Sinners and Saints: A Review of the Returns of Ethical and Unethical Companies Diana Evans Jeremy Johns Kyle Kelleher Sashidhar Mylavarapu Catherine Nguyen Karen Richards Natalie Schlitt Ami Suchde Brian Tackenberg Michael Wilkinson Katya Zulaica
Sinners and Saints: A Review of the Returns of Ethical and Unethical Companies Abstract This paper examines the returns of companies listed on the 100 Best Corporate Citizens list, published annually since 2000 by Business Ethics Magazine We created a matched sample based on book-to-market value similar to the evaluation performed by Filbeck, Gorman, and Zhao in The Best Corporate Citizens: Are th ey good for their shareholders (2008) Our results show that the ethical companies out performed the book-to-market matched sample. We created a second match based on SIC code, beta, and ma rket capitalization. Ethi cal companies again outperformed the Â“extendedÂ” matched sample. Accord ing to our results, the ethical firms in the extended match also outperformed the market index since the cons truction of the list. Finally, we created an Anti-index of companies excl uded from consideration to join the 100 Best Corporate Citizens list (such as tobacco, alc ohol, and defense). The Anti-i ndex performed rather poorly relative to the S&P 500 and very poorly when comp ared to the index of ethical companies.
1Sinners and Saints: A Review of the Returns of Ethical and Unethical Companies 1. Introduction The practices of list making and ranking system s have gained popularity over time and are particularly prevalent in the business wo rld. The variety of lists ranges from the 100 Best Companies to Work For published by Forbes to the Most Admired Companies published by Fortune Magazine. Many lists of significance fa ll in categories other than business including society, lifestyle and educati on to name a few. For example, U.S News and World Report publish AmericaÂ’s Best Colleges list and Thomas Reuters publishes the Journal Citation Report (JCR) These lists are used by many different pe ople for a variety of different reasons. The high regard and attractiveness of lists raises the question of whether these lists have an impact on the items included in the rankings, ei ther qualitatively or qua ntitatively. Another point of interest explores if lists go beyond merely ranking. More specifically, this means attempting to determine if these lists provide users with info rmation beyond the ranking in a specific area and establishing significant relati onships to factors excluded from the list itself. Our study focuses on the 100 Best Corporate Citizens list published annually by Business Ethics Magazine The overall goal is to determine if ther e is a relationship between inclusion on the 100 Best Corporate Citizens list and investorsÂ’ retu rns. We based our orig inal study off of the working paper by Filbeck, Gorman, and Zhao (2008) where they conducted a matched sample based on book to market value of each of the fi rms on the list to examine whether the firms on
2the list outperformed the matched firms. We fe lt that the matched sample performed by Filbeck, Gorman, and Zhao was too vague; therefore, we ran an extended matched sample based on a number of factors including the fi rst two digits of the SIC code, be ta, and market capitalization to see if the firms on the list would outperform the firms in our extended matched sample. In addition, we created an Anti-index using compan ies that were purposefully excluded from the 100 Best Corporate Citizens list to see whether the companies in the Anti-index could outperform an index of the companie s that were included on the list. Section two of this paper provi des a history of lists accounting for their rise to prominence, followed by a discussion of non-business lists, busin ess related lists, and finance related business lists. Section three contains a de scription of the methods used by Business Ethics Magazine to select the companies that are included annually on the 100 Best Corporate Citizens list, a description of the Filbeck, Gorman, and Zhao (2008) working paper with which we plan to compare our results, a description of our exte nded matched sample, and a description of the Anti-index we created. The results of each of the matches we perf ormed are detailed in section four of the paper. Finally, a summary of our findings along with concluding remarks compose section five. 2. Literature Review 2.1 History of Lists Â– Broad Overview
3A list is a type of organizational tool. A list is defined as, Â“a series of names or other items written or printed together in a meaningful grouping or sequence so as to constitute a record (Random House).Â” Human nature tends to use lis ts because they are an attempt to organize a world of chaos into something understandable. Li sts are usually short, concise, and easy to follow making them a convenient resource (Matos 2008). There are several different types of lists all with different purposes. Some lists are used as informational tools, such as a list of restaurants in a town; while others are used for comparison. Seeing the popularity lists have gained, it is reasonable to wonder where this trend began and how it ha s evolved in society. The practice of list making date s back hundreds, even thousands, of years. People have been utilizing lists for various reasons for as long as we know. Some of the earliest lis ts include the Seven Wonders of the Ancient World and the Te n Commandments that were created thousands of years ago (Matos, 2008). These early forms of ranking systems have laid the foundation for the vast array of lists today. It goes without saying that Americans are drawn to rankings and lists. This heavy dependence on reviews, ratings and recommendations has created a list-driven society. The origins of how lists gained popularity in the U.S. can be dated b ack to 1891 when Congress passed the International Copyright Act. This law prohibite d piracy of literary works of authors not only from the U.S., but also foreign nations (Laurs, 1960). Before this act was passed, authors were hesitant to make public the number of books they had sold because piracy was a common practice. After this law was adapted, authors began publicizing their wo rks and the success these works had achieved. These events initiated the evoluti on of what we now know as Â‘The Best Sellers ListÂ’. These lists
4were slow to gain popularity; Publishers Weekly did not start printing a be st sellerÂ’s list until 1912, and the New York Times following in 1942 (Matos, 2008). Since then, book lists have become one of the most common and widely used types of lists we have today. Soon after book lists were first seen, music lists were the new buzz. Music charting began with Phonoscope in 1896, a journal that listed the most popular phonogr aph records (Matos, 2008). In 1913, Billboard music began ranking the top songs based on popularity and number of times they were performed (Billboard.com). B illboard ranking lists continued to gain prominence, and were soon held as the standard for music best lists. Music charting then hit the radio and te levision soon after through the program, Your Hit Parade (Matos, 2008). This program aired on the radio in1935 and then switched between NBC and CBS until the beginning of 1953. This broadcast s how began playing the top 15 songs of the week and eventually narrowed the list down to the top 10 songs of the week ( National Radio Hall of Fame & Museum ). To determine these lists, this radio institution used a national tabulation system of record and sheet music sale s. These hit singles were not played by the original artist, yet instead, th ey were performed by the showÂ’s own staff to bring a more innovative perspective. (Felsenthal, 2008). After fifteen years on the air, Your Hit Parade then made the transition to the new standard of television from 1950 until 1959 (Matos, 2008). After the Â‘50s, people wanted to go back to listening to their favorite songs by the original artists. At the sa me time, a new genre of music
5was formed: rock Â’nÂ’ roll. In 1957, Dick Clark embraced this new emergence through American Bandstand a live show that featured teenagers dancing to top r ecorded hits (Buxton, 2008). In 1970, Casey Kasem launched American Top 40 through the radio from the Billboard Hot 100 list as we know it today. (Keit h, 2002). Our societyÂ’s desire to rank music and other lists concerning the media ventured into new form s through television, such as, Casey KasemÂ’s AmericanÂ’s Top 10 MTVÂ’s Total Request Live: Top 20 Countdown and E!Â’s Daily 10 top stories. The Â‘80s and Â‘90s and continued this releva nce of ranking popularity in the media through the mediums of the television show, Entertainment Tonight and the magazine, Entertainment Weekly (Matos, 2008). Most commonly known for their celebrity interviews, ET also exploits ratings in the media through movies, fashion, musi c, celebrities, and television shows. Top box office movies of the weekend and top earning Hollywood couples are just some of the many examples of both whom and what are ranked (etonline.com). EW later on embraced a similar prospective on what is popular in the media th rough the form of a ma gazine. Yet, now through our continuous innovation of technology, both ET and EW reveal the most recent popularity charts through the means of the internet. With the numerous technological ad vances over the years, the info rmational resource that most people rely on today is the inte rnet. Computers and now, more recently, cellular phones are the two mediums that can access the internet. Thus, since most of our society spends an abundant
6amount of time on both of these mediums dail y, we can stay well informed by constantly checking the charts (Matos, 2008). 2.2 Impact of Non-Business Lists Two of the most well known t ypes of non-business lists are college rankings and journal rankings based on impact factors. Understandi ng how these rankings effect college admissions and the popularity and use of journals may provide insight as to whether or not lists have an impact beyond use as an organizational tool. College rankings first appeared in the United States in the 1870s (H olub, 2002). Newsmagazines such as the U.S News and World Report, and various other sources, such as BarronÂ’s and Princeton Review often produce lists that rank colleges based on a numbe r of different criteria. The popularity of these types of lists can be seen in the rise in sales of the issue of U.S News and World Report containing its ranking of AmericaÂ’s Best Colleges (Ehrenberg and Monks, 1999). Also, Time magazine reports that the public spe nds approximately $400 million on college-prep materials, which would include money spent on college rankings information (Holub, 2002). Although rankings are popular, many studies have been performed to discover whether these college rankings have any impact on collegebound students and the colle ges themselves. Monks and EhrenbergÂ’s study (1999) found that a decrease in rank on the U.S News and World ReportÂ’s college ranking list was correlated with decrease in selectivity of the down ranked university that resulted from a having to admit a larger percen tage of a smaller applic ant pool. They also found that the average SAT score of the entering fres hman class decreased as rank decreased.
