Differential impact of state tobacco control policies among race and ethnic groups

Authors

  • John A. Tauras

    Corresponding author
    1. Department of Economics, University of Illinois at Chicago, IL, USA and
    2. National Bureau of Economic Research, Cambridge, Massachusetts, USA
    Search for more papers by this author

  • The author has no conflicts of interests.

John A. Tauras, Department of Economics (m/c 144), 601 S. Morgan, Chicago, IL 60607-7121, USA. E-mail: tauras@uic.edu

ABSTRACT

Aims  This paper describes patterns of racial and ethnic cigarette use in the United States and discusses changes in state-level tobacco control policies. Moreover, this paper reviews the existing econometric literature on racial and ethnic smoking and discusses the limitations of that research. Finally, this paper outlines an agenda for future research.

Methods  Patterns of racial and ethnic smoking and changes in state-level tobacco control policies in the United States were obtained from a variety of sources, including surveys and government and private documents and databases. After an extensive literature search was completed, the existing research was scrutinized and recommendations for much-needed future research were put forth.

Findings  Despite the fact that certain racial and ethnic minorities bear a disproportionate share of the overall health burden of tobacco, less than a handful of econometric studies have examined the effects of state-level public policies on racial and ethnic smoking. The existing literature finds Hispanics and African Americans to be more responsive to changes in cigarette prices than whites. Only one study examined other state-level tobacco policies. The findings from that study implied that adolescent white male smoking was responsive to changes in smoke-free air laws, while adolescent black smoking was responsive to changes in youth access laws.

Conclusions  While much has been learned from prior econometric studies on racial and ethnic smoking in the United States, the existing literature suffers from numerous limitations that should be addressed in future research. Additional research that focuses on races and ethnicities other than white, black and Hispanic is warranted. Furthermore, future studies should use more recent data, hold sentiment toward tobacco constant and control for a comprehensive set of tobacco policies that take into account not only the presence of the laws, but also the level of restrictiveness of each policy.

INTRODUCTION

Cigarette smoking is the single most preventable cause of death and disability in the United States, responsible for more than 430 000 premature deaths each year [1]. Despite the deleterious health consequences of cigarette smoking, an estimated 44.5 million adults in the United States aged 18 years and over were current smokers in 2004, representing 20.9% of the total United States adult population [2]. While tobacco use causes disease and preventable death in all segments of the population, certain racial and ethnic minorities bear a disproportionate share of the overall health burden [3]. A recent Surgeon General's report entitled ‘Reducing Tobacco Use’ identifies the elimination of tobacco-related disparities among racial and ethnic minority groups as a major goal in the campaign to reduce the health and economic burden of tobacco use [4]. If that goal is to be met, additional research examining the effectiveness of alternative tobacco control policies on patterns of smoking among members of racial and ethnic minority groups is needed.

This paper discusses patterns of cigarette use in the United States, paying special attention to the dynamics patterns of racial and ethnic consumption. It also provides information on state-level tobacco control policies, placing special emphasis on the recent flurry of tobacco policy enactment at the state level. Moreover, this paper reviews previous econometric studies examining the impact of state-level tobacco control policies on cigarette consumption by individuals of different race and ethnic groups. Finally, this paper discusses the limitations of the previous research and makes recommendations for much-needed future research.

CIGARETTE CONSUMPTION IN THE UNITED STATES

Between the years 1900 and 1964, the use of cigarettes increased dramatically in the United States. Per capita consumption of cigarettes rose from 54 cigarettes per year in 1900 to a peak of 4345 in 1963, the year prior to the release of the first Surgeon General's report on the health consequences of cigarette smoking [5]. Throughout the following three decades, significant progress was made in reducing cigarette smoking in all segments of the US population. By 2004, per capita cigarette consumption declined to 1747 [6]. Furthermore, overall smoking prevalence in the United States declined from 40.4% in 1965 to 24.7% in 1995 [7]. Much of this success has been attributed to tobacco control strategies, including widespread dissemination of information on the risks of smoking, antismoking advertisements, limits on cigarette manufacturer's advertisements, restrictions on smoking in private work-places and public places, increases in cigarette excise taxes, restrictions on access to minors and various others [5]. Recent survey data, however, suggest that the rate of decline in smoking prevalence has slowed in recent years. According to the National Health Interview Surveys, the prevalence of cigarette smoking among adults declined by only 3.4% between 1993 and 2003, compared to 7.1% between 1983 and 1993 [8]. The definition of current smoking changed in 1992 to include some-day smokers. The Centers for Disease Control and Prevention (CDC) suggests that including some-day smoking in the definition of current smoking increases the smoking prevalence estimate by approximately 1.0% [9].

