Journal of Applied Business and Economics

North American Business Press Atlanta - Seattle – South Florida - Toronto

Editors Dr. Adam Davidson Dr. William Johnson

Editor-In-Chief Dr. David Smith

NABP EDITORIAL ADVISORY BOARD

Dr. Andy Bertsch - MINOT STATE UNIVERSITY Dr. Jacob Bikker - UTRECHT UNIVERSITY, NETHERLANDS Dr. Bill Bommer - CALIFORNIA STATE UNIVERSITY, FRESNO Dr. Michael Bond - UNIVERSITY OF ARIZONA Dr. Charles Butler - COLORADO STATE UNIVERSITY Dr. Jon Carrick - STETSON UNIVERSITY Dr. Mondher Cherif - REIMS, FRANCE Dr. Daniel Condon - DOMINICAN UNIVERSITY, CHICAGO Dr. Bahram Dadgostar - LAKEHEAD UNIVERSITY, CANADA Dr. Deborah Erdos-Knapp - KENT STATE UNIVERSITY Dr. Bruce Forster - UNIVERSITY OF NEBRASKA, KEARNEY Dr. Nancy Furlow - MARYMOUNT UNIVERSITY Dr. Mark Gershon - TEMPLE UNIVERSITY Dr. Philippe Gregoire - UNIVERSITY OF LAVAL, CANADA Dr. Donald Grunewald - IONA COLLEGE Dr. Samanthala Hettihewa - UNIVERSITY OF BALLARAT, AUSTRALIA Dr. Russell Kashian - UNIVERSITY OF WISCONSIN, WHITEWATER Dr. Jeffrey Kennedy - PALM BEACH ATLANTIC UNIVERSITY Dr. Jerry Knutson - AG EDWARDS Dr. Dean Koutramanis - UNIVERSITY OF TAMPA Dr. Malek Lashgari - UNIVERSITY OF HARTFORD Dr. Priscilla Liang - CALIFORNIA STATE UNIVERSITY, CHANNEL ISLANDS Dr. Tony Matias - MATIAS AND ASSOCIATES Dr. Patti Meglich - UNIVERSITY OF NEBRASKA, OMAHA Dr. Robert Metts - UNIVERSITY OF NEVADA, RENO Dr. Adil Mouhammed - UNIVERSITY OF ILLINOIS, SPRINGFIELD Dr. Roy Pearson - COLLEGE OF WILLIAM AND MARY Dr. Veena Prabhu - CALIFORNIA STATE UNIVERSITY, LOS ANGELES Dr. Sergiy Rakhmayil - RYERSON UNIVERSITY, CANADA Dr. Robert Scherer - CLEVELAND STATE UNIVERSITY Dr. Ira Sohn - MONTCLAIR STATE UNIVERSITY Dr. Reginal Sheppard - UNIVERSITY OF NEW BRUNSWICK, CANADA Dr. Carlos Spaht - LOUISIANA STATE UNIVERSITY, SHREVEPORT Dr. Ken Thorpe - EMORY UNIVERSITY Dr. Robert Tian – SHANTOU UNIVERSITY Dr. Calin Valsan - BISHOP'S UNIVERSITY, CANADA Dr. Anne Walsh - LA SALLE UNIVERSITY Dr. Thomas Verney - SHIPPENSBURG STATE UNIVERSITY Dr. Christopher Wright - UNIVERSITY OF ADELAIDE, AUSTRALIA Volume 14(3) ISSN 1499-691X

Authors have granted copyright consent to allow that copies of their article may be made for personal or internal use. This does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Any consent for republication, other than noted, must be granted through the publisher:

North American Business Press, Inc. Atlanta - Seattle – South Florida - Toronto

©Journal of Applied Business and Economics 2013

For submission, subscription or copyright information, contact the editor at: [email protected]

Subscription Price: US$ 360/yr

Our journals are indexed by UMI-Proquest-ABI Inform, EBSCOHost, GoogleScholar, and listed with Cabell's Directory of Periodicals, Ulrich's Listing of Periodicals, Bowkers Publishing Resources, the Library of Congress, the National Library of Canada. Our journals have been used to support the Academically Qualified (AQ) faculty classification by all recognized business school accrediting bodies.

This Issue

What Did the Hair Dryer Cost?...... 11 Thomas E. Buttross, George Schmelzle

This case demonstrates the various types of “costs” that students will be exposed to in a cost/managerial accounting course. This case, a simple example of a student replacing a hair dryer, can be presented on the first day of class and can introduce students, in an interesting way, to many different cost calculations along with the justification for each calculation. The purpose of the case is to stimulate discussion the first day of class and to help students appreciate the complexity and usefulness of cost accounting systems.

Investigating the Impact of Economic Uncertainties on Attendance of Premier League Soccer in the United Kingdom and Major League Soccer in the United States ...... 16 Ross Peterman, Nichaya Suntornpithug

This research analyzes correlation of attendance of English Premier League and Major League Soccer to their respective nations’ household disposable incomes from 2002 through 2008. By examining professional soccer attendance in countries with a vast discrepancy in soccer popularity throughout a period of economic uncertainty, fan behavior is greater understood.

Evaluation of Mechanisms for Globalizing Information Systems Programs ...... 24 Anu Gokhale

As US businesses continue to grow their international presence, they need professionals who understand the practices in other parts of the world. This is most applicable for Information Technology and Information Systems related professionals who frequently communicate with vendors and clients outside the US. Research establishes that US educational programs in information systems and technology would be strengthened by enhancing student and faculty knowledge of technology deployment models and business practices as well as the culture and customs in other parts of the world. This Delphi study surveyed faculty and students to investigate more effective mechanisms to accomplish this goal.

Virtues of Investment Fund Recommendations: Consumer Reports and Money Magazine ...... 35 C. Edward Chang, Thomas M. Krueger

In February 2007, Consumer Reports (CR) and Money magazine (MM) published investment fund recommendations. CR recommended only mutual funds while about one fifth (13 of 67) of MM selections were exchange-traded funds. Enough time has now passed to allow for a rigorous evaluation of these two selections. The findings of this study consistently laud the performance of both recommendations, because both significantly outperform category averages. The abnormal performance of CR funds appears to arise from minimizing risk, whereas MM funds’ strength lies in reducing costs and making selections that will benefit more from market advances.

