The researchers found that Given the mixed evidence, however, the first six waves of the US Health alcohol-related traffic fatalities rose in more research is needed to better and Retirement Survey, a longitudinal counties covered by smoke-free bar understand the impact on problem survey of adults aged 51 to 61 at policies, but did not change in counties gambling of smoke-free policies that baseline conducted from 1992 without such policies, attributing this cover gaming establishments. through 2002, were used to examine to an increase in driving by smokers the impact of smoking restrictions who seek out bars where smoking is Domestic violence and other factors on self-reported allowed (either in other jurisdictions drinking (Picone et al., 2004). It was not covered by a smoke-free bar Reductions in drinking that might concluded that more comprehensive policy, or those that are covered but result from smoke-free policies could restrictions on smoking (those that do not comply with the policy) that reduce domestic violence, given the cover more venues, including bars) more than offsets any reductions in established relationship between significantly reduce drinking among drinking caused by the policy. This alcohol consumption and violence women, but have little impact on seems to be a short-term, transitional (Markowitz, 2000). For example, it was drinking among men. effect that will eventually disappear as found that Irish smokers reported less More recently, state level data these policies diffuse throughout the drinking following the implementation were used on beverage-specific USA and as compliance increases. of the comprehensive smoke-free alcohol consumption in the USA to Clearly, more research is needed policy in Ireland, than did smokers in look at the impact of state smoke-free to fully understand the impact of Scotland and the rest of the United policies on drinking (Gallet & Eastman, smoke-free policies on drinking, other Kingdom, when comparable policies 2007). Relatively crude indicators of substance use, and their related con- were not in effect (Hyland et al., the policies were used: an indicator sequences. 2008b). To date, there is no substantive for any smoking ban and an indicator evidence that smoke-free policies for a ban on smoking in restaurants Problem gambling have increased domestic violence, and/or bars. Estimates indicated that but some evidence suggests that they the policies resulted in reductions in As described above, the evidence are likely to reduce such violence. consumption of and spirits, but is mixed on the impact of smoke- an increase in wine consumption. free policies on business activity Noise The researchers concluded that in gaming establishments. To the “the benefits from reducing tobacco extent that such policies do result Anecdotal reports suggest that consumption by enacting smoking in a decline in gambling revenues, smoke-free policies, particularly

89 IARC Handbooks of Cancer Prevention those that cover bars and clubs, the implementation of a smoke-free If the loss of cigarette tax revenues increase noise outside of these policy, or on the effectiveness of is of particular concern, adoption establishments and result in greater different approaches to reduce it. of smoke-free policies as part of a complaints from neighbors. There comprehensive approach to reducing are, however, very limited, solid data Summary tobacco use that includes increases to support this. Data from repeated in cigarette and other tobacco tax cross-sectional surveys of pub/bar Little reliable evidence exists on the revenues, is an effective means of and restaurant employees in Norway, impact of smoke-free policies on other both preserving the revenue stream following the implementation of the problem behaviours, including other generated by these taxes, for the country’s comprehensive smoke-free substance use and its consequences, short- to medium-term, and reducing policy, indicate that almost half of bar problem gambling, domestic violence, tobacco use. employees and one-third of restaurant noise, and litter. While concerns about Those opposed to tobacco control employees ‘agree completely’ that the potential for increased problem policies and programmes also raise there is more noise outside of the behaviours (with the exception of concerns about the impact of these premises, with about one in five bar gambling) have been raised, there efforts on employment, arguing that employees and one in ten restaurant are almost no data to support these the resulting reductions in tobacco employees reporting an increase in concerns. use will lead to job losses in tobacco- complaints from neighbors (Lund, related farming, manufacturing, 2006). To the extent that there are Smoke-free policies and , as well as in other real concerns about increased noise, and the macroeconomy sectors of the economy. Again, adoption or enforcement of existing any reductions in tobacco-related anti-noise policies is appropriate. As noted above, opponents of tobacco employment that result from smoke- control efforts often raise concerns free policies, or other tobacco control Litter about the broader, macroeconomic activities, will be offset by increased effects of tobacco control policies. employment in other sectors as the Anecdotal reports also suggest that With respect to smoke-free policies, money once spent on cigarettes is there may be an increase in litter these concerns are most relevant spent on other and services. following the implementation of to the impact on tax revenues and This is particularly true in many smoke-free policies, as smokers are employment. These issues are brief- developing countries where smoking forced outside to smoke and drop ly discussed in this section (see is increasing, and where the short- their cigarette butts on sidewalks Jha & Chaloupka, 1999; Prabhat & term impact of the policies is more and streets. This is supported by Chaloupka, 2000 for more complete likely to be slowing the growth in the findings from the Norwegian discussions of these issues). tobacco use rather than significantly bar/restaurant employee survey To the extent that smoke-free reducing it. described above, with the majority of policies reduce cigarette smoking, employees indicating an increase in as described in Chapter 7, the Summary and conclusions cigarette litter following the country’s implementation of these policies smoke-free policy (Lund, 2006). will reduce revenues generated by Smoke-free policies impact bus- Some have suggested that supplying cigarette excise and other taxes. inesses in numerous ways, from smokers with portable ashtrays However, these reductions in improving the health and productivity would be an effective approach to revenues are likely to be offset by of their employees to reducing their reducing potential litter, while others increases in other tax revenues, as insurance, cleaning, maintenance, recommend better enforcement of the money that smokers once spent and potential litigation costs. existing litter laws. There is no reliable on cigarettes is now being spent Experience to date suggests that evidence, however, on the extent on other goods and services which there are minimal short-term costs to which litter increases following are subject to VAT and other taxes. to businesses of implementing

90 Impact of smoke-free policies on businesses, the hospitality sector, and other incidental outcomes comprehensive smoke-free policies. sound research studies from smoke-free policies are adopted in a Existing evidence from developed developed countries consistently growing number of these countries. countries indicates that smoke- conclude that smoke-free policies Few studies exist on the impact free workplace policies have a do not have an adverse economic of smoke-free policies on gaming net positive effect on businesses; impact on the business activity of establishments, and their results are the same is likely to be the case in restaurants, bars, or establishments mixed; more research is needed on developing countries. Establishing catering to tourists, with many studies these venues. and maintaining designated indoor finding a small positive effect of There is insufficient evidence or outdoor smoking areas is more these policies. These studies include about the effects of smoke- costly to implement than a completely outcomes such as official reports of free policies on various problem smoke-free policy. There are minimal sales, employment, and the number behaviours, such as other substance costs to governments related to of businesses. Very limited evidence use and its consequences, problem enforcement and education. from South Africa, an upper middle- gambling, domestic violence, noise, Much of the debate over the resource country, is consistent with and litter. No credible evidence exists economic impact of smoke-free these findings. It is likely that the same to support claims that smoke-free policies, and as a result, much of would be true in other developing policies will negatively affect the the research, has focused on the countries; nevertheless, research overall economy. hospitality sector. Methodologically confirming this would be useful as

91