A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Business Administration Marketing Doctor of Philosophy 2021 ABSTRACT By Xiaoyun Zheng ................................ ................................ ................................ ................. ................................ ................................ ................................ ................ ................................ ........... ................................ ................................ ................................ ............................ ................................ ................................ ................................ ............................. ................................ ................................ ................................ ....................... ................................ ................................ ................................ .. ................................ ................................ ................................ ..................... ................................ ................................ ................................ ......................... ................................ ................................ ................................ ........................... ................................ ................................ ................................ ............................. ................................ ................................ ................................ ....................... ................................ ................................ ................................ .... ................................ ................................ ................................ .. ................................ ................................ ................ ................................ ................................ ................................ ............................... ................................ ................................ ................................ ......................... ................................ ................................ ...................... ................................ ................................ ..................... ................................ ................................ ................................ ......................... ................................ ................................ ................................ .. ................................ ................................ .............................. ................................ ................................ ........................ ................................ ................................ ................................ ...................... LIST OF TABLES ................................ .............................. ................................ .................. ................................ ................... ........................... ................................ ................................ ............. ................................ ................................ ............. ................................ ................................ ..................... ................................ ................................ ................................ ............ ................................ ................................ ............................... ................................ ................................ ...................... LIST OF FIGURES ......... ............................... ................................ .......................... ................................ ................................ ............. ................................ ................................ .......... A BRIEF BACKGROUND ON RESEARCH IN BRAND PROTECTION, RESEARCH QUESTIONS AND OVERALL DISSERTATION STRUCTURE 1 ESSAY ONE 2 Saporito Sample: Brand Protection Action Announcements I compile a significant amount of data collected from a wide range of sources to construct my sample of legal and marketing brand protection action announcements . Specifically, to obtain data for news announcements about a broad range of brand protection actions, I used Factiva and Lexis - Nexis, two databases that include a large selection of busin ess news and publications. Consistent with prior research (Sorescu, Warren, and Ertekin 2017) , I use these databases to search for indexes of all major publications worldwide. I selected consumer goods and pharmaceutical products to test my hypotheses for two reasons. First, according to information from U . S . Customs and Border Protection , consumer goods (including handbags/wallets, apparel/accessories , footwear , and personal care products ) and pharmaceutical products were the top two categories of products seized in the United States in 2018 because of IPR infringement . Therefore, brand owners from these two categories are more likely to take action considering their heavy losses. Second, these two categories offer a large number of publicly traded firms that have adequate variation in resources and marketing activities but relatively similar industr y backgrounds (Sorescu, Chandy , and Prabhu 2007). For the consumer goods sample, I used the filter function in Factiva to select news before downloading. In the first step, I focused on the companies listed in the S&P 500 Index. According to Cesareo and S töttinger (2015), brand protection actions require a great deal of investment in both time and effort from the brand owners. Therefore, companies in the S&P 500 are more likely to have the financial resources necessary to make investments to protect their brands, which make s them a representative sample for my investigation. In the second step, I category to be either co . For the pharmaceutical sample, I selected 20 global companies according to their total prescription sales in 2017 (Osinga et al. 2011) 3 . These companies are representative of this industry for my purposes because 10 firms in my sample already control 41.58% of the U.S. pharmaceutical market (Pharmaceutical Technology 2019). Pharmaceutical companies are the principal targets of counterfeits as their products are in high demand and the large profits from the counterfeits are one of the incentives for counterfeiters to engage in this illicit tra de. For both industries, my sampling frame comprises all brand protection actions announced by publicly traded firms without time restrictions. I include all articles published in all news sources included in my sample, the news needed to include at least the brand name, brand protection action (i.e., content, target, or other information), and source (Bayus, Jain, and Rao 2000) . For multiple announcements of the same brand, I selected the announcement that appeared the earliest to ensure that no other mentions of the same brand protection actions were made before this date, to eliminate any confounding effect. M y final sample consisted of 293 brand protection announcements across 34 companies listed in the S&P 500 during the 1989 - 2018 period. This sample size is sufficient as in the previous event studies, sample sizes have varied from 170 (Tipton, Bharadwaj, and Robertson 2009) , to 206 (Swaminathan, Murshed, and Hulland 2008) , and 3,552 (Borah and Tellis 2014). For short - term event studies, I examine abnormal stock returns to the announcing firms during several days centere d on the event (i.e., brand protection action announcement) to minimize potential confounding effects (Brown and Warner 1985; Tipton, Bharadwaj, and Robertson 2009). For long - term event studies, I use the cumulative abnormal returns (CARs) over at least si x months after the event date ( Jacobson and Mizik 2009; Wiles et al. 2010). Short - term event studies . In line with previous research (e.g., Tellis and Johnson 2007; Talay, Akdeniz, and Kirca 2017), I employ both the market - adjusted model (MAR) and the Fama - French - Carhart four - factor model (FF4) as benchmark asset pricing models for robustness check purpose. The m arket - adjusted model uses the average return of the entire stock market, R mt , as the proxy for expected returns and can be expressed as follows : (1) E (R it ) = R mt , where R it denotes the rate of return on the stock price of firm i on day t, and R mt represents the daily returns on the equally - weighted stock market index on day t. For t he Fama - French - Carhart four - factor model, I estimate the expected abnormal returns using four risk factors, which can be expressed as follows: (2) E ( R it ) = R ft + 1 (R mt - R ft ) + 2 (SMB t ) + 3 (HML t ) + 4 (UMD t ) w here R mt is described as previously, R ft is the risk - free rate of return at time t, SMB t is the difference between the rates of return of small - and large - market capitalization stock portfolios on day t, HML t is the difference between the returns of high and low bo ok - to - market stock portfolios on day t, and UMD t is the momentum factor. Using the stock returns data from Center for Research in Security Prices, I calculated the abnormal return by choosing 255 trading days (from 300 days to 46 days before an event) as the estimation window. I used ordinary least squares to obtain the parameter estimates, which I then used to calculate the expected return ( Zhao, Calantone, and Voorhees 2018). I computed CARs for each event by summing the daily abnormal returns across the event window. L ong - term event