While overall smoking has been declining throughout the past four decades, marked differences have been observed in smoking trends among adolescents and adults of different demographic subgroups. The most dramatic trend differences among adolescent occurred along the dimension of race and ethnicity. White, Hispanic and African American 12th grade students all had similar smoking rates in 1976. According to the Monitoring the Future Surveys, the 30-day prevalence rate for white, Hispanic and African American high school seniors was 38.2%, 35.7% and 36.7%, respectively, in 1976 [10]. Smoking rates among African Americans declined steadily until 1992, whereas smoking rates among whites and Hispanics declined until 1981 and 1980, respectively, although the smoking decline was much more pronounced among Hispanics between the years 1976–80 than it was for whites during this time. By 1991, very large differences in smoking rates were observed for the three racial groups. The 30-day prevalence rates for whites, African Americans and Hispanics were 32.2%, 10.6% and 24%, respectively, in 1991 [10]. After 1992, all three racial/ethnic groups exhibited steady increases in smoking rates into the late 1990s followed by sharp declines in the early 2000s. In 2004, 28.2%, 10.1% and 18.5% of white, African American and Hispanic 12th graders, respectively, were current smokers [10]. Figure 1 shows trends in smoking among white, African American and Hispanic 12th graders for the period 1977–2005 in the United States. Moreover, American Indian and Alaskan native high school seniors had higher rates of smoking than all other race/ethnicity groups throughout the past 30 years, while Asian American and Pacific Islander high school seniors had the lowest smoking rates in the 1970s, but by the 1980s the Asian Americans and Pacific Islanders were replaced by the African Americans as having the lowest 30-day prevalence rates.

Figure 1.

The prevalence data are from the Monitoring the Future (MTF) surveys. In deriving the percentages for each racial subgroup, MTF researchers combined data for the specified year and the previous year to increase subgroup sample size and thus provide more stable estimates. Moreover, prior to 2004, respondents were asked to select the one race/ethnicity category that they thought best described them. In 2005 the race/ethnicity question was changed in half the questionnaire forms. For the revised MTF question, respondents were given a list of race/ethnicity options and instructed to mark all of them that applied. MTF researchers found a very low occurrence of respondents (about 6% in 2005) selecting more than one ethnic/racial group

Significant changes in smoking patterns have also been observed among adults of different race and ethnic backgrounds. According to the National Health Interview Surveys, the prevalence of current smoking among American Indians and Alaskan Natives, whites, African Americans, Hispanics and Asians in 2005 was 32%, 21.9%, 21.5%, 16.2% and 13.3%, respectively [11]. The percentage declines in current smoking prevalence between 1997 and 2005 were 6.2%, 13.4%, 19.5%, 20.6% and 21.3% for American Indians and Alaskan Natives, whites, African Americans, Hispanics and Asians, respectively. These trends imply that an inverse relationship exists between current smoking prevalence rates in 2005 and the percentage decline in prevalence between 1997 and 2005—a period characterized by unprecedented changes in state excise taxes and tobacco control policies. Figure 2 displays trends in adult smoking prevalence rates for various race and ethnic groups in the United States between 1978 and 2005. Indeed, a potentially significant contributor to the small decline in American Indian smoking prevalence rate over this recent period is that American Indians residing on reservations do not have to abide by state regulations, and are therefore not required to pay state cigarette excise or sales taxes or adhere to other state tobacco control policies such as smoke-free air laws or youth access laws.

