The financial impact of product recall on electronic companies

(1)

The Financial Impact of Product Recall on Electronic Companies

Author:

Yang Zhong

Student Number:

10825835

Thesis Supervisor:

Evgenia Zhivotova

Finish Date:

25

th

January 2017

UNIVERSITY OF AMSTERDAM

COLLEGE OF ECONOMICS AND BUSINESS

BSc ECONOMICS AND BUSINESS

Bachelor Specialization Economics and Finance

(2)

2

Statement of Originality

This document is written by Yang Zhong who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

ABSTRACT

Product recall has significant negative impact on companies. Different characters in the recall, such as the recall strategy, the remedy provided to consumers with a recalled product, the cost of recall, and the defect type could contribute to the negative effect. Unlike previous researches that investigated a mix of all the industries, this study primarily emphasized on the consumer electronics industry since the development and usage of electronic devices are growing at an ascending pace. In this study, 70 recall announcements about consumer electronics that were published on the Consumer Product Safety Commission (CPSC) from 2009 to 2015 were selected. Employing the event study method, the results showed that product recall had a negative impact on the stock return on the announcement day and the day after. A regression analysis on the abnormal return against the four factors was done and the results indicated that a proactive recall strategy and the cost associated with the recall contributed negatively to the stock market significantly. A refund and a design flaw also resulted in a negative stock return, but the result was not very significant.

Keywords: Product recalls, event studies, efficient market hypothesis, stock market, firm strategy

(4)

4

The public burst into an uproar when the world’s most profitable mobile phone maker, Samsung Electronics Co., announced the recall of 1 million Galaxy Note 7 devices on September 15, 2016 due to the overheating problem. One month after the initial announcement, the recall expanded to another 1 million devices caused by additional incidents with the replacement phones. The final update on January 22, 2017, as posted on the company’s website (Samsung, 2017), provided an ultimate solution to consumers with a Galaxy Note 7. The series of recall of the mobile phone led to a decrease of 16.8% and 30% in the company’s net profit and overall operating profit respectively in the third quarter in 2016 as compared to the same period in 2015 (Wall Street Journal, 2016). Furthermore, half a month after the second announcement, another wave of recall involving 2.8 million of 34 models of top-load washing machines further damaged the company’s reputation. Samsung received 733 reports on excessive vibration or detachment issues of the washing machine and nine reports of injuries before the recall was published on Consumer Product Safety Commission (CPSC) on November 4, 2016. The market value of Samsung was further reduced by 0.7% on this day as compared to that on September 15, 2016.

These recall announcements reduce a company’s market value through increasing costs and decreasing sales revenue. Costs are raised since the company has to bear the direct costs that are related to the recall, such as advertisement, transportation and product refund costs (Chu, Lin, & Prather, 2005, p.34). Indirect costs that are associated with consumers’ disapproving behaviors like negative words and boycotts also reduce the company’s value (Ni et al, 2014, p.312). The resale value of the recalled products falls if the quality problem signaled by the recall may not or cannot be solved during the recall (Vassilikopoulou et al, 2009, p. 175). Market value of other products that are sold by the same company could be adversely affected as well if the recall marks a stain on the capability of the company to produce good quality products (Hartman, 1987, p. 367). The reduction in the company’s value

(6)

6

intensifies with the pervasiveness of the quality problem signaled (Hartman, 1987, p. 367). Moreover, Reilly and Hoffer (1983, p.447) discovered a negative correlation between product recall and market demand in the automobile industry. From the investors’ side, the damaged brand reputation and the altered quality perception destruct their confidence in the company. They might invest less or cease investing in the firm, which will further hamper the firm’s financial value (Chen, Ganesan & Liu, 2009, p. 214).

It is thus clear that product recall has an adverse effect on companies. Various results from previous empirical researches have supported this statement. Chen, Ganesan & Liu (2009, p. 219), Chu, Lin & Prather (2005, p. 40), Davidson & Worrell (1992, p. 470) and Pruitt & Peterson (1986, p. 119) found out that negative abnormal return was associated with product recall announcements in all non-automobile industries. In the automobile industry, which the product quality is heavily regulated before introducing the product, Jarrell & Peltzman (1985, p. 513) and Bromiley & Marcus (1989, p. 238) also agreed with the negative abnormal return during the event period.

Nevertheless, other than the automobile industry, few researches focused extensively on industry-wide investigations. Among all non-automobile industries, the consumer electronics industry draws great attention since electronic devices, ranging from power adapters to mobile phones to air conditioners, have gained increasing existence in people’s everyday life due to the development of high technology (Loch, Stein & Terwiesch, 1996, p. 4; Maidique & Zirger, 1984, p. 192). On the one hand, there is increasing number of companies aiming at designing and building new and better functioning electronic devices to provide more convenience to people’s everyday life. On the other hand, the probability of product recall in the consumer electronics industry is exceptionally high. Seven out of ten top consumer electronics companies as listed by Yahoo Finance (2016) had recall announcements published on the CPSC from 2009 to 2015. Chu, Lin & Prather also found out that electronic

(7)

7

companies had the highest recall incidences in their research (2005, p. 46). Therefore investigation into the consumer electronics industry can shed light on how product recall and the specific characters within the recall affect existing and potential electronic companies. They could gain insights on the way to manage such product-harm crisis efficiently so as to provide consumers with safer products and more holistic purchasing experience. Hence this essay aims to evaluate the financial impact of product recall announcements on the consumer electronics industry in a more recent time period, and the factors within the recall that contribute to the effect. This central research question will be answered in the form of an empirical research. In the next section, a literature review on information about product recall and relevant factors of the recall that have impacts on the company will be discussed. This section will also raise the relevant hypotheses. Section III will describe the data and introduce the methodology that will be employed in the analysis. The research results will be presented in Section IV and the major conclusions will be listed in the final section.

