Settling the smoke: public policy and shareholder wealth in the cigarette industry.
Although tobacco smoking was touted in earlier years as an aid for a variety of ailments, in light of the medical evidence on the detrimental health effects of smoking, the federal government took steps throughout the 1960s to raise consumer awareness of the dangers of smoking. For example, the 1964 Report of the Surgeon General was one of the earliest public acknowledgments that smoking increases the risk of developing lung cancer. This was followed by the passage of the Federal Cigarette Labeling and Advertising Act in 1965, which required warning labels to be placed on cigarette packages. Then in 1967 the Federal Communications Commission (FCC) applied the Fairness Doctrine to cigarettes by mandating that broadcasters provide airtime for antismoking messages. However, application of the Fairness Doctrine soon ended after Congress passed the 1970 Public Health Cigarette Smoking Act, which banned cigarette advertising on TV and radio from 1971 onward. (1) In subsequent years, new warnings from the surgeon general, lawsuits levied against the industry, and recommendations to regulate tobacco as a drug all contributed to further declines in cigarette demand.
Although antismoking policies are expected to reduce the demand for cigarettes, several studies posit a relationship between such policies and the supply side of the market. For instance, Barnett et al. (1995) find that competition increased in the cigarette industry following the release of the 1964 surgeon general's report, whereas Eckard (1991) finds a reduction in competition after the 1971 advertising ban. Also, Farr et al. (2001) and Gallet (2003) find the Fairness Doctrine and the advertising ban intensified the effect of advertising on market power.
Nonetheless, an unfortunate limitation of these studies is that arguments for antismoking policies affecting cigarette supply are based on models of firm behavior, yet industry-level data is typically used in the estimation of the empirical model. (2) As an alternative, event study methodology relies on firm-level security price data to investigate the effects of antismoking policy "events" on cigarette producers, with the argument being that policies enhancing (diminishing) profit expectations will have positive (negative) effects on shareholder wealth. However, because existing event studies of the cigarette industry either focus on different antismoking policies or define regulatory event dates in different ways, it is difficult to reach a consensus regarding the relative importance of policies. For instance, Scheraga and Calfee's (1996) study of the early 1960s finds little impact of the Surgeon General's Report on the stock returns of impacted firms; studies that focus on the late 1960s (e.g., Johnson et al., 1991; Lamdin, 1999; Mitchell and Mulherin, 1988) find abnormal returns associated with the 1971 advertising ban. To the best of these authors' knowledge, though, a study encompassing all relevant policies throughout the entire decade has not been done.
Comparisons across existing event studies are further stymied by differences in empirical procedures. For example, Mitchell and Mulherin (1988) test for effects on shareholder wealth by regressing a portfolio of cigarette company returns on a market portfolio of returns and a dummy variable (set equal to one for every month between July 1968 and December 1970). The authors interpret the estimated positive coefficient of the dummy variable as an indicator of the positive wealth effects associated with the 1971 advertising ban. However, this interpretation may be inaccurate because the authors aggregate the wealth effects of multiple regulatory events into a single cumulative abnormal return measure. Indeed, when the event window is reduced to less than two and a half years, Lamdin (1999) fails to find positive abnormal returns with monthly data in the cigarette industry. Hence, to control for such ambiguity, a consistent procedure should be used across all events.
In light of these limitations, the purpose of this article is to reexamine the effect of antismoking policies on shareholder wealth by expanding on the current literature in several ways. First, to minimize the influence of events not related to antismoking policies on the estimated wealth effects, this article uses daily stock price data for the five U.S. publicly traded cigarette producers (i.e., American Brands, Liggett, Lorillard, Philip Morris, and R.J. Reynolds) between 1964 and 1971. (3) Second, the authors identify a comprehensive set of 23 important dates pertaining to the use of information as a deterrent to smoking or legal measures to regulate the industry's advertising practices. Obtaining significant abnormal returns across these dates, the analysis then draws comparisons between the effects of different antismoking initiatives on shareholder wealth. Finally, given that Scheraga and Calfee (1996) find differences in abnormal returns across firms, the authors pay attention to firm-level nuances. In particular, the authors use a variety of firm characteristics (i.e., market share, advertising intensity, percentage of sales associated with filter-tip cigarettes, percentage of advertising expenditures devoted to TV and radio, and tobacco leaf inventories) to further analyze differences in the wealth effects across firms.
In the remainder of the article, section II offers several arguments for how the profit of cigarette producers may have been affected by antismoking policies. This is followed in section III with a discussion of the event dates chosen for the study. The estimation methodology is provided in section IV, followed by the estimation results in section V. The article concludes with a summary in section VI.
II. THEORY CONSIDERATIONS
The literature provides several arguments for how antismoking policies of the 1960s may have affected firm profits and therefore shareholder wealth. Each of these is discussed separately in this section.
A. Demand Reductions and Oligopoly Coordination
Models of oligopoly behavior over the business cycle provide insights into the potential effects of antismoking policies on firm profits. In particular, with the intended effect of antismoking policies being to reduce cigarette demand, a rational forecast in earlier years would be to expect lower future demand and profit. As a consequence, following several game-theoretic studies (i.e., Bagwell and Staiger, 1997; Haltiwanger and Harrington, 1991; Rotemberg and Saloner, 1986), if firms expect future profit to be lower this creates the temptation to cheat on any cooperative solution in the industry, resulting in more competitive pricing and lower profit. (4) In the end, therefore, policies for which the predominant effect is to reduce the future size of the market should coincide with negative stock price reactions.
B. Advertising and Competition
With several studies finding an impact of advertising on competition in the cigarette industry, another avenue exists for advertising-related policies to affect shareholder wealth. Yet the evidence is mixed in the literature. For example, studies that estimate supply relationships for the cigarette industry (e.g., Farr et al., 2001; Gallet, 2003; Tremblay and Tremblay, 1995) find that increases in advertising lead to increased market power in the industry. Such results support the anticompetitive view of advertising, which predicts that if incumbent firms use advertising to deter entry (by forcing entrants to aggressively advertise to secure market share), then policies that reduce advertising will stimulate competition and therefore weaken profit. In such cases, the industry would not welcome advertising restrictions, because they would dampen shareholder wealth.