Figure 2.

The prevalence data are from the National Health Interview Surveys, United States, 1978–2005. For 1978–91, current cigarette smokers include people who reported smoking at least 100 cigarettes in their lives and who reported at the time of the survey that they currently smoked. For 1992–2005, current smokers include people who reported smoking at least 100 cigarettes in their lives and who reported at the time of the survey that they currently smoked every day or on some days. Data from 1978 to 1991 were obtained from the 1998 Surgeon General's Report. Data from 1992+ were obtained from various Morbidity and Mortality Weekly Reports published by the CDC

The current smoking prevalence rate among Asians is lower than the other races primarily because the prevalence rate among adult Asian females is extremely low—4.8% in 2004 [12]. Moreover, since 1964 African American males had consistently higher smoking prevalence rates than their white male counterparts. Forty years later, African American male smoking prevalence rates were finally less than white male rates. Current smoking prevalence in 2004 was 23.9% and 24.1% for African American and white males, respectively [12]. Smoking prevalence rates among African American and white females have historically been very similar; however, African American female smoking prevalence rates have declined faster than white female smoking prevalence rates since 1997 [12].

There are also racial and ethnic differences in smoking intensity. According to the National Health Interview Surveys, white current smokers have the highest age-adjusted mean number of cigarettes consumed on days smoked, whereas Hispanics have the lowest. In particular, using average annual smoking intensity figures for 2002–2004, Adams and colleagues found that conditional on being a current smoker, the mean number of cigarettes consumed on days smoked by whites, American Indians or Alaskan Natives, Native Hawaiians or Pacific Islanders, African Americans, Asian Americans and Hispanics were 15.8, 14.4, 11.7, 11.1, 10.1 and 9.2, respectively [13].

Further, there are also racial and ethnic differences in the age of smoking initiation. According to the 2002–2004 National Health Interview Surveys, among current smokers 18 years of age and older, Hispanics had the highest age-adjusted prevalence of fairly regular smoking prior to the age of 16 (32.8%), whereas Asian Americans had the lowest prevalence (13.3%) [13]. During the same period, among current smokers 18 years of age and older, the age-adjusted prevalence of fairly regular smoking prior to age 16 was 32.1% for whites, 32% for American Indians and Alaskan Natives and 23.2% for African Americans.

Trends in immigration rates are likely to affect the aforementioned racial and ethnic smoking trends. Differences in attitudes and acceptance of smoking, differences in tobacco control policies and differences in smoking-related health knowledge between destination and origin countries alter migrant's costs and benefits of cigarette consumption. The extent of the effects of immigration on smoking rates will depend not only on the rates of immigration along racial and ethnic lines, but also on the duration of assimilation post-migration among racial and ethnic migrants.

The aforementioned trends in smoking by individuals of different race and ethnic groups suggest that subgroups of the population in the United States are likely to respond differently to changes in external stimuli. From a government perspective, it is important to quantify the impact of public policy in affecting the recent aforementioned trends in smoking by individuals of different races and ethnicities. Once armed with this knowledge, policy makers will have a better sense of which policy levers will be most influential in decreasing smoking related disparities. The next section provides a brief overview of recent developments in state-level tobacco control.