studies . The effectiveness of brand protection actions involves several factors that require a long - term period to assess. In other words, it may take a longer time for the effect of brand protection actions to be visible and for investors to update their expectations. C ons istent with previous literature, I compute buy - and - hold abnormal returns (BHARs) by stock over a long - term window . The benchmark comprises stocks whose risk profile is analogous to the firm over the same period. (3) = where i represents the firm, t is the month following the announcemen t, is the return of firm i in month t following the announcement, and is the return of a matched portfolio that includes all stocks with the same size, book - to - market, and momentum quintiles as firm i. Dependent Variable: CAR Event studies widely adopt CAR measured as cumulative percentage changes in stock prices across the event window, to capture the changes caused by the events (Skiera, Bayer, and Schöler 2017) . (4) C AR it = where AR it is the abnormal return of firm i o n day t (event day) and k and l are the number of days before and after the event day, respectively, used to compose the event window. For the cross - sectional analysis, I use the short - term abnormal stock returns as t he dependent variable. Independent Variable I categorized news that mention s as legal actions. Marketing brand protection actions included product - , price - , place - , and promotion - - dimensional code - place - related actions. it is coded as price - related actions. consumers to identi f y - related actions. The independent variable is a dummy variable that takes the value of 1 if the marketing brand protection actions are promotion - related and 0 otherwise. Moderators Online versus offline context. Consistent with Ertekin, Sorescu, and Houston (2018), I operationalize threat context as a dummy variable that takes the value of 1 if the threat occurs in the online context and 0 if the threat occurs offline. I PR protection. Because of access restrictions and data staleness issues in the other indices, I selected the Law and Order and the Park index to reflect the protection strength of IPR in host countries 4 . I assigned each country a score, which is the avera ge of the standardized scores of the two indices; this is consistent with the calculation method in Zhao ( 2006) . B rand p rotection c ommitment. I measure b rand protection commitment as the degree of a investment in tackling the counterfeiting , rated on a seven - point Likert scale, with 1 being the l ow est commitment , 4 being a m oderate commitment , and 7 being the h igh est commitment . One of the authors and a research assistant rated the brand protection commitment for the pharmaceutical industry subsample. I evaluate the reliability and consistency of the coding using the intra - class correlation coefficient (ICC). Both ICC scores a re considered high (pharmaceutical subsample: .91; consumer goods subsample: .91) indicating a high level of intercoder reliability and consistency (Bliese 2000). Control Variables Action multiplicity. It is not uncommon for firms to send multiple signals to the market, either intentionally or unintentionally. According to signaling theory, the aim of such behavior is to increase signal observability, or the extent to which outsiders are able to notice the signal (Connelly et al. 2011). A higher signal freque efforts to make the signals clearer. I operationalized action multiplicity by counting the actions mentioned in the news. For example, a news story mentioned that Procter & Gamble was filing lawsuits for th ree defendants; thus, this news includes three signals. Firm size. Firm size indicates the scale and scope of operations (Aldrich 1972). Previous s future financial performance in this context (e.g., Ertekin, Sorescu, and Houston 2018). Consistent with the prior measurement, I operationalize firm size as the natural logarithm of the total assets of the corresponding firm. L everage . uses debt to acquire additional assets. Previous research suggests that varying degrees of financial leverage affect stock market returns (e.g., Wiles, Morgan, and Rego 2012) . I measure leverage as the ratio of long - term debt to total assets . Cash flow common determinants of changes in stock prices associated with marketing events ( Mazodier and Rezaee 2013). I measure cash flow as the net operating income before depreciation adjusted for working capital accruals (Luo 2009). M arket value. Market value is a firm - capitalization (Wiles, Morgan, and Rego expectation s financial performance . I operationalize market value by multiplying the price of a stock by its total number of outstanding share s (Kumar and Shah 2009). News specificity. Consistent with Talay, Akdeniz, and Kirca (2017), I operationalize the specificity of brand protection actions by a count variable for the number of words in each story. A list of constructs, measures, and data sources is provided in Table 2 - 1. M odel Development Next, to examine how brand protection actions and the moderation effects of different factors influence stock market reactions, I used the CARs from short - term event studies as the dependent variable and conducted a cross - sectional analysis. I adopted a two - stage Heckman (1979) model to address the potential self - selection bias caused by systematic differences between firms that planned to take brand protection actions and those that did not. In the first stage, I ran a probit selection model to estimate the probability that a firm would take a brand protection action. T he value of the dependent variable was 1 if the firm took action and 0 if otherwise . I supplemented my original sample with a matched sample of firms that did not engage in brand protection actions during the sample time frame. Consistent with previous research ( e.g., Wiles, Morgan and Rego 2012), I applied inclusion criteria to ensure that the matched firms are similar to the firms in the original sample in terms of being targeted by counterfeiters: these matched firms needed to be publicly traded firms that belong to the same industry and have simi lar firm value (within 2 sample). The resulting sample has 7 , 903 observations, which include 398 focal firm - year observations and 7 , 505 matched firm - year observations. In terms of the exclusion variable, I sel number of brand protection actions undertaken in the same industry in the previous year. This instrumental variable meets the requirement of relevance assumption such that the action intensity of the pee also meets the exclusion restriction, as this industry - level variable would be the same for all focal ific firm should not be affected. In addition, I added several firm - level variables to the first - stage model. The first - stage model is as follows: I i ,t = 0 + 1 Industry intensity i ,t + 2 Cash flow i ,t + 3 Market value i ,t + 4 Leverage i ,t + 5 Firm size i ,t i ,t where i denotes the firm, t denotes the time, and I i ,t denotes whether the firm has taken the brand protection actions or not. The second stage of the Heckman procedure involved a least squares regression on the CARs. I included the inverse Mills ratio obtained from the first - stage selection model, the hypothesized independent variables, and a set of control variables in the regression model of CARs. CAR (0,0)i , k 0 1 Promotion dummy i , k 2 Commitment i , k 3 IPR i , k 4 Context i , k + 5 Promotion dummy × Commitment i , k 6 Promotion dummy × IPR i , k 7 Promotion dummy × Context i , k 8 News specificity i , k 9 Acion multiplicity i , k 10 Cash flow i , k + 11 Market value i , k 12 Firm size i , k 13 Leverage i , k 14 Inverse Mill Ratio i , k i , k where i denotes the firm and k denotes the event. Results and Discussion Table 2 - 2 presents means, standard deviation, and correlations for all continuous variables and control variables. In general, the correlation between varia b les was lower than the upper threshold (r = .50) for low correlation conditions (Voorhees et al. 2016) . To assess the potential threats from multicollinearity, I checked the average and maximum variance inflation factor (VIF) values and f ou nd the VIFs well below the acceptable cutoff of 10 (average VIF =1. 70, maximum VIF = 4 . 