In contrast, the procompetitive view of advertising maintains that better information to consumers regarding price, quality, and so on, increases consumer awareness of the availability of substitutes, and thus policies that reduce advertising block an effective means of entering the market. Such policies would therefore be welcome by the industry, because they would weaken competition, enhance the profit of incumbents, and therefore boost shareholder wealth. The procompetitive view of advertising finds support in studies by Mitchell and Mulherin (1988) and Eckard (1991) who focus on the 1971 advertising ban. In particular, Mitchell and Mulherin's (1988) event study finds that market reactions to regulatory events leading up to the adoption of the advertising ban are associated with value creation. This suggests that financial markets viewed the advertising ban in a positive light (i.e., profits were expected to increase after the ban), which would be consistent with a reduction in competition following the immediate drop in advertising resulting from the ban. (5) Eckard (1991), on the other hand, finds that the ban coincided with greater stability of market shares, less brand entry, and higher price-cost margins, which are all indicative of a reduction in competition and therefore also support the procompetitive view of advertising. (6)
In light of the opposing evidence on the role of advertising in the cigarette industry, it remains unclear whether events leading up to advertising-related policies had positive or negative effects on shareholder wealth. The empirical results presented in section V shed further light on this issue.
C. Objectives of Regulation
Depending on the objective function of regulators, antismoking policies may have been viewed positively or negatively by the industry. For instance, if regulators were driven solely by a concern for public health, then policies designed to reduce tobacco consumption would presumably have an ill effect on profits and shareholder wealth in the industry. Indeed, in the case of the Surgeon General's Report, it is hard to find evidence that firms benefited from the health warning. (7)
Yet Stigler (1971), Peltzman (1976), and others offer an alternative motivation behind regulation that may be applicable to advertising-related policies. Namely, if regulators become "captured" by the industry (i.e., operate on behalf of the industry), provided advertising is procompetitive, the industry would find it in its best interest to lobby for restrictions on advertising. By doing so, competition would be reduced and profits enhanced. Indeed, this is suggested by Mitchell and Mulherin's (1988) findings of a positive stock reaction to the advertising ban. Moreover, Gormley (1979) finds that members of the FCC, who are responsible for administering the Fairness Doctrine, often voted in favor of the broadcast industry, particularly if they were formerly employed in that industry. Such proindustry behavior may have had a positive effect on the cigarette industry, perhaps by postponing the application of the Fairness Doctrine to cigarette advertising or enforcing a weaker version of the doctrine.
III EVENT DATES
When conducting an event study, key dates in the adoption of a policy change must first be identified. Because this study focuses on antismoking policies initiated in the period between the release of the 1964 Surgeon General's Report and the 1971 advertising ban, the authors follow the approach taken by existing studies. For example, with respect to the legislation of 1965 (i.e., cigarette warning labels) and 1970 (i.e., advertising ban), this article follows Mitchell and Mulherin (1988) and Lamdin (1999) by defining event dates according to the process by which each bill became law. Event dates were then determined by a search through the major publications of the time. The discussion to follow presents the events and their corresponding dates in chronological order.
The first event was the release of the 1964 Surgeon General's Report on January 12, 1964. Soon after, on April 28, 1964, producers in the cigarette industry adopted a Code of Advertising to self-regulate the industry, such that producers whose advertisements were directed toward youths or made unproven health claims regarding a particular brand could be fined up to $100,000.
A series of eight events ultimately led to requiring that health warnings be placed on cigarette packages. To begin, on June 25, 1964, the Federal Trade Commission (FTC) voted in favor of requiring warning labels on cigarette packages. However, it postponed moving on this pending legislation that was being proposed in Congress. Indeed, the Senate Commerce Committee voted in favor of warning labels legislation on April 28, 1965. This was followed by a similar vote of the House Commerce Committee on May 28, 1965. The full Senate approved its version of the bill on June 17, 1965; the full House approved its version on June 23, 1965. The final bill was approved in conference on July 2, 1965, and signed into law on July 28, 1965. Warning labels then became law effective on January 2, 1966. (8)
The next 13 events dealt directly with the issue of advertising. For instance, on June 3, 1967, the FCC ordered that the Fairness Doctrine be applied to cigarettes. That is, although not required to match on an equal time basis, the FCC required all TV and radio stations to broadcast antismoking messages. Then, following the dates provided by Mitchell and Mulherin (1988) and Lamdin (1999), the FTC and FCC voted on July 1, 1968, and February 6, 1969, respectively, to recommend that Congress ban cigarette advertisements from TV and radio. Subsequently, the House Commerce Committee voted on May 29, 1969, in favor of a bill to ban cigarette advertising, which was then voted in favor by the full House on June 19, 1969. (9) The Senate Commerce Committee voted in favor of a similar bill on November 6, 1969, which the full Senate approved on December 13, 1969. A House-Senate conference reached an agreement on the bill on March 4, 1970; it was signed into law on April 2, 1970. Then on November 13, 1970, the industry ended its Code of Advertising, and with prosmoking advertisements soon to end on TV and radio, the FCC announced an end to the application of the Fairness Doctrine to cigarettes on December 16, 1970. Finally, the advertising ban became effective on January 2, 1971, one day after the heavily advertised college football bowl games.
Table 1 provides a summary of the 23 events, as well as the expected effects on shareholder wealth (indicated as positive +, negative -, or ambiguous -/+), depending on the most likely applicable theory from sections II. A few points are worth noting. First, because most events are expected to lead to an eventual reduction in cigarette demand (except for the elimination of the Fairness Doctrine on December 16, 1970), the discussion of stock price reactions to demand reductions (see section IIA) is applicable to many of the events. Events dealing predominantly with advertising (such as the advertising ban) or regulations (such as warning labels, the Fairness Doctrine, and the advertising ban) are less common. Second, events that focus on informing consumers of the health consequences of smoking (such as the release of the surgeon general's report or the initial proposals for requiring warning labels) are likely driven by regulators' desires to reduce cigarette demand (see sections IIA and IIC), and hence the authors expect a negative reaction of shareholder wealth to such events. (10) Third, for other events the expected reaction of shareholder wealth is ambiguous. This is due to the different arguments of the competing theories. For example, because several events address advertising issues, following section IIB, if advertising is procompetitive (anticompetitive), then events that seek to lower advertising will have positive (negative) stock price reactions. Moreover, per the discussion in section IIC, if regulators are driven mainly by concerns for consumer (producer) well-being, then regulations will have negative (positive) stock price reactions. For example, associated with the events of June 17, 1965, through January 2, 1966, it may be that the weaker version of warning labels that emerged is an indicator of the industry's ability to effectively lobby Congress for less intrusive warning labels. Also, depending on the lobby abilities of the industry, because the application of the Fairness Doctrine may have been less severe than expected, it is difficult to sign the expected reaction of returns to events surrounding the Fairness Doctrine, with regard to the objectives of regulators.