II. Literature review II.1 Product recall

Product recall typically occurs when the item is considered to generate danger to consumers, contain hazardous material that could expose to consumers, cause possible injury or death due to improper use, or disobey consumer product safety standards (Chu, Lin & Prather, 2005, p.34). In the consumer electronics industry, the most common hazard is the overheating problem. Such problem in smaller devices, such as the Galaxy Note 7, will cause the phone to catch fire or explode. Since the average daily usage time of mobile phones is increasing nowadays, such hazard poses great safety concerns. The damage is more severe when the overheating problem is situated in a larger device like the air conditioner. Both humans and properties will be burnt and damaged. Since most families possess more than one

(8)

8

electronic device, the hazard posed by these devices accumulates and magnifies. A defect in the product can be discovered and reported by all parties related to the product: manufacturers, distributors, importers, retailers, consumers or federal agency that is responsible for the quality check (Chu, Lin & Prather, 2005, p.34). After the discovery of a defect, the product recall decision will be made by either the company or the responsible agency or both and will be announced by both the company and the agency (Pruitt & Peterson, 1986, p.114). The likelihood of product recall has continued to soar due to factors like increased globalization, greater complexities in the production process and stricter supervision from firms and agencies (Berman, 1999, p. 69, Dawar & Pillutla, 2000, p. 215).

This research chooses to evaluate the recall of electronic devices in the United States. Being a top developed country in the world, electronic devices have covered almost the entire U.S. compared to South Asia where some areas do not have access to electricity. Meanwhile, the US government exerts increasing concern on product safety and the US market is highly regulated by various kinds of agencies. A large consumer base and under supervision of well-rounded policies make the US electronic industry a perfect investigation target.

In the United States, agencies that deal with product recall circumstances include: the National Highway Traffic Safety Administration (NHTSA), the Food Safety and Inspection Services (FSIS), the Food and Drug Administration (FDA), the Consumer Product Safety Commission (CPSC), and the Environmental Protection Agency (EPA). Since this research focuses on electronic products, which fall under the consumer products category, the CPSC is relevant.

The Consumer Product Safety Act (CPSA) of 1972, which was the foundation of the CPSC, and the subsequent Consumer Product Safety Improvement Act (CPSIA) of 2008 were implemented to protect consumers from possible harm posed by consumer products and to set up safety standards and regulations for consumer products (Beamish & Bapuji, 2008, p. 200). The CPSC manages more than 15,000

(9)

9

kinds of consumer products ranging from toys to recreational products (Ni et al, 2014, p.310). Companies are required by law to report to the CPSC within twenty-four hours after their detection of defect or below-safety-standard products (Consumer Product Safety Act, 1972, p. 58). Consumers are also given the opportunity to directly report to the CPSC about a defective product through a 24-hour hotline or an online form on the website (CPSC, 2017). After making the decision to recall, an announcement will be published by the CPSC including details like product name, company name, date, description of the product, hazards, consumer contact details, number of units, price sold, remedy, incidents, importer and manufacture place. A sample of product recall announcement published by the CPSC is provided in the Appendix A.

The information of the recall published by the CPSC affects both the consumer and the investors’ view of the company in a negative way. Since Fama (1976, p. 383) stated in his efficient-market hypothesis that the security price fully reflects all available information at any time, the stock return is used as an indication of the company’s status before and after the recall (Chen, Ganesan & Liu, 2009, p. 215). Therefore, it is proposed that a product recall announcement will destruct the company and the stock market will incorporate this information by showing a negative stock return.

Hypothesis 1(0): Product recall information is not related to stock return in the consumer electronics industry.

Hypothesis 1(1): Product recall information is negatively related to stock return in the consumer electronics industry.

II.2 Factors that contribute to the negative stock return

Product recall affects a company adversely, however, different factors embedded in the recall may contribute to the overall negative effect differently. Such factors include recall strategies the company adopts, remedies that the company provides to compensate consumers, costs of recall, and the type of product defect.

(10)

10

II.2.1 Recall strategies

Various studies agreed that companies differ substantially on how they manage a product recall circumstance (Dawar & Pillutla, 2000, p. 216; Laufer & Coombs, 2006, p. 383; Siomkos & Kurzbard, 1994, p. 32; Chen, Ganesan & Liu, 2009, p. 216). Chen, Ganesan & Liu classified recall strategies into two categories based on the company’s reaction to the incident: proactive recalls and passive recalls (2009, p. 216).

Proactive recalls are normally issued in the process of internal testing before any consumer reports on incidents or injuries (Chen, Ganesan & Liu, 2009, p. 216). When the firm or the federal agency discovers a defective or potentially hazardous product that might lead to a product recall, a firm that adopts a proactive recall strategy responds by either a voluntary recall or exerting super effort. Siomkos & Kurzbard (1994, p.32) defined a voluntary recall as an action before government intervention. Super effort is mightier than a voluntary recall and companies provide more compensation and communication than what is required by law (Laufer & Coombs, 2006, p. 384). For instance, Olympus America publishes recall information on its product support website of every malfunctioned product even though the problem is as minimal as a small crack on the LCD monitor cover (Olympus America, 2017).

On the contrary, passive recalls are issued in the late investigation stage after serious incidents or injuries are reported (Chen, Ganesan & Liu, 2009, p. 216). Companies adopting passive recall strategies might either refuse to take any responsibility for the production of a defective product or try to shift this responsibility to other companies or entities like the consumers (Chen, Ganesan & Liu, 2009, p. 216). Laufer & Coombs (2006, p. 384) described a case of Chipotle refusing the accuse from a university student claiming that the food purchased from Chipotle was bacteria-contaminated. The company’s spokesperson stated that Chipotle has never failed a health inspection and the food contamination could happen in ways other than food handling. A milder response than refusing or shifting the responsibility is an involuntary recall under the demand of federal agencies (Siomkos & Kurzbard, 1994,

(11)

11

p. 32).

Consumers tend to consider companies employing proactive recall strategies as being more responsible compared to those using passive recall strategies since actions are taken prior to government intervention. Proactive recall strategy indicates that the company cares for consumer surplus and demonstrates corporate social responsibility (Siomkos & Kurzbard, 1994, p. 32). The potential negative effect on the company might be mitigated and the company might even be perceived as being trustworthy (Siomkos & Malliaris, 2011, p. 64). Nevertheless, investors see proactive and passive recalls from a different perspective. In the period of product recall, they care more about the capability of the firm to keep a healthy cash flow in the short run as well as the impact on product sales (Chen, Ganesan & Liu, 2009, p. 216). Generally speaking, firms are not well prepared for such product-harm crisis such that most firms indeed use passive recall strategy (Dawar & Pillutla, 2000, p. 215). This is because many senior executives fail to understand the importance of crisis management and early responses since they believe that their companies are relatively immune to crises (Pearson & Clair, 1998, p. 69). Investors hence consider a proactive recall strategy as a last resort since they believe that the company must foresee a destructive financial consequence associated with the discovery of defective products (Chen, Ganesan & Liu, 2009, p. 217). They also see super effort as an overreaction since it costs more money and attracts unnecessary consumer attention on defect products (Laufer & Coombs, 2006, p. 384). Furthermore, since consumers’ reaction to recall strategy is ambiguous to investors, they weigh possible losses more heavily than potential gains due to loss aversion (Tversky & Kahneman, 1991, p. 1040).