TOBACCO CONTROL INITIATIVES

Efforts to reduce cigarette smoking in the United States began in the mid-1960s following the release of the first Surgeon General's report, which linked cigarette smoking causally to lung cancer and a host of other diseases. While consumer education and information dissemination were the mainstay early in the campaign against tobacco, the arena shifted to public policy interventions in an attempt to decrease the health and financial burden associated with cigarette smoking. Prior to 1970, restrictions on smoking in public places were motivated mainly by concerns over smoking as a potential fire hazard [14]. Starting in the 1970s, states started enacting smoking restrictions in private work-sites and public places in an effort to decrease the disease risk associated with smoking [5]. During the 1970s, 31 states either enacted new restrictions on smoking in private work-sites or pubic places or extended existing restrictions [4]. By 1990, 45 states had enacted some type of restriction on smoking in public places, although very few of these restrictions could be considered comprehensive enough to provide any significant protection [4]. While minimum purchase age laws have been in place since the 1890s, these laws were seldom enforced. For example, in 1988 minors purchased approximately 1 billion cigarettes but there were only 32 reported violations of youth access laws in all 50 states during that year [15]. The Synar Amendment, passed by Congress in 1992, put into place a national youth-access policy. It specifies that states lose Federal funding for mental health programs if they fail to adopt and enforce stringent regulations adequately. As part of Synar, states are required to set the minimum age for the legal purchase of tobacco products at 18 or higher. The Synar amendment also required states to enforce youth access tobacco laws by auditing retail tobacco outlet behavior. In particular, a state agency (or a private organization under contract) was to conduct random, unannounced inspections using decoy minors. States were required to reduce the retailer violation rate to less than 20% by a specific time, that depended on the initial rate of non-compliance determined in the base year, 1997. All states were to be in compliance by 2003. By 2005, Kansas was the only state with a random, unannounced non-compliance rate in excess of 20% [the Synar data for years 1997–2005 can be found on the Substance Abuse and Mental Services Administration (SAMSHA) website at http://prevention.samhsa.gov/tobacco/01synartable.aspx].

Beginning in the late 1990s a new era of tobacco control emerged. State governments stepped up their antismoking efforts significantly by increasing cigarette excise taxes, implementing complete bans on smoking in public places and private work-sites and enacting more restrictive youth access laws. The 1990s also witnessed 46 states suing 11 major tobacco companies to recoup Medicaid costs for the care of people injured by tobacco use. The suit was settled in 1998, when the tobacco industry agreed to pay the states $206 billion over a 25-year period. This settlement became known as the Master Settlement Agreement (MSA). In addition, the four states that were not part of the MSA (Mississippi, Florida, Texas and Minnesota) settled their lawsuits separately with the tobacco industry. In the wake of the MSA and other settlements, states have been able to take advantage of prevailing antitobacco sentiment as an opportunity to increase excise taxes dramatically and strengthen tobacco control policies.

Since 1 January 1998, 43 states and the District of Columbia have increased their excise taxes on cigarettes at least once. As of 11 November 2006 state excise tax rates on cigarettes ranged from $2.575 per pack in New Jersey to $.07 per pack in South Carolina. Even Tennessee, Georgia, Kentucky, Virginia and North Carolina, tobacco-producing states that have traditionally resisted raising tobacco taxes, have recently increased their tax rates on cigarettes. The state excise tax increases, in conjunction with the industry initiated price increases (to finance the MSA and other settlement payments), have resulted in among the most dramatic cigarette price increases ever observed in the United States. The average price of a brand-name pack of cigarettes in the United States increased from $2.06 per pack in November 1997 to $4.22 per pack by November 2005 [16]. The average inflation-adjusted cigarette excise tax increase across all states over the past decade (1997–2006) was substantially larger than the average inflation-adjusted cigarette excise tax increases of previous decades. For example, the average nominal excise tax increase across all states between 1997 and 2006 was $0.37 per pack [16]. This compares to an average nominal excise tax increase of only $0.084 per pack for the previous decade (1987–1996) [16].

Beginning in the 1990s and continuing into the mid-2000s, many states also strengthened their smoke-free air laws. Currently, all 50 states and the District of Columbia have smoke-free air laws restricting smoking in certain indoor locations. These laws range from simple restrictions on where people can smoke inside buildings to total bans on smoking inside buildings.

The number of states implementing complete bans on smoking, the most restrictive provisions, has increased markedly over the past decade. For example, the number of states to ban smoking in all private work-sites has increased from one in 1995 to 17 in 2007. Similarly, the number of states to ban smoking in restaurants has increased from one in 1995 to 22 in 2007, and the number of states to ban smoking in bars has increased from one in 1998 to 16 in 2007. Moreover, Maryland, Minnesota, Montana, New Hampshire, Oregon, and Utah have passed legislation to ban smoking in workplaces, restaurants, or bars that is yet to go into effect.