Because of the variety of possible reactions of shareholder wealth to these events, it remains of interest to examine the change in shareholder wealth throughout this period. The next section outlines the estimation methodology.
IV ESTIMATION METHODOLOGY
The application of event study methodology in welfare analysis of regulatory changes is a common approach to evaluating the firm-level market impact of regulatory changes in economics and finance (see, for example, Binder, 1985), Lamdin, 2001; MacKinlay, 1997). Assuming markets are efficient, information contained in the public announcement of a regulatory policy will be capitalized in the security prices of firms affected by that policy. If investors perceive the regulation to result in a higher stream of current and future profits, then the security prices of the impacted firms will rise and vice versa. The abnormal return on a firm's security is thus a positive or negative change in the normal return measured after controlling for general market volatility.
The discussion in section II offers several scenarios for how antismoking policies can affect the profits of cigarette companies, implying variation in abnormal returns across events. In addition, differences in firm characteristics, such as company size, inventories, and advertising intensity, make it unlikely that the abnormal returns associated with a given event will be uniform across companies. Although evidence on the impact of a given policy on shareholder wealth can, on average, be summarized by significance, signs, and magnitude of the abnormal returns for that event, variation in abnormal returns across events and firms can be further exploited to examine what factors contribute to gains or losses in shareholder wealth from antismoking policies.
Thus the estimation methodology presented next is in two stages. In the first stage, the authors employ event study methodology to estimate abnormal stock returns for each of the five publicly traded cigarette companies over the 23 events outlined in section III. In the second-stage analysis, the authors use the estimated abnormal returns to further investigate how firm-specific factors determine the observed market effects associated with the 23 events.
A. First-Stage Estimation
Estimation of abnormal returns in the first stage begins with the identification of the dates on which the 23 events discussed in section III were publicly announced. Because the events of interest occur on the same date for all five firms (i.e., there is total clustering), the authors adopt the methodology outlined in MacKinlay (1997) and obtain the abnormal returns by estimating a multivariate regression model with unaggregated security data and dummy variables for each event: (11)
(1) [R.sub.it] = [[alpha].sub.i] + [[beta].sub.i][R.sub.mt] + [j=23.summation over (j=1)][[delta].sub.ij][D.sub.jt] + [[epsilon].sub.it].
In equation (1), [R.sub.it] denotes the return on the security of firm i on day t, [R.sub.mt] is the market return on day t as measured by the value-weighted index compiled by the Center for Research in Security Prices, [D.sub.jt] is a dummy variable set equal to one for event date j and zero otherwise, and [[epsilon].sub.it] is a zero mean disturbance term. Estimates of [[delta].sub.ij] in equation (1) produce the desired measure of the abnormal return for firm i corresponding to event j.
To capture as much of the event's impact as possible, it is standard to include in the estimation of abnormal returns more than one trading day's worth of security returns information. Specifically, to better capture when the market learned of new information related to the passage of antismoking policies, the authors estimate abnormal returns using a three-day window including the day before, the day of, and the day after the event (i.e., [D.sub.jt-1] = [D.sub.jt] = [D.sub.jt+1] = 1 and 0 otherwise). With a three-day event window the interpretation of [[delta].sub.ij] is that of a cumulative abnormal return (CA[R.sub.ij]). Because this window most closely coincides with events immediately related to the dates discussed in the preceding section, it is the preferred window to identify abnormal returns.
Using a three-day window is an approach similar to Johnson et al. (1991), but others such as Mitchell and Mulherin (1988) and Lamdin (1999) have used average monthly returns to estimate abnormal returns from a regulatory event announced within that month. Yet extending the event window beyond three trading days, although not unusual in event studies, is costly as the probability of capturing stock price reactions to unrelated events increases with the size of the window. Nevertheless, the authors also estimate equation (1) with a one-month event window as a point of comparison to those studies that have used monthly data to estimate abnormal returns.
B. Second-Stage Estimation
In the second-stage analysis, the authors use the estimated abnormal returns from the first stage to explain the firm-level market effects associated with antismoking policies. (12) In particular, the authors pool the estimated abnormal returns from the three-day event window across the five firms and use them as the dependent variable in a panel model to evaluate whether firm-specific characteristics help explain the estimated wealth effects. The model is given by:
(2) CA[R.sub.ij] = [^.[delta].sub.ij] = [[gamma].sub.1] + [[gamma].sub.2]M[S.sub.ij] + [[gamma].sub.3]AS[R.sub.ij] + [[gamma].sub.4]PERFIL[T.sub.ij] + [[gamma].sub.5]PERT[V.sub.j] + [[gamma].sub.6]IS[R.sub.ij] + [u.sub.ij],
where [^.[delta].sub.ij] is the ith firm's estimated cumulative abnormal return for event j, M[S.sub.ij] is the ith firm's market share corresponding to the year that event j occurred, AS[R.sub.ij] is the ith firm's advertising-to-sales ratio in the year that event j occurred, PERFIL[T.sub.ij] is the percentage of firm i's sales that are filter-tip cigarettes for the year that event j occurred, PERT[V.sub.j] is the industry percentage of advertising devoted to TV and radio during the year that event j occurred, IS[R.sub.ij] is the ith firm's tobacco leaf inventory-to-sales ratio for the year corresponding to event j, and [u.sub.ij] is a zero mean error term. (13)
Several estimation issues regarding equation (2) need mentioning. First, three versions of the model are estimated. Model 1 pools the positive and negative abnormal returns, whereas models 2 and 3 only use the negative and positive abnormal returns, respectively, as the dependent variable. Accordingly, results from the estimation of models 2 and 3 allow the authors to identify differences in firm reactions depending on whether the abnormal return is positive or negative. Second, because the data are pooled over the five firms, the authors estimate a fixed-effects version of equation (2) by including firm dummy variables to account for time-invariant and unobserved firm characteristics. (14) Third, because the dependent variable in equation (2) is estimated in equation (1), this introduces a heteroscedasticity problem in the second-stage regression (see Blonigen and Wilson, 1999; Harris and Ravenscraft, 1991; Saxonhouse, 1976). To control for this, the analysis uses weighted least squares to estimate equation (2), where each observation of the dependent and independent variables is weighted by the inverse of the standard error of the estimated coefficient [^.[delta].sub.ij].