It is thus proposed that a proactive recall strategy attracts more attention from investors and results in a more negative change in the stock market as compared to a passive strategy.

(12)

12

return than a passive recall strategy.

Hypothesis 2(1): A proactive recall strategy is more negatively related to stock return than a passive recall strategy.

II.2.2 Remedies

Regardless of the recall strategies employed, companies must provide remedies to retrieve and remove defective products from the market since inadequate actions could cause liability and punitive damages (Ross & Prince, 2009, p. 964). Both consumers’ satisfaction and investors’ reaction are closely related to the remedies provided by the company (Ni et al., 2014, p. 313).

There are generally two categories of remedies: repair or replacement and discard or refund. In the first category, companies provide consumers with repair kits, or ask them to return the product for repair or replacement. After this kind of remedy, the same product or a similar model is still in the consumers’ hands. In the second category, consumers are asked to discard the product immediately or receive a refund of the purchase price (Ni et al., 2014, p. 313). Under this category, consumers prefer refund rather than discard since discard provides absolutely no form of compensation.

Consumers perceive the second category of remedies as more effective than the first category (Ni et al., 2014, p. 313). This is mainly because they are not sure whether the root causes of the defectiveness will be successfully solved or not when the company offers a repair or replacement. Potential future product failures may reduce consumers’ confidence in the company. In contrast, the second category of remedies removes the product entirely from the consumers’ life. Further damages are therefore reduced to zero (Ni et al., 2014, p. 313). Hence when the company provides several different remedies, most consumers will choose a refund. Similar to recall strategies, investors hold a different opinion on remedies. They believe that firms are forced to provide refunds because the potential financial damages associated with the

(13)

13

incident are estimated to be very huge (Ni et al., 2014, p. 313). For example, the total cost associated with the recall of Galaxy Note 7 was estimated to be $5.3 billion by the Wall Street Journal (2016). Employing the most effective remedy from the consumers’ perspective can retrieve as many defective products as possible (Ni et al., 2014, p. 313). After getting the refund, consumers tend to reduce their demand for the company’s product. Investors prefer companies that offer remedies within the first category since they see these companies as being confident in the repair or replacement they provide (Ni et al., 2014, p. 313). They believe that these remedies are effective in minimizing future damages and can thus retain or increase consumer demand.

It is therefore reasonable to propose that investors hold a critical view on the second category of remedies because these strategies bring negative information to the stock market.

Hypothesis 3(0): A refund does not affect the stock return negatively under a product recall circumstance as compared to repair or replacement.

Hypothesis 3(1): A refund affects the stock return negatively under a product recall circumstance as compared to repair or replacement.

II.2.3 Costs as a percentage of market value

Another character in the product recall announcement that is of great importance is the cost that is related to the recall process. Although the cost that a company has to bear is not relevant for consumers, investors care deeply about the cost since it affects the company’s cash flow (Chen, Ganesan & Liu, 2009, p. 216).

The total costs involved in the recall include both direct costs and indirect costs. It is often assumed that all the defect products that are listed by the CPSC are successfully recalled and become worthless thereafter (Jarrell & Peltzman, 1985, p. 518; Pruitt & Peterson, 1986, p. 121; Govindaraj, Jaggi, & Lin, 2004, p. 40).

(14)

14

Companies sometimes publish direct costs that are related to the recall. In case this information is not provided, the direct costs are estimated by multiplying the retail price by the total number of products that are recalled (Jarrell & Peltzman, 1985, p. 518; Pruitt & Peterson, 1986, p. 121). Indirect costs include potential costs of additional recall, damaged reputation due to consumers’ negative words and boycotts, legal lawsuit expenses and other costs that could not be directly estimated (Govindaraj, Jaggi, & Lin, 2004, p. 40).

The total costs spent in the product recall process could be the same in real money terms for different firms, however, the costs may constitute different proportion of the firms’ market value and influent cash flow to a varying extent. It is hence more accurate to use cost in dollars as a percentage of market value rather than the real money terms (Pruitt & Peterson, 1986, p. 121; Jarrell & Peltzman, 1985, p. 519). When the percentage is higher, the company is left with less cash flow and the investors are less confident in the company. This information is then reflected in the stock price, making it rational to propose that the recall costs as a percentage of the market value is negatively related to the stock return.

Hypothesis 4(0): The recall costs as a percentage of the market value is not negatively related to the company’s stock return.

Hypothesis 4(1): The recall costs as a percentage of the market value is negatively related to the company’s stock return.

II.2.4 Defect type

The fundamental reasons of product recall must be detected and solved to prevent further problems. Beamish & Bapuji categorized two defect types: manufacture flaw and design flaw (2007, p.3; 2008, p. 197). Hora et al. agreed with this categorization and defined manufacture flaw as problems within raw materials or any step in the production process (2011, p. 768). Manufacture flaw in electronic devices includes unsuitable choice of materials or unstable connection between different parts

(15)

15

(Viswanadham & Singh, 1998, p. 51). Use of unqualified materials is also a manufacture flaw and can cause corrosion or rupture of the device. This problem is usually found in electric water kettles. Design flaw is rooted in the structure of the product and might be present before manufacturing (Hora et al., 2011, p. 768). The most common design flaw in electronic devices is the failure in heat transfer causing overheating problem (Jamnia, 2008, p. 5). Such overheating problem can be found in all kinds of devices ranging from batteries to refrigerators. Other design flaws can cause short circuit or electrical charges to human beings. For example, Apple recalled travel adapter kits and plugs in February 2016 due to the risk of electric shock (CPSC, 2016).