Throughout the 1990s and early 2000s, many states also enacted laws that penalize minors for tobacco-related offenses. For example, the number of states that prohibited minors from possessing tobacco products increased from eight in 1990 to 36 in 2005. Similarly, the number of states that prohibit the use of tobacco products by minors increased from eight in 1990 to 19 in 2005. Finally, the number of states that prohibit minors from purchasing cigarettes increased from 14 in 1990 to 41 by 2003.

PRIOR ECONOMETRIC STUDIES ON THE IMPACT OF STATE TOBACCO CONTROL POLICIES

The discipline of economics has made substantial contributions to understanding the determinants of cigarette demand, particularly as it relates to the impact of government interventions on cigarette consumption.

Economic theory of demand

Empirical models of cigarette smoking are based upon the economic theory of demand. In order to derive cigarette demand equations, an individual's utility function must first be assumed. This utility function has cigarettes, other goods and tastes as arguments. An individual maximizes his or her utility subject to a budget constraint, which is comprised of the price of cigarettes, income and the prices of all other goods. This constrained maximization yields demand functions for cigarettes. The demand functions show that cigarette consumption is related to the price of cigarettes, prices of related goods, income and the individual's tastes.

Clean indoor air laws, youth access laws and other tobacco control policies affect cigarette demand through what economists call the ‘full price’ of cigarettes. The ‘full price’ of cigarettes can be thought of as not only the monetary purchase price of cigarettes, but also includes the costs associated with obtaining and consuming cigarettes. Therefore, any policy that increases the monetary price of cigarettes (for example, increased excise taxes), makes cigarettes more difficult to obtain (for example, minimum legal purchase age restrictions and/or prohibitions on sale), or raises the expected costs associated with consumption (for example, clean indoor air laws) will increase the ‘full’ price of cigarettes. According to the basic law of economics, as the ‘full price’ of a good rises the quantity demanded of that good will decrease.

Previous econometric studies—all races/ethnicities combined

The most consistent finding from previous empirical studies is that cigarette consumption is related inversely to the price of cigarettes. Many of the previous studies have employed aggregate-level data (either time–series for one geographic unit or pooled cross-sectional time–series for multiple geographic units) in their investigation of cigarette demand. Price elasticity estimates obtained from these studies range from −0.14 to −1.12, with a majority of the estimates falling in a narrower range of −0.30 to −0.50 [4]. Differences in the price elasticity estimates can be attributed to differences in data and modeling techniques [4].

A growing number of studies have employed micro-level (individual level) data to examine the determinants of cigarette consumption. Using micro-level data has several advantages over using aggregate-level data. First, the use of micro-level data allows researchers to analyze the effects of prices and other policy variables on the probability that an individual smokes and on average cigarette consumption among smokers, and not simply average cigarette consumption, which is the primary focus of aggregate studies. Secondly, if the micro-level data are longitudinal in nature, they can be used to examine the determinants of smoking transitions in the smoking uptake and cessation continuums. In general, the total price elasticity estimates obtained using individual level data are comparable to those that employed aggregate level data. For a comprehensive review of these studies see Chaloupka & Warner [17] and the various Surgeon General's reports [4,14,18].

While numerous studies of the effects of price on cigarette smoking have been completed in recent years, a much smaller number of studies have examined the impact other tobacco control policies, such as smoke-free air laws and youth access laws, on smoking behavior. Chaloupka & Grossman [19] found that strong restrictions on smoking reduce significantly both the propensity and intensity with which youth smoke. Chaloupka & Wechsler [20] and Tauras & Chaloupka [21] concluded that strong smoking restrictions reduce significantly both smoking prevalence and average daily cigarette consumption among young adults. Tauras, Markowitz & Cawley [22] concluded that purchase, use and possession laws are related inversely to youth and young adult smoking prevalence. Tauras [23] found private work-site restrictions and restrictions on smoking in other public places to decrease moderate smoking uptake among young adults. After accounting for the potential endogeneity of smoking restrictions, Ohsfeldt, Boyle & Capilouto [24] concluded that the strongest restrictions on cigarette smoking lead to significant decreases in smoking prevalence. After accounting for the potential self-selection of workers, Evans, Farrelly & Montgomery [25] concluded that work-place smoking bans reduce the probability of adult smoking by 5% and reduce the average daily cigarette consumption of smokers by 10%. Finally, after controlling for the possibility that unobserved state-level sentiment towards smoking may be driving both the creation of new smoke-free air laws and adult smoking rates to decrease, Tauras [26] found that more restrictive smoke-free air laws decreased average smoking by adult smokers, but had little impact on the prevalence of smoking by adults.