Expectations on how firm-specific variables can affect cumulative abnormal returns are summarized in Table 2. First, bigger firms (higher MS) with higher market power may be better insulated from the effects of antismoking regulations, thereby experiencing smaller abnormal returns in absolute value; or it could be that bigger firms are bigger targets, having larger absolute abnormal returns. Second, firms that advertise more are particularly sensitive to antiadvertising events. Hence, increases in the advertising-to-sales ratio may magnify the abnormal returns. Alternatively, it may be that firms that advertise more have greater brand loyalty, which allows them to more easily ride the antismoking wave (i.e., abnormal returns are dampened). Third, because filtered cigarettes were viewed as being less harmful, firms with higher filter-tip sales should be less affected by antismoking events. Fourth, because advertising-related events eventually led to the elimination of advertising on TV and radio, a higher percentage of advertising expenditures devoted to broadcast media should intensify the abnormal return. Finally, firms with higher leaf inventories may be more susceptible to antismoking policies that result in demand reductions, because this will leave the firm with possible unwanted inventories.
V ESTIMATION RESULTS
With the exception of Lorrilard, equation (1) was estimated for each firm for the 2,331 trading days spanning the period between October 30, 1962, and March 8, 1972. (15) This sample period includes 300 trading days before the first and after the last event date, respectively--an extension that allows the sample to include trading days that are relatively unaffected by the events of interest. (16)
A. First-Stage Estimation Results
First-stage estimates of the abnormal returns for the five cigarette producers (i.e., [^.[delta].sub.ij]) are presented in Table 3. The authors begin with a discussion of the three-day event window results. As expected, the Surgeon General's Report negatively affected the industry. With respect to the Code of Advertising, one might expect that its adoption would be a positive event for the industry, because its purpose was likely to dissuade outside regulation of advertising content. However, the estimated negative wealth effects do not support this. Rather, it may be that investors viewed the code as an indication of future advertising regulation. Largely negative wealth effects are also associated with early events corresponding to the warning labels legislation (i.e., June 25, 1964, through June 17, 1965), possibly reflecting investor concerns over the detrimental impact of warning labels. Yet it is interesting to note that for later events (i.e., June 23, 1965, through January 2, 1966) the abnormal returns become positive. As mentioned earlier, one possible explanation for this is that as the legislation moved closer to becoming law, it incorporated restraints on future regulation (particularly limits on the FTC's authority to regulate the industry), which investors viewed in a positive light. (17)
Events surrounding the Fairness Doctrine are on average associated with positive wealth effects, although these effects across firms are mixed. In contrast, events surrounding the advertising ban tended to have a negative impact on returns in the cigarette industry, with significance most pronounced during the early stages of the legislation effort. (18) Such a finding is consistent with the anticompetitive view of advertising in that the reduction in advertising following the ban was expected to reduce profit (due to the increase in competition).
Table 3 also reports results from the monthlong event window, which has been widely used in previous literature. A comparison of the 3-day and 31-day results reveals that even though CARs estimated with the 31-day window remain largely negative, their significance drops off appreciably. Indeed, across all 104 firm-event abnormal returns, only 21 are significantly different from zero at conventional levels, which is not surprising given the increased noisiness of longer event windows. In addition, interpretation of the 31-day results has two important shortcomings. First, extending the event window (from 3 to 31 days) increases the probability of capturing stock price movements unrelated to announcement of antismoking policies (such as announcement of dividend payouts, earnings reports, etc.). Second, the 31-day window results aggregate the wealth effects from antismoking events that took place within the same month, making it difficult to draw conclusions about the relative importance of individual policies.
To gain further insight into the relative impact of antismoking policies, the authors compute average CARs (denoted [mu]), across all firms and all events corresponding to each policy. Ranking these according to size, the authors find that the advertising ban had the largest negative impact on the industry ([mu] = -0.38), followed by the Code of Advertising ([mu] = -0.34), then the 1964 Surgeon General's Report ([mu] = -0.23), and last the advent of warning labels ([mu] = -0.19); whereas the average abnormal return for the Fairness Doctrine was marginally positive ([mu] = 0.08). (19) Finally, it should be noted that the average CARs are substantially lower when computed using a month-long event window.
As further indication of the impact of these policies on shareholder wealth, Table 4 reports for each of the 23 events the dollar amount lost/gained by each firm and the industry as a whole. On average, industry losses were greatest for early antismoking events corresponding to the imposition of warning labels and the advertising ban. Across firms, the accumulated loss in market value was greatest for R.J. Reynolds (-$136.2 million), followed by American Brands (-$84.63 million), Philip Morris (-$43.16 million), Liggett (-$5.08 million), and last, Lorillard (-$2.32 million). Such variation is not too surprising, however, given that R.J. Reynolds, American Brands, and Philip Morris were the three largest cigarette producers at the time, with the market shares of Liggett and Lorillard being much smaller. Nonetheless, this does highlight the importance of accounting for differences in stock price reaction across the firms. Overall, cumulative losses to the industry from antismoking policies initiated in the period 1964 to 1971 amounted to $217 million, which translates into approximately $1.5 billion when converted into 2004 dollars.
B. Second-Stage Estimation Results
Equation (2) was estimated using the 3-day CARs, and the results are reported in Table 5. With respect to Model 1, which pools the positive and negative CARs, little can be said of the importance of firm characteristics in explaining differences in wealth effects across firms. In this model, the [R.sup.2] is relatively low and PERFILT is the only variable with a marginal degree of significance. Perhaps this is due to the positive and negative abnormal returns offsetting each other somewhat, making it difficult to reach any conclusions. Yet when the positive and negative abnormal returns are separated, results from the estimation of Models 2 and 3 show substantial improvement. (20) For instance, with respect to the negative abnormal returns, the authors find larger firms (higher MS) who aggressively advertised (higher ASR), particularly on TV and radio (higher PERTV), were most sensitive to antismoking events (i.e., the abnormal return was larger in absolute value). (21) Also, firms with higher filter-tip sales were less affected by antismoking policies. Such results are consistent with several of the authors' earlier predictions in that (1) bigger firms were bigger targets of antismoking campaigns, (2) firms that advertised much were more sensitive to antismoking (particularly antiadvertising) policies, (3) antismoking policies geared toward the ban of advertising on TV and radio were most harmful when the share of broadcast advertising was high, and (4) higher filter-tip sales somewhat insulated firms from the detrimental effects of the antismoking campaign.
Finally, firm characteristics remain important in model 3, although explanatory power is lower, where the CARs tend to be larger for firms that aggressively advertised, devoted less advertising to TV and radio, and carried smaller stocks of tobacco leaf.
Studies of the effects of antismoking policies on the cigarette industry have largely focused on the demand side of the market with only a handful of papers analyzing wealth effects from such policies mostly using monthly security price data. In this article, the authors identify 23 dates associated with antismoking policies initiated in the period 1964-71. Using daily data on stock returns of five U.S. publicly traded cigarette producers, the authors conduct a two-stage analysis of the impact of antismoking policies on cigarette producers. In the first stage the analysis estimates the wealth effects associated with health warnings and regulatory changes, and in the second stage the authors conduct a firm-level analysis of factors that explain the magnitude of wealth effects across tobacco companies.