7 HolubÂ’s (2002) study found that co lleges tend to rely on the co llege rankings as a form of marketing to attract students, donors, and alumni. College rank ings also impact universities because these rankings influence a universityÂ’s reput ation, and ability to be more selective. This impact has resulted in an effort by many univers ities to influence their rank (Ehrenberg, 2003). For example, student selectivity is one of the parameters used by the U.S. World and News Report to rank the universities on its AmericaÂ’s Best Colleges list. Student selectivity accounts for 15% of the overall calculation of the rank. Ac cording to Ehrenberg (2003), universities have begun to decline qualified st udents in order to increase their sc ore on selectivity, but this activity has resulted in a larger pool of applicants the admissions office rs have to sort through, and a greater percentage of students being accepted by early decision. Although many studies have shown that the rank ings may be biased in some way (the U.S. News and World Report has been under a lot of scrutiny fo r its methodology of calculating rank), students still rely on the rankings for information and universities still vie for the top positions. According to Ehrenberg (2003), the rankings have continued to flourish despite obvious biases because it is in everyoneÂ’s best interest to us e the rankings since they are so influential. Another popular type of ranking is done on acad emic journals. Academic journals are ranked based on impact factors and these im pact factors can then be used to evaluate individual papers, authors, and institutions. An impact factor meas ures the number of citatio ns a journal receives over two years and divides that number by the numb er of articles that journal published during the same time period (Kai, 2008). Thomson Reuters publishes the Journal Citation Report (JCR)
8each year. The JCR is a metric used to show Â“t he worldÂ’s leading journals and their impact and influence in the global research community Â” (Thomson Reuters, 2008). The information provided by the JCR is used to determine if a journal is performing well in comparison to its competitors and if a journal needs to change it s market strategy so that it is better known and respected and will receive more citations in the future. According to Kai (2008) impact factors can also be used to rank unive rsities and research in stitutions, and could affect an authorÂ’s chance of being hired or promoted at one of th ese institutions. According to Garfield (2006) journal impact factors are infl uenced by the size of the journa l community. He asserts that a relatively small number of articles account for a majority of the citations. Impact factors, just like college rankings, have criticisms. According to Kai (2008), impact factors can be manipulated by the journals to show improvement. One way journals accomplish increasing impact factor is by publishing more revi ews than actual papers because the reviews will receive more citations. Garfie ld (2006) offers the criticism that if a journal article is published outside of the two year time frame used by the JCR the citations in the article will not be counted towards the impact factor. Although impact factors have faults, th ey are still used to evaluate journals because it is currently the be st system available, and because people rely on impact factors they have become important to both those who use them and those who are evaluated by them. The research done on college and journal impact f actors has shown that a lis t or ranking can have a significant impact on the universities or jour nals on the list and that moving up or down a ranking on a list can have an impact as well. However, the magnitude of the impact is
9determined by the reputation a nd respectability of the list creator, and the popularity and proliferation of the list. This research has also shown that the lists or rankings with the most influence are those that have a methodical way of calculating rank. Although both measurement systems are flawed people still use them because a better system has yet to be developed and if people believe these rankings ar e important the rankings will have some amount of influence. 2.3 Impact of Business Related Lists This portion of the paper will focus on busines s related but non-finance lists and the impact these lists have ob the companies th ey rank. Business lists, in genera l, are often used to evaluate companies in comparison to one another. These li sts are often compiled using quantitative rather than qualitative data and those who use these list s will often use qualitative data to evaluate the reliability of these lists. We w ill explore a companyÂ’s financial performance after they are added to a list. We will examine Fortune Magazine's 100 Best Companies to Work For and Forbes' 75 Most Reputable Com panies in the U.S. We will evaluate the each company's performance by calculating the difference in net income in the year after the company was added to the list. Criteria for being on Fortune Magazine's 100 Best Companies to Work For include number of employees, percentage of job growth, benefits, he alth, work life, and diversity. After examining the list, we pulled up the most recent income stat ements for the top ten companies on the list. For the full list see figure 1. The list was released at the beginning of February 2008, so we looked at statistics for the nine-month period ending Sept ember 2008. We were able to find the necessary income statements for five of the companies out of the top ten. We found that of the companies
10whose financial statements were available, only one (Qualcomm) reported a net income less than the nine-month period ending September of the prev ious year. Was the fact that these companies were added to this list the reason that their net income for the most part went up? It is impossible to tell exactly what effect the list had on the co mpany, but it is a correlation that cannot be ignored nonetheless. The next list we will examine is ForbesÂ’ 75 Most Reputable Companies in the U.S The Reputation Institute provided the criteria for this list from their consumer survey results. Google again topped this list, for the full list see figure 2. Income statements were available for all companie s on this list, and all but 3 reported an increase from the net income for the same period of the pr evious year. We looked at quarterly statements ending September 2008 for this list because it wa s released in early June of 2008. Companies reporting lower net incomes than th e previous year were General Mills, United Parcel Service, and Texas Instruments. Again, the correlation is obvi ous, but the specific im pact that the list had is difficult to read. In conclusion, the impact of the lists themselves is near impossible to tell, but there is a definite correlation between companies being put on the list and their performance improving thereafter. Whether or not the reason for the performance sp ike was inclusion on the list cannot possibly be understood. What we do know is that lists are very important to us. We also know how difficult it is to distinguish correlation from causation once lists are published.
112.4 Impact of Finance Related Lists Among the many types of lists there are business re lated lists that fall within the financial industry. These lists rank companies whose main product is the service of financial management in all capacities from credit, ba nking, investing, and any type of service dealing with the overall management of money, both institutional and consumer ba sed. The rankings are based on quantitative financial indicators, versus qualitative factors, such as business volume, debt issuance, total assets, or revenue s. In addition, the bus iness related lists wi thin the financial industry differ in composition from general busine ss related lists because all of the companiesÂ’ products revolve around the management of mone y while the general business related lists companiesÂ’ products range from internet serv ices, goods, technology, money, healthcare, and an unlimited range of industries. Whether these finance-industry specific rankings impact the financial performance of the corporation is the question. We w ill evaluate two different lists and the financial performance in terms of annual net income increases and decrease s after the release of th e rankings. The first list ranks the top U.S. commercial banks by total a sset value and the second ranks the top U.S. commercial banks by total revenue by EconomyWatch. The list of the Top 10 Commercial Banks by Assets was released in 2005 and of the ten banks listed all but one reported increases in net inco me from 2005 to 2006 (For the full list see Figure 3). This strong correlation between being listed as a top 10 company a nd increases in revenue could indicate that the list is influential on net income because ne t income is not calculated using
12total assets. That assumption may be flawed though because asset values could be an indicator of financial strength, thus explai ning the financial strength in increases of net income. Evaluating the net income increases and decreases in the Top 10 Commercial Banks by Revenue in 2005, we also find that all of the companies but one found increases in net income between 2005 and 2006 (For the full list see Figure 4). The ba nks with decreases in net income are not the same between the two lists, but once again the st rong correlation shows a link between the list and financial performance of the company. The ranking by revenue plays a significant role in this list though because revenue is a direct component for net income calculations. This calculation means that the increases or decreases in net income can not be attributed to the timing of the list publication since the companies are ranked by quantitative financial indicators (revenue) that are not independe nt of our financial performance metric of net income. Overall when evaluating the effect of business related lists ranked by quantitative financial indicators on the financial performance of a company, one can not deduce that publication of a finance list is the causal factor in the financial performance of a company. Since the two are not independent of one another, the relationship may be one of revers e causality creating significant bias in the evaluation. In addition, many external short and long term factors play into net income results. Measuring those effects are comp lex and citing the impact of business related finance lists is limited. 3. Sample Selection
13 3.1 Business Ethics List Description The benefits from participating in corporat e citizenship go far beyond earning a good return for your stockholders. In the first issue of 100 Best Corporate Citizens the editors of Business Ethics indicated that there were several intangible benefits that a company received when it became a so called "good corporate citizen". Some of these benefits include better employees, improved customer loyalty, lower cost of cap ital and lower cost of mitigation. How does a company attain the status of being a "good corporate citizen"? According to Filbeck, Gorman, and Zhao (2008), a firm must not only se rve their shareholders but it must also serve a variety of other stakeholders to the best of its ability. In order to accurately compare companies, Business Ethics identified 7 stakeholder groups used as measurement criteria to determine how well each company serves the stakeholders around it. These stakeholders include shareholders, the community, minorities and women, employees, the environment, non-U.S. stakeholders and customers. The data garnered from each st akeholder group would be used to indicate the strengths and concerns of each company. The overall score of each category would then be determined by subtracting the concerns from the st rengths. These scores w ould be the criteria by which companies could be compared fairly. In order to obtain the data necessary to make valid comparisons, Business Ethics first collected information on companies that were listed on the Russell 1000 Index, Domini 400 Index, and the S&P 500 index. However, before 2003 only the enti re S&P 500 list was used in conjunction with
14150 companies from the Domini 400 index. In add ition to this, KLD Research and Analytics has been a consistent source of comparative data fo r each stakeholder categor y since the creation of the list in 2000. KLD excludes all companies that derive more th an two percent of sales from military weapons, derive any revenues from the sale of alcohol or tobacco products, and derive any revenues from the providing of gaming services or products. It also excludes a ny firms that own interests in power plants or derive electricity from nuclear power plants in wh ich they have an interest. As a result, the original Business Ethics sample also excludes companie s that meet any of the above criteria. 3.2 Business Ethics Sample The purpose of creating the sample was so that it could be compared against a sample of matched companies whereby several hypothesis' c ould be tested, for example, whether a firm experiences improved returns following their inclusion to the Business Ethics list. Filbeck, Gorman, and Zhao (2008) mention that because Business Ethics list companies are members of three specific indices then the Â“whole sampleÂ” they use will also be comprised from those same indices (S&P 500, Russell 1000, and Domini Soci al 400). After passing through the KLD exclusionary screen mentioned above, a firm had to meet three more criteria to even qualify for the list. The first criterion was that the company must be listed in the Center for Research on Stock Prices (CRSP) database im mediately before the date of its addition to the ethics list was announced publicly. It must have also ha d records on CRSP immedi ately following that
15announcement date. The last criterion was that the company must have comprehensive data on the Standard and PoorÂ’s database called Resear ch Insight. This database includes essential financial and statistical market data for companies within the United States and Canada. The 660 companies that remained were those that make up the Business Ethics sample from each of the seven years of data (2000 to 2006). Most of the resulting companies for both sample s are part of the manuf acturing sector, followed by the finance, insurance, and real estate sector s. It is significant to note that both samples purposely have similar sector compositions. The t op five industry sectors make up about 93% of each sample. And lastly, companies on the Business Ethics list make up around 5 to 12% of each of these sectors. The paper goes on to explain th e procedure used to create th e matched sample. This matched sample is based on market capitalization and th e book value of common e quity-to-market value of common equity or BE/ME ratios. Filbeck, Go rman, and Zhao (2008) claim that using this basis to create the matched sample will safeguard against mis-specification and will produce sufficiently specific and usable results. They did not use market capita lization coupled with industry size to create the sample because th ere would be a limited number of matchable companies within each industry. Furthermore, it is important to note that the samples are compared directly against the S&P 500 Index for good measure. Using CRSP, the previous year end market cap italization was record ed. Using Insight the BE/ME ratios were calculated. At that point, anything with a ra tio less than zero was removed.