What Drives Employee Stock Options Programs? Safeguarding Human Capital and Recruiting Wanted Skills ...... 53 Yu Peng Lin

We believe that the investments in firm-specific human capital are at risk from employee turnovers and that a firm’s productivity is improved when capital and skill are better matched. We extend the attraction and retention justifications associated with employee stock options programs by hypothesizing that such programs are used more often by firms that invest in employee training and new physical capital. Our hypotheses are largely borne out in our panel of 219 U.S. firms between 1990 and 1999. The empirical results support a positive association between the likelihood of employee stock option programs adoption and investments in employee training and physical capital.

Empirical Analysis of Bicultural Border College Students’ Attitudes Toward Money ...... 70 Yeong Nain Chi, Gaurango Banerjee

The objectives of this study were to investigate bicultural border college business students’ attitudes toward money, and to use the results to educate students about the possible impacts of their money attitude dimensions on their financial behavior patterns. A questionnaire survey using five-point Likert- type Money Attitude scale developed by Yamauchi and Templer (1982) was employed. Empirical results based on the K-means cluster analysis identified three groups of respondents. Statistical analyses revealed that there were significant differences between the money attitude dimensions with respect to cluster, gender and student classification.

The New Economic Reality and the Unemployment Rate: Will It Ever Get Below 5% Again? ...... 83 Robert L. Howard, Belinda P. Shipps

The unemployment rate has been at 8% or higher during most months since June 2009, when the last recession ended. In the past, the unemployment rate rose during recessions, continued to rise after the official end of the recession, but then declined substantially. Some changes have occurred in the current economy, however, which may prevent the normal decline in the unemployment rate after our most recent recession. We evaluate these and make a conclusion of their impact on the unemployment rate and society.

Foreign Direct Investment Flows: An Examination of Its Distribution Among Middle- and Low-Income Countries ...... 108 Gabriel Manrique, Val Vlad, Yanping Chong

The potential effects foreign direct investment (FDI) has on growth and employment have led many developing countries to compete for FDI. In a period of rapid globalization the competition for FDI inflows has become intense. We examine factors that affect the distribution of relative shares of FDI flows among low- and middle-income developing countries using panel data. We use a simultaneous equations model to account for the bidirectional determination between the FDI stock and FDI flow. Estimation results demonstrate that countries with higher GDP growth rates, lower tax rates, smaller cost of business start-up, less corruption and higher secondary school enrollment are more successful in attracting FDI.

Risk, Return, and Income Mix at Commercial Banks: Cross-Country Evidence ...... 123 Rifat Gorener, Sungho Choi

The paper examines whether and how increased reliance on non-interest income affects the financial performance of banks, as measured by stock market return data for publicly traded commercial banking companies in 42 countries. In general, we find that non-interest income is associated with riskier stock returns at commercial banking companies, due primarily to increased market, or systematic, risk. This finding is new to the literature, and suggests that fee-based banking activities increase banks’ exposure to the business cycle. In contrast, we find almost no evidence linking non-interest income to changes in the total risk, interest rate risk, or idiosyncratic risk. Our results also suggest that the stock markets efficiently price the increased risk associated with the non-interest income market. That is, after controlling for cross-sectional differences in risk, market returns do not fluctuate with the mix of bank income. This result offers a potential explanation for the initial conventional wisdom among industry participants that expansion into non-interest activities would result in an improved risk–return trade-off at commercial banks. Finally, we find that cross-country differences in regulatory practices, economic conditions, and social institutions influence our main results in important ways, but on average our risk– return results appear to be robust across countries.

GUIDELINES FOR SUBMISSION

Journal of Applied Business and Economics (JABE)

Domain Statement

The Journal of Applied Business and Economics is dedicated to the advancement and dissemination of business and economic knowledge by publishing, through a blind, refereed process, ongoing results of research in accordance with international scientific or scholarly standards. Articles are written by business leaders, policy analysts and active researchers for an audience of specialists, practitioners and students. Articles of regional interest are welcome, especially those dealing with lessons that may be applied in other regions around the world. This would include, but not limited to areas of marketing, management, finance, accounting, management information systems, human resource management, organizational theory and behavior, operations management, economics and econometrics, or any of these disciplines in an international context. Focus of the articles should be on applications and implications of business, management and economics. Theoretical articles are welcome as long as their focus is in keeping with JABE’s applied nature.

Objectives

. Generate an exchange of ideas between scholars, practitioners and industry specialists

. Enhance the development of the Business and Economic disciplines

. Acknowledge and disseminate achievement in regional business and economic development thinking

. Provide an additional outlet for scholars and experts to contribute their ongoing work in the area of applied cross-functional business and economic topics.

Submission Format

Articles should be submitted following the American Psychological Association format. Articles should not be more than 30 double-spaced, typed pages in length including all figures, graphs, references, and appendices. Submit two hard copies of manuscript along with a disk typed in MS-Word.

Make main sections and subsections easily identifiable by inserting appropriate headings and sub-headings. Type all first-level headings flush with the left margin, bold and capitalized. Second-level headings are also typed flush with the left margin but should only be bold. Third- level headings, if any, should also be flush with the left margin and italicized.

Include a title page with manuscript which includes the full names, affiliations, address, phone, fax, and e-mail addresses of all authors and identifies one person as the Primary Contact. Put the submission date on the bottom of the title page. On a separate sheet, include the title and an abstract of 200 words or less. Do not include authors’ names on this sheet. A final page, “About the authors,” should include a brief biographical sketch of 100 words or less on each author. Include current place of employment and degrees held.

References must be written in APA style. It is the responsibility of the author(s) to ensure that the paper is thoroughly and accurately reviewed for spelling, grammar and referencing.

Review Procedure

Authors will receive an acknowledgement by e-mail including a reference number shortly after receipt of the manuscript. All manuscripts within the general domain of the journal will be sent for at least two reviews, using a double blind format, from members of our Editorial Board or their designated reviewers. In the majority of cases, authors will be notified within 60 days of the result of the review. If reviewers recommend changes, authors will receive a copy of the reviews and a timetable for submitting revisions. Papers and disks will not be returned to authors.

Accepted Manuscripts

When a manuscript is accepted for publication, author(s) must provide format-ready copy of the manuscripts including all graphs, charts, and tables. Specific formatting instructions will be provided to accepted authors along with copyright information. Each author will receive two copies of the issue in which his or her article is published without charge. All articles printed by JABE are copyrighted by the Journal. Permission requests for reprints should be addressed to the Editor. Questions and submissions should be addressed to:

North American Business Press 301 Clematis Street, #3000 West Palm Beach, FL USA 33401 [email protected] 866-624-2458

What Did the Hair Dryer Cost?