In addition to finding significant abnormal returns at both the firm and industry level, the empirical results reveal significant variation in the wealth effects from the various types of policies. Indeed, the authors find that on average, the advertising ban had the largest negative impact on the industry, followed by the Code of Advertising, then the 1964 Surgeon General's Report, and finally, the advent of warning labels. In contrast, the average cumulative abnormal return for the Fairness Doctrine was marginally positive. Further investigation into differences in wealth effects across firms reveals that losses in market value were higher for the larger firms who advertised aggressively, particularly on TV and radio. Overall, the authors estimate that antismoking regulatory policies initiated in the period 1964-71 had a predominantly negative impact on the cigarette industry with inflation-adjusted industry losses in market value and shareholder wealth amounting to approximately $1.5 billion.
TABLE 1 Antismoking Policies Initiated between 1964 and 1971: Dates and Expected Signs Calendar Date Event Jan. 12, 1964 1964 Report of the Surgeon General Apr. 28, 1964 Code of Advertising adopted Jun. 25, 1964 FTC proposes health warnings be placed on cigarette packages Apr. 28, 1965 Senate Commerce Committee approves warning labels bill May 28, 1965 House Commerce Committee approves warning labels bill Jun. 17, 1965 Senate approves bill (FTC prohibited from regulating advertising for 3 years) Jun. 23, 1965 House approves bill (FTC prohibited from regulating advertising forever) Jul. 2, 1965 House-Senate conference approves common bill (FTC prohibited from regulating advertising for 4 years) Jul. 28, 1965 Federal Cigarette Labeling and Advertising Act signed into law Jan. 2, 1966 Warning labels become effective Jun. 3, 1967 FCC orders the Fairness Doctrine be applied to cigarette advertising Jul. 1, 1968 FTC votes in favor of recommending to Congress that all cigarette advertisements be banned from TV and radio Feb. 6, 1969 FCC offers similar proposal May 29, 1969 House Commerce Committee approves bill to ban advertising on TV and radio Jun. 10, 1969 Supreme Court upholds the constitutionality of the Fairness Doctrine Jun. 19, 1969 House approves bill Nov. 6, 1969 Senate Commerce Committee approves bill to ban advertising on TV and radio Dec. 13, 1969 Senate approves bill Mar. 4, 1970 House-Senate conference approves common bill Apr. 2, 1970 Public Health Cigarette Smoking Act signed into law Nov. 13, 1970 Code of Advertising ended Dec. 16, 1970 Fairness Doctrine ended Jan. 2, 1971 Advertising ban becomes effective Expected Sign of CARs Calendar Date IIA IIB IIC Jan. 12, 1964 - Apr. 28, 1964 - -/+ Jun. 25, 1964 - - Apr. 28, 1965 - - May 28, 1965 - - Jun. 17, 1965 - -/+ Jun. 23, 1965 - -/+ Jul. 2, 1965 - -/+ Jul. 28, 1965 - -/+ Jan. 2, 1966 - -/+ Jun. 3, 1967 - -/+ Jul. 1, 1968 - -/+ -/+ Feb. 6, 1969 - -/+ -/+ May 29, 1969 - -/+ -/+ Jun. 10, 1969 - -/+ Jun. 19, 1969 - -/+ -/+ Nov. 6, 1969 - -/+ -/+ Dec. 13, 1969 - -/+ -/+ Mar. 4, 1970 - -/+ -/+ Apr. 2, 1970 - -/+ -/+ Nov. 13, 1970 - -/+ -/+ Dec. 16, 1970 + -/+ Jan. 2, 1971 - -/+ -/+ Notes: Expected impact on shareholder wealth captured by CARs. Expected signs of CARs are provided, depending on the most applicable theories from section II. Section IIA coincides with demand reduction and oligopoly coordination theories, IIB coincides with advertising and competition theories, and IIC coincides with objectives of regulation theories. TABLE 2 Predicted Impacts of Second-Stage Explanatory Variables on Absolute Value of CARs Variable Prediction Market share (MS) Negative (if bigger firms are insulated from anti-smoking policies); Positive (if bigger firms are bigger targets of anti-smoking policies) Advertising-to-sales ratio Negative (if firms with greater advertising (ASR) have greater brand loyalty); positive (if firms with greater advertising are more susceptible to anti-advertising events) Percent filter cigarettes Negative (if filter cigarettes are viewed as (PERFILT) being less harmful) Percent TV and radio Positive (if higher broadcast advertising advertisements (PERTV) coincides with firms being more affected by advertising policies geared toward TV and radio) Tobacco leaf inventory-to- Positive (if firms with higher inventories sales ratio (ISR) are more susceptible to anti-smoking policies which reduce demand) Note: CARs for each firm measure the change in shareholder wealth from antismoking regulatory events initiated between 1964 and 1971. TABLE 3 Three-Day and Thirty-One-Day Cumulative Abnormal Returns from Antismoking Policies between 1964 and 1971 American Brands Liggett Lorillard (N = 2,331) (N = 2,330) (N = 1,509) Event Date 3-Day 31-Day 3-Day 31-Day 3-Day Jan. 12, 1964 -0.22 (a) -0.12 0.80 -0.09 -0.95 (a) Apr. 28, 1964 -0.69 (b) -0.16 -0.76 (a) -0.10 0.33 Jun. 25, 1964 -1.88 (a) -0.03 -0.93 (a) 0.08 -0.95 (a) Apr. 28, 