16The remaining companies from the whole sample made up the potential matches for the Business Ethics list sample. Of course, if a ny companies appeared in both samples they were removed from the whole sample. Next, a matching sc ore (MS) was calculated for each of the Business Ethics sample companies against each of the whole sample companies. The companies from the whole sample with the lowe st MS to a corresponding Business Ethics company were then chosen as the matched companies. This procedure was c onducted for each of the seven survey years. 3.3 Matched Firm Sample and Index To compare the returns of companies listed in the Top 100 Ethical Companies ( Business Ethics list) to companies not listed, we created two matched samples. The first matched sample created was solely based on book to market value, in or der to compare our results with those obtained by Filbeck, Gorman, and Zhao (2008). Then, a more focused matched sample was created to narrow down the comparable companies by first looking at the Standard International Classification (SIC) codes of each company listed in the Business Ethics list. The matched companies share the first two digits of the SIC codes of the companies listed to allow for direct comparison within the same industries. We narrowed down the compan ies even further with in the SIC match by selecting those companies with a beta closest to that of the listed co mpany. Comparing beta ensures that the returns observed are not affected by large differe nces in risk between the listed company and the matched company. If there we re still multiple companies in the matched sample pool after comparing by SIC codes and be ta, we matched by the firm with a market capitalization closest to that of the listed company. A similar market capitalization of the two companies shows that both the listed company and the matched company have a comparable
17level of influence on the market. Once we crea ted this matched sample, we compared each company to the market using the S&P 500 inde x. We believe the S&P 500 index is a good representation of the total mark etÂ’s performance and so use th is index as a benchmark to compare companies. As mentioned previously, th e study done by Filbeck, Gorman, and Zhao (2008) created a matched sample based on book value of common e quity to market value of common equity. Filbeck, Gorman, and Zhao (2008) argue that this method Â“minimize[ s] possible industry misclassification.Â” While we agree that the book to market ratio is a good variable for determining a matched sample, we feel strongl y that the matched sample must also be a reflection of the listed companyÂ’ s industry. Therefore we created two indexes, one which follows Filbeck, Gorman, and Zhao (2008) matching sole ly on book to market ratio, and another which has the criteria that we feel best reflects a true match of the companies listed. For our true match index we used SIC codes to compare companies listed on the Business Ethics list with companies not listed by industry. An alte rnative classification system that we opted not to use was North American Industry Classifica tion System (NAICS) codes. The NAICS codes were developed in 1997 to replace the SIC codes to allow for greater compatibility with the United Nations Statistical OfficeÂ’s Internationa l Standard Industrial Classification system. However, the Securities and Exchange Commi ssion, which is responsible for regulating the securities industry and overseei ng the disclosure of financial records made by publicly traded companies, presently still uses the SIC coding syst em and thus this system of classification better complements our data.
18 We define beta as a factor th at describes the degree of risk undertaken by a company to meet their investorsÂ’ required rates of return. In calculating companiesÂ’ beta, we took the monthly returns of each company on the Business Ethics list and regressed these returns against the monthly average returns observed in the S&P 500 for the same respective dates. The slope obtained through this regression re presents the change in the fi rmsÂ’ required rates of return, given a change in the required re turn of the market (S&P 500). We interpret this slope (beta) to represent the level of risk undertaken by a company relative to the risk premium of the market. We selected companies for the matched sample which have the closest betas to that of the company on the Business Ethics list to ensure comparison betw een companies with a similar level of risk. Investors use market capitalization to determine the size of a company based on the current price of the companyÂ’s stock and the number of shar es outstanding. Market capitalization is an important factor to determine the size and amount of influence a company may have on its industry. In order to further maintain the simila rity between the listed company and the matched company, we took the remaining companies derived from the above criteria and selected a single company with the closest market capitalization to the company listed on the Business Ethics list. 3.4 Anti-Ethics Sample and Index
19In order to analyze whether firms on the Business Ethics list outperformed their counterparts excluded from the list, we created our own Anti-i ndex composed of firms in the beer and liquor (SIC Codes, 2080-2085), tobacco (SIC Codes, 2100-2199) and defense (SIC Codes, 3760-3769, 3795, 3480-3487) industries. The index is equal weighted and value we ighted to allow for direct comparison. We will start the analysis by giving an overview of each industry. Beer & Liquor (SIC Codes: 2080-2085) SIC Code Industry 2080 Beverages 2082 Malt Beverages 2083 Malt 2084 Wines, brandy, and brandy spirits 2085 Distilled and blended liquors *SIC Code List The alcohol industry in the United States has seen nothing but success and consolidation since prohibition ended in the early 1930Â’s. With $137 b illion in sales in 2002, beer accounting for 86% of sales, and the top ten companies controll ing 95% of the beer market, most recently with the buyout of Anheuser Busch by InBev, industr y revenues are growing but the number of players is shrinking (Y oast, 5-10). The consolidation of the industry ca n be attributed to the vertical integration strategies used by many of the companies. Previously, a three-tiered system was utilized to separate producers, distributors, and sell ers; however, the distinctions between the three led to conflicts of interest in terms of laws, regulations, and consumer pricing. Now, companies that are vertically integrated
20are able to align strategies, pr icing, and most importan tly, lobby for laws and regulations to be passed that benefit the industry as a whole (Yoast, 7). This type of consolidation is not just apparent in the beer market, but in the spirits (strong alcoholic liquor) mark et as well, with the top five companies controlling 59% of the U.S. mark et (Yoast, 11). With less conflict of interest and increasing consumption despite the economic downturns of recent times, both markets have seen an increase in sales. To keep consumption high and sales increasing, the alcohol industry spends $4.8 billion a year on advertising in the U.S. alone (Yoast, 16). Ethi cal problems arise, however, with such a high budget. For advertising alone, the budget for the National Institute on Alcohol Abuse and Alcoholism is 10% of the marketing budget for the alcohol industry in its entirety (Yoast, 16). Reinvesting revenues in market re search; using mediums such as movies, television, print, radio, web; and creating a Â“lifestyleÂ” br and has allowed the industry to pe rmeate almost every aspect of American culture (Yoast, 17). Such a wide imp act has ultimately led to numerous anti-alcohol organizations calling for a curb in advertising as well as the elimination of targeted youth marketing. A portion of the advertising efforts has been focused on attracti ng the brand loyalty of younger and younger consumers; for example, a study done by the FTC (Federal Trade Commission) found that ei ght alcohol companies made product placements in PG and PG-13 movies (Yoast, 19). While targeting younger consum ers is prevalent today, the most apparent Â“lifestyleÂ” type marketing campaign by the indust ry can be seen by its attempt to capitalize on the healthy and fit lifesty le trend. Recently, there has been sign ificant investment to research the health benefits of alcohol, even though they might be limited. Furthermore, the products themselves have been marketed not just as a so cial facilitator but as a way to stay healthy
21through moderate consumption. The entire advertis ing strategy of the alc ohol industry has been to implant the drinking of its beverages into th e lifestyle of consumers regardless of age. Therefore, just taking the success and profitability of the alcohol industry from face value we believe it is significant to examine the returns of said firms against so called "ethical firms". Tobacco (SIC Codes: 2100-2199) SIC Codes Industry 2100 Tobacco Products 2110 Cigarettes 2120 Cigars 2130 Chewing and Smoking Tobacco and Snuff 2140 Tobacco Stemming and Redrying *SIC Codes In line with the alcohol industry, the tobacco industry has seen a rise in its revenues even through increased regulation, economic woes, and a negativ e brand image. As an example, the revenues of the worldÂ’s largest tobacco firm, Altria Group, rose by 37% from 1994 to 2007 and the revenues of the second largest firm, British Am erican Tobacco, rose by 67% from 1999 to 2006 (Colbert). A portion of this significant increas e in revenues can be attributed, specifically in the U.S., to the rising costs of cigarettes. Consumers have had to deal with tobacco firms raising their prices, in addition to price hikes as a result of federal a nd state taxes (Colbert). Despite this, intensive marketing campaigns have kept the industr y strong. In 2005, cigarette companies spent $13.11 billion and smokeless tobacco manufacturers sp ent $250.8 million on advertising and promotions