04 ). Therefore, I conclude that multicollinearity is not a threat to the validity of my findings. S hort - T erm Stock Market Reaction I tested several event windows surrounding the brand protection action announcement date and report the results of CARs for window (0,0), ( - 1,2), and ( - 1,0) in Table 2 - 3. I chose CARs for the (0,0) window as they have the most significant t - statistic (Swamin athan and Moorman 2009). Although my hypotheses focus on the impact of marketing brand protection actions, my empirical analysis examines the stock market reactions to both marketing and legal actions to obtain additional insights. The results indicate tha t the average CAR for all brand protection actions is positive and significant (MAR: CAR = .20%, p < .05; FF4 : CAR = .21%, p < .01). Consistent with my - term stock market reaction is positive and signif icant (MAR: CAR = .17%, p < .10; FF4 : CAR = .21%, p < .05), in support of H 1 (a) . Regarding the legal brand protection actions, the CAR results are positive and marginally significant, opposite to my expectation and findings from previous research (Ertekin, Sorescu, and Houston 2018)(MAR: CAR = .27%, p < .10; FF4 : CAR = .21%, n.s.). There are several reasons why my findings differ from the results documented in Ertekin, Sorescu, and Houston (2018). First, their results are based on a more heterogeneous sample containing 1,918 legal brand protection cases filed by 540 firms from 214 di fferent industries , whereas my sample focuses on cases in consumer goods and pharmaceuticals, the two most vulnerable and targeted industries. While investors in other industries may need to downgrade nexpected negative information regarding brand infringement and related consequences, investors of consumer goods and pharmaceutical industries may be already fully aware of existing counterfeiting issues. As such, their way of weighing competing positive and negative signals associated with legal brand protection actions should be different from average investors, such that they tend to view these legal actions much more positively. my arguments. They investigated stock marke t reactions to patent infringement litigations in the IT industry. Their results suggest that the news of patent infringement litigation is associated with significantly positive abnormal returns for plaintiff firms . These findings indicate that although legal departments realize the high costs of patent nts are undoubtedly one of the most important intellectual properties and competitive advantages; the significance of patent protection in the IT industry is akin to the critical role of trademark protection plays in consumer goods and pharmaceutical indus tries. Second, Ertekin, Sorescu and Houston (2018) focus on trademark infringement lawsuits that include seven type s of brand threats (i.e., counterfeiting, gray market, brand misappropriation, copycats, false advertising, cross - industry brand misappropri ation , and cross - industry imitation ) . Counterfeiting is the most severe infringement (account ing for 31.13% of the total sample) and cross - industry imitation the least severe. Different level s of trademark infringement indicate the different potential damage levels to brand equity, which are critical cues in helping brand owners a nd investors assess the potential damage to brand equity and determine whether actions are necessary. The more sever e the trademark infringement, the more investors are concerned about my study, I focus on counterfeiting only; therefore, investors should r important market - based assets from the worst infringement crisis. These signals ensure investors that the brand owner is taking the counterfeiting issue seriously. Thus, legal actions in this context may well be associated with a positive short - term stoc k market reaction. Long - T erm Stock Market Reaction To test the hypotheses regarding the long - term effects of brand protection actions on stock market reactions, I conducted separate long - term event studies for all brand protection actions, legal actions, and marketing actions (see Table 2 - 4). My results show that twelve months after the brand protection actions, firms experience positive average monthly BHARs of 1.25% ( p < .10). While I expected that legal actions are associated with a positive stock market abnormal return, the results show that the average monthly BHARs for legal actions are - .3.46% ( p < .05) in the first 12 months. For marketing brand protection actions, the results co nfirm my expectation that they lead to a positive stock market reaction in the first 12 months with an average monthly BHAR of 2.93% ( p < .01). Thus, H 1 (b) is supported. Heckman Model R esults In Table 2 - 5, I reported the results of Heckman Model. Model 1 is the base model that only includes the independent variable of interest (promotion dummy) in the second stage. Model 2 is a full model that includes the independent variable of its interaction with three moderators. The choice of taking brand protection action is the dependent variable in the selection model and Short - term CAR for window (0,0) is the dependent variable in the regression model. I focus my 2 = 25.34, p <.05) than the main effect model (Model 1). In the first stage, the exclusion variable = .034, p < .01). This suggests that the more peer firms take brand protection actions, the higher is the chance that the focal firm will follow and take similar actions. The results also indicate that decisions to take action. Th e leverage level of the firms does not show a significant relationship with the decision of taking brand protection actions. In the second stage, the coefficient of the main effect is positive and significant ( = .028, p < .01), suggesting that among the four types of marketing - mix responses to brand infringement, promotion - related actions are significantly associated with more positive stock market reactions than the other three types of actions, in support of H 2 . Regarding the moderating effects, H 3 pr edicts that the b rand threat context would moderate the relationship between marketing brand protection actions and firm value, such that the relationship is weaker for promotion - related actions when the threat involves the online context. I found marginal ly significant evidence from the result ( = - .009, p < .10), therefore, H 3 is partially supported. For the second moderator, I predict that the IPR protection strength would moderate the relationship between marketing actions and firm value, such that the relationship is stronger for promotion - related actions when the protection strength is higher. I found a positive coefficient for the interaction term ( = .005, p < .05), in support of H 4 . Finally, for H 5 , I propose that brand protection commitment moderates the relationship between marketing actions and firm value, such that the relationship is weaker for promotion - related actions when commitment is higher. The interaction effect is significant ( = - .006, p < .05), in support of H 5 . As an additional analysis, I conducted Heckman model analysis using long term CAR with various windows, including 6 months, 12 months, and 24 months after the event. The results are consistent across windows. I report the results of BHARs of 24 months in Model 3 of Table 5. The findings show that the main effect of the promotion dummy is not significant ( = - .078, p > .10), which suggests that promotion - related actions exert a similar effect as other types of marketing actions in the long - term. However, two out of three moderators have a significant moderating effect on the brand protection - stock market response relationship. Specifically, host ion dummy and stock market reaction ( = - .212, p < .01), which means t hat when protection strength is greate r, the relationship is weaker for promotion - related actions compared to other types of marketing actions. For the effect of threat context, the rel ationship is weaker for promotion - related actions when the threat involves the online context ( = - .307, p < .05), consistent with short - term results. R obustness Tests In my sample, I have cases in which one news article involved several brand protection actions/ events . To ensure that my results are robust, I conducted the analysis exclud ing the cases with multiple protections and then estimate similar models using this reduced sample. The results are consistent with Model 2 in terms of the direction and significance of the coefficients. In addition to the current control variables, prior li terature suggests that R&D expenditure ( e.g., Borah and Tellis 2014) , advertising expenditure (e.g. Chen et al. 2012) , and industry type (dummy variable ) may affect the magnitude of abnormal returns. My robustness test results suggest that with these addit ional control variables, the previous findings still hold. In 2010s, the total cost of counterfeits for firms was already $250 billion per year in the United States alone (Chaudhry and Zimmerman 2013). Despite the theoretical and practical (Yang and Sonmez, 2017, p. 423). Thus, both scholars and practitioners have called for more rigorous analysis and a better understanding of the impact of brand protection efforts ( e.g., Ertekin, Sorescu and Houston 2018 ; Wilson, Grammich and Chan 2016 ; Yang and Sonmez 201 7 ). Brand protection efforts require significant firm - level investments, and both managers and investors hope to ensure that these investments will strengthen brand equity and create value for the firm. This study extends prior literature by focusing on the re lationship between brand protection actions and their short - and long - term stock market reactions from a marketing perspective. Previous research has shown that legal brand protection actions are an effective tool to tackle counterfeiting issues. My event counterfeit activity can indeed be an effective tool to avoid the erosion of brand equity from to because the s and help investors actions, they evaluate not only the contents but also the intent of the signals ( Stuart and Muzellec 2004). Indeed, I find that the short - and long - term impacts of brand protection actions differ for legal versus marketing brand protection actions. Although legal actions have a positive impact on short - term stock market prices, in the long run , they hurt firm value as their impacts become negative. Marketing actions, on the other hand, are favored by investors both in the short - term and long - term. As such, my findings provide a novel insight into the favorable stock market response to m arketing brand protect ion actions . that bran Implications for R esearch Brand protection is an important issue that is under - investigated in the marketing literature. Specifically, extant research focuses extensively on the demand side, which examines the implications of counterfeiting for consumers (for a meta - analysis on the consumer responses to counterfeited products, see Eisend 2016 ). Nevertheless, demand - side studies neglect the role of the supply side i.e., the eff orts of brand owners to protect their brands and the impact of these efforts on firm performance. In this regard, Ertekin, Sorescu, and Houston (2018) are one of the few studies that contribute to this research stream by investigating the consequences of t My study extends this research stream by exploring an alternative solution --- using marketing mix variables to combat counterfeiting. I compare the effects of both l egal and of marketing actions by examining their typologies and contingency effects on stock market reaction. In general, my findings indicate that the effects of protecting brand equity using marketing actions have both a positive contemporaneous impact and a positive carryover effect over time, in line with related brand equity research (Mizik 2010). Categorizing different marketing signals based on marketing - mix , this study reveals that not all marketing brand protection actions are equally influential under all circumstances from - related brand protection actions to the other types of action s. Furthermore, this study reveals that under specific boundary conditions, the strength of the relationship between marketing brand protection actions and stock market reaction is altered. In particular, investors may adjust their interpretations of firms protection efforts depending on firm - country - IPR protection strength) and environment - level (e.g., brand infringement context) factors. Implication for Practice M y research findings are important for managers who are facing rampant counterfeit activity in recent years. Brand owners are often worried about disclosing their brand protection actions out of concerns about informing investors that the brand is under att ack. However, this study commitment and ability to protect brand equity one of their most important firm assets. Specifically, I empirically confirm that marketing brand protection actions can elicit more positive stock market reactions than legal brand protection actions in the long run. As I discussed, marketing brand protection actions have their own advantages; therefore, brand managers should use marketing tools to combat counterfeit activity before legal actions. level, and the legislative and cultural environment when employing brand protection actions. Stumpf, Chaudhry, and Perretta ( 2011) argue that brand managers must experiment with different anti - counterfeiting actions by country and brand; what works will be determined empirically. This suggestion complements my findings. As my data indicate, firms most widely adopt promotion - related actions, which account for almost 71.33% of all marketing brand protection actions used. Analytic res ults also show that investors favor promotion - related actions more than the other types of marketing brand protection actions. This finding might not be a coincidence. A possible explanation is that brand protection managers are most familiar with promoti on - related actions; therefore, they tend to employ these actions repeatedly. In turn, investors are more likely to be more confident about these actions because they possess more knowledge about them than about the other types of actions. The consequence i s that managers will be biased toward promotion - related actions as they are associated with more positive stock market reactions and perhaps neglect the other brand protection actions. Managers should use promotion - related brand protection action more judi ciously as my commitment level, the IPR environment, and the threat context all significantly affect the impact infringement involves an online th reat, the impact of promotion - related actions on firm value will be undermined more significantly than the impacts of other types of marketing actions. This impl ies that under this condition, implementing product - or place - related actions at the very begin ning might lead to a - market performance than promotion - related actions thereby restricting their slack resources and decision flexibility to support subsequent brand protection actions. Thus, I recommend that brand owners select appropriate brand protection actio ns depending on the circumstance under which the firm will implement these actions, rather than use one specific action only. Finally, the findings highlight the importance of taking the most suitable communication strategy. This study indicates that key judgment s . When brand owners articulate their brand protection actions, details such as the brand threat severity, the level of monetary/nonmonetary investment commitment, and their implementation plans all carry critical i nformation about their willingness and ability to protect the brand. The clearer the s . This could largely eliminate information distortion and avoid unnecessary negative con sequences. Limitations and Future Research Despite its contributions, my study has several limitations that additional research could address. First, with the limited sample size, my analyses had limited statistical power that prevented me from examining s ome potentially interesting moderators such as the number of suggested that marketing alliances have a significant impact on firm value since an alliance enables firm s to access new knowledge and new markets ( Swaminathan and Moorman 2009). Future research could examine how different characteristics of partnerships among brand owners would influence the relationship between marketing brand protection actions and firm va lue. Similarly, innovation is a double - innovation helps firms gain competitive advantages against counterfeiters by offering customers more products/services that better satisfy their needs; on the other hand, innovation is risky and would alter the link between marketing brand protection actions and firm value. Further research could expand the sample size to all publicly traded firms available in a database . Doing so might provide additional important observations. S econd, given the little attention that research on the brand protection topic has attracted, marketing scholars can build on my study by exploring additional characteristics of brand protection action signals. For example, signal fit (the extent to which the signal is correlated with unobservable quality), signal timing (the time span between multiple signals or repeating signals) , and signal consistency (agreement between signals from one source) all potentially Besides, researchers could explore how marketing brand protection action s may affect other product - or firm - level performances, such as existing product sales, new product success, corporate social performance, or debt - holder risk. L ike other marketing investments, brand protection investments require evidence to prove thei r productivity. The impact of strategic marketing investment on related financial benefits has profitability ( e.g., Rust et al. 2004; Katsikeas et al. 2016). However, due t o the sensitivity and limited access to data regarding financial losses caused by counterfeiting, I could not further examine how brand protection actions could contribute to firm performance financially (using other financial metrics such as s ales revenue , p rofit , EVA, and ROI). If such data are available, future research could better address the effectiveness of brand protection actions. E SSAY TWO A considerable number of studies has looked at the links between various marketing phenomena and equity - holder risk. For systematic risk, t op ics that have been investigated in the past include service innovation (Dotzel and Shankar 2019), marketing alliance ( Thomaz and Swaminathan 2015), strategic orientation ( Bhattacharya, Misra, and Sardashti 2019 ), and corporate social performance ( McAlister, Srinivasan, and Kim 2007) ; in terms of idiosyncratic risk, examples include customer satisfaction ( Tuli and Bharadwaj 2009) emphasis ( Han, Mittal, and Zhang 2017) and consumer negative voice (Luo 2007). Research that r elates to brand management also gained increasing recognition. Impact of consumer - based brand equity ( Rego, Billett, and Morgan 2009 ), b rand q uality ( Bharadwaj , Tuli , and Bonfrer 2011), b rand architecture strategy ( Hsu, Fournier, and Srinivasan 2016 ), and b rand r ating d ispersion ( Luo, Raithel, and Wiles 2013 ) on firm risk have all been examined. brand owners need to comply with evolving regulations of packaging / labeling and safety protocols in any case uncertainty about future profitability, which threatens the cash flow consistency ( Tuli and Bharadwaj 2008) . The consequence of this would be a harder justification for decision - makers to invest in these costly brand protection actions, as they may consider them to be riskier. Sample: Brand protection action announcement . brand protection actions. I utilized both Factiva and Lexis - Nexis, two databases that include a broad selection of business and news publications, to search the inde xes of all major publications worldwide, consistent with prior research (Sorescu, Warren, and Ertekin 2017) . For consumer goods samples, I utilized t he filter function in Factiva to select news before downloading. My first step is to focus on