1965 -1.23 (a) 0.13 -0.30 (a) 0.11 -0.76 (c) May 28, 1965 -0.21 (a) -0.19 -0.40 (a) -0.22 (b) -0.20 Jun. 17, 1965 0.06 0.16 -0.17 (a) 0.09 -0.46 (a) Jun. 23, 1965 0.06 0.08 0.05 0.54 (a) 0.82 (a) Jul. 2, 1965 0.28 (b) -0.18 0.64 (a) -0.45 (a) 0.01 Jul. 28, 1965 0.24 0.21 (c) 0.15 -0.1E-5 0.12 Jan. 2, 1966 0.43 (a) -0.02 0.38 (b) 0.05 0.50 (b) Jun. 3, 1967 -0.43 (c) -0.07 0.10 0.04 0.01 (c) Jul. 1, 1968 -1.09 (a) 0.18 -0.82 (b) 0.27 (c) 0.65 (c) Feb. 6, 1969 -1.35 (b) 0.13 -0.47 -0.22 -- May 29, 1969 -1.59 (a) 0.29 -0.28 (a) 0.05 -- Jun. 10, 1969 -0.52 (a) -0.27 -0.59 (a) -0.05 -- Jun. 19, 1969 -0.32 (a) -0.08 0.14 -0.13 -- Nov. 6, 1969 0.41 (a) 0.35 (a) 0.11 -0.07 -- Dec. 13, 1969 -0.06 -0.25 (c) -0.63 (a) -0.7E-5 -- Mar. 4, 1970 -0.12 0.08 0.76 (a) 0.58 (a) -- Apr. 2, 1970 -0.07 0.26 0.35 (a) -0.09 -- Nov. 13, 1970 -0.27 (c) 0.38 (b) -0.31 0.41 (a) -- Dec. 16, 1970 -0.01 -0.30 0.46 -0.27 (c) -- Jan. 2, 1971 -0.06 0.03 0.22 (b) 0.09 -- Lorillard Philip Morris R.J. Reynolds (N = 1,509) (N = 2, 331) (N = 2,331) Event Date 31-Day 3-Day 31-Day 3-Day 31-Day Jan. 12, 1964 -0.26 -0.63 (a) -0.29 -0.15 -0.22 Apr. 28, 1964 -0.33 (b) -0.16 (b) -0.01 -2.02 (a) -0.08 Jun. 25, 1964 -0.03 -1.16 (a) -0.33 (a) -1.46 (a) -0.14 Apr. 28, 1965 -0.03 -0.20 (c) 0.18 (c) -0.64 (a) 0.10 May 28, 1965 -0.05 -0.91 (a) -0.17 0.13 (b) -0.14 Jun. 17, 1965 0.03 -0.39 (b) -0.13 -0.30 (a) 0.10 Jun. 23, 1965 -0.07 0.03 0.27 0.11 0.17 Jul. 2, 1965 0.28 (b) -0.60 (a) -0.31 0.31 (c) -0.29 (c) Jul. 28, 1965 -0.10 0.52 0.25 0.19 (b) 0.25 Jan. 2, 1966 0.06 -0.27 (c) -0.03 0.65 (a) -0.02 Jun. 3, 1967 0.07 1.097 (a) 0.32 0.24 -0.04 Jul. 1, 1968 0.67 (a) -1.59 (a) -0.02 -1.36 (a) 0.11 Feb. 6, 1969 -- -1.56 (a) -0.21 -0.59 -0.17 May 29, 1969 -- -0.05 0.69 (a) -0.01 0.25 (b) Jun. 10, 1969 -- -0.79 (a) -0.10 -0.07 0.01 Jun. 19, 1969 -- 0.14 -0.20 0.42 (b) -0.08 Nov. 6, 1969 -- -0.10 (b) 0.42 0.19 0.16 Dec. 13, 1969 -- 0.06 (c) 0.26 0.22 (c) -0.11 Mar. 4, 1970 -- 0.06 0.11 -1.50 (a) 0.10 Apr. 2, 1970 -- -1.96 (a) -0.09 -0.36 (a) 0.23 Nov. 13, 1970 -- 0.46 (b) 0.21 (b) 0.35 0.29 Dec. 16, 1970 -- 0.18 -0.46 (b) 0.06 -0.03 Jan. 2, 1971 -- -0.27 (a) 0.24 -1.58 (a) 0.10 Note: a, b, and c denote significance at the 1%, 5%, and 10% levels, respectively. Results coincide with estimates of [[delta].sub.ij] in equation (1). TABLE 4 Wealth Effects from Regulatory Changes in the Cigarette Industry: 1964-71 (Millions of Dollars) Event date American Brands Liggett Lorillard Philip Morris Jan. 12, 1964 -1.64 2.35 -2.83 -1.69 Apr. 28, 1964 -6.19 -2.42 1.02 -0.47 Jun. 25, 1964 -16.31 -2.82 -2.72 -3.29 Apr. 28, 1965 -12.43 -0.99 -2.29 -0.65 May 28, 1965 -2.06 -1.29 -0.59 -2.93 Jun. 17, 1965 0.56 -0.55 -1.28 -1.17 Jun. 23, 1965 0.57 0.16 2.27 0.09 Jul. 2, 1965 2.62 2.07 0.03 -1.79 Jul. 28, 1965 2.29 0.48 0.35 1.53 Jan. 2, 1966 4.31 1.04 1.46 -0.86 Jun. 3, 1967 -4.02 0.27 0.04 4.83 Jul. 1, 1968 -10.57 -2.62 2.22 -9.39 Feb. 6, 1969 -15.08 -1.68 -- -11.17 May 29, 1969 -16.56 -0.83 -- -0.33 Jun. 10, 1969 -5.09 -1.70 -- -5.39 Jun. 19, 1969 -3.06 0.39 -- 0.91 Nov. 6, 1969 4.30 0.29 -- -0.77 Dec. 13, 1969 -0.61 -1.57 -- 0.46 Mar. 4, 1970 -1.07 1.98 -- 0.46 Apr. 2, 1970 -0.63 1.01 -- -15.15 Nov. 13, 1970 -3.12 -1.11 -- 4.70 Dec. 16, 1970 -0.12 1.65 -- 1.93 Jan. 2, 1971 -0.72 0.81 -- -3.02 Total -84.63 -5.08 -2.32 -43.16 Event date R.J. Reynolds Industry Jan. 12, 1964 -2.61 -6.41 Apr. 28, 1964 -40.24 -48.30 Jun. 25, 1964 -26.84 -51.99 Apr. 28, 1965 -11.12 -27.47 May 28, 1965 2.18 -4.69 Jun. 17, 1965 -4.93 -7.37 Jun. 23, 1965 1.82 4.90 Jul. 2, 1965 5.04 7.96 Jul. 28, 1965 3.12 7.77 Jan. 2, 1966 11.49 17.44 Jun. 3, 1967 3.55 4.67 Jul. 1, 1968 -23.26 -43.61 Feb. 6, 1969 -11.07 -39.00 May 29, 1969 -0.16 -17.89 Jun. 10, 1969 -1.11 -13.30 Jun. 19, 1969 6.51 4.75 Nov. 6, 1969 3.50 7.31 Dec. 13, 1969 3.97 2.25 Mar. 4, 1970 -24.07 -22.70 Apr. 2, 1970 -5.74 -20.51 Nov. 13, 1970 6.62 7.09 Dec. 16, 1970 1.27 4.74 Jan. 2, 1971 -34.09 -37.03 Total -136.17 -271.39 Note: Figures correspond to accumulated loss (or gain) in stock value over 3-day window for each event. TABLE 5 Second-Stage Estimation Results: Explaining Variation in Wealth Effects from Antismoking Policies Initiated between 1964 and 1971 Model 3: Positive Model 1: Pooled Model 2: Negative Abnormal Variable Abnormal Returns Abnormal Returns Returns Constant -4.43 (3.92) -0.35 (0.33) 3.56(b) (1.37) Market share (MS) -0.46 (6.41) -2.52(a) (0.36) 1.36 (5.39) Advertising-to- 35.05 (25.68) -5.11(b) (2.33) 19.33(b) (7.53) sales ratio (ASR) Percent filter 7.53(c) (3.95) 1.68(a) (0.11) 0.67 (1.70) cigarettes (PERFILT) Percent TV and -0.05 (0.04) -0.01(c) (0.004) -0.05(a) (0.10) radio advertisements (PERTV) Tobacco leaf 1.49 (2.38) -0.26 (0.27) -2.71(a) (0.71) inventory-to- sales ratio (ISR) Firm effects Yes No No included Sample size 104 60 44 [R.sup.2] 0.32 0.98 0.90 F-statistic 2.90 510.00 49.65 Note: a, b, and c denote significance at the 1%, 5%, and 10%, respectively. Robust standard errors in parentheses.