22(Fact Sheet). However, ethical issu es have arisen due to the speci fic target of these advertising dollars. Women are particularly over-targeted by advertisements through the creation of tailored brands which instill an independent/strong brand im age. This type of disproportionate marketing exists when targeting specific races as well. For example, brands have been created using Native American names to target a that specific popul ation, and more marketing dollars are spent in African American magazines relative to other publications (Fact Sheet). To answer the outcry by anti-tobacco groups, firms have begun marketing against youth smoking. In addition, the types of cigarettes produced are changi ng by advancements in science and technology, as the materials us ed can reduce the inhalation of toxins. Also, other type of materials used are limiting the environmental imp act caused by cigarettes; for example, in New York, only burn free cigarettes are allowed whic h automatically extinguish if not being puffed (Tobacco). In its search to delve into every consumer niche, however, the indus try has increasingly concentrated on international promotions, specifi cally in China, which will most likely be the trend for the future (Colbert). In the United States, the tobacco industry has managed to cement a strong footing. Even with intense pressure fr om different non-profit groups, coalitions, the government, and consumers the industr y has kept steady in terms of sales, which is why we also deem this industry useful for our Anti-index sample. Defense (SIC Codes: 3760-3769, 3795, 3480-3487) SIC Codes Industry
23 3760 Guided Missiles and Space Vehicles and Parts 3761 Guided Missiles and Space Vehicles 3764 Guided Missile and Space Vehicle Propulsion Units and Propulsion 3795 Tanks and Tank Components 3480 Ordnance and Accessories, Except Vehicles and Guided Missiles 3482 Small Arms Ammunition 3483 Ammunition, Except for Small Arms 3484 Small Arms 3489 Ordnance and Accessories, Not Elsewhere Classified *SIC Codes The defense industry, although vital for national secu rity, has always faced ethical questions due to its nature of business and secrecy. The industr y faced tough times since the end of the Cold War. But due to the ensuing wars in Afghanistan and Iraq after 2001, there has been a revival of the production, research, and development of defense products. With all that in mind, and the United States l ooking to scale back on the operations in Iraq, defense contractors are searching for new areas of growth for the industry. Specifically, the industry sees the new area of growth to be in Africa. Since 2004, some of the top defense companies in the U.S. have received upwards of $1.1 billion to train African soldiers, and currently, are bidding on $1.2 billi on worth of contracts (Delevingne ). However, ethical issues arise from these contracts as well. Industry e xperts cite that the tr aining is not used to professionalize militaries for sovereign safe keepi ng, but instead it is used as a tool for guerilla warfare, terror, and internal strife among the African countries (Delevingne).
24Strictly in regards to U.S. spending on defense, there will most likely only be a small increase, if at all, in the budget over the next few years. Due in large part to concerns of domestic infrastructure, health care, and the economy, the GEIA (Government Electronics Industry Association) forecasts a $8.7 billi on increase over the next ten years for spending on fuel costs, food, clothing; a decrease in th e research budget by $5.5 billion th rough 2019; and a decrease of $1.4 billion on the acquisition of new aircraft, ships, electronics, and upgrade/re pairs (Keller). On the whole, the defense industry raises ma ny ethical questions. A lthough its products are necessary and vital for the protec tion of national security, many see the industry as promoting warfare and conflict. Due to the recent revival of the industry we again will use these applicable firms to create the Anti-index. 4. Stock Performance 4.1 Within List Ranking Â– Top 15 vs. Bottom 15 In this section we examine the yearly returns of the top 15 ethical companies from the Business Ethics List and compared them to the yearly retu rns of the bottom 15 ethical companies from the list. We created both an equal weighted averag e and an average based on market capitalization. We then ran regressions to determine the extent that factors other than owning a spot on the list affected the returns of the ethical companies. Th e factors that we ran regressions to check for include differences in betas, differences in ma rket capitalization, differences in book to market value and a multiple regression that tested for a ll three. The main goal of this section was to
25determine whether a companyÂ’s position on the list was more important than simply being on the list. By comparing the top 15 companies on the li st to the bottom 15 companies on the list we should be able to determine whether the ra nkings themselves have any importance. Our study found that the average monthly returns for the equal weighted top 15 companies list was .6895% with a standard deviation of 4.65%. In comparison the average monthly returns for the equal weighted bottom 15 companies list was 1.192% with a standa rd deviation of 4.85%. When we ran the two sample difference-betw een-means t-test, we could not say with 95% confidence that the mean returns of the top 15 list outperformed the bot tom 15 list on the equal weighted index. Furthermore, our study found the average mont hly returns for the value weighted top 15 companies list was .4236% with a standard de viation of 4.49%. In comparison the average monthly returns for the value weighted bottom 15 companies list was .7142% with a standard deviation of 4.93%. When we ran the two sample difference-between-means t-test, we could not say with 95% confidence that the mean returns of the top 15 list outperformed the bottom 15 list based upon average value weighted monthly returns. Both the difference-between-means tables can be found in Tables 12 Â– 13. In our study we created indexes which can be used to compare the annuali zed returns of certain groups of stocks over a specified time period. Fo r each index we calculate d the total value of investing $1000 from April 1, 2001 until March 31, 2007. We created four indexes in total, including a top 15 companies index based upon e qual weighted monthly returns, a bottom 15
26companies index based upon equal weighted mont hly returns, a top 15 companies index based upon value weighted monthly returns, and a bottom 15 companies index based upon value weighted monthly returns. If we would have invested $1000 in the top 15 company equal weighted index our investment would have grown to $1516.83 with an annualized re turn of 7.19%. If we would have invested $1000 in the bottom 15 equal value weighted inde x for the same time period our investment would have grown from $1000 to $2157.73 with an annualized return of 13.68%. These annualized rates of return were calculated with the XIRR functi on in Microsoft Excel with one cash outflow on April 1st 2001 and one cash inflow on March 31st, 2007. All of these results are illustrated in Graphs 12 Â– 15. If we would have invested $1000 in the top 15 company value weighted index our investment would have grown to $1262.01 with an annualized re turn of 3.95%. If we would have invested $1000 in the bottom 15 company value weighted i ndex our investment would have grown to $1526.91 with an annualized return of 7.3%. In bot h weighting scenarios the bottom 15 company index actually outperformed the top 15 company i ndex by a substantial amount In order to take into account other factors that may have create d these results we did re gression analysis that would account for variances due to market cap italization, beta, book to market value and a combination of the three. We ran four regressions in tota l that compared differences in returns between the top and bottom 15 companies to specific variables of interest. Al l four of the regressions returned results that
27indicated incredibly weak correl ations. Our highest coefficient of determination for regression was .0886 which existed between a combination of th e three variables of in terest and a difference of returns of the top 15 companies and bottom 15 co mpanies. This is an extremely weak result that is indicating that there is little to no corre lation between the variables of interest and the differences of returns. All of our other regressions returned data that was weaker than the above mentioned regression. These regressions can be found in Tables 14 Â– 17. To find the difference between the returns of th e top and bottom 15 companies from the Business Week Ethical Companies list we first needed to find a yearly average return to use for each company. We did this by summing the monthly retu rns starting in April an d ending in March of each year for each company. This is not a conventional approach and may have introduced some level of error into our research. However it was the simplest approach and we did not think that the level of error introduced by this method would have much weight on the research as a whole. We then found the difference between the returns of the top 15 companies for each year and the bottom 15 companies for each year. We also calcu lated standard deviations for each yearly average return to judge the amount of variance th at existed between each i ndividual yearly return and the overall average return that we used to compare top 15 companies with bottom 15 companies. These standard deviations were calculated for each year from 2001 to 2006 in Microsoft Excel so they did ta ke into account degrees of freed om within the calculation. In order to run regressions we decided to use average monthly returns for the top 15 index and the bottom 15 index for each year. This was done in order to improve the quality of the regressions by greatly incr easing the sample size from what it w ould have been if the regressions