Previous econometric studies on racial and ethnic smoking

While numerous econometric studies have examined the determinants smoking for the US population as a whole, less than a handful of econometric studies have examined the impact of state tobacco policies on racial and ethnic smoking.

Studies on adolescent smoking

Chaloupka & Pacula [27] were the first to examine racial differences in the impact of cigarette prices, youth access laws and smoke-free air laws on adolescent smoking prevalence. Using the 1992, 1993 and 1994 Monitoring the Future (MTF) surveys, the researchers found that African American adolescents are more responsive to price changes than are white adolescents. They estimated the prevalence price elasticity of demand for African Americans to be −1.11 compared to a prevalence price elasticity of demand of −0.64 for white adolescents. These results suggest that African American adolescent smoking prevalence is nearly twice as responsive to changes in cigarette prices, as is white adolescent smoking prevalence. Furthermore, the authors found smoke-free air laws to decrease smoking prevalence rates by adolescent white males, but have no effect on any other racial group. In addition, they found youth access restrictions to have a negative impact on adolescent black smoking prevalence, but have no impact on white youth. The authors caution that they controlled only for the existence of youth access laws and clean indoor air laws and not the level of enforcement of these policies. They suggest that it is possible for some tobacco control policies to be enforced differentially in a manner that is correlated to race, yielding the aforementioned clean indoor air and youth access results.

Gruber & Zinman [28] conducted some sensitivity analyses that separated youths into white and non-white youths and white and black youths. Using the (MTF) surveys, the researchers found that neither white nor non-white younger teenagers were responsive to price changes. Younger teenagers were defined as individuals in 8th or 10th grade. On the other hand, they found older non-white teenagers to be extremely responsive to changes in cigarette prices (the total non-white price elasticity of demand is estimated to be −4.35), while older white teenagers are not affected significantly by price changes. Older teenagers were defined as those in 12th grade. Due to data limitations, they were not able to decompose the non-white racial category into different race/ethnic categories to examine what was driving the implausibly high price elasticity among the older non-white group. Furthermore, using the Youth Risk Behavior Surveys (YRBS), Gruber & Zinman [28] compared the price responsiveness of black and white youths. Again, the researchers found both younger and older white teenagers to be unresponsive to price changes. Younger teenagers were defined as those in 8th, 9th or 10th grade, whereas older teenagers were defined as individuals in 12th grade using the YRBS surveys. Moreover, they found younger black teenagers to be unresponsive to price changes. However, they found older black teenagers to have an unreasonably high price elasticity of demand of −17.51 (prevalence elasticity of −9.26 and conditional demand elasticity of −8.25). Finally, using the Vital Statistics Natality (VSN) data, the researchers came to the opposite conclusion with respect to race and price responsiveness. They found both younger and older black teenage mothers to be unresponsive to cigarette price changes, whereas they found both younger and older white teenage mothers to decrease their consumption of cigarettes significantly when prices increase. Younger teenagers were defined as individuals aged 13–16 years, whereas older teenagers were defined as individuals aged 17–18 years in the VSN data. The researchers attribute the mixed results and unreasonably high price elasticities for non-whites and blacks on the lack of changes in cigarette taxes (i.e. a lack of within-state variation in cigarette prices) during the time that their data were collected (1991–1997).