People have different perceptions on these two kinds of flaws. As mentioned by Hora et al. (2011, p. 769), the manufacture process of a product is often outsourced to contract manufacturers with the company supervising and checking the quality. When a product is recalled due to a manufacture problem, although the company might be blamed for not inspecting the product quality carefully, it can transfer the responsibility to the contract manufacturers and minimize the negative effect associated with the recall (Hora et al., 2011, p. 769). However, the design of the product is limited within the firm so as to maintain coordination and protect intellectual property (Novak & Eppinger, 2001, p. 193; Ulrich & Ellison, 2005, p. 318). Hence when the company publishes a product recall specifying a design flaw, both the consumers and the investors see this action as an admission of its own mistake (Hora et al., 2011, p. 769). Companies understand that this will damage their reputation, so they tend to delay or refuse the recall until the recall is finally published (Rhee & Haunschild, 2006, p. 103). The recall announcement indicating design flaws therefore brings tremendous effect to companies.

Hypothesis 5(0): A recall that is caused by design flaws is not more negatively related to stock return than one that is caused by manufacture flaws.

(16)

16

Hypothesis 5(1): A recall that is caused by design flaws is more negatively related to stock return than one that is caused by manufacture flaws.

III. Data and methodology III.1 Prerequisites

Before analyzing the financial impact of product recall, a few conditions need to be satisfied. The first condition is that the asset returns should be normal and independently and identically distributed throughout the time (MacKinlay, 1997, p. 17). Since this is empirically reasonable and important for the model setup, it is generally assumed that this condition holds (MacKinlay, 1997, p. 17). The next necessary condition is that the product recall has to be an unanticipated event for all traders (McWilliams & Siegel, 1997, p. 634). It is possible that there is information leakage about the event prior to the formal announcement, such as leakages from the insiders or public rumors. This makes it hard to determine the exact date that the traders are aware of the recall and hence poses problems to event studies (McWilliams & Siegel, 1997, p. 627). In this study, it is assumed that there is no information about the recall prior to the day on which the CPSC publishes the announcement. The final condition states that the product recall must be isolated from other confounding events such that the effects of these events will not affect each other (McWilliams & Siegel, 1997, p. 634). Such confounding events include changing management teams, brand expansion or signing government contracts. If there are events other than product recall in the event window, it is hard to determine which event causes the abnormal return (McWilliams & Siegel, 1997, p. 634). It is therefore important to control for confounding events by consulting various financial news sites like the Wall Street Journal and discard cases involving unanticipated events other than product recall (Chen, Ganesan & Liu, 2009, p. 218). The financial crisis during 2007 and 2008 is considered as a major confounding effect since the stock market was severely hampered. It is therefore hard to isolate the effect of product recall during that period.

(17)

17

Hence the investigation period in this study is from 2009 to 2015.

III.2 Data

The next step is to gather data that satisfy the conditions as stated above. Product recall announcements for electronic companies that are listed on the New York Stock Exchange were selected. Among all these publicly traded companies, those that had confounding events during the event window were discarded. The sample included 70 announcements from 2009 to 2015 in total.

The daily stock return data could be obtained from the Center for Research in Security Prices (CRSP) at the University of Chicago and the recall details could be found on the CPSC website. It was also possible to find information on three out of four relevant factors mentioned earlier from the website. When the recall announcements mention injuries or incidents, the company was considered to employ passive strategies. Otherwise it was considered as adopting proactive strategies. Whether the company provides repair, replacement or refund is clearly listed in the recall information. As mentioned earlier, if a company offers several remedies, it is assumed that consumers will choose a refund prior to a repair or replacement. When calculating costs as a percentage of market value, since there is only information on price and quantity sold on the CPSC website, cost was assumed to contain only the value of the recalled products, i.e. price multiplied by quantity. Market value of the company on the event day could be obtained from DataStream. Unlike these three factors, it is not straightforward to categorize a manufacture flaw and a design flaw. This requires careful reading and analysis of the text under “Hazard” on the CPSC recall page. As suggested by Hora et al. (2011, p. 771) and Beamish & Bapuji (2008, p. 202), two volunteers were asked to code the defect type as “manufacture flaw” or “design flaw” for the product recall announcements independently. They were given suggestions on how to determine the defect type as indicated in the previous section. The agreement level of the coders was 96% and the coding was considered to be reliable (Cohen’s Kappa=0.89 and p<0.001). The full list of companies included in the

(18)

18

sample can be found in Appendix B.

Summary statistics of the overall 70 product recall announcements are presented in Table 1. Panel A showed that the number of product recall announcements in the electronic industry was generally evenly spread in each year. As indicated in Panel B, 84% of the companies in the sample employed passive product recall strategy. This figure proved that most companies are unprepared for product-harm crisis like a product recall and tend to use a passive recall strategy. Among all the announcements, only 30% of them distributed a refund for the recalled product. This could be explained by the fact that a refund costs huge amount of money and most companies could not afford to do so. The large proportion of companies providing repair or replacement could also be attributed to the companies’ confidence in the repair kits or the replaced products. Panel D indicated that 76% of the product defects

Table 1. Descriptive Statistics of the Product Recall Sample Panel A Recalls by Year

2015 11 16% 2014 13 19% 2013 10 14% 2012 10 14% 2011 8 11% 2010 6 9% 2009 12 17% Total 70 100%

Panel B Recalls by Strategy

Proactive 11 16%

Passive 59 54%

Total 70 100%

Panel C Recalls by Remedy

Repair/Replace 49 70%

Refund 21 30%

Total 70 100%

Panel D Recalls by Defect Type

Manufacture 17 24%

Design 53 76%

Total 70 100%

Panel E Frequency of Product Recalls

One Recall 25

Two Recalls 7

Three Recalls 5

(19)

19

were due to design flaws. Panel E represented an analysis of the number of times an individual company is included in the product recall sample. Except for a few companies like Whirlpool Corporation and Newell Brands that consists of many subdivisions, product recall appears to be a relatively infrequent incident.