the companies that are listed on the S&P 500 Index. As literature has indicated, brand protection actions require a large number of investments in both time and e ffort from the brand owners (Cesareo and Stottinger 2015). Therefore, I believe that companies that belong to the S&P 500 acquire the necessary resources to make investments to protect their brands, which makes them very representative samples in the initi al investigation of my research questions. Within to be either co . For pharmaceutical products, I selected the top twenty global companies based o n their total prescription sales in 2017, similar to what Osinga et al. (2011) had used for their selection criteria. They are Pfizer (US), Novartis (Switzerland), Roche (Switzerland), Merck & Co (US), Johnson & Johnson (US), Sanofi (Franc), GlaxoSmithKlin e (UK), Abbvie (US), Gilead Sciences (US), Amgen (US), AstraZeneca (UK), Bristol - Myers Squibb (US), Eli Lilly (US), Teva Pharmaceutical Industries (Israel), Bayer (Germany), Novo Nordisk (Denmark), Allergan (US), Shire (Ireland), and Takeda (Japan). Boehri nger Ingelheim (Germany) is not the US public traded company, therefore I excluded it. These companies are representative of this industry; just ten firms out of my samples already control the top 47.9 percent of the US market ( Medical Marketing & Media 2017). These pharmaceutical companies are the biggest targets as their products are highly demanded in the market, and the huge profits from counterfeit medicines are one of the incentives for counterfeiters to conduct illicit tra des. For both industries, my sampling frame comprises any brand protection action statements announced by publicly traded firms without the restriction of timelines. The reason for targeting public cratic risk, which is measured as the standard deviation of a . I include all articles across all news sources target or other information) and sources to be considered as a credible signal (Bayus, Jain, and Rao 2000) . When multiple announcements for the same brand protection actions are identified, I selected the announcement that appeared at the earliest date to ensure that no other mentions of the same brand protection actions were made prior to this date. M y final sample consisted of 293 brand protection announcements across 34 companies listed in the S&P 500 during the 1989 - 2018 period. Dependent Variable: Firm - Idiosyncratic Risk I estimate idiosyncratic risk using the Fama - French - Carhart four - factor model, which is widely adopted by previous scholars ( e.g., Srinivasan and Hanssens 2009) . Idiosyncratic risk accounts for the part of the risk associated with firm - specific factors; it daily return volatility that cannot be explained by changes in average market portfolio returns. The model can be expressed as below: E ( R it ) = R ft + 1 (R mt - R ft ) + 2 (SMB t ) + 3 (HML t ) + 4 (UMD t ) Where R mt represents the daily returns on the equal - weighted stock market index on day t, R ft is the risk - free rate of return at time t, SMB t is the difference between the rates of return of small - and large - market capitalization stock portfolio on day t, HML t is t he difference between the returns of high and low book - to - market stock portfolios on day t, and UMD t is the momentum factor. Using data from CRSP, I calculated the idiosyncratic risk for each firm based on daily stock returns. Independent Variable considered legal actions. In terms of marketing brand protection actions, they are categorized into product - related, price - related, promotion - related, and place - re lated. News mentioning identi f y - promotion - make a direct purchase from place - - related actions. Lastly, if - related actions. The independent variable is a dummy variable that captures whether the marketing brand prote ction actions are promotion - related. The value takes 1 if the news is product - related; otherwise, the value takes 0. All price - related, promotion - related, and place - related brand protection actions are categorized as the non - product - related actions. Moder ating Variable IPR protection. Because of access restrictions and data staleness issues in the other indices, I selected the Law and Order and the Park index to reflect the protection strength of IPR in host countries. I assigned each country a score, wh ich is the average of the standardized scores of the two indices; this is consistent with the calculation method in Zhao ( 2006) . Long - term orientation . L ong - term orientation reflects the fostering of virtues oriented toward future rewards in particular, perseverance and thrift Hofstede, Hofstede , and Minkov 2010) . I adopted the scores from Hofstede, Hofstede , and Minkov ( 2010 ), which is ranging from 1 - 1 00. The higher the score, the more long - term oriented a country is considered to be. Regulatory quality . I adopted the scores from Kaufmann, Kraay, and Zoido - Lobatón ( 1999) to measure regulatory quality. This indicator measures the incidence of policies such as trade - related or foreign investment - related , as well as perceptions of those regulations . Higher ratings correspond to better outcomes. Control variable Action Multiplicity. It is not uncommon for firms to send multiple signals to the market, either intentionally or unintentionally. Using terms from signaling theory, such behavior from firms is aiming to increase the signal observability, which refers to the extent to which ou tsiders are able to notice the signal (Connelly et al. 2011). Higher signal frequency (Janney and Folta I operationalized action multiplicity by counting the actions mentioned in the news. F or example, a news story mentioned that Procter & Gamble is filing the lawsuits for three defendants; this news is considered as including three signals. Two raters read each news carefully and counted the number of actions mentioned in the news. News Spec ificity. Although there are no studies suggesting that more words in the news are significantly associated with the higher signal credibility, I assume that it is possible that news with longer lengths typically carr ies more information. In Talay, Akdeniz, and Kirca (2017), they operationalized specificity of a new product pre - announcement for automobile model i by using a count variable for the number of characteristics about the car mentioned in the preannouncement. I, therefore, ad o pted this operationalization and measured the news specificity as a count variable for the number of words of each news. Online vs. offline context. Consistent with Ertekin, Sorescu, and Houston (2018), I operationalize threat context as a d ummy variable that captures whether the infringement occurs online or offline. The value takes 1 if the threat involves an online context and takes 0 if it is a pure offline threat. Advertising spending . This variable is measured as annual advertising ex penditure of the corresponding firm. Research and development (R&D) spending . This variable is measured as annual R&D expenditure of the corresponding firm. Industry type. In my sample, I have data from two industries only. Hence, I operationalize industry type as a dummy variable that takes the value of 1 if the focal brand owner is from the pharmaceutical industry; otherwise, it takes the value of 0. Other V ariable Cash flow common determinants of changes in stock prices associated with marketing events ( Mazodier and Rezaee 2013; Pruitt et al. 2004 ). Cash flow is m easured as the net operating income b efore depreciation adjusted for working capital accruals (Luo 2009). M ar ket value. Market value is a firm - financial performance ; therefore, it is important to control it. Market value is operationalized as the product of the price of a stock by its total number of outstanding share s (Kumar and Shah 2009). Firm Size. Firm size indicates the scale and scope of operation (Aldrich 1972 ). Previous and Kirca 2012; Talay, Akdeniz and Kirca 2017). Cons istent with prior measurement, I operatio nalized firm size as the natural logarithm of the total assets of the corresponding firms. L everage uses debt to acquire additional assets. Previous literature suggests that varying degrees of financial leverage impact the stock market returns (e.g. Wiles et al. 2012). I measured leverage as the ratio of long - term debt to total assets . A list of constructs, measures, and dat a sources is provided in Table 3 - 1. M odel Development I adopted a two - stage Heckman (1979) model to address the potential self - selection bias caused by systematic differences between firms that planned to take the brand protection actions and those who did not. In the first stage, a probit selection model is used to estimate the probability that a firm would take a brand protection action. T he value of the dependent variable was 1 if the firm took the actions and 0 if it did not. I supplemented my original sample with a matched sample of firms that have not en gaged in brand protection actions. Consistent with previous literature (e.g. Ertekin, Sorescu and Houston 2018; Wiles, Morgan and Rego 2012), the following inclusion criteria are applied to make sure that matched firms are similar to the firms in my origin al sample when it comes to the possibility of being targeted by the counterfeiters: these matched firms are public - traded firms that belong to the same industry, and have similar firm value (within 2 The resulting sample has 7903 observations, which include 398 focal firm - year observations and 7505 