28had been run with average yearly returns. Average monthly returns were calculated for each individual month by using a simple average of all the individual monthly returns from each company within the list. We then found the diffe rence between the average monthly returns of the top 15 companies and the bottom 15 companies for each month. These differences were what were used as the independent variable in the regression analysis. Severa l regressions were run based upon various explanatory variables. These explanatory variables in clude differences in average yearly betas between the top and bottom 15 companies, differences in average market capitalizations between the top and bottom 15 comp anies, differences in average book to market values between the top and botto m 15 companies, and a regression that utilizes all three variables. In order to determine average beta for the top and bottom 15 et hical companies, we found the yearly average of the individual firms for each year. We then found the difference between the average annual betas of the top 15 companies and the bottom 15 companies. The same method was used to determine the book to market values and the market capitaliza tion values to use in regression analysis. It was these differences that were used wh en conducting regression analysis. 4.2 Ethics Firms vs. Book to Market Matche d Sample, Extended Matched Sample, and AntiEthics Sample In this section we look at the re turns of the comp anies listed on the Business Ethics List compared to the returns of th e companies in the book to market matched sample, the extended matched sample and the anti-ethics sample. We created an equal weighted and a value weighted index of the Business Ethics companies and compared the returns of the Business Ethics index to
29four benchmark portfolios, which we also broke down into equal weighted and value weighted indexes: the book to market ratio matched sample (to follow the method of Filbeck, Gorman, and Zhao (2008)); the extended matched sample (usi ng SIC, beta, and mark et capitali zation); the anti-matched sample; and the S&P 500 Index. Our study found that the monthly returns of the e qual weighted ethics list had a mean of 1.2% and a standard deviation of 4.5%. In comparis on, the equal weighted book to market matched sample had a mean return of 1.5% and a standard deviation of 5.5%. The value weighted ethics list had mean monthly returns of 0.57% and a standard deviation of 4.1%, while the value weighted book to market matched sample had m ean returns of 0.24% and standard deviation of 3.1%. When we ran the two sample difference-be tween-means t-test, we could not say with 95% confidence that the mean returns of the ethics list outperformed the book to match on either the equal or value weighted indices. One noticeable observation was that the st andard deviations of all our t-tests were extraordinarily large compared to their respective sample means. All difference-between-means tables can be found in Tables 1-4. We then compared the monthly returns of th e ethics list to the extended matched sample described in section 3.3. We found that the equal weighted exte nded matched index had a mean of 0.52% and a standard deviation of 5.1%. Altho ugh this mean is lower than the mean of the equal weighted ethics list index, we can not say with 95% confid ence that the ethics list has a significantly higher mean. When comparing the valu e weighted indices of the extended matched sample to the ethics list, we found that the va lue weighted extended matched sample had a mean return of 0.55% and a standard deviation of 4.2%. The t-test of the difference-between-means
30showed that we can not say with 95% confidence that the returns of the ethics list companies outperformed the extended matched sample. We ran the same tests for the anti-ethics sample and found that the equal weighted sample had a mean return of 0.035%, and a standard deviati on of 0.261%. The value weighted sample had a mean return of 0.04% and a standard deviation of 0.48%. The two sample t-test again showed that the results are not statistic ally significant within a 95% conf idence interval to say that the ethics list outperformed the anti-ethics list. For each of the benchmark portfolios we calcula ted the final value of investing $1,000 on April 1, 2001 until March 31, 2007. $1,000 invested in the S&P 500 would have grown to $1,224.53. The annualized return for the S&P 500 is 3.43%. A chart of this growth can be found in Graph 1. The equal weighted Business Ethics list index would have grown $1,000 into $2,159.67. The annualized return of this index is 13.69%. A ch art of this growth can be found in Graph 2. $1,000 invested in the value weighted Business Ethics list index would have grown to $1,416.55. The annualized return for this index is 5.98% A chart of this gr owth can be found in Graph 3. When we look at the book to market matched samp le indices, we see that $1,000 invested in the equal weighted index would have grown to $2,637.71. The annualized return for this index is 17.6%. However, the value weighted inde x would have grown $1,000 to only $1,151.19, which is an annualized return of 2.37%. Charts of this growth can be found in Graphs 4 and 5 respectively. The extended matched sample indices show that $1,000 invested in the equal
31weighted index would have grown to $1,327.48, whic h has an annualized return of 4.83%. In the value weighted extended matched index, $1,000 would have grown to $1,394.37. This is an annualized return of 5.7%. Charts of the extend ed matched sample indices growth can be found in Graphs 6 and 7, respectively. If we had invested in the unethical companies list ed in the equal weighted anti-ethics sample, our $1,000 would have grown to $1,025.58, with an annua lized return of 0.42%. In the value weighted anti-ethics sample index, our $1,000 would have grown to $1,028.91, which is an annualized return of 0.48%. As can be seen by th is data these returns are very low, indicating that some degree of the unethical nature of th ese companies may have impacted their long term returns. These findings can be seen in the ch arts in Graphs 8 and 9, respectively. Graph 10 tracks the growth of $1,000 invested in each of th e indices. The annual re turns for all of these indices both equal and value weighted, can be f ound in Table 5. Graph 11 shows the annualized returns of each of the indices. We noticed a conspicuous difference between the e qual and value weighted returns in the data. We know that the period of time we are exam ining involved the Â“dot-comÂ” bubble which sent many companies stocks skyrocketing above traditiona l price to earnings ratios. Indices that were capitalization weighted are much more prone to over-weighting these overpriced securities and under-weighting the underpriced securities. As suc h, value weighted indices during that period of time, including the S&P 500 and our value we ighted ethics list index, would have underperformed the equal weighted index. An inte resting observation that corresponds with this analysis is that the indices th at have a fundamental component, namely the equal weighted book
32to market matched sample index and the e qual weighted extended matched sample index, outperformed all other indices in the ensuing months after the Â“dot-comÂ” bubble burst. The fundamental component built into these indices en sured that they did not overweight the Â“dotcomÂ” sweethearts and instead bough t the stocks with lower price to earnings ratios. We ran six regressions of one index to anothe r to evaluate whether any two indices were correlated. The indices that had the highest degree of correlation we re the value weighted ethics list index and the S&P 500 index. The coefficien t of determination for the regression was 0.719, and the equation that related the two indices was Y = 0.93X + 0.002, with X = S&P 500 returns and Y = ethics list returns. The other regressions which we ran did not sh ow any correlation that would suggest that the returns be tween any other two indices woul d be linked. The scatter plots of each regression, the regression statistics, and the residual plots can be found in Tables 6-11. 5. Conclusion Our results showed that during the period that was tested, in vesting in the ethics list would have yielded greater returns than i nvesting in the book-to-market matched sample, the extended matched sample, the anti index, and the S&P 500. The Anti-index performed rather poorly relative to the S&P 500 and very poorly when comp ared to the index of ethical companies. When regressions were run between each of the indices, we found th at only the regression of the Ethics List on the S&P 500 produced results that showed any correla tion. All of our other regressions indicated that ever y other index had little predicte d power over another index. This was expected because it is intuitive that compan ies that behave in a fundamentally different
33manner will have returns that have low correlatio n. An interesting follow up to this paper would be to examine the combinations of ethical compan ies and unethical companie s in order to create a diversified portfolio. Lastly, ev en though our ethics list outper formed the other indices, the results were not conclusive enough to say with 95% confidence that any one index had a greater mean then any other index. We noticed distri butions that had small mo nthly mean returns and large standard deviations. When we compared the top 15 ethical companies with the bottom 15 ethical companies we returned mixed results from the different calcula tions that we ran. The average monthly results showed a lot of variance between the two groups and also between the equal-weighted returns and the average-weighted returns. There was no clear answer to which group performed better over the time period in question. In addition, when we ran the two sample difference-betweenmeans t-test, we could not say with 95% confid ence that the mean returns of the top 15 list outperformed the bottom 15 list on the equal-wei ghted index. When we created indices, the bottom 15 consistently performed better than the top 15 for the time period tested. Lastly, our regressions showed very little correlation and did not offer anything of importance to the research. All in all, our research did not produce significant results, and di d not deviate from what we had originally hypothesized. The companies on the 100 Best Corporate Citizens list would have yielded greater returns than i nvesting in the book-to-market matched sample, the extended matched sample, the anti index, and the S&P 500.