CAR: Cumulative Abnormal Return
FCC: Federal Communications Commission
FTC: Federal Trade Commission
1. Several studies (e.g., Baltagi and Levin, 1986; Hamilton, 1972; Schneider et al., 1981) find reductions in cigarette demand resulting from antismoking policies of the 1960s. See Chaloupka and Warner (2000) and Gallet and List (2003) for surveys of this literature.
2. An exception is Eckard (1991), who uses firm-level data to analyze market share instability before and after the 1971 advertising ban.
3. Except for Johnson et al. (1991), previous studies of wealth effects associated with antismoking policies have used monthly stock price data to identify abnormal returns.
4. The game-theoretic literature addresses strategies that can be used to sustain cooperation among members of a cartel. In particular, within the framework of a repeated game, cheating on a collusive agreement today is deterred by the threat of punishment tomorrow. Hence, if profit is expected to be lower tomorrow, the cost of cheating today (in terms of the sacrifice of future collusive profit during the punishment phase) is lower. With the relative benefit from cheating increasing, this stimulates firms to cheat today (i.e., undercut the cooperative price or produce beyond the quota level of output). Therefore, to mitigate the temptation to cheat, the cartel must behave more competitively today by reducing price or increasing cooperative output.
5. However, with modifications to the empirical procedure, Johnson et al. (1991) and Lamdin (1999) fail to find similar stock price reactions surrounding the advertising ban.
6. Interestingly, according to Adams and Brock (1999), even policies not specifically directed toward the marketing of cigarettes may have affected the role of advertising in the industry. In particular, prior to the volume of health information, producers often directed advertising toward the relative safety of their product. Hence, it was not uncommon for a cigarette producer to claim its brand was healthier than rival brands. As the severity of the health effects of smoking became more publicized, however, agreements were reached in the industry to redirect advertising away from claims about the health benefits of one brand over another.
7. In fact, for years the industry sought to discredit any references to smoking and health. For example, in an internal memo dated January 29, 1964 (taken from http://tobaccodocuments.org) written by the president of Philip Morris (George Weissman), in response to the release of the surgeon general's report, at one point Mr. Weisman states "We must in the near future provide some answers which will give smokers a psychological crutch and a self-rationale to continue smoking. These answers must also point up the weaknesses in the Report and the path for future research."
8. Interestingly, both the House and Senate versions of the bill were relatively weak, because they refused to require that health warnings also be attached to advertisements. They also barred the FTC from regulating cigarette advertising for several years.
9. Also important, related to the FCC's ability to regulate the cigarette industry, on June 10, 1969, the U.S. Supreme Court upheld the constitutionality of the Fairness Doctrine.
10. Similar to the Surgeon General's Report, the industry's position was against the placement of health warnings on cigarette packages. For example, in Weissman's (1964) memo, he suggested that the industry fight any labeling action. Therefore, early on one can expect event dates corresponding to warning labels to be viewed negatively.
11. Another method to accommodate cross-correlation of abnormal returns is to aggregate the abnormal returns into a portfolio. As discussed in MacKinlay (1997), the advantage of the multivariate approach over the portfolio approach is that the former allows for the possibility that some firms have positive abnormal returns and some firms have negative abnormal returns for a given event. This information is vital for the second-stage regression, in which the authors explore the variation in abnormal returns in more detail.
12. A second-stage analysis is a common approach to identifying factors that explain variation in wealth effects; see, for example, Harris and Ravenscraft (1991), Blonigen et al. (2004), and Desai and Hines (2004). To the best of these authors' knowledge, however, this approach has not been used in previous studies of antismoking policies.
13. Data on the percentage of advertising devoted to TV and radio is only available at the industry-level.
14. Due to violations of the rank condition, however, the authors were not able to include firm dummy variables in models 2 and 3. Also, across all three models the authors initially included year dummy variables but found them to be individually and jointly insignificant. Hence, they were excluded from the reported results.
15. Loews acquired Lorillard in 1968, and hence the authors do not have abnormal returns for Lorillard after 1968.
16. The data used to estimate equations (1) and (2) came from various sources. Data on stock returns were obtained from the Center for Research in Security Prices database. With respect to the second-stage regression, data on MS and PERFILT came from Maxwell (1998). Data on ASR and ISR came from Standard and Poor's Industry Surveys (various years). Finally, data on PERTV were collected from the Federal Trade Commission (1998).
17. Similar to Mitchell and Mulherin's (1988) analysis of the advertising ban, by proposing to self-regulate cigarette advertising via the code, the industry may have "finessed" Congress into passing a warning labels bill that left advertising untouched. Also, Weissman's (1964) memo states that he was initially in favor of mild warning labels "to thwart the efforts of various states." Furthermore, according to Senator Moss at the time, the final version of the law was merely a warning "in whispered tones" (New York Times, July 14, 1965, p. 23). All of these suggest the ill effects of the warning labels legislation were much accounted for early on.
18. This result is similar to Lamdin (1999), who also finds negative wealth effects strongest for early events associated with the advertising ban. Yet such results contradict Mitchell and Mulherin (1988).
19. To place these averages in context, the value of -0.38 corresponding to the average abnormal return for the advertising ban clarifies that the events surrounding the advertising ban led to an average per-event abnormal drop in stock prices of slightly over one-third of 1%.