matched firm - year observations. total number of brand protection actions conducted in the same industry in the previous year. This instrumental variable meets the requirement of relevance assumption such that the action al so meets the exclusion restriction as this industry - level variable would be the same for all focal impacted by this kind of variable. In top this exclusion variabl e, I added some firm - level variables to the first stage model. The first stage model is as follows: I i ,t = 0 + 1 Industry intensity i ,t + 2 Cash flow i ,t + 3 Market value i ,t + 4 Leverage i ,t + 5 Firm size i ,t i ,t , where i denotes the firm, t denotes the time, I i ,t denotes whether the firm has taken the brand protection actions. The second stage of the Heckman procedure involved a least squares regression on the firm idiosyncratic risk , and I included the Mills lambda from the first - stage selection model, hypothesized independent variable as well as the control variables. Risk i , k 0 1 Promtion dummy i , k 2 IPR i , k 3 Long - term orientation i , k 4 Regulatory quality i , k 5 Promtion dummy * IPR i , k 6 Promtion dummy* Long - term orientation i , k 7 Promtion dummy* Regulatory quality i , k 8 Acion multiplicity i , k 9 News specificity i , k 10 Brand threat context i , k 11 Adverting i , k 12 R&D i , k 13 Industry type i , k 14 Inverse Mill Ratio i , k i , k , where i denotes the firm, and k denotes the event. Table 3 - 2 presents means, standard deviation, and correlations for all variables in my hypotheses as well as the control variables. In general, the correlation between varia b les was lower than the upper threshold (r = 0.50) for low correlation conditions (Voorhees et al. 2016) . To assess the potential threats from multicollinearity, I checked the average and maximum variance inflation factor (VIF) values to find that VIFs are well below the acceptable cutoff of 10. Therefore, I concluded that multicollinearity is not a threat to the validity of my findings. Heckman Model R esults Overall , m 2 = 113.98, p < .01). Table 3 - 3 presents the estimation outputs. I report the results of two m odel s . Model 1 is the base model that only includes the independent variable of interest ( product dummy) and the control variables in the second stage. Model 2 is a full model that includes the independent variable and its interaction with three moderators. The choice of taking brand protection action is the dependent variable in the selection model and firm - idiosyncratic risk is the dependent variable in the regression model. I will focus on the interpretation of Model 2 as it is the full model. For Model 2, in the first stage, the exclusion variable industry intensity has a significant impact = .020, p < .01). This means brand In the second stage, I find strong evidence that the main effect is negative an d significant ( = - .005, p < .01), which supports H1 that promotion - related brand protection actions are associated with less firm - idiosyncratic risk compared to the other three types of actions. H2 negative ly moderates the relationship between marketing brand protection action announcements and idiosyncratic risk, such that the relationship is stronger for pro motion - related actions when IPR protection is higher . I found that the test result is consistent wit h the hypothesis ( = - .002, p < .05), therefore, H2 is supported. Regarding H3, the estimation result reveals that the moderating effect of long - term orientation positively impacts the relationship between promotion - related actions and firm - idiosyncratic risk ( = 6.28e - 5 , p < .05), so H3 is supported. In terms of the last moderator, as predicted in H4, a = .004, p < .01). Robustness Checks I also conducted a few robustness checks to ensure that my results are robust. First of all, I tried a few different set s of control variables based on the prior literature, and the results still hold. Second, I utilized a different type of standard error in my model specification, and this does not change my main model results either. Finally, I used idiosyncratic risk that are based on the period of two months after the event date as the dependent variable, and the model results are still consistent with my current model. The result s of the hackman model suggest a significant impact of marketing brand protection actions on firm risk and an interesting interplay between the contextual factors and brand Prior research has delved much into how consumers perceive and react to counterfeit products as well as the consequences of the counterfeiting on genuine brands ; my research switches the focus from the individual - level to a firm - level and country - level effect , and particularly examines different aspects of institutional contexts regulatory, cognitive, and normative component of a country institutional profile. In the brand protection scenario, I argue - term orientation , and regulatory quality respectively capture the three components of this construct. The empirical tests lent support to my hypotheses. First of all, f rom the view of investors , promotion - related actions among four types of brand protection actions seem to be a better choice to start with when conducting brand protection activities in foreign countrie s . However, this effect is contingent on institutional factors. When conducting brand protection actions in countries that feature strong IPR protection , firms stock price returns will be higher for promotion - related actions than for other three types of actions, indicating that investors do believe that certain type of actions would be more effective or appropriate based on the regulatory environment of different markets. Furthermore, host cultural characteristics also play a key role in the development and implementation of brand protection actions globally. My research is able to show that the level of long - term orientation in each country will modify the risk reduction effect of promotion - related actions , as the stock prices more fluctuate in countries that possess higher long - term orientation . Finally, regulatory quality as the reflection of the The more regulative a society, the less advantageous promotion - related actions wou ld be considered by investors over the other three type s of actions. The a bove new insights offer both theoretical and managerial implications, and I will discuss them in the following section. Implications for Research This study contributes to the branding literature by investigating brand protection activities and their impact on firm performance, a very managerially - relevant but under - investigated phenomenon that often keeps brand managers up at night. Specifically, I approached this topic from the marketing - finance interface perspective and explored the financial impact of brand protection actions on firm - idiosyncratic risk. Prior literature has established the relationship between brand protection actions and abnor mal stock returns (Ertekin, Sorescu, and Houston 2018), however, I propose that firm risk is another critical financial dependent variable that merits further investigation. As Luo and Bhattacharya ( long - term shareholde r value is influenced not only by the expected size and growth of stock returns (i.e., the first moment) but also by stock price volatility (i.e., the second moment cash flow and higher risk ( Srivastava, Shervani, and Fahey 1998 ). Since investments for brand protection actions are nontrivial, it would be a good supplement for current literature to understand how to better mitigate risk. My results suggest that compared to other brand protection actions, pro motion - related actions seem to have a stronger mitigation effect. Furthermore, my research provides meaningful knowledge of brand protection activities from an institutional perspective. Although we have rich literature regarding perception / reaction to counterfeit issues, and corresponding underlying mechanisms that explain prior findings, we still lack understandings about the potential impacts from a more holistic view. I t is well recogn ized that differences in country - level institutional environments intensify the information asymmetry among parties nested in different countries (e.g., Roth & O'Donnell, 1996). My results resonate with this view and confirm that regul ations and culture of different societies would H ence, I add a critical layer of granularity to comprehending the contingency impacts of brand protection efforts in the global market . Drawn on the theore tical augments about the country institutional profile, my work identified three variables that capture the domain - specific nature of this construct in the context of brand protection . Consequently, this research is the first of its kind to discuss the characteristics of institutional environments and their contingency effect s in the brand protection context . Implication for Managers There is no doubt about the necessity to protect a br and. The critical question is how to effectively manage the risk associated with the uncertainty of outcome and rapid - changing counterfeiter behaviors. T his research provides guidance to managers for a deeper understanding reaction toward brand probation actions, and a better decision - making process brand protection var ies by the types of actions. Pro motion - related actions are co nsidered to be less risky, compared to other types of brand protection actions. But this is contingent on country - level factors such