34References Buxton, Rodney. Museum of Broad cast Communications. 2008. Chicago, IL.. Colbert, Catherine. "Tobacco." Hoovers. D&B Company. 5 Dec. 2008 . Deevingne, L. (2008, October 24). Defense C ontractors' Growing African Business. Business Week Online Retrieved December 5, 2008, from Academic Search Premier database. Ehrenberg, Ronald G., 2003. Met hod or Madness? Inside the USNWR College Rankings, ILR Collection Working Papers. Ehrenberg, Ronald G. and James Monks, 1999. The Impact of U.S. News & World Report College Rankings on Admissions Outcomes and Prici ng Policies at Selectiv e Private Institutions, NBER Working Paper Series. Entertainment Tonight CBS Interactive, Inc. 28 Mar, 2008. . "Fact Sheet: Tobacco Industry Marketing." Smoking & Tobacco Use. Apr. 2007. Centers for Disease Control and Prevention. 5 Dec. 2008 . Felsenthal, Norman. Museum of Broa dcast Communications. 2008. Chicago, IL. . Filbeck, G., Gorman, R., and Zhao, X.,(2008). Th e "Best Corporate Citizens:" Are they good for their shareholders? Working Paper. Fortune Magazine's 100 Best Companies to Work For. 02-04-2008. . Garfield, Eugene, 2006. The History and M eaning of the Journal Impact Factor. The Journal of the American Medical Association 295, 90-93. Holub, Tamara, 2002. College Rank ings, ERIC Digests. Keller, J. (2008, November). Defense budgets headed down, no matter who's in the White House. (Cover story). Military & Aerospace Electronics 19 (11), 1-7. Retrieved December 5, 2008, from Academic Search Premier database. Keith, Micheal. Casey Kasem: Reaching for the Stars Journal of Radio Studies, 2002. 9: 92.
35Kirdahy, Michael. The 75 Most Reputable Companies in the U.S. Reputation Institution Report, 2008. . List. Dictionary.com Unabridged (v 1.1) Random House, Inc. 10 Nov. 2008. Dictionary.com . Matos, Michaelangelo, 2008. Hot 100: C ounting Â‘Em Down Through the Years. Billboard Online September. National Radio Hall of Fame & Museum The Museum of Broadcast Communications. Chicago, IL. 2008. . "SIC Code List." EHSO. 26 Feb. 2008. Environment, Health and Safety Online. 5 Dec. 2008 . Simmons, Kai. 2008. The Misused Impact Factor, Science 322, 165. "Tobaco Industry Profile." Industry Cent er Tobacco. Yahoo! Finance. 5 Dec. 2008 . U.S. Securities and Exchange Commission. EDGAR Database November 2008. . Yoast, Richard A., and Janet Williams. Alcohol Industry 101: Its Structure and Organization. Rep.No. American Medical Association. Chicago, IL: Office of Alcohol and Other Drug Abuse, 2004. 2008. Thomson Reuters Releases the 2007 Citation Reports, Information Today 25, 37.
36 Appendix Table 1: T-Test Difference Between Means of Va lue Weighted Ethics List and Value Weighted Extended Match (assumes equal population variances). Hypothesized Difference0 Level of Significance0.05 Population 1 Sample Sample Size73Confidence Level95% Sample Mean0.00568 Sample Standard Deviation0.04109 Population 2 Sample Degrees of Freedom144 Sample Size73 t Valu e 1.976575034 Sample Mean0.00547 Interval Half Width0.013521959 Sample Standard Deviation0.04157 Interval Lower Limit-0.013311959 Population 1 Sample Degrees of Freedom72 Interval Upper Limit0.013731959 Population 2 Sample Degrees of Freedom72 Total Degrees of Freedom144 Pooled Variance0.001708 Difference in Sample Means0.00021 t Test Statistic0.030697 Two-Tail Test Lower Critical Value-1.976575 Upper Critical Value1.976575 p -Value 0.975554 Do not reject the null hypothesis Intermediate Calculations DataConfidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations Confidence Interval
37 Table 2: T-Test Difference in Means Between Va lue Weighted Ethics List and Value Weighted Book to Market Match (assume s equal population variances). Hypothesized Difference0 Level of Significance0.05 Population 1 Sample Sample Size73Confidence Level95% Sample Mean0.00568 Sample Standard Deviation0.04109 Population 2 Sample Degrees of Freedom144 Sample Size73 t Valu e 1.976575034 Sample Mean0.00244 Interval Half Width0.011950941 Sample Standard Deviation0.03131 Interval Lower Limit-0.008710941 Population 1 Sample Degrees of Freedom72 Interval Upper Limit0.015190941 Population 2 Sample Degrees of Freedom72 Total Degrees of Freedom144 Pooled Variance0.001334 Difference in Sample Means0.00324 t Test Statistic0.535866 Two-Tail Test Lower Critical Value-1.976575 Upper Critical Value1.976575 p -Value 0.592878 Do not reject the null hypothesis Intermediate Calculations DataConfidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations Confidence Interval
38 Table 3: T-Test Difference Between Means of Equal Weighted Ethics List and S&P 500 (assumes equal population variances). Hypothesized Difference0 Level of Significance0.05 Population 1 Sample Sample Size73Confidence Level95% Sample Mean0.01174 Sample Standard Deviation0.04472 Population 2 Sample Degrees of Freedom144 Sample Size73 t Valu e 1.976575034 Sample Mean0.00352 Interval Half Width0.013494082 Sample Standard Deviation0.03745 Interval Lower Limit-0.005274082 Population 1 Sample Degrees of Freedom72 Interval Upper Limit0.021714082 Population 2 Sample Degrees of Freedom72 Total Degrees of Freedom144 Pooled Variance0.001701 Difference in Sample Means0.00822 t Test Statistic1.204042 Two-Tail Test Lower Critical Value-1.976575 Upper Critical Value1.976575 p -Value 0.230549 Do not reject the null hypothesis Intermediate Calculations DataConfidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations Confidence Interval
39 Table 4: T Test Difference Between Value Weig hted Ethics List Mean and Value Weighted Anti-Index Mean (assumes equal population variances). Hypothesized Difference0 Level of Significance0.05 Population 1 Sample Sample Size73Confidence Level95% Sample Mean0.00568 Sample Standard Deviation0.04109 Population 2 Sample Degrees of Freedom144 Sample Size73 t Valu e 1.976575034 Sample Mean0.00041 Interval Half Width0.009569889 Sample Standard Deviation0.00478 Interval Lower Limit-0.004299889 Population 1 Sample Degrees of Freedom72 Interval Upper Limit0.014839889 Population 2 Sample Degrees of Freedom72 Total Degrees of Freedom144 Pooled Variance0.000856 Difference in Sample Means0.00527 t Test Statistic1.088471 Two-Tail Test Lower Critical Value-1.976575 Upper Critical Value1.976575 p -Value 0.278205 Do not reject the null hypothesis Intermediate Calculations DataConfidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations Confidence Interval
40 Graph 1: S&P 500 Growth of $1,000 $ $200.00 $400.00 $600.00 $800.00 $1,000.00 $1,200.00 $1,400.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 S&P 500 Growth of $1000
41 Graph 2: Equal Weighted Ethics List Growth of $1,000 $ $500.00 $1,000.00 $1,500.00 $2,000.00 $2,500.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 Equal Weighted Ethics List Growth of $1000 Graph 3: Value Weighted Ethics List Growth of $1,000
42 $ $200.00 $400.00 $600.00 $800.00 $1,000.00 $1,200.00 $1,400.00 $1,600.00 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 J a n u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 F e b r u a r y 0 0 M a r c h 0 0 M a r c h 0 0 M a r c h 0 0 M a r c h 0 0 Value Weighted Ethics List Growth of $1000 Graph 4: Equal Weighted Book to Market Matched Growth of $1,000 $ $500.00 $1,000.00 $1,500.00 $2,000.00 $2,500.00 $3,000.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 EW Book To Market Match Growth of $1000 Graph 5: Value Weighted Book to Market Matched Growth of $1,000
43 $ $200.00 $400.00 $600.00 $800.00 $1,000.00 $1,200.00 $1,400.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 VW Book To Market Match Growth of $1000 Graph 6: Equal Weighted Extended Matched Growth of $1,000 $ $200.00 $400.00 $600.00 $800.00 $1,000.00 $1,200.00 $1,400.00 $1,600.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 EW Extended Match Growth of $1000 Graph 7: Value Weighted Exte nded Matched Growth of $1,000