A third paper by DeCicca and colleagues [29] examined the impact of cigarette prices on smoking among white, black and Hispanic adolescents employing the National Educational Longitudinal Study of 1988. Unlike the previous studies that examined the determinants of average smoking and/or smoking prevalence, the study by DeCicca and colleagues examined the determinants of smoking initiation. The researchers found that price was an insignificant determinant of smoking onset between waves of data among white adolescents but had a negative effect on black and Hispanic smoking onset. The results implied that a small tax increase of 20 cents would yield a very large reduction in the probability of initiation among Hispanics and a substantial reduction in the probability of initiation for African Americans. In particular, a 20-cent increase in the price of cigarettes was found to reduce the starting hazard rate from 17.7% to 13.33% among Hispanics and from 7.8% to 6.3% for African Americans.

Studies on adult smoking

A paper by Farrelly et al. [30] examined the impact of cigarette prices on adult smoking propensity and intensity by race and ethnicity. The paper employed 14 years of National Health Interview Surveys between 1976 and 1993. The researchers found cigarette prices to be an important determinant of African American and Hispanic smoking prevalence and average consumption. In addition, they found cigarette prices to be related inversely to average smoking among whites, but to be an insignificant determinant of smoking prevalence among whites. The overall price response suggested that African Americans were more than twice as price-responsive as whites and Hispanics were found to be more than six times as price-responsive as whites. The total estimated price elasticity of demand for whites, African Americans and Hispanics was −0.15, −0.35 and −0.93, respectively. Moreover, the paper by Farrelly and colleagues [30] found lower-income adults to be much more price-elastic than higher-income adults, a finding that has been confirmed in other studies by Hersch [31] and Townsend, Roderick & Cooper [32]. The larger response to price changes by individuals of lower income is an interesting cross-cutting issue, because individuals of some minority subgroups are more likely to be below poverty thresholds than the general populations.

To summarize, the previous econometric research that examined the determinants of cigarette smoking by race and ethnic groups generally finds Hispanic and African American adolescents and adults to be more responsive to changes in cigarette prices than white adolescents and adults, respectively. This is consistent with the large percentage decreases in Hispanic and African American smoking prevalence rates that have occurred between 1997 and 2004—a period of dramatic price increase in the United States. In some instances the estimated price response was found to be implausibly large, such as Gruber & Zinman for African American adolescents and DeCicca and colleagues for Hispanic adolescents. In another instance, African American mothers were found to be unresponsive to cigarette prices while white mothers responded significantly. Furthermore, the only study to investigate the effect of other tobacco control policies (Chaloupka & Pacula) found smoke-free air laws to decrease smoking prevalence rates by adolescent white males, but have no effect on any other racial group. Moreover, the researchers found youth access restrictions to have a negative influence on adolescent black smoking prevalence, but have no effect on white youth.

LIMITATIONS OF PRIOR ECONOMETRIC STUDIES

While the econometric studies described above made substantial contributions to the literature and gave us a better understanding of the differential response to tobacco control policies by individuals of different race and ethnic groups, the studies suffer from several limitations. First, all the studies use data that predate the dramatic changes in cigarette prices and tobacco control policies that have occurred post-1998. In fact, Gruber & Zinman point out that it may be a lack of variation in cigarette prices during the time their data were collected (1991–1997) that yielded both the insignificant price effects for whites and the unreasonably high price elasticities for non-whites and blacks. Secondly, none of the previous studies examined the impact of prices or policies on cigarette smoking among individuals of different races, including Asian American/Pacific Islander and Native American/Native Alaskan groups that have among the lowest and highest smoking rates in the country, respectively. Thirdly, three of the four studies ignored other tobacco control policies entirely in their demand equations and no previous study has examined the influence of state spending on tobacco control, which has been found to be a very important determinant of both adolescent and adult cigarette demand in the general population [33,34]. The omission of other policies may bias the price coefficients away from zero if states that have higher cigarette taxes also have stronger tobacco control policies and more spending on tobacco control. Fourthly, no prior study has examined the effect of the strength of state smoke-free air laws and youth access restrictions. Chaloupka & Pacula simply controlled for whether or not these policies existed, but did not control for the level of restrictiveness of each of these policies. Fifthly, only two studies (Gruber & Zinman and DeCicca et al.) controlled for state tobacco sentiment in their demand equations. This is an important factor to hold constant, because state sentiment towards tobacco may be driving both policy enactment and decreased smoking. If tobacco sentiment within the state is influential in forming new tobacco policies, than not controlling for such sentiment will lead to biased price and policy estimates (away from zero).