as rules and laws, cultur al tradition, and social norms . Thus, managers should weigh their stakes and make a less risky res ource allocation decision accordingly. My work discusses three aspects of a country institutional profile, and interestingly, their influences for the same type of brand protection actions vary a lot. For example, the risk reduction effect of promotion - related actions would be enhanced in a country that ha s strong IPR protection; however, such an effect would be mitigated if a country possesses a high level of long - term orientations. Therefore, in different countrie s, brand managers need to diligently design their brand protection strategy and make trade - off s based on various factors so as to make the optimal decisions. Furthermore, there may well be a diversification within the same country when it comes to social norms and cultur al differences. Take Switzerland for instance, German - and French - speaking regions of this country demonstrate different sets of formal and informal rules, norms, and value systems ; such regional culture heterogeneity is due to the history and language divergence, which is something that deserves more attention as well ( Hofstede, Hofstede and Minkov 2010) . Given the complexity and diversity of brand protection tasks, it may be wise to empower subsidiaries to make strategic decisions in their operating environment. This is supported by my data , such that a significant amount of brand protection actions is carried out by local subsidiaries of multinational enterprises. Subsidiary autonomy and been found to be positively related to the subsidiary performance ( Geleilate, Andrews, and Fainshmidt 20 20; Slangen and Hennart 2008 ). This is extremely beneficial for brand protection efforts for two reasons: on the one side, local teams are more familiar with the institutions that the organization is embedded, and they also have a better idea of the particular counterfeiting challenges within the region, therefore , they can react more promptly and accurately. On the other hand, higher a utonomy tend s to elicit stronger team morale and motivat es them to work more devotedly and diligently ( Lazarova, Peretz, and Fried 2017). Fighting with counterfeiters is tough. Empowering subsidiaries to take more controls/ responsibilities will ensure that they have more resources needed and are willing to engage in such challenging work. The limitations of this study provide opportunities for future research. First of all, because I only selected the pharmaceutical industry and consumer goods industry to conduct the data collection, this might limit the generalizability of my findings. Future research could expand the sampling frame to all public traded firms across the industries; by doing this, more observations will be available, which can introduce more variation in the dataset. According to marketing literature, besides country - level facto rs, industrial - level factors such as industry type ( Srinivasan, Lilien, and Sridhar 2011) , the d emand Instability ( Han, Mittal, and Zhang 2017), market/environmental turbulence, innovativeness , and competitive intensity , may possess a differential effect on firm performance ( e.g., Jaworski and Kohli 1993 ; Kirca, Jayachandran, and Bearden 2005). It would be interesting to see what roles these factors could play in the brand protection context. Second, the present study substantiates the effect of anti - counterfeiting efforts and how they could impact firm - idiosyncratic risk. Future research could draw on prior theoretical work in strategic alliance and identify various characteristics of the alliance that could further mitiga te the firm risk. For example, Swaminathan and Moorman (2009) showed that m arketing alliance capability , which refers to the ability of firms to generate higher returns from marketing alliances over time , may positively influence a firm's value creation. M ani and Luo (2015) empirically tested that more alliance activities will reduce both firm systematic risk and idiosyncratic risk. In addition, Thomaz and Swaminathan (2015) demonstrated that repeat partnering, and nce partners significantly impacts the firm risk following a marketing alliance announcement . In brand protection literature, alliances could be made among different stakeholders, such as other companies, non - profit organizations, governments, and law enfo rcement. I will leave the issue of incorporating richer information on alliance type (alliances formed across different partner categories) to future research. T hird, like other marketing investments, brand protection investments require evidence to prove their productivity. The impact of strategic marketing investment on related financial and profitability ( e.g., Rust et al. 2004; Katsikeas et al. 2016). Howe ver, due to the sensitivity and limited access to data regarding financial losses caused by counterfeiting, I could not further examine how brand protection actions could contribute to firm performance financially (using other financial metrics such as sal es revenue, profit, EVA, and ROI). If such data are available, future research could better address the effectiveness of brand protection actions. L astly, in my dissertation I am using long - term orientation from the H regarding national cultural dimensions to represent the cognitive aspect of country institutional profile. This operationalization might face some pushback because of some concerns about the relevancy and even the theoretical foundations of H related work ( Minkov and Hofstede 2012; Venaik and Brewer 201 3 ). Following the suggestions from prior literature ( Beugelsdijk , Kostova and Roth 2017 ; Kirkman, Lowe, and Gibson 2006 ), future research may build on my findings and further explore other possible alternative concepts that could better capture cultural effects in international business. To be specific, i n Essay one, results activity can indeed be perspective. These bra nd protection efforts provide market signals that investors attend to s and help investors predict actions, they evaluate not only the contents but also the intent of the signals (Stuart and Muzellec 2004). Indeed, I find that the short - and long - term impacts of brand protection actions differ for legal versus marketing brand protection actions. Althoug h legal actions have a positive impact on short - term stock market prices, in the long run, they hurt firm value as their impacts become negative. Marketing actions, on the other hand, are favored by investors both in the short - term and long - term. As such, my findings provide a novel insight into the favorable stock market response to marketing brand protect ion actions . that bran In Essay two, I explore the impact of marketing brand protection actions on firm idiosyncratic risk. Particularly , i n this essay I parse out this relationship by incorporating a more holistic view h ighlighting the impact from signaling environment. The results suggest that promotion - related actions among four types of brand protection actions seem to be a better choice to start with when conducting brand protection activities in foreign countries. However, this effect is contingent on institutional factors. When conducting b rand protection actions in countries that feature strong IPR protection, firms stock price returns will be higher for promotion - related actions than for other three types of actions, indicating that investors do believe that certain type of actions would b e more effective or appropriate based on the also play a key role in the development and implementation of brand protection actions globally. My research is able to show that the level of long - term orientation in each country will modify the risk reduction effect of promotion - related actions, as the stock prices more fluctuate in countries that possess higher long - term orientation. Finally, regulatory quality as the reflection of protection efforts as well. The more regulative a society, the less advantageous promotion - related actions would be considered by investor s over the other three types of actions. My research findings are important for managers who are facing rampant counterfeit activity in recent years. Brand owners are often worried about disclosing their brand protection actions out of concerns about info rming investors that the brand is under attack. However, this dissertation shows that publicity of brand protection actions is a good way to demonstrate brand one of their most important firm assets. Also , m y dissertation offers some useful and actionable implications for brand managers regarding taking the most suitable communication strategy. On the one side, the communication channel of brand protection messages should be able to reach out to as muc h audience as possible, as it may effectively reduce information asymmetry, cultivate trusting relationships with external stakeholders and raise funding from stock markets. On the other side, w hen brand managers articulate their brand protection actions, details such as the brand threat severity, the level of monetary/nonmonetary investment commitment, and their implementation plans all carry critical information about their willingness and ability to protect the brand. The clearer the message, the better Finally , my dissertation finds that managers should take the contextual factors into consideration when employing brand protection actions, such as the legislative and cultural environment