44 $ $200.00 $400.00 $600.00 $800.00 $1,000.00 $1,200.00 $1,400.00 $1,600.00 $1,800.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 VW Extended Match Growth of $1000 Graph 8: Equal Weighted Anti-Index Growth of $1,000 $960.00 $970.00 $980.00 $990.00 $1,000.00 $1,010.00 $1,020.00 $1,030.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 Equal Weighted Anti Index Growth of $1000 Graph 9: Value Weighted Anti-Index Growth of $1,000
45 $930.00 $940.00 $950.00 $960.00 $970.00 $980.00 $990.00 $1,000.00 $1,010.00 $1,020.00 $1,030.00 $1,040.00 A p r i l 0 1 A u g u s t 0 1 D e c e m b e r 0 1 A p r i l 0 2 A u g u s t 0 2 D e c e m b e r 0 2 A p r i l 0 3 A u g u s t 0 3 D e c e m b e r 0 3 A p r i l 0 4 A u g u s t 0 4 D e c e m b e r 0 4 A p r i l 0 5 A u g u s t 0 5 D e c e m b e r 0 5 A p r i l 0 6 A u g u s t 0 6 D e c e m b e r 0 6 Value Weighted Anti Index Growth of $1000 Graph 10: Growth of $1,000 in All Indices 0 200 400 600 800 1000 1200 1400 1600 1800 14710131619222528313437404346495255586164677073 Ethics List Value Weighted Growth of $1000 S&P Growth of $1000 Anti Index Value Weighted Growth of $1000 Value Weighted Growth of $1000 Extended Match Value Weighted Growth of $1000
46 Table 5: Annualized Return Data for All Indices EqualValueEqualValueEqualValue Standard Deviation Mean 1.06610%0.15225% 1.66937% 0.77869%3.68687%2.11042% Standard Deviation Mean 3.217440% 0.048680% 1.358672% 0.426839%4.627793%1.563098% Standard Deviation Mean 1.498542%0.969614% 1.771123% 0.844489% 2.202663%1.625824% Standard Deviation Mean 0.135362% 0.042901% 2.548932% 1.614794%2.482737%2.194007% 1.908253 0.138131 0.139983 0.128196 0.164561 5.319032 Anti 200120022003 0.088794 0.127851 0.134983 BM Extended Ethics EqualValueEqualValueEqualValue Standard Deviation Mean 1.03851%0.04204%1.61586%1.18470%0.62655%0.69612% Standard Deviation Mean 0.222576% 0.078725% 2.632809%0.618059% 4.640990% 0.164047% Standard Deviation Mean 4.544295%0.651182%2.008882%1.245984% 10.279100% 3.144652% Standard Deviation Mean 1.549045%0.932935%0.711573%0.523687%0.872384%0.577606% Anti 20052006 1.9069780.1027562.694592 0.0869460.065295 0.1338181.9111961.908866 0.072843 2004 BM Extended Ethics
47 Ethic s B M ExtendedAnt i Ethic s B M ExtendedAnt i 2001 1.06610 % 3.217440 % 1.498542 % 0.135362 % 2001 0.15225 % 0.048680 % 0.969614 % 0.042901 % 2002 1.66937 % 1.358672 % 1.771123% 2.548932 % 2002 0.77869% 0.426839 % 0.844489 % 1.614794 % 200 3 3.68687 % 4.627793 % 2.202663%2.482737 % 200 3 2.11042 % 1.563098 % 1.625824 % 2.194007 % 200 4 1.03851 % 0.222576 % 4.544295%1.549045 % 200 4 0.04204 % 0.078725 % 0.651182 % 0.932935 % 200 5 1.61586 % 2.632809 % 2.008882 % 0.711573 % 200 5 1.18470 % 0.618059 % 1.245984 % 0.523687 % 2006 0.62655 % 4.640990 % 10.279100 % 0.872384 % 2006 0.69612 % 0.164047 % 3.144652 % 0.577606 % Equal WeightedValue W eighted Weighted Graph 11: Annualized Return Data for A ll Indices: Equal and Value Weighted 12.00000% 10.00000% 8.00000% 6.00000% 4.00000% 2.00000% 0.00000% 2.00000% 4.00000% 6.00000% 200120022003200420052006Average Yearly Equal Weighted Returns Ethics BM Extended Anti
48 4.00000% 3.00000% 2.00000% 1.00000% 0.00000% 1.00000% 2.00000% 3.00000% 200120022003200420052006Average Yearly Value Weighted Returns Ethics BM Extended Anti Table 6: Equal Weighted Anti-I ndex Returns Regressed on S&P 500 Regression Statistics Multiple R0.512523401 R Square0.262680236 Adjusted R Square0.252147097 Standard Error0.002261036 Observations72 ANOVA dfSSMSFSignificance F Regression10.0001274920.00012749224.938456094.17364E 06 Residual700.000357865.11228E 06 Total710.000485352 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept0.0002285050.0002676540.8537352320.396163841 0.0003053130.000762324 S&P Return0.035784910.0071658084.9938418174.17364E 060.0214931580.050076663
49 y = 0.035x + 0.000 R = 0.262 0.012 0.01 0.008 0.006 0.004 0.002 0 0.002 0.004 0.006 0.008 0.01 0.15 0.1 0.0500.050.1 Y XScatter Diagram Anti Index Returns Linear (Anti Index Returns) Table 7: Equal Weighted Ethics Returns Regressed on S&P 500
50Regression Statistics Multiple R0.433516438 R Square0.187936502 Adjusted R Square0. 176335595 Standard Error0.062589143 Observations72 ANOVA dfSSMSFSignificance F Regression10.0634625040.06346250416.200155710.000142375 Residual700.274218060.003917401 Total710.337680563 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept0.0146630420.007409091.9790611850.051743838 0.0001139210.029440006 S&P Return0.7983923160.1983612124.0249417030.0001423750.4027733581.194011274 Table 8: Equal Weighted Anti Ethics List Regressed on Equal Weighted Ethics List
51Regression Statistics Multiple R0.38588538 R Square0.148907527 Adjusted R Square0.136749063 Standard Error0.002429226 Observations72 ANOVA d f SSMSFSignificance F Regression17.22726E 057.22726E 0512.247231870.000814819 Residual700.000413085.90114E 06 Total710.000485352 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept9.87293E 050.0002954550.3341602010.739257759 0.0004905370.000687996 Ethics List Returns0.0146296450.0041803713.499604530.0008148190.0062921580.022967131 Table 9: Value Weighted Ethics Re turns Regressed On S&P 500 Returns
52Regression Statistics Multiple R0.848215369 R Square0.719469312 Adjusted R Square0. 715461731 Standard Error0.021916752 Observations72 ANOVA dfSSMSFSignificance F Regression10.0862347560.086234756179.52706775.32507E 21 Residual700.0336240830.000480344 Total710.119858839 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept0.0024063150.0025944310.9274927040.35685781 0.0027681130.007580744 S&P Return0.9306768640.06945986713.398771135.32507E 210.7921435291.069210199 Table 10: Value Weighted Anti-Index Regressed on S&P 500
53Regression Statistics Multiple R0.51945862 R Square0.269837258 Adjusted R Square0.259406362 Standard Error0.004114182 Observations72 ANOVA d f SSMSFSignificance F Regression10.0004378720.00043787225.869038452.93042E 06 Residual700.0011848551.69265E 05 Total710.001622727 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept0.0001740840.0004870230.3574461560.721833292 0.0007972520.001145421 S&P Return0.0663179950.0130389095.0861614662.93042E 060.0403127120.092323278 Table 11: Value Weighted Ethics Returns Re gressed on Value Weight ed Anti-Index Returns
54Regression Statistics Multiple R0.521675355 R Square0.272145176 Adjusted R Square0.26174725 Standard Error0.004107675 Observations72 ANOVA d f SSMSFSignificance F Regression10.0004416170.00044161726.173024792.61289E 06 Residual700.0011811091.6873E 05 Total710.001622727 CoefficientsStandard Errort StatP valueLower 95%Upper 95% Intercept6.25642E 050.0004887590.128006130.898511163 0.0009122360.001037364 Ethics List Returns0.0606998990.0118648165.1159578562.61289E 060.037036270.084363529
55Table 12: T-Test Difference Between Means of Equal Weighted Returns of Top 15 Ethical Companies and Bottom 15 Ethical Companies (assumes equal popula tion variances). T-Test Difference Between Means of EW Top 15 List and Bottom 15 List (assumes equal population variances) Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 72Confidence Level95% Sample Mean 0.006895133 Sample Standard Deviation0.046823522 Population 2 Sample Degrees of Freedom142 Sample Size 72 t Value1.976810963 Sample Mean 0.011923193 Interval Half Width0.015758656 Sample Standard Deviation0.048816788 Interval Lower Limit-0.020786715 Population 1 Sample Degrees of Freedom71Interval Upper Limit0.010730596 Population 2 Sample Degrees of Freedom71 Total Degrees of Freedom142 Pooled Variance0.00228776 Difference in Sample Means-0.005028059 t Test Statistic -0.630734194 Two-Tail Test Lower Critical Value-1.976810963 Upper Critical Value1.976810963 p -Value 0.529228162 Do not reject the null hypothesis Intermediate Calculations Data Confidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations Confidence Interval Table 13: T-Test Difference Between Means of Value Weighted Returns of Top 15 Ethical Companies and Bottom 15 Ethical Companies (assumes equal popula tion variances). T-Test Difference Between Means of VW Top 15 List and Bottom 15 List (assumes equal population variances) Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 72Confidence Level95% Sample Mean 0.004785862 Sample Standard Deviation0.044856281 Population 2 Sample Degrees of Freedom142 Sample Size 72 t Value1.976810963 Sample Mean 0.007711268 Interval Half Width0.015506437 Sample Standard Deviation0.049174623 Interval Lower Limit-0.018431843 Population 1 Sample Degrees of Freedom71Interval Upper Limit0.01258103 Population 2 Sample Degrees of Freedom71 Total Degrees of Freedom142 Pooled Variance0.002215115 Difference in Sample Means-0.002925406 t Test Statistic -0.372940329 Two-Tail Test Lower Critical Value-1.976810963 Upper Critical Value1.976810963 p -Value 0.709748958 Do not reject the null hypothesis Confidence Interval Intermediate Calculations Data Confidence Interval Estimate for the Difference Between Two Means Data Intermediate Calculations