Thomas E. Buttross The Pennsylvania State University at Harrisburg

George Schmelzle Missouri State University

CASE

The Problem This morning, the hair dryer of one of the students in this class broke. The student, Jane, discussed the problem with her older brother Dick.

Dick: Hi Jane. What happened to your hair dryer? Jane: It broke this morning in the shower. I was multitasking. Dick: I do not think you should use it in the shower. Wal-Mart has some on sale for $10.00. Jane: $10.60 with sales tax. And I will have to use my credit card. Dick: You will have to pay 19% interest; that is $2.00 for one year. You should sign up for a personal finance course. Jane: Well, I can figure out that I will have some extra driving. According to Google Maps, Wal-Mart is 10 miles from here and the school is 18 miles. But since Wal-Mart is sort of on the way it only adds two miles to the round trip. Dick: That is also about 15 minutes that you cannot talk to your friends. Are you going to drive or will you have our neighbor Mr. Spot drive you? Mr. Spot is cheap and he will charge you $2.00. Jane: I prefer to drive. Dick: You sure have gotten a lot of use out of the Ford Escort I sold you four years ago. Jane: Yes, that was a good $6,500 investment. In another year, I will have gotten five years use and put 100,000 miles on it, and it will still have a $500 resale value. Since I am graduating then, I can use that $500 as a down payment for a new car. With gas at $3.00 per gallon, I’ll need to make sure my new car gets 30 miles to the gallon, like this one. Dick: What are you doing after school?

Journal of Applied Business and Economics vol. 14(3) 2013 11 Jane: I am going to work at Joe’s Bistro. I now get $8.00 per hour and I am up to 20 hours per week. This little problem with the hair dryer will not affect my work hours. Dick: The economy is booming. By the way, I will give you a dollar for the old hairdryer. I might be able to fix it up and use it to dry the new puppy when it rains. Jane: Sure. Well, I better get going or I will be late for school.

Assignment Question What did the hair dryer cost? Present at least three different possible costs that you feel could be defended and present them starting with the most desirable and ending with the least desirable. If you wish to include some additional costs that are not included above, simply add a descriptive statement of the cost you would add; you do not need to assume any monetary amounts for costs where you are not given sufficient information to compute one.

Required You will work in small groups and each group will prepare one group solution. For purposes of this exercise, all amounts are material (i.e., do not justify excluding an amount because it is immaterial). While many of the amounts would clearly be immaterial in the “real world” the amounts were kept small so that teams could concentrate on the concepts and not number crunching. For each of your selected answers, be prepared to explain the rational for each of the above costs you would include in the cost of the hair dryer and for each you would exclude.

TEACHING NOTES

Purpose The purpose of the case is to show that, absent generally accepted accounting principles (GAAP), cost is a very nebulous concept. If multiple costs can be obtained in this simple retail purchase of a single product, the complexity of determining cost in a multiple-product manufacturing environment becomes apparent. A second purpose is to set the tone for the course on the first day of class. Students are being told that this is not going to be another memorize GAAP course. While many problems in the course will be structured to have single right answers, a certain level of comfort with ambiguity will be necessary for success in the course. And the case gets students actively involved in their own learning. A third purpose is to introduce the concept of different costs for different purposes. One cost computation may be the best for GAAP, while a different cost computation is best for making the purchase/do not purchase decision. Two secondary purposes of the case are to serve as an icebreaker on the first day of class and to get the students to think about teamwork.

Case Learning Objectives 1. To make students realize that there is no single “right” answer when determining cost. The cost of an asset, product, service or other cost object will be dependent on the objective of the user of the cost information (e.g., satisfying generally accepted accounting principles (GAAP) versus decision making), as well as the philosophy of the individuals calculating the cost (e.g., preference for full versus variable cost). 2. Related to number 1, to get students to realize that costing products (services, etc.) is complex, and multiple computations are often needed to properly capture useful information. 3. Related to numbers 1 and 2, to allow students to discover several of the topics and techniques to be used in the course (such as cost allocation, full costing, and variable/incremental costing), making these topics more intuitively understandable prior to their course coverage. 4. To get students to articulate their justification for determining cost in a particular way (to help them start the habit of having a justification for other decisions during the course). 5. To get students enthused about cost or managerial accounting by providing a case that they can easily relate to their own lives.

12 Journal of Applied Business and Economics vol. 14(3) 2013 6. To provide students with an icebreaker exercise, so that, by working in teams, they will become more familiar with one another. 7. To get students actively involved in their own learning, both by working the case and by the subsequent class discussion of it.

Possible Courses Where This Case Should Be Used This case can be used on the first day of class in an introductory managerial accounting course (either undergraduate or MBA) and as an introductory case in cost accounting to bring these various cost topics back to the surface.

Analysis of Question/Teaching Plan Possible responses, some of which will not show up, are presented below. Some comments that may be used by students or the instructor are included in the box with the cost number. The instructor needs to avoid any temptation to indicate to students that any one of these is the true or most correct cost. As indicated in the comments below, the instructor can indicate circumstances where a particular cost factor or computation may be more useful than some alternatives.

Cost 1: Provided by the instructor as the minimum cost. Expense sales tax as a consumption/use tax, instead of including it in the purchase cost. No attempt is made to justify this as a good cost, nor is it labeled a bad cost. It is simply presented as the simplest alternative that might be considered appropriate. Purchase price $10.00

Cost 2: The sales receipt amount. The person on the street would usually give this answer when asked what something cost. The instructor might indicate that this is a good GAAP cost computation, but not inclusive enough to be the best computation to use for the purchase/do not purchase decision. Purchase price $10.00 Sales tax 0.60 Total $10.60

Cost 3: The instructor can discuss interest as a financing cost that is only capitalized by GAAP for self-constructed long-lived assets and certain discrete inventory projects. However, interest can and should be considered a legitimate cost for internal decision making, such as deciding whether to purchase an asset. Purchase price $10.00 Sales tax 0.60 Interest 2.00 Total $12.60