of the host country, and brand threat context. These contextual factors may interact with b rand protection signal characteristics and exert influence on firm value. Stumpf, Chaudhry, and Perretta (2011) argue that brand managers must experiment with different anti - counterfeiting actions by country and brand; what works will be determined empiric ally. This suggestion complements my findings. Brand protection endeavors require dedicated investment and holistic examination globally. Differences in the national norm, value, culture, and ethnic beliefs between countries represent the challenges from i nstitutional contexts. I recommend that brand owners proactively take actions depending on the circumstance, while use cautions when implanting their strategies. APPENDICE S Authors Keywords Data Sources Review Focus Level of Research Focus Review Focus Review Article # Staake, Thiesse, Fleisch (2009) ProQuest ABI/INFORM, EBSCOhost Business Source Premier Management Literature Phenomenon 1. general descriptions of the phenomenon 2. impact analyses 3. investigation about illicit actors 4. investigation about customer behavior and attitudes No Eisend and Schuchert - Guler 2006 No No Consumer Purchase Intention Consumer - level The determinants of intention to purchase counterfeit products No Hoecht and Trott (2014) No No General Business Firm - level The success conditions of 11 anti - counterfeiting strategies No Eisend et al. (2017) counterfeit*, pirate*, fake, and illicit* Google Scholar, Business Source Complete, JSTOR, Psy - INFO, and ProQuest Dissertations & Theses) Consumer Morality Consumer - level A meta - analysis about the influence of morality on attitudes, intentio ns, and behavior toward counterfeit and pirated products. 196 T able 1 - 1 ( c Li and Yi (2017) No No Supply Chain Literature P henomenon 1 . introduce the social acceptance of counterfeiting and piracy 2 . the negative effect of counterfeiting and piracy on supply chain management and society N o Yang and Sonmez (2017) counterfeiting, ACS, strategies against counterfeiting, and strategy effectiveness EBSCO, PROQUEST (ABI Inform Complete), JSTOR, Emerald, Science Direct, Web of Science, Social Science Research Complete, and Business Research Complete. Multiple Discipline Phenomenon Anti - counterfeiting strategies (ACS), and examines their strategic effectiveness 51 This study ProQuest ABI/INFORM Complete, EBSCOhost Business Source Complete Marketing Literature Firm - level The firm - level brand protection, anti - counterfeiting and/or piracy prevention issues 78 Conceptual Studies/Commentaries Empirical Studies Consumer - level Firm - level Diamond (1962) JM Miaoulis and D'Amato (1978) JM Cohen (1991) JM Peterson, Smith and Zerrillo (1999) JAMS Commuri (2009) JM Wilcox, Kim & Sen (2009) JMR Amaral and Loken (2016) JCP Wang, Stoner and John (2019) JCP Eisend, Hartmann, and Apaolaza (2017) JIM Bhagat and Umesh (1997) JMFM Ertekin, Sorescu and Houston (2018) JM This Dissertation Constructs Operationalizations Data Source CAR i Short - term abnormal returns estimated for one day around the announcements of a firm initiated brand protection action Center for Research on Security Prices BHAR i Buy - and - Hold abnormal returns estimated for months after the announcements of a firm initiated brand protection action Center for Research on Security Prices Promotion related brand protection actions Dummy variable that takes the value of 1 if this nonlegal news is promotion - related and zero otherwise Factiva Intellectual Property Right (IPR) protection intensity The average of the standardized scores of The Law and Order index from the Gallup (Gallup 2018) and Park (2008) index Gallup (2018); Park (2008) Threat context Dummy variable that takes the value 1 if the threat involves online context and takes 0 if it is a pure offline threat. Factiva BP Action Commitment/Cost 1 - 7 Likert scale that measures the investment (both financially or emotionally) that the firm is putting to tackle the counterfeiting Factiva Industry intensity The total number of lawsuits filed in the same industry in the previous year Factiva T able 2 - 1 ( Control Variables Leverage Ratio of the long - COMPUSTAT Firm size Natural logarithm of the annual total assets of the firm COMPUSTAT Cash flow Net operating income before depreciation adjusted for working capital accruals COMPUSTAT Market value Product of common shares outstanding and annual closing price COMPUSTAT News specificity Total number of each news' word count Factiva Actions multiplicity Count number of the brand protection actions mentioned in each news. Factiva Mean Cumulative Abnormal Returns (Market Adjusted Model) Type of Brand Protection Action ( - 1.0) ( - 1,2) (0,0) Sample size Total actions .26% ** .24% * .20%** 349 Legal actions .32% .53% * .27% * 88 Nonlegal actions .24% * .14% .17% * 261 Place - related actions - .06% - .08% - .07% 48 Product - related actions .35% - .60% .31% 23 Promotion - related actions .32% ** .32% .24% ** 188 Price - related actions - 1.54% - 2.95% * - 1.50% * 2 Mean Cumulative Abnormal Returns (Fama - French Four Factor Model) Type of Brand Protection Action ( - 1.0) ( - 1,2) (0,0) Sample size Total actions .29% *** .36% ** .21% *** 349 Legal actions .27% .51% * .21% 88 Nonlegal actions .30% ** .31% * .21% ** 261 Place - related actions .10% .09% .07% 48 Product - related actions .40% - .53% .38% 23 T able 2 - Promotion - related actions .35% ** .49% ** .23% ** 188 Price - related actions - .65% - 1.46% - .85% 2 Mean Cumulative Abnormal Returns (Fama - French Four Factor Model) Type of Brand Protection Action (0,12) (0,9) (0,1) (0,3) (0,6) Sample size Total actions 1.25% * 2.39% *** .31% .71% * .97% * 380 Legal actions - 3.46% ** 1.83% .42% .83% 2.07% * 100 Marketing actions 2.93% *** 2.59% *** .27% .66% .58% 280 Place - related actions - 5.08% * - 4.46% * - .32% .07% .07% 52 Product - related actions .48% .88% - .12% - .55% .38% 25 Promotion - related actions 5.35% *** 4.74% *** .20% .37% ** .23% ** 201 Price - related actions - 2.04% - 9.57% - 1.56% - 1.67% - .85% 2 A: Results of the First - Stage Heckman Selection Model 1 Model 2 Model 3 Variable Coefficient SE Coefficient SE Coefficient SE Constant - 5.008*** .402 - 4.994*** .404 - 6.435*** .456 Industry intensity .034*** .007 .034*** .006 .045*** .007 Cash flow 1.349*** .256 1.344*** .256 1.810*** .264 Market value 1.34e - 5 *** 1.25e - 6 1.34e - 5 *** 1.25e - 6 6.33e - 6 *** 9.56e - 7 Leverage 1.10e - 3 4.29e - 3 1.19e - 3 4.35e - 3 4.09e - 4 3.96e - 3 Firm size .273*** .044 .269*** .044 .433*** .048 2 1.52 25.34** 39.57** B: Results of the Second - Stage Determinants of CARs Model 1 Model 2 Model 3 Variable Coefficient SE Coefficient SE Coefficient SE Constant - .5.29e - 5 .002 - .023 .039 - 2.635 1.078 Promotion dummy .002 .002 .028*** .010 - .078 .224 Commitment .005** .002 - .057 .052 IPR - .004* .002 - .181*** .058 Context .008* .005 .261* .137 Promotion*Commitment - .006** .002 .020 .058 Promotion*IPR .005** .002 - .212*** .065 Promotion*Context - .009* .005 - .307** .151 News specificity - 5.62e - 6 * 3.06e - 6 6.21e - 5 7.94e - 5 T able 2 - ) Action multiplicity 8.89e - 5 .003 - .150** .066 Cash flow .021 .019 .967** .489 Market value - 3.05e - 9 3.43e - 8 - 1.25e - 6 8.19e - 7 Firm size 1.98e - 4 .003 242*** .080 Leverage .001 5.11e - 4 - .010 .012 Inverse Mills Ratio 1.35e - 4 .001 1.72e - 4 .005 .451*** .139 A: Results of the First - Stage Heckman Selection Model 1 Model 2 Variable Coefficient SE Coefficient SE Constant .010*** .001 - 4.678*** .239 Industry intensity .020*** .005 .018*** .005 Cash flow 1.392*** .279 1.366*** .274 Market value 1.21e - 5 *** 9.87e - 7 1.22e - 5 *** 9.86e - 7 Leverage .001 .002 - .001 .002 Firm size .273*** .026 .263*** .026 2 48.92*** 113.98*** B: Determinants of Firm Idiosyncratic Risk Model 1 Model 2 Variable Coefficient SE Coefficient SE Constant .010*** .000 .013*** .001 Promotion dummy .001** .000 - .005** .002 IPR .001*** .001 Long - term orientation - 5.48e - 6 .000 Regulatory Quality - .003*** .001 Promotion*IPR - .002** .001 Promotion*LTO 6.28e - 5 ** .000 Promotion*RQ .004*** .001 News specificity - 2.22e - 6 *** 4.12e - 7 - 1.85e - 6 *** 3.98e - 7 Ta ble 3 - Action multiplicity .003** 0.00 .003** .001 Brand Threat Context .001 .001 .001 .001 Advertising - 2.30e - 7 - 1.88e - 7 - 1.38e - 7 2.00e - 7 R&D - 2.27e - 7 ** - 1.14e - 7 - 2.03e - 7 * 1.12e - 7 Industry type - 0.001 0.001 - .001 .001 Inverse Mills Ratio .002** .000 .002** .000 * p <.10 ** p <.05 *** p <.01 0 5000 10000 15000 20000 25000 30000 35000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Total Number of Seizures Year 15 Years Seizure Total Stock Return Promotion - related Actions Vs. Nonpromotion - related Actions Moderators Brand Threat Context ( - ) IPR Protection (+) Brand protection commitment ( - ) Action Multiplicity Firm Size Leverage Cash flow Market value News Specificity Controls Firm Idiosyncratic Risk Promotion - related Actions Vs. Nonpromotion - related Actions Moderators IPR Protection (+) Long - term Orientation ( - ) Regulatory Quality ( - ) Action Multiplicity B r and Threat Context Advertising Expenditure R&D Expenditure Industry Type News Specificity Controls REFERENC ES Chikada, Akino , and Anil Gupta (201 7 ), Online Brand Protection , In Handbook of Research on Counterfeiting and Illicit Trade . Cheltenham : Edward Elgar Publishing. Handbook of Research on Counterfeiting and Illicit Trade . Cheltenham: Edward Elgar Publishing.