56Table 14: Equal Weighted Differences of retu rns of top 15 and bottom 15 companies regressed on differences in Beta SUMMARY OUTPU T Regression Statistics Multiple R0.047039777 R Square0.002212741 Adjusted R Square 0.012041363 Standard Erro r 0.032856426 Observations72 ANOVA dfS S M S FSignificance F Regression10.0001675830.0001675830.1552353420.694779349 Residual700.0755681290.001079545 Total710.075735713 CoefficientsStandard Erro r t StatP valueLower 95%Upper 95%Lower 95.0%Upper 95.0% Intercept 0.0033264510.005800502 0.573476330.568160138 0.0148951870.008242286 0.0148951870.008242286 X Variable 1 0.0087390570.022180394 0.3939991650.694779349 0.0529764580.035498344 0.0529764580.035498344 Table 15: Equal Weighted Differences of retu rns of top 15 and bottom 15 companies regressed on differences in Market Capitalization SUMMARY OUTPU T Regression Statistics Multiple R0.227185793 R Square0.051613384 Adjusted R Square0.038065004 Standard Erro r 0.032032737 Observations72 ANOVA dfSSM S F Significance F Regression10.0039089760.0039089763.809561260.054962327 Residual700.0718267360.001026096 Total710.075735713 CoefficientsStandard Erro r t StatP valueLower 95 % Upper 95 % Lower 95.0 % Upper 95.0 % Intercept0.0253678180.0160242061.5830936620.117908085 0.0065914520.057327089 0.0065914520.057327089 X Variable 1 1.00649E 095.15669E 10 1.951809740.054962327 2.03496E 092.19816E 11 2.03496E 092.19816E 11
57Table 16: Equal Weighted Differences of retu rns of top 15 and bottom 15 companies regressed on differences in book to market Regression Statistics Multiple R0.254043705 R Squar e 0.064538204 Adjusted R Square0.051174464 Standard Error0.031813714 Observations72 ANOVA dfS S M S FSignificance F Regression10.0048878470.0048878474.8293519960.031289726 Residual700.0708478660.001012112 Total710.075735713 CoefficientsStandard Errort StatP valueLower 95%Upper 95%Lower 95.0%Upper 95.0% Intercept0.0024660420.0050681680.4865746110.62807982 0.00764210.012574184 0.00764210.012574184 X Variable 10.0804261910.0365976392.1975786670.0312897260 .0074345020.153417880.0074345020.15341788 Table 17: Equal Weighted Differences of retu rns of top 15 and bottom 15 companies regressed on all 3 of the variables of interest SUMMARY OUTPU T Regression Statistics Multiple R0.297680504 R Square0.088613682 Adjusted R Square0.048405462 Standard Erro r 0.031860101 Observations72 ANOVA dfS S M S FSignificance F Regression30.006711220.0022370732.2038698160.095544422 Residual680.0690244920.001015066 Total710.075735713 CoefficientsStandard Erro r t StatP valueLower 95%Upper 95%Lower 95.0%Upper 95.0% Intercept 0.0266967920.039636311 0.6735438320.502884347 0.1057898180.052396233 0.1057898180.052396233 Beta0.0365115850.0280921441.2997080370.198089868 0.0195454150.092568585 0.0195454150.092568585 Market Cap1.06918E 091.60879E 090.664588460.508561285 2.14111E 094.27947E 09 2.14111E 094.27947E 09 Book to Market0.190275310.1251000871.5209846310.132900359 0.0593580240.439908644 0.0593580240.439908644
58Graph 12 Equal Weighted Top 15 Company Index Growth of $1,000 Graph 13 Equal Weighted Bottom 15 Company Index Growth of $1000
59Graph 14 Value Weighted Top 15 Company Index Growth of $1000 Graph 15 Value Weighted Bottom 15 Company Index Growth of $1000
60Figure 1: FortuneÂ’s 100 Best Companies to Work For Rank Company Rank Company 1 Google 51 Paychex 2 Quicken Loans 52 FactSet Research Systems 3 Wegmans Food Markets 53 Vision Service Plan 4 Edward Jones 54 CH2M HILL 5 Genentech 55 Perkins Coie 6 Cisco Systems 56 Scripps Health 7 Starbucks 57 Ernst & Young 8 Qualcomm 58 Scottrade 9 Goldman Sachs 59 Mayo Clinic 10 Methodist Hospital System 60 Alcon Laboratories 11 Boston Consulting Group 61 Chesapeake Energy 12 Nugget Markets 62 American Express 13 Umpqua Bank 63 King's Daughters Medical Center 14 Network Appliance 64 EOG Resources 15 W. L. Gore & Associates 65 Russell Investments 16 Whole Foods Market 66 Nixon Peabody 17 David Weekley Homes 67 Valero Energy 18 OhioHealth 68 eBay 19 Arnold & Porter 69 General Mills 20 Container Store 70 Mattel 21 Principal Financial Group 71 KPMG 22 American Century Investments 72 Marriott Inte rnational 23 JM Family Enterprises 73 David Evans & Associates 24 American Fidelity Assurance 74 Granite Construction 25 Shared Technologies 75 Southern Ohio Medical Center 26 Stew Leonard's 76 Arkansas Children's Hospital 27 S.C. Johnson & Son 77 PCL Construction Enterprises 28 QuikTrip 78 Navy Federal Credit Union 29 SAS Institute 79 National Instruments 30 Aflac 80 Healthways 31 Alston & Bird 81 Booz Allen Hamilton 32 Rackspace Managed Hosting 82 Nike 33 Station Casinos 83 AstraZeneca 34 Recreational Equipment (REI) 84 Stanley 35 TDIndustries 85 Lehigh Valley Hospital & Health Network 36 Nordstrom 86 Microsoft 37 Johnson Financial Group 87 Yahoo 38 Kimley-Horn & Associates 88 Four Seasons Hotels 39 Robert W. Baird 89 Bright Horizons Family Solutions 40 Adobe Systems 90 PricewaterhouseCoopers 41 Bingham McCutchen 91 Publix Super Markets 42 MITRE 92 Milliken 43 Intuit 93 Erickson Retirement Communities 44 Plante & Moran 94 Baptist Health South Fla.
61 45 Children's Healthcare of Atlanta 95 Deloitte & Touche USA 46 CarMax 96 Herman Miller 47 J. M. Smucker 97 FedEx 48 Devon Energy 98 Sherwin-Williams 49 Griffin Hospital 99 SRA International 50 Camden Property Trust 100 Texas Instruments Figure 2: ForbesÂ’ 75 Most Reput able Companies in the U.S. Rank Company Rank Company 1 Google 39 Aflac 2 Johnson & Johnson 40 Home Depot 3 Kraft Foods 41 NIKE 4 General Mills 42 CVS 5 Walt Disney 43 Microsoft 6 United Parcel Service 44 Anheuser-Busch Cos. 7 3M 45 Kroger 8 Xerox 46 J.C. Penney 9 Colgate-Palmolive 47 Honeywell International 10 Texas Instruments 48 Kohl's 11 Eastman Kodak 49 Enterprise Rent-A-Car 12 General Electric 50 Motorola 13 Sara Lee 51 Safeway 14 FedEx 52 Express Scripts 15 Deere & Co. 53 Tyson Foods 16 Goodyear 54 Lockheed Martin 17 Apple 55 Cisco Systems 18 Hewlett-Packard 56 Alcoa 19 Intel 57 Meijer 20 Publix Super Markets 58 Target 21 Caterpillar 59 Union Pacific 22 Whirlpool 60 Goldman Sachs Group 23 Boeing 61 The Travelers Cos. 24 Costco Wholesale 62 CBS 25 Dell 63 State Farm Insurance 26 Coca-Cola 64 Northrop Grumman 27 Marriott International 65 Hartford Finl Service 28 Berkshire Hathaway 66 Rite Aid 29 Walgreen 67 Circuit City Stores 30 Toys "R" Us 68 Southern Co. 31 Procter & Gamble 69 Merrill Lynch 32 PepsiCo 70 Eli Lilly & Co. 33 Office Depot 71 Morgan Stanley 34 Fidelity Investments 72 MetLife 35 IBM 73 Raytheon 36 Best Buy 74 Chubb
6237 Staples 75 American Express 38 Lowe's Cos. Figure 3: Top 10 Commercial Banks by Assets in 2005 Rank Company 1 Citigroup 2 Bank of America Corp. 3 J.P. Morgan Chase & Co. 4 Wells Fargo 5 Wachovia Corp. 6 U.S. Bancorp 7 Capital One Financial 8 National City Corp. 9 SunTrust Banks 10 Bank of New York Figure 4: Top 10 Commercial Banks by Revenue in 2005 Rank Company 1 Citigroup 2 Bank of America Corp. 3 J.P. Morgan Chase & Co. 4 Wells Fargo 5 Wachovia Corp. 6 U.S. Bancorp 7 Capital One Financial 8 National City Corp. 9 SunTrust Banks 10 Bank of New York