20. The authors also estimated model 1 with an interacting dummy variable accounting for the negative returns (i.e., dummy variable equals one if return is negative, zero if positive). In this case, the [R.sup.2] increased to 0.85. However, because there are sizable differences between the results for models 2 and 3, the authors chose to report those separately in Table 5.
21. Consistent with the results in Table 4, therefore, stock values for the larger firms fell substantially more in response to the antismoking policies.
Adams, W., and J. Brock. The Tobacco Wars. Cincinnati: South-Western College Publishing, 1999.
Bagwell, K., and R. Staiger. "Collusion over the Business Cycle." Rand Journal of Economics, 28, 1997, 82-106.
Baltagi, B. H., and D. Levin. "Estimating the Dynamic Demand for Cigarettes Using Panel Data: The Effects of Bootlegging, Taxation, and Advertising Reconsidered." Review of Economics and Statistics, 68, 1986, 148-55.
Barnett, P., T. Keeler, and T. Hu. "Oligopoly Structure and the Incidence of Cigarette Excise Taxes." Journal of Public Economics, 57, 1995, 457-70.
Binder, J. "Measuring the Effects of Regulation with Stock Price Data." Rand Journal of Economics, 16, 1985, 167-83.
Blonigen, B., and W. Wilson. "Explaining Armington: What Determines Substitutability between Home and Foreign Goods?" Canadian Journal of Economics, 32, 1999, 1-21.
Blonigen, B., K. Tomlin, and W. Wilson. "Tariff-Jumping FDI and Domestic Firms' Profits." Canadian Journal of Economics, 37(3), 2004, 656-77.
Chaloupka, F. J., and K. E. Warner. "The Economics of Smoking." Pp. 1539-627 in The Handbook of Health Economics, edited by J. Newhouse and A. Cuyler. New York: North Holland, Elsevier Science Publishers, 2000.
Desai, M. A., and J. R. Hines Jr. "Market Reactions to Export Subsidies." NBER Working Paper 10233, 2004.
Eckard, E. W. Jr. "Competition and the Cigarette TV Advertising Ban." Economic Inquiry, 29, 1991, 119-33.
Farr, S., C. Tremblay, and V. Tremblay. "The Welfare Effect of Advertising Restrictions in the U.S. Cigarette Industry." Review of Industrial Organization, 18, 2001, 147-60.
Federal Trade Commission. Report to Congress Pursuant to the Federal Cigarette Labeling and Advertising Act. Washington, D.C., 1998.
Gallet, C. A. "Advertising and Restrictions in the Cigarette Industry: Evidence of State-by-State Variation." Contemporary Economic Policy, 21, 2003, 338-48.
Gallet, C. A., and J. A. List. "Cigarette Demand: A Meta-Analysis of Elasticities." Health Economics, 12, 2003, 821-35.
Gormley, W. "A Test of the Revolving Door Hypothesis at the FCC." American Journal of Political Science, 23, 1979, 665-83.
Haltiwanger, J., and J. Harrington. "The Impact of Cyclical Demand Movements on Collusive Behavior." Rand Journal of Economics, 22, 1991, 89-106.
Hamilton, J. L. "The Demand for Cigarettes: Advertising, the Health Scare, and the Cigarette Advertising Ban." Review of Economics and Statistics, 54, 1972, 401-11.
Harris, R., and D. Ravenscraft. "The Role of Acquisitions in Foreign Direct Investment: Evidence from the U.S. Stock Market." Journal of Finance, 46, 1991, 825-44.
Johnson, M. S., R. C. Mittelhammer, and D. P. Blayney. "The Impact of Regulation on Shareholder Wealth in the Tobacco Industry: An Event-Study Approach." Agricultural Finance Review, 51, 1991, 21-34.
______. "Event Studies of Regulation and New Results on the Effect of the Cigarette Advertising Ban." Journal of Regulatory Economics, 16, 1999, 187-201.
Lamdin, D. J. "Implementing and Interpreting Event Studies of Regulatory Changes." Journal of Economics and Business, 53, 2001, 171-83.
MacKinlay, C.A. "Event Studies in Economics and Finance." Journal of Economic Literature, 35, 1997, 13-39.
Maxwell, J. C. Historical Sales Trends in the Cigarette Industry. Richmond, Va.: Davenport and Company, 1998.
Mitchell, M. L., and J. H. Mulherin. "Finessing the Political System: The Cigarette Advertising Ban." Southern Economic Journal, 54, 1988, 855-62.
New York Times. "Cigarette Warning Bill Is Passed by Congress." July 14, 1965, 23.
Peltzman, S. "Toward a More General Theory of Regulation." Journal of Law and Economics, 19, 1976, 211-40.
Rotemberg, J., and G. Saloner. "A Supergame-Theoretic Model of Price Wars during Booms." American Economic Review, 76, 1986, 390-407.
Saxonhouse, G. "Estimated Parameters as Dependent Variables." American Economic Review, 66, 1976, 178-83.
Scheraga, C., and J. E. Calfee. "The Industry Effects of Information and Regulation in the Cigarette Market: 1950-1965." Journal of Public Policy and Marketing, 15, 1996, 216-26.
Schneider, L., B. Klein, and K. Murphy. "Government Regulation of Cigarette Health Information." Journal of Law and Economics, 24, 1981, 575-612.
Standard and Poor's. Industry Surveys. Various years.
Stigler, G. "The Theory of Economic Regulation." Bell Journal of Economics, 2, 1971, 3-21.
Tremblay, C., and V. Tremblay. "The Impact of Cigarette Advertising on Consumer Surplus, Profit, and Social Welfare." Contemporary Economic Policy, 13, 1995, 113-24.
Weissman, G. Memo to Mr. Joseph P. Cullman titled "Surgeon General's Report. January 29, 1964. Available online at www.tobaccodocuments.org.
Wooster: Assistant Professor, Department of Economics, 6000 J Street, California State University, Sacramento, CA 95819. Phone 1-916-278-7078, Fax 1-916-278-5768, Email email@example.com
Gallet: Assistant Professor, Department of Economics, 6000 J Street, California State University, Sacramento, CA 95819. Phone 1-916-278-6099, Fax 1-916-278-5768, Email firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Wooster, Rossitza B.; Gallet, Craig A.|
|Publication:||Contemporary Economic Policy|
|Date:||Apr 1, 2005|
|Previous Article:||New evidence on youth smoking behavior based on experimental price increases.|
|Next Article:||Housework and the wages of young, middle-aged, and older workers.|