AN AGENDA FOR FUTURE RESEARCH

While much has been learned about the economic determinants of racial and ethnic smoking from previous studies, there is much more that needs to be learned. It is imperative for researchers to examine the effectiveness of state-level policies on cigarette smoking by individuals of different races and ethnicities using more recent data. As mentioned above, a new era of tobacco control emerged in the late 1990s. The era has been characterized by an unprecedented number of states increasing cigarette excise taxes and implementing stronger tobacco control restrictions. It has also been characterized by significant changes in smoking patterns along racial/ethnic divides. These changes in smoking patterns and tobacco control policies will provide much more variation on both the left- and right-hand sides of the demand equations, very probably yielding the most precise estimates of the true impact of policy on smoking by demographic groups.

It is also imperative that researchers consider other race and ethnic groups, not only white, African American and Hispanic populations. On the two extremes of smoking prevalence are American Indians and Alaskan Natives and Asians and Pacific Islanders. Nothing is known about the influence of state tobacco control policies on these two subgroups.

More research on tobacco control policies is desperately needed. All but one study have neglected to see how racial and ethnic groups respond to changes in smoke-free air laws, youth access laws and other state policies. Moreover, in the one study that did look at youth access laws and smoke-free air laws, only the existence, not the strength, of these laws was examined. Furthermore, nothing is known about the impact of state spending on tobacco control on smoking by race and ethnic groups—a void that needs to be addressed.

It is also important to quantify the inter-relationship between cigarette consumption and other substances such as alcohol and smokeless tobacco and even illicit substances such as marijuana and cocaine by individuals of different races and ethnicities. It may be the case that public policies aimed at these other substances may be having a differential effect on demand for cigarettes along racial and ethnic grounds. Conversely, it may be the case that stronger tobacco control policies and higher cigarette prices may have a differential effect on the consumption of other substances by different race and ethnic groups. Nothing is known currently about these inter-relationships, and future research is desperately needed.

Large differences in smoking prevalence rates along gender lines are observed within racial and ethnic groups. It is important to know whether or not public policy is attributed to these disparities, or whether these disparities were formed solely for other cultural reasons.

In addition, more research that examines the effect of public policies on racial and ethnic smoking intensity, smoking initiation, smoking progression, smoking regression and smoking cessation are warranted. Only one previous study has attempted to examine how prices and state policies affect daily consumption by smokers of different races and ethnicities. Similarly, only one previous study has looked at the effectiveness of state policies in deterring smoking initiation among individuals of different race and ethnic groups. Moreover, no prior studies have investigated the impact of state tobacco control policies on smoking progression, regression or cessation.

Finally, it is important to investigate whether or not new immigrants respond differently to state tobacco control efforts than do individuals who have acculturated, and if these responses vary by racial and ethnic groups.

CONCLUSION

State public policies aimed at tobacco have probably had a significant differential effect on cigarette smoking by race and ethnic groups in the United States, leading to the current observed disparities in smoking prevalence rates. This paper has discussed prior econometric studies that examined the impact of cigarette prices and state-level tobacco control policies on cigarette smoking among individuals of different race and ethic backgrounds, pointed out the limitations of the prior studies and set an agenda for future research. More research employing recent data is warranted in order to yield more precise estimates of the true influence of policy on cigarette demand and to avoid many of the limitations found in prior studies. Armed with this type of knowledge, state policy makers will have a better sense of which policy levers are most effective in curtailing smoking along racial and ethnic lines so that smoking-related health disparities can be eliminated.

Ancillary