Cost 4: For this option, he instructor can discuss incremental or marginal costing. The instructor can point out that this makes some sense if there is idle capacity (sunk costs). Purchase price $10.00 Sales tax 0.60 Interest (may not appear; OK if it does or does not) 2.00 Gas: 2 extra miles at $0.10 per mile. Students could choose to use 20 miles ($2.00) for a round trip to Wal-Mart or 38 miles ($3.80) for total round trip to school and back. If the student chooses the 38 miles, they could also reduce

Journal of Applied Business and Economics vol. 14(3) 2013 13 the gas cost by the $2.00 a neighbor would have charged to take the student to school, making it $1.80 net. These nuances get students to think, as long as they are simply presented as defensible alternatives. 0.20 Total $12.80

Cost 5: For this cost, the instructor can discuss full costing. The instructor can indicate that, in the long run, a business must recover all of its costs before a profit can be generated. This cost also makes sense when dealing with a capacity constraint, i.e., where additional capacity will be needed. Purchase price $10.00 Sales tax 0.60 Interest (may not appear; OK if it does or does not) 2.00 Gas: alternative amounts discussed in Cost 4 0.20 Depreciation: Some students may argue that the car cost will be there anyway, allowing the introduction of the sunk cost concept. Student who depreciate generally use units of output: ($6,500-$500)/100,000 miles = $0.06 per mile. Use the same mileage as used for gas. So, this could be 20 or 38 miles, as shown in Cost 4 under gas. Straight-line depreciation can also be used, allowing a discussion of whether to charge for one full day or a partial day, such as fifteen minutes. When discussing these possibilities, do not get bogged down in the minute computations. 0.12 Total $12.92

Cost 6: This cost provides an option to discuss opportunity cost. Purchase price $10.00 Sales tax 0.60 Interest (may not appear; OK if it does or does not) 2.00 Gas: alternative amounts discussed in Cost 4 0.20 Depreciation: alternatives discussed in Cost 5 0.12 Time Lost: This allows discussion of the concept of opportunity cost. If students fail to discuss time lost, the instructor can raise the issue near the end of the case. The computation is 15 minutes at $8.00 per hour. While the student lost hallway talk time, the job provides one way to monetize the value of the time. Can also mention the cost of the owner’s capital as another sometimes overlooked opportunity cost. 2.00 Total $14.92

The case requirements say, “If you wish to include some additional costs that are not included above, simply add a descriptive statement of the cost you would add; you do not need to assume any monetary amounts for costs where you are not given sufficient information to compute one.”

Additional Possible Considerations for Case The case (and the above cost computations) omits some relevant maintenance costs that some students may wish to add: oil change, tune up, new tires. If students do not bring this up, immediately before ending the case, the instructor can ask the class if there are any relevant costs omitted from the case. Occasionally, some students will suggest using the Internal Revenue Service mileage rate (instead of gas and depreciation), in order to capture all of the

14 Journal of Applied Business and Economics vol. 14(3) 2013 relevant automobile costs, in which case the above point is moot. Some students may wish to use the $1.00 sale of the old hair dryer as an offset (1.00) to the cost of the new hair dryer, using a net cost approach. Here, the instructor can discuss the idea of separate economic events, but point out that a $1.00 “trade-in” of the old hair dryer would trigger the like-kind exchange rules used by GAAP and taxes, where an offset is considered relevant.

Strategy for Using the Case in Class The case should not be rushed; however, as shown below, it could be completed more quickly if the faculty member presenting the case allots less time for students to create a solution and less time to reviewing solutions.

Activity Minutes Hand out the case and read it to the class (can save a couple of minutes by handing out case before class as students arrive) 5 Have students turn chairs around to form ad hoc teams of 3-to-5; first row with second row, third row with fourth row, etc. 2 Team time to create a solution (can reduce to 10, if necessary, but 15 allows students to work without feeling rushed and to engage in small talk a few minutes after creating their solutions) 15 Class time to review solutions (can reduce to 20, if necessary) 25 Debrief 2 Total (can reduce 10 minutes fairly easily, as shown above) 49

While students are asked to rank their solutions in the case requirements, it is not necessary to go over solutions in rank order. Instead, when presenting the case, instructors are advised to start with one simple solution by asking one team chosen at random for its simplest solution. Then each remaining team is asked for its simplest solution that is not already on the board, cycling thorough the teams until there are no new responses.

Journal of Applied Business and Economics vol. 14(3) 2013 15

Investigating the Impact of Economic Uncertainties on Attendance of Premier League Soccer in the United Kingdom and Major League Soccer in the United States

Ross Peterman Indiana University – Purdue University Fort Wayne

Nichaya Suntornpithug Indiana University – Purdue University Fort Wayne

INTRODUCTION

Soccer is one of the most popular sports with a great spectator base throughout the world (Sagatomo and Greenwell, 2011). In the United States, however, despite the increasing popularity of soccer as a sport to play among youth, the spectator base is still far behind that of other sports such as American football, basketball, and baseball (Carlin, 2010). As a result, Major League Soccer (MLS) is still considered as a second-tier sport league (Collins, 2006). Nevertheless, MLS has been trying to expand their spectator base through various marketing means such as sponsoring many grassroots tournaments nation-wide to spur interest to the league among soccer participants (Warfield, 2004). However, an economic condition that the US is facing may confound professional soccer attendance. The recent phenomenon in the US and many countries has been littered with economic instability. This instability has had affects on local, regional, national, and global levels. The resultant impacts from the economy have been analyzed on many facets and in many different lights. But, the economic fluctuation’s impact on professional soccer attendance has yet to be studied. This research aims to provide the previously missing insight on correlation between professional soccer attendance and economic uncertainty. Since soccer is the world's most popular sport, it is beneficial to examine from an international perspective. This paper aims to analyze the impact of the economic uncertainties in the United States and the United Kingdom on their relative professional soccer match attendance over the years. While the professional soccer league representing the US is Major League Soccer (MLS), the professional soccer league representing the UK is the English Premier League (EPL). These two soccer leagues are distinctly different in terms of popularity and history. It is valuable to compare and contrast the impact the economy has had on soccer match attendance in a country where it is the most popular sport, the United Kingdom, with a country where it is not one of the most popular sports, the United States, in order to garner more

16 Journal of Applied Business and Economics vol. 14(3) 2013 encompassing and complex conclusions. By analyzing the impact that the economy has on professional soccer attendance, a greater acumen of consumer behavior will be forthcoming. The implications will impact soccer team executives, venue management, and in-venue marketing sponsors. Since the current economic instability seems rather persistent, and periods of uncertainty are predictably recurrent, developing a deeper understanding of consumer behavioral patterns in times of economic inconsistency will be beneficial.