III.3 Methodology

III.3.1 Estimation window and event window

After identifying the necessary conditions, the next crucial thing is to determine the estimation window and the event window. The estimation period covers a length of one year and starts on the fifteenth month before the announcement day and ends on the third month before the announcement day. Assuming that there are twenty trading days per month and 252 trading days per year, the estimation period is thus day (-312, -60). This period is chosen to separate the estimation window and the event window so as to isolate the impact of product recall (Zhao et al., 2013, p. 119). When deciding the event window, Brown and Warner (1980, 1985) did not recommend a long event window since the power of the test statistic would be significantly reduced and consequently affect the interpretation of the result. Whereas a short event window is always preferred since it not only captures the effect of the event but also excludes confounding effects (McWilliams & Siegel, 1997, p. 636). In general, the appropriate event window is one to two days (McWilliams & Siegel, 1997, p. 652). Since product recall is assumed to be an unanticipated event as mentioned in the previous section, the first trading day conditional on the recall information is the event day itself, i.e. day 0, and the event window defined in this research is hence day (0, 1). McWilliams & Siegel (1997, p. 652) and Zhao et al (2013, p. 119) both agreed with this event window. Other than these two days, the change in the stock market from day -5 until day 5 is also examined for comparison purpose.

III.3.2 Event study

(20)

20

change in corporate policy (McWilliams & Siegel, 1997, p. 626). Since product recall is considered as a major incident for a company, various researchers have agreed on employing the event study method to examine the financial consequences of such event (Chen, Ganesan & Liu, 2009, p. 218; McWilliams & Siegel, 1997, p. 626; Pruitt & Peterson, 1986, p.116). The step-by-step process is described below.

Step 1: Obtain the estimated parameters for calculation of the expected normal return for firm i on day t.

There are various models available to estimate the normal returns of a company’s stock. The market model on daily stock returns is the most frequently used since this model can isolate the impact of market-related factors and control for systematic risk (Hendricks & Singhal, 2003, p. 508; Govindaraj, Jaggi, & Lin, 2004, p. 36; Zhao et al., 2013, p. 119). The market model is based on the idea that the return of any given security is linearly related to the market return (MacKinlay, 1997, p. 15). The relation between these two returns can be demonstrated in the following equation:

𝑅_!" = 𝛼_!+ 𝛽_!×𝑅_!" + 𝜀_!"

where 𝑅_!" is the actual return of the stock i on each day t within the estimation

window, 𝑅!" is the S&P 500 market return on each day t, 𝛼! and 𝛽! are the

intercept and slope of the relationship between stock i and the market respectively, and 𝜀_!" is the error term with 𝐸 𝜀_!" = 0 and 𝑉𝑎𝑟 𝜀_!" = 𝑆_!!_._𝛼

!, 𝛽! and 𝑆!! can be

estimated using the ordinary least squares (OLS) regression on data over the estimation window.

Other than the market model, the constant mean return model is used to ensure the robustness of the results. In this model, the relation between the return of the security and the mean return of the security is:

𝑅_!" = 𝜇_!+ 𝜁_!"

where 𝑅!" is the actual return of the stock i on each day t within the estimation

window, 𝜇! is the mean return for stock i during the estimation window and 𝜁!" is the

(21)

21

Step 2: Compute the expected normal return for firm i on day t.

The expected normal return for stock i on day t in the market model is the return without the occurrence of the unanticipated event and can be calculated using the following equation:

𝐸 𝑅!" = 𝛼!+ 𝛽!×𝑅!"

When considering the constant mean return model, the expected normal return

remains constant and equals the mean return for the stock 𝜇! during the estimation

window.

Step 3: Calculate the abnormal return (AR) for firm i on day t.

Upon the occurrence of the product recall, the actual ex post return differs from the expected normal return and the abnormal return measures the difference between them. The AR of stock i on day t during the event window day 0 and day 1 is calculated as:

𝐴𝑅!" = 𝑅!"− 𝛼!− 𝛽!×𝑅!"

Step 4: Compute the cumulative abnormal return (CAR) for firm i in the event period.

The abnormal returns for firm i on every day within the event period are then aggregated and resulted in CAR for each firm.

𝐶𝐴𝑅! 𝑡!, 𝑡! = 𝐴𝑅!" !!

!!!!

Step 5: Compute the average cumulative abnormal return (ACAR) for all firms in the sample during the event period.

The cumulative abnormal return for each firm during the event period is summed up and averaged to get ACAR.

𝐴𝐶𝐴𝑅! 𝑡!, 𝑡! =

1

𝑁 𝐴𝑅!"

!

!!!

A t-test is then used to test the statistical significance of ACAR upon the product recall announcements:

(22)

22 𝑡 = 𝐶𝐴𝑅!" !!!! !!!! / 𝑆𝐷!" ! !!!! !!!! 𝑁 ! !!! where 𝑆𝐷_!" = {𝑆_!!_{× 1 +}! !+ (!!"!!!)! (!!"!!!)! !!!!" !!!!"# } !

!, T and 𝑅_! are the total number of

days and the average market return in the estimation period respectively.

IV. Results and Analysis

Table 2 presents the mean abnormal return and the associated t test statistics from day -5 to day 5. The p-values of the one-tailed test is in parentheses under the test statistics (*, ** and *** denote statistical significance at 1%, 5% and 10% levels). Results from the market model are shown in column two and three and results using the constant mean return model are shown in column four and five.

The results shown in Table 2 provide evidence that product recall

Table 2. Abnormal Return and Test Statistics

Mean AR t test statistics Mean AR t test statistics

-5 0.14% 0.5062 0.18% 0.4915 (0.6928) (0.6877) -4 0.04% 0.1889 -0.10% -0.3248 (0.5746) (0.3732) -3 0.03% 0.1224 0.15% 0.6202 (0.5485) (0.7314) -2 0.08% 0.4555 0.22% 0.7159 (0.6749) (0.7618) -1 -0.27% -1.0787 -0.29% -0.7890 (0.1422) (0.2164) 0 -0.68% -2.1973 -0.46% -1.7775 (0.0157**) (0.0399**) 1 -0.40% -1.8095 -0.84% -1.2592 (0.0347**) (0.1061) 2 0.05% 0.3235 0.24% 0.9672 (0.6264) (0.8316) 3 0.00% -0.0081 0.00% 0.0225 (0.4968) (0.5059) 4 -0.03% -0.1216 -0.01% -0.0645 (0.4518) (0.4744) 5 0.21% 0.7159 0.15% 0.7181 (0.7618) (0.7624)

Market Model Constant Mean Return Model

(23)

23

announcements have negative impact on a company’s stock market. The mean abnormal return obtained from the market model on the event day, day 0, was -0.68% and was significantly smaller than zero at 5% significance level. The mean abnormal return on day 1 was -0.40%, which was less negative than that on the event day itself, and was also significant at 5% level. Besides these two days, the mean abnormal return on the other days was not significantly smaller than zero. This result was not altered by the choice of methodology. The magnitude and statistical significance of the mean AR obtained from the market model and the constant mean return model were similar to each other. The mean AR on day 0 obtained from the constant mean return model was -0.46%, and was significantly smaller than zero at 5% significance level. None of the methods indicated any significant mean AR on the days before day 0 or after day 1. The consistency of results from the two models ensured the robustness of the result. This result provided evidence that the market reacts to the recall announcement on the publication day of the announcement on the CPSC and the trading day after the publication. Since traders could only place a trade when the market is open, they are forced to wait until the next trading day if the recall announcement will be published near or after the market closes. Therefore the market reacts to the new information on day 0 and day 1, and the direction is the same.

The conclusion derived from the mean CAR was similar to that from mean AR. Table 3 shows the mean cumulative abnormal return during the event window and the associated t test statistics. The numbers in parentheses under the t test statistics are the p-values of the one-tailed test. The values surround the event day were significantly negative at 1% significance level in the market model. The insignificance of the results during the period (-5, -1) and (2, 5) further proved that the market reacts only on the event day and the day after. The results were robust and not sensitive to the models used.

To further investigate the factors that cause the negative abnormal return, a multiple ordinary least squares (OLS) regression is done on CAR of the event window

(24)

24

(0, 1) against recall strategy, remedy, cost as a fraction of market value and defect type. The regression model looks like the following:

𝐶𝐴𝑅! = 𝛽!+ 𝛽!×𝑆𝑡𝑟𝑎𝑡𝑒𝑔𝑦!+ 𝛽!×𝑅𝑒𝑚𝑒𝑑𝑦!+ 𝛽!×𝐶𝑜𝑠𝑡!+ 𝛽!×𝐷𝑒𝑓𝑒𝑐𝑡 𝑇𝑦𝑝𝑒!+ 𝜀!

where 𝐶𝐴𝑅_! stands for the cumulative abnormal return for company i and is the

dependent variable in the model. This value is the average of the two CARs from the

market model and the constant mean return model. 𝑆𝑡𝑟𝑎𝑡𝑒𝑔𝑦_!, 𝑅𝑒𝑚𝑒𝑑𝑦_! and

𝐷𝑒𝑓𝑒𝑐𝑡 𝑇𝑦𝑝𝑒! are dummy variables. 𝑆𝑡𝑟𝑎𝑡𝑒𝑔𝑦! is 1 if the company adopts a

proactive recall strategy and 0 if passive. Similarly, 𝑅𝑒𝑚𝑒𝑑𝑦! is 1 if the company

offers refund and 0 if repair or replacement is provided. 𝐷𝑒𝑓𝑒𝑐𝑡 𝑇𝑦𝑝𝑒_! is 1 if a design

flaw is detected and is 0 if a manufacture flaw. 𝐶𝑜𝑠𝑡_! is a continuous variable and

measures the cost of recall as a percentage of the market value. The result of the regression analysis is shown in Table 4.

Event Day Mean CAR t test Stastics Mean CAR t test statistics

-1 to 0 -0.96% -2.2491 -0.75% -1.5229 (0.0138**) (0.0662***) 0 to 1 -1.07% -2.8377 -1.30% -1.7808 (0.003*) (0.0397**) -1 to 1 -1.34% -2.4707 -1.59% -1.9469 (0.008*) (0.0278**) -2 to 2 -1.21% -2.0174 -1.12% -1.2593 (0.0238**) (0.1061) -3 to 3 -1.18% -1.7595 -0.96% -1.0782 (0.0415**) (0.1424) -4 to 4 -1.16% -1.6222 -1.08% -1.0782 (0.0547***) (0.1780) -5 to 5 -0.81% -0.9826 -0.74% -0.6289 (0.1646) (0.2657) -5 to -1 0.03% 0.0496 0.17% 0.1738 (0.5197) (0.5687) 2 to 5 0.23% 0.4328 0.38% 0.7676 (0.6667) (0.7773)

Market Model Constant Mean Return Model

(25)

25

As indicated by the negative coefficients on the four variables, a proactive strategy, a refund, a higher cost as a percentage of recall and a design flaw contributed to a negative abnormal return. To be specific, a proactive recall strategy resulted in a 1.82% reduction in the abnormal return and this value was significant at 5% significance level. Even though a proactive recall strategy shows that the company cares for the consumers, it actually hurts the company’s stock market more negatively than a passive recall strategy. The stock market and the investors interpret this strategy as a signal of tremendous financial loss and this information is reflected in the negative stock return. The coefficient on the continuous variable, cost as a percentage of market value, was the most negative among all the variables and was significant at 5% significance level. This result confirmed the theory that the higher the cost associated with the recall, the more negative the stock return will be. A refund and a design flaw resulted in 0.41% and 0.76% reduction on stock return respectively, however, these results were not significant at a 10% significance level. This could be explained by the rapid change in the characteristics of consumer electronics. The short product cycles and lifetimes, evolving usage patterns and dynamics and rapid technology development give rise to dramatic change in consumer electronics (Urban et al., 2014, p. 11). Consumers might have switched to other better-functioning products before the defect ones are scheduled for a recall. So the type of remedy the company provides and the defect type in the recalled products do not result in a significant decrease in the abnormal return. Nevertheless, this does not necessarily mean that the remedy and defect type are irrelevant. Consumers can make inference

Factors Coefficient Standard Err. t test statistics p-value

Strategy -0.0182 0.0088 -2.07 0.042**

Remedy -0.0041 0.0066 -0.63 0.533

Cost -0.0555 0.0263 -2.11 0.039**

Defect Type -0.0076 0.0088 -0.86 0.392

Constant -0.0200 0.0101 -1.97 0.053

Table 4. OLS Regression Analysis

(26)

26

on the company’s attitude and product quality from the information and make decision on whether to buy future products or not. Since product quality is considered to be the second most important factor after product price, it is vital for companies to design good quality products and provide decent remedies to consumers if a product is recalled (Chao, Iravani, & Savaskan, 2009, p. 1122).

V. Conclusion

Technology increases the number of products people can use in their everyday life. However, this is always accompanied with more defects in the products due to factors like increasing globalization and greater complexity. Previous literatures have discussed the effect of product recalls on companies from all the industries or focused on industries like automobiles that are more heavily regulated. Since the penetration rate of electronic devices keeps increasing every year, this research focused on the financial impact of recall announcements on electronic companies and evaluated which factors contributed to the effect.

A sample of 70 announcements from 2009 to 2015 was collected. An event study was conducted to see the effect of product recall announcements on the stock return. Results showed that such announcements are accompanied by negative stock return on the event day and the day after. This conclusion is robust and is not sensitive to the model employed. A multiple Ordinary Least Square regression was then done to examine the possible factors that could lead to the negative effect. The result indicated that a proactive recall strategy contributed to a more negative effect on the abnormal return than a passive recall strategy. Also, the higher the cost as a percentage of market value, the more negative the abnormal return is. However, although the remedy that a company provides and the defect type situated in the recalled product contribute to the negative abnormal return, the result was not very significant. These conclusions provide important insights for electronic companies. They should pay great attention to product quality before selling the products to

(27)

27

consumers so as to reduce the product recall incidence to minimum. In case a product is defect, the company should decide carefully on the strategy it should employ and manage the recall as efficiently as possible.

What needs to be mentioned is that the data chosen in this research was from 2009 to 2015 in order to isolate the effect of product recall from the financial crisis. Further research could choose to examine the period of financial crisis to see if a similar result applies. Further research could also investigate the long run effect of product recall since only the short run effect was measured in this thesis. Finally, it would be interesting to see if other industries, such as food or drugs, have similar characters.

(28)

28

Bibliography:

Beamish, P. W. & Bapuji, H. (2008). Toy Recalls and China: Emotion vs. Evidence, Management and Organization Review, 4(2), p. 197-209.

Berman, B. (1999). Planning for the Inevitable Product Recall, Business Horizons, 42(2), p. 69-78.

Bromiley, P. & Marcus, A. (1989). The Deterrent to Dubious Corporate Behavior: Profitability and Safety Recalls, Strategic Management Journal, 10(3), p. 233-250.

Brown, S. J. & Warner, J. B. (1980). Measuring Security Price Performance, Journal of Financial Economics, 8, p. 205-258.

Brown, S. J. & Warner, J. B. (1985). Using Daily Stock Returns, the Case of Event Studies, Journal of Financial Economics, 14, p. 3-31.

Chao, G. H., Iravani, S. M. R., & Savaskan, R.C. (2009). Quality Improvement Incentives and Product Recall Cost Sharing Contracts, Management Science, 55(7), p. 1122-1138.

Chen, Y., Ganesan, S., & Liu, Y. (2009). Does a Firm’s Product-Recall Strategy Affect Its Financial Value? An Examination of Strategic Alternatives during Product-Harm Crises, Journal of Marketing, 73(6), p. 214-226.

Chu, T. H., Lin, C. C. & Prather, L. J. (2005). An Extension of Security Price Reactions around Product Recall Announcements, Quarterly Journal of Business and Economics, 4(3/4), p. 33-48.

Consumer Product Safety Act (1972). Public Law 92-573; 86 Stat. 1207.

Davidson III, W. N. & Worrell, D. L. (1992). The Effect of Product Recall Announcements on Shareholder Wealth, Strategic Management Journal, 13(6), p. 467-473.

Dawar, N. & Pillutla, M. (2000). Impact of Product-Harm Crises on Brand Equity: the Moderating Role of Consumer Expectations, Journal of Marketing, 37(2), p.215-226.

(29)

29

Fama, E. F. (1976). Foundations of Finance: Portfolio Decisions and Securities Prices. Basic Books, New York.

Govindaraj, S., Jaggi, B. & Lin, B. (2004). Market Overreaction to Product Recall Revisited-the Case of Firestone Tires and the Ford Explorer, Review of Quantitative Finance and Accounting, 23, p. 31-54.

Hartman, R. S. (1987). Product Quality and Market Efficiency: the Effect of Product Recalls on Resale Prices and Firm Valuation, the Review of Economics and Statistics, 69(2), p. 367-372.

Hendricks, K. B. & Singhal, V. R. (2003). The Effect of Supply Chain Glitches on Shareholder Wealth, Journal of Operations Management, 21, p. 501-522. Hora, M., Bapuji, H., & Roth, A. V. (2011). Safety Hazard and Time to Recall: the Role

of Recall Strategy, Product Defect Type, and Supply Chain Player in the U.S. Toy Industry, Journal of Operations Management, 29, p. 766-777.

Jamnia, A. (2008). Practical Guide to the Packaging of Electronics: Thermal and Mechnical Design and Analysis, Second Edition, CRC Press.

Jarrell, G. & Peltzman, S. (1985). The Impact of Product Recalls on the Wealth of Sellers, Journal of Political Economy, 93(3), p. 512-536.

Laufer, D. & Coombs, W. T. (2006). How Should a Company Respond to a Product Harm Crisis? The Role of Corporate Reputation and Consumer-based Cues, Business Horizons, 49, p. 379-385.

Loch, C., Stein, L. & Terwiesch, C. (1996). Measuring Development Performance in the Electronics, Journal of Product Innovation Management, 13(1), p. 3-20. MacKinlay A. C. (1997). Event Studies in Economics and Finance, Journal of

Economic Literature, XXXV, P. 13-39.

Maidique, M. A. & Zirger, B. J. (1984). A study of Success and Failure in Product Innovation: The Case of the U.S. Electronics Industry, IEEE Transactions on Engineering Management, EM-31(4), p. 192-203.

(30)

30

Theoretical and Empirical Issues, the Academy of Management Journal, 40(3), p. 626-657.

Ni, J. Z., Flynn, B.B., & Jacobs, F. R. (2014). Impact of Product Recall Announcements on Retailers’ Financial Value, Int. J. Production Economics, 153, p. 309-322.

Novak, S. & Eppinger, S. D. (2001). Sourcing by Design: Product Complexity and the Supply Chain, Management Science, 47(1), p. 189-204.

Pearson, C. M. & Clair, J. A. (1998). Reframing Crisis Management, the Academy of Management Review, 23(1), p. 59-76.

Pruitt, S. W. & Peterson, D. R. (1986). Security Price Reactions around Product Recall Announcements, the Journal of Financial Research, IX(2), P. 113-122. Reilly, R. J. & Hoffer, G. E. (1983). Will Retarding the Information Flow on Automobile

Recalls Affect Consumer Demand? Economic Inquiry, 21(3), p. 444-447.

Rhee, M. & Haunschild, P. R. (2006). The Liability of Good Reputation: A Study of Product Recalls in the U.S. Automobile Industry, Organization Science, 17(1), p. 101-117.

Ross. K. & Prince, J. D. (2009). Post-Sale Duties: the Most Expansive Theory in Product Liability, Brooklyn Law Review, 74(3), p. 963-986.

Siomkos, G. J. & Kurzbard, G. (1994). The Hidden Crisis in Product-harm Crisis Management, European Journal of Marketing, 28(2), p. 30-41.

Siomkos, G. J. & Malliaris, P. G. (2011). Consumer Response to Company Communications During a Product Harm Crisis, Journal of Applied Business Research, 8(4), p. 59-65.

Tversky, A. & Kahneman, D. (1991). Loss Aversion in Riskless Choice: A Reference-Dependent Model, The Quarterly Journal of Economics, 106(4), p. 1039-1061.

Ulrich, K. T. & Ellison, D. J. (2005). Beyond Make-Buy: Internalization and Integration of Design and Production, Production and Operations Management, 14(3), p.

(31)

31

315-330.

Urban, B., Shmakova, V., Lim, B., & Roth, K. (2014). Energy Consumption of Consumer Electronics in U.S. Homes in 2013 – Final Report to the Consumer Electronics Association.

Vassilikopoulou, A., Siomkos, G., Chatzipanagiotou, K., & Pantouvakis, A. (2009). Product-harm Crisis Management: Time Heals All Wounds? Journal of Retailing and Consumer Services, 16, p. 174-180.

Viswanadham, P. & Singh, P. (1998). Failure Modes and Mechanisms in Electronic Packages, Chapman & Hall.

Zhao, X., Li, Y., & Flynn, B. B. (2013). The Financial Impact of Product Recall Announcements in China, Int. J. Production Economics, 142, p. 115-123.

Consulted Websites:

Consumer Product Safety Commission (2017). Homepage, https://www.cpsc.gov/ Consumer Product Safety Commission (2017). File a Report,

https://www.saferproducts.gov/CPSRMSPublic/Incidents/ReportIncident.aspx Consumer Product Safety Commission (2017). Apple Recalls Travel Adapter Kits and

Plugs Due to Risk of Electric Shock,

https://www.cpsc.gov/Recalls/2016/Apple-Recalls-Travel-Adapter-Kits-and-Plugs Olympus America (2017). Important Announcement for All e-m5 Camera Owners,

http://cache.olympusamerica.com/static/cpg_section/service_announcement/E-M5_Service_Announcement.pdf

Samsung (2017). Samsung Expands Recall to All Galaxy Note 7 Devices, http://www.samsung.com/us/note7recall/

Wall Street Journal (2016). Galaxy Note 7 Recall Sinks Samsung Profit,

http://www.wsj.com/articles/samsung-profit-falls-on-galaxy-note-7-recall-1477526 701

Yahoo Finance (2017). Consumer Electronics & Appliances Industry Profile, https://biz.yahoo.com/ic/prof/16.html

(32)

32

Appendix A. CPSC Announcement Example

Retrieved from the CPSC at:

(33)

33

Appendix B. Sample Recall Announcements

In this appendix, a full list of companies that were included in the sample was provided based on the date of recall announcement published on the CPSC.

Company Name Date of Announcement Strategy Remedy Defect Type

United Technologies 22/12/2015 passive repair design

Black & Decker 19/11/2015 passive replace design

Turtle Beach 20/10/2015 passive replace manufacture

NVIDIA 31/07/2015 passive replace design

Whirlpool 29/07/2015 passive repair design

Target Corp 21/07/2015 passive refund design

Cooper Lighting 15/07/2015 passive replace design

Olympus 15/07/2015 proactive repair manufacture

CREE 04/06/2015 passive refund design

Apple 03/06/2015 passive refund design

A.O. Smith 02/06/2015 proactive replace design

Keurig 23/12/2014 passive repair design

Ethan Allen 10/12/2014 passive refund design

NRG energy 25/11/2014 passive replace design

HP 26/08/2014 passive replace design

middleby 06/08/2014 passive repair design

Sony 10/07/2014 passive refund design

TRANE 24/06/2014 proactive repair manufacture

Nokia 18/06/2014 proactive refund manufacture

Newell Brands 29/05/2014 passive refund design

NACCO Industries 28/05/2014 passive replace design

Acuity Brands 28/05/2014 passive repair design

Canon 06/03/2014 passive repair manufacture

Sears 15/01/2014 passive refund design

Google 17/12/2013 passive replace design

Middleby 19/11/2013 passive repair design

Southern Co 30/10/2013 passive replace design

Emerson 24/10/2013 passive repair design

Home Depot 25/07/2013 passive refund design

Best Buy 19/06/2013 passive refund design

Dollar Tree 09/05/2013 passive refund design

Williams-Sonoma 23/04/2013 passive refund manufacture

3M 26/03/2013 passive replace design

HEICO Lighting 08/01/2013 proactive replace design

Newell Brands 30/08/2012 passive replace design

Cooper Lighting 29/08/2012 passive replace design

Energizer 22/08/2012 passive refund design

Emerson 16/08/2012 passive repair design

Canon 14/08/2012 passive replace manufacture

Williams-Sonoma 02/02/2012 passive refund design

(34)

34

Company Name Date of Announcement Strategy Remedy Defect Type

Helen of Troy 12/01/2012 proactive refund design

Big Lots 12/01/2012 passive refund design

Zagg 30/11/2011 passive replace design

Acuity Brands 09/08/2011 passive replace design

Honeywell 28/07/2011 passive replace design

Sony 30/06/2010 passive repair design

Cooper Lighting 30/04/2010 passive repair manufacture

American Electric Lighting 03/03/2010 proactive repair design

Franklin Electric 18/02/2010 proactive replace manufacture

Home Depot 17/12/2009 passive refund design

IBM 10/11/2009 proactive replace manufacture

Sony 28/10/2009 passive replace design

Energizer 14/07/2009 proactive refund design

Middleby 16/06/2009 passive repair design

Big Lots 11/03/2009 proactive refund design

The financial impact of product recall on electronic companies