LITERATURE REVIEW

The soccer popularity in the United States (US) and soccer popularity in the United Kingdom (UK) has been apparent. As for the UK, “it's the English that modernized soccer and started the international phenomenon that it's become today” (Soccer Fans Info, 2011). Not only that, but “they were the first nation to have professional soccer clubs and it's the English that created the Laws of the Game as early as 1863, the "constitution" that is still the book of soccer rules nowadays” (Soccer Fans Info, 2011). As a result, the UK’s soccer roots run deep and their ties are heartfelt. As for the US, “the US were generally regarded as a country where soccer was scarcely popular, with basketball, baseball and American football coming on top of the people's favorite team sports” (Soccer Fans Info, 2011). Even though “the past two decades sparked a major interest in soccer in the US,” one can conclude that the popularity divide of soccer between the US and the UK is very sizable (Soccer Fans Info, 2011). Researcher holds different opinions on the impact of economic troubles in the UK and in the US on soccer attendance in their respective leagues, the US’s MLS and the UK's EPL. For examples, a Reuters/Zogby survey found that almost 15 percent of those polled said they are attending fewer sporting events this year, and most of those people cited the weak economy as the reason,” (Klayman 2008). Nevertheless, evidences showed that most sports were still able to set attendance record (Klayman, 2008). According to Zimbalist, "the evidence from past recession is indeed that sports is one of the last things people cut back on…they need their distractions and they need their obsessions" (cited in Klayman, 2008). Klayman’s article quoted a sports patron who said, “the tickets cost a lot, but that's what you have savings for, for special events" (Klayman 2008). As for MLS, Klayman (2008) found that “smaller leagues,” similar to the MLS, “may do even better than the big ones” (Klayman, 2008). As for the EPL, it is one of the top-tier sports in the UK with a long history of fan hood and nostalgia, similar to the US’s Major League Baseball. Klayman’s article cites a professional sports executive who said, "Baseball was there during the Depression, during World War II, as an element of the country's recovery from 9/11, and will continue to be available at affordable prices,” a sentiment that may be shared by the English people about their storied history and intimate bond with their beloved EPL (Klayman, 2008). Peter (2008) asserted that in relation to Major League Baseball, “gas prices are up. Food prices are up. So, oddly enough, is attendance at Major League Baseball games…” (Peter, 2008). Zimbalist explained that “people seek the comfort of their sports addiction when there’s bad economic times,” and that “there’s something very deeply rooted about our sports attachments. People tend to give that up last” (Peter, 2008). Dalakas believes that tickets to sporting events have real value, although not seen as functional value (cited in Peter, 2008). He stated that “a product can offer functional benefits and psychological benefits…sometimes perceived value may not be real value. But sometimes perceived value, the fact that it’s about emotions, doesn’t make it any less valuable” (cited in Peter, 2008). Other evidences, however, suggest a different perspective. Wieberg (2009), for example, found that “every one of college basketball's 12 top-drawing conferences saw attendance fall during the just- completed men's regular season, a likely reflection of the nation's economic slide,” (Wieberg, 2009). Basketball is one of the top-tier sports in the US. The top-tier popularity is more comparable to soccer in the UK. Humphreys (2010) believes the financial crisis has adversely affected sport attendance in North America. Humphreys found that “attendance and franchise values declined slightly” (Humphreys, 2010). Humphreys later stated that “increasing reliance on revenues from businesses in the form of premium

Journal of Applied Business and Economics vol. 14(3) 2013 17 seats, luxury suites, and sponsorship may lead to future problems if the downturn continues for a prolonged period,” which has been the case (Humphreys, 2010). One can see that there are vastly contradicting perspectives on how economic conditions affect professional soccer attendance in the US and UK. That being the case, this research was conducted to better understand the relationship between the two. The goal is to better comprehend how economic condition, on a personal income level, translates into professional soccer match attendance figures.

HYPOTHESES

Despite various views in the literature on the impact of the economic downturn on sport attendance in the UK and in the UK, the popularity difference in the two countries is clear. Drawn from existing literature, it can be hypothesized that the economic situation in the UK and the US over the years are likely to minimally impact soccer attendance in the EPL but slightly more impact MLS attendance because of the soccer popularity disparity between the two nations, the difference in size of the consumer base, and the difference in the availability of substitutes. It is evident that soccer is vastly popular in the UK with a storied history and an avid fan base. In contrast, soccer is not one of the most popular sports in the US, the history is bleak, and the fan base is limited. The popularity of soccer is incomparable between the two nations. As a result of the soccer popularity disparity between the two countries, the fan behavior amidst an economic recession is likely to differ. The UK soccer fans are more likely to be more loyal and budget their soccer passion even if it necessitates other financial sacrifices. Fans in the US with lesser passion for soccer are less likely to make financial sacrifices to budget soccer match attendance. Therefore, it is proposed that sport popularity is likely to have a positive relationship with sport attendance elasticity. To expound on that theory, greater sport popularity is likely to lead to a lesser attendance figure elasticity and a lesser popularity would result in greater attendance figure elasticity. The next issue is that of the difference in size of consumer base for soccer match attendance. In the UK, soccer is the most popular sport, the passion runs deep, the history is storied, and the fan base is colossal. Resultantly, if some fans cannot afford to attend matches because of the suffering economy, there are a plethora of replacement candidates. Contrastingly in the US, soccer is not the most popular sport, the history is lacking, the passion is spotty, and the soccer fan base is limited. Consequently, if some fans cannot afford to attend matches because of economic factors, there are less available soccer fans for replacement. That being said, the UK is greater protected to have more inelastic soccer attendance and the US is likely to have more elastic soccer attendance. Therefore, a weak economy should impact MLS attendance more so than it will impact EPL attendance. Finally, the discrepancy between the availability of substitutes for consumers in the UK and the US is worth noting. In the US, soccer is listed as the eleventh most popular sport in the nation (Most Popular Sports, 2010). On the contrary, in the UK, soccer is distinctively recognized as the nation’s most popular sport (Most Popular Sports, 2010). Excluding all other means of entertainment and focusing on sport substitutes, there is a tremendous differentiation between availability of sport substitutes for sport consumers in the two countries. So, if sports fans are buckling down on their discretionary spending due to economic instability and decide to reduce their sport attendance frequency, soccer attendance is more likely to be reduced in the US. Resultantly from the differing availability of substitutes, the EPL will have maintained more stable and inelastic attendance despite the English economic downturn, whereas MLS will have seen more attendance fluctuation because of the American economic downturn. It is therefore hypothesized that the economic downturn in the UK and the US is likely to minimally affect EPL attendance in the UK, while moderately affect MLS attendance in the US. The US’s MLS attendance is likely to have more elastic in regards to the economic state, whereas the UK’s EPL attendance is likely to be more inelastic to the economic state due to the differences between soccer in the respective countries. The economic situation in the UK and in the US is likely to have minimal impacts on soccer attendance in the EPL in the UK but slightly more impacts on MLS attendance in the US

18 Journal of Applied Business and Economics vol. 14(3) 2013 because of the soccer popularity disparity between the two nations, the difference consumer base size, and the difference in the availability of substitutes.

METHODOLOGY

The two statistical measures used for the research were attendance and economic conditions. Furthermore, attendance was represented by annual average attendance of the EPL and of MLS for seasons 2001-2002 through 2007-2008. Since there was a slight discrepancy in the season timetable for two leagues, seven seasons that end in years 2002 through 2008 were chosen. As for the economic state, the metric used was Disposable Personal Income, also known as Household Disposable Income (HDI). The annual average attendance data for the EPL was accumulated through the official website of league, Official Site of the Premier League. This website contained the average annual attendance figures for the league by team, whereby the mean team attendance was then averaged for a league-wide average by year and was used as the attendance metric. It was best to use the official website of the EPL as the attendance data source since it is the original source of the desired data. It was best to employ the average attendance for the league as the attendance metric for a more representative and comprehensive sample. It was best to separate by year since the economic metric was annual. The average annual attendance for MLS was compiled from the attendance statistics of the official website of the league, Full Season Stats. Team attendance means were averaged into a league-wide average for each year and were used as the attendance metric. It was best to utilize the official MLS website as the source for attendance data since it is the original data source. It was best to assign the league-wide average attendance as the attendance in order to employ a representatively comprehensive sample. It was best to separate by year since the economic metric was annual. The economic metric used was Household Disposable Income (HDI), equating to “the maximum amount that households can afford to spend on consumption goods or services without having to reduce their financial or non-financial assets or to increase their liabilities” (OECD Factbook, 2010). HDI is the most suitable metric for economic condition and its impact on soccer attendance since it shows consumers discretionary spending budgets, which is where the money for soccer attendance would be derived. Resultantly, changes to HDI directly impact consumer ability to attend soccer matches, so once correlated with attendance, was telling of the elasticity of fan hood. The Google public data exporter and the OECD Factbook 2010 were utilized to gather HDI data. These two sources were reliable and well recognized. With the attendance and HDI data, correlations between the two variables were calculated. By doing so, the relationship between attendance in the EPL and the UK economy in terms of HDI, as well as MLS attendance and the US economy in terms of HDI were tested. The correlations for both were expected to be low, but slightly higher for the MLS than the EPL since consumer demand was estimated to be more elastic.

DATA ANALYSIS AND CONCLUSION

The seven season (2002-2008) attendance averages with the year’s HDI of the respective countries were shown in Figure 1.

Journal of Applied Business and Economics vol. 14(3) 2013 19 FIGURE 1 HOUSEHOLD DISPOSABLE INCOME IN THE UK AND US

Table 1 shows the correlation results on the relationship between EPL attendance and UK’s HDI for the years 2002-2008. Assigned were the variables “X” for HDI and “Y” for average attendance.

TABLE 1 THE CORRELATIONS BETWEEN THE UK’S EPL ATTENDANCE AND HDI

Year HDI (X) Average 2002 1.750 34381.400 2003 3.110 35465.950 2004 0.440 35019.800 2005 2.180 33891.900 2006 0.220 33875.350 2007 -0.080 34379.150 2008 1.750 35990.500 Sum 9.370 243004.050 Mea 1.339 34714.864 St 1.173 805.225 N 7.000 R 0.38237 r2 0.14621

Table 2 provides the correlation results on the relationship between MLS attendance and US HDI for the years 2002-2008. Assigned were the variables “X” for HDI and “Y” for average attendance.

20 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 2 THE CORRELATION RESULTS BETWEEN THE US’S MLS ATTENDANCE AND HDI

Year HDI Average (X) Attendance (Y) 2002 3.59 15821.60 2003 2.92 14898.30 2004 3.04 15558.70 2005 1.38 15108.00 2006 3.93 15504.30 2007 1.87 16770.30 2008 0.75 13755.70 Sum 17.480 Mean 2.497 15345.27 St 1.184 922.97 N 7.000 R 0.43058 r2 0.18540

For the UK - EPL analysis, the above correlation coefficient value of 0.382 showed that the UK’S HDI was a weak predictor of EPL attendance, but that there was a positive correlation between the variables. The above coefficient of determination revealed that 14.62% of EPL attendance variability could be explained by the changes in the UK’s HDI. For the US – MLS analysis, the above correlation coefficient value of 0.431 showed that the US’ HDI was a weak predictor of MLS attendance, but that the two variables were correlated. The above coefficient of determination explained that 18.54% of MLS attendance variability could be explained by the changes in the US’ HDI. One can now see that for the years 2002-2008 the correlation between the US’s HDI and MLS attendance was slighter stronger than the correlation between the UK HDI and EPL attendance, but the difference was all but insignificant. This correlation discrepancy revealed a slight differentiation in attendance elasticity between the two societies. In the UK, people are slightly less apt to sacrifice soccer match attendance, even as their spending money diminishes. In the US, people were slightly more willing to scrap soccer match attendance when their disposable income dropped. In the UK, attendance was slightly less elastic, partly, due to the nation’s deep soccer roots, vast fan base, and unparalleled popularity to other sports. In the UK, there was a weaker availability of substitutes and a lesser bargaining power of consumers. In the US, attendance was slightly more elastic, partly, due to the nation’s weaker soccer following and lesser popularity in comparison to other sports. In the US, there was a greater availability of substitutes and a stronger bargaining power of consumers. Overall, there was a weak correlation between HDI and professional soccer match attendance. The country-specific differences were minimal and nearly insignificant. Despite a sporadic HDI for the UK and US over the years 2002-2008, attendance for the EPL and MLS both stayed relatively constant. Therefore, it can be concluded that patrons in the US and UK did not relate the allotment of spending money with the ability to afford and attend professional soccer matches. It is possible that individuals saw their affiliation with their soccer team as a part of their social identity and as a result found it hard to depart from their attendance regardless of economic condition. Henri Tajfel’s research found that “in a situation devoid of the usual trappings,” people “still act in terms of their ingroup membership and of an intergroup categorization” (Tajfel, 1971). This could be the explanation of consumer behavior resulting in rather inelastic attendance figures in times of economic fluctuation.

Journal of Applied Business and Economics vol. 14(3) 2013 21 LIMITATIONS

This research was limited to the impact each country’s HDI has had on its respective professional soccer league attendance. For future research, it would be beneficial to analyze the impact of multiple factors on attendance, and to establish a hierarchy in terms of variable correlation to attendance. Other economic measures could also be utilized to measure correlation with attendance. It could also be very insightful to compare attendance of a plethora of professional sports within multiple countries to see if there was any consumer hopping amidst economic downturns or whether leagues are relatively constant. If there were consumer hopping, it would be interesting to research the root cause. It would also be beneficial to analyze the correlation figures in comparison to each other. The findings would be more helpful if the number of seasons analyzed was expanded to cover a larger time period, but this research was limited by the unavailability of data.

IMPLICATIONS

There are several implications from this research. This research showed that HDI was not strongly correlated to professional soccer match attendance in the UK or the US. In both nations, soccer team owners could rather fearlessly maintain investment in their teams, for example, by way of new player acquisition or stadium renovation, knowing that attendance will be rather constant regardless of the economic state, by way of maintained ticket revenue, in-venue sales, and in-venue advertising. In- venue soccer advertisers in both nations could comfortably maintain investments or secure long-term deals knowing that attendance will be considerably inelastic, thereby maintaining desired reach. Due to more inelastic attendance, team management could do less short-term promotions in terms of money savings and could utilize promotions such as meeting the players or providing signature opportunity instead of ticket price discounts, since ticket sales are predictably sufficient. Marketers can also see that since rather inelastic attendance is so far borderless, they can confidently market internationally. This research disclosed that although soccer was less popular in the US than it was in the UK, professional soccer match attendance was impacted similarly by HDI. It is interesting to observe that HDI had little impact on professional soccer attendance, regardless of the sport’s national popularity. Socially, it is interesting to see that although the two nations have a vast disparity in popularity of the sport, patrons from both found a way to attend sporting events despite erratic economies. The implications in the US could lead to continued MLS expansion into new markets amidst an economic recession. This research could help MLS executives pitch expansion to potential owners and markets despite economic uncertainty by pointing out consistent attendance figures. Contrastingly, this research could help potential owners and new markets sell MLS executives on further expansion regardless of national economic state by emphasizing the steadfast MLS attendance.

REFERENCES

Carlin, J. (2010), “The Global Game,” Time, 175, 60-71.

Collins, S. (2006), “National sports and other myths: The failure of the US soccer,” Soccer and Society, 7(2-3), 353-363.

"Full Season Stats," (accessed October 28, 2011), [available at http://www.mlssoccer.com/content/ 2000-full-season-stats].

Humphreys, B. (2010), “The Impact of the Global Financial Crisis on Sport in North America," Optimal Strategies in Sports Economics and Management, 2010, (accessed October 14, 2011), [available at http://www.springerlink.com/content/ mu7642664u501645/].

22 Journal of Applied Business and Economics vol. 14(3) 2013

"International Soccer," Soccer Fans Info, (accessed November 1, 2011), [available at http://www.soccer-fans-info.com/international-soccer.html].

Klayman, B. (2008), "Sports Attendance Up, Hot Dog Spending down," Business & Financial News, Breaking US & International News, (accessed September 14, 2011), [available at http://www. reuters.com/article/2008/06/30/us-economy-usa-sports-idUSN2439252820080630].

"Most Popular Sports by Country (2011), Most Popular Sports, (accessed June 14, 2011), [available at http://www.mostpopularsports.net/by-country].

"OECD Factbook 2010 - Google Public Data Explorer," (accessed June 7, 2011), [available at http://www.google.com/publicdata/explore?ds=ltjib1m1uf3pf_].

"Official Site of the Premier League,” Official Site of the Premier League, (accessed June 7, 2011), [available at http://www.premierleague.com/ page/Statistics].

Peter, J. (2008), "Attendance Soars Even as Economy Sags," (accessed April 21, 2011), [available at http://sports.yahoo.com/mlb/news? slug=jo-mlbandrecession043008].

Tajfel, H., M.G. Billig, R.P. Bundy, and C. Flament (1971), "Social categorization and intergroup behaviour", European Journal of Social Psychology, 1(2), 149-78.

Tokuyama, S. and C.T. Greenwell (2011), “Examining Similarities and Differences in Consumer Motivation for Playing and Watching Soccer,” Sport Marketing Quarterly, 20 (3), 148-156.

Wieberg, S. (2009), "Economy Hits the Hardwood: Attendance down at NCAA Games." (accessed October 25, 2011), [available at http://www.usatoday.com/sports/college/mensbasketball/2009-03-12- attendances-down_N.htm].

Journal of Applied Business and Economics vol. 14(3) 2013 23

Evaluation of Mechanisms for Globalizing Information Systems Programs

Anu Gokhale Illinois State University

INTRODUCTION

Different IS programs fulfill different missions and serve different student populations (Avgerou, 2008; Ceccucci et al., 2008). There are significant differences between the baccalaureate IS programs that are currently offered; therefore, it is difficult to speak of a generic IS baccalaureate degree (Gorgone, 2006; Reichgelt et al., 2004). For the purpose of this article, all information technology (IT) related programs housed in the College of Business, including information systems (IS) programs will be referred to by a common title of Information Systems (IS). A recent study from international accreditation body Association to Advance Collegiate Schools of Business (AACSB) highlighted the immense challenges before business schools in today's increasingly global and competitive business world asserting that business and society require graduates with global competencies (Stephens et al., 2001; Zammuto, 2008). Bloomberg Businessweek ranking of undergraduate business schools endorses this assertion; while the methodology used to rank the top undergraduate programs was based on measures of student satisfaction and academic quality, international exposure and experience were prioritized by all the top-ranked schools.

REVIEW OF LITERATURE

IS educational programs in the US grew rapidly in the last decade, to fulfill employment needs of businesses, both big and small that embraced the Y2K update followed by a burst in online applications (King, 2005). The offshore model came of age due to the urgency and necessity of Y2K, when numerous offshore IT companies were hired to assist in the massive enterprise of updating critical system code (Weber, 2004). In a Delphi study on offshore outsourcing, the author observed that the practice of exporting IT work has grown manifold, both in terms of labor volume as well as diversity, depth, and complexity of projects undertaken, and management of these outsourced IT projects has become very critical to project success (Gokhale, 2007; Wolk, 2008).

24 Journal of Applied Business and Economics vol. 14(3) 2013 In the last decade, research studies have identified a link between student employability skills and international experiences. Employers are increasingly looking for graduates who have had an international experience (Dar et al., 2006; Stanek, 2000). Although the outsourcing of business processes and operations around the globe has blurred the boundaries of enterprise computing, IS applications are most effective when the systems are tailored to the needs of the local community in which they will be used (Buckley et al., 2004). In addition, working in multicultural teams has become inevitable for businesses with worldwide operations (Kvasny, 2006).

Effects of Global Economy on IS Education The Association of American Colleges and Universities studied desired student learning outcomes of an undergraduate education, and published a table that was drawn together from a variety of sources: the standards of regional and specialized accreditation agencies from across the country, best practices articulated by educational associations, qualities sought by employers, and contributions of faculty and administrators at various colleges and universities (Wergin, 2005). The table demonstrates a widespread and growing consensus; two relevant outcomes include: 1) Intercultural knowledge and collaborative problem-solving skills, and 2) Integrative thinking and the ability to transfer knowledge from one setting to another. In a global economy, it is no longer sufficient for a state to compare itself with the state next door; the comparison is against world standards. Investments in IS education need to benchmark best practices wherever they are found. More and more universities seek to pursue agreements for collaborative programs with international partners to nurture certain essential qualities that will give their graduates an edge in a fast-changing global economy. Global education implies that students are educated across disciplinary and geographical boundaries — beyond the content knowledge of a particular discipline in a specific country (Hatakenaka, 2004). Benefits go far beyond academic learning — enabling students to develop broader perspectives on their academic field of study and an ability to develop skills in cross- cultural communication, adaptability, and critical thinking applicable to everyday life (Stier, 2003). Colleges and universities are encouraged (and sometimes mandated) to find ways to introduce their students to diverse racial, ethnic, and cultural situations and learning, but most educators know little about education in other countries (Fuller et al., 2005).

Status of Globalizing IS Programs at US Institutions In this era of globalization, internationalization — both as an idea and an agenda — is receiving widespread attention at academic institutions across North America (Dewey et al., 2009). Although faculty are necessarily key participants in initiatives to internationalize academia, surprisingly little work has been published that addresses the roles, responsibilities, and problems faced by the faculty on an operational level. The three main ways in which an international dimension is typically added to student learning are student and faculty exchanges, joint projects involving in-person or virtual collaboration, and study abroad. There is wide variety of international experiences which fall under the general category of study abroad programs; these include: summer or semester abroad, dual-degree, short immersive experience in a foreign country, internship in a foreign country, and course taught by foreign faculty.

Status of Globalizing IS Programs at Institutions Outside the US A Dutch study which found a large increase in the provision of internationalized curricula over the last ten years, largely in economics and business studies, the humanities, and social sciences (Stronkhorst, 2005). They include curricula with international subjects, curricula with international comparative approaches, and interdisciplinary regional and area studies. The study also examined the process of internationalizing the curriculum and concluded that it is lengthy and complex, that individual academics play a vital role, and that it requires support from a combined bottom-up and top-down strategy that is consistent with the institution's policy on internationalization. The international IS literature includes an increasing number of studies of IS innovation experiences in other regions of the world, mainly the developing countries of Asia, Africa, and Latin America

Journal of Applied Business and Economics vol. 14(3) 2013 25 (Avgerou, 2008; Festervand, 2002). Although there are many researchers that present case studies of internationalization processes, there is little work that evaluates programs across institutions (Hatakenaka, 2004; Joseph et al., 2006; Reichgelt, 2002; Von Konsky et al., 2006). There is need for a study that researches multiple universities and focuses on mapping internationalization for IS majors, addressing barriers to internationalization, and improving structures and systems to enhance internationalization in IS-related academic departments. A critical analysis of internationalization mechanisms will benefit institutions that seek a framework for navigating diverse tensions and responsibilities implicit in an internationalization imperative (Voogt et al., 2008).

PURPOSE OF THE STUDY

The purpose of this study is to assess international academic experiences of students and faculty and determine most effective mechanisms for enhancing global education for IS-related majors. The study was limited to US students and faculty with experience in an academic environment in Asia, because Asia is significantly diverse with respect to the US culture and educational setting. The following research questions were developed to establish a basis for the methodology: 1. What are the most effective mechanisms for enhancing global education for faculty in IS-related disciplines? 2. What are the most effective mechanisms for enhancing global education for students in IS-related majors?

METHODOLOGY

The Delphi technique for a qualitative study was used to develop a range of possible outcomes of international experiences and seek out information that generated consensus among respondents. The Delphi study technique is a means of achieving consensual validity among raters by providing them feedback regarding other raters’ responses and, if possible, the reasons for such (Young, 2006). A panel of raters independently completes a rating task; the results are then tabulated and returned to the panel members for rerating (Linstone, 1995). The new results are tabulated, and the process continues until consensus or near consensus is achieved. As panel members work independently and are not in a group setting, group dynamics are not a factor. The Delphi technique is a valid technique for long-term forecasting (Czinkota et al., 2005). Delphi items are typically broad-based and convergence of medians is effectively achieved in three or four rounds, after which it reaches the point of diminishing returns (Linstone et al., 1995). The role of the researcher is to act as a neutral facilitator when conducting a Delphi study. Much of the popularity and acceptance of Delphi rests on the claim of superiority of group over individual opinions, and preferability of private opinion over face-to-face confrontation (Gokhale, 2001). On the other hand, numerous examples are cited where discussions among participants integrated with the Delphi technique clarified the issues and made honest communication possible (Dillman, 2000). This study used an anonymous setting in the first three phases to arrive at a meeting of the minds or consensus among the experts. In the fourth and final stage, there was a teleconference followed by a four- day long online discussion which facilitated greater insight into the issues. A questionnaire was developed and utilized as the primary data collection tool. Specific items and lists for particular questions were originally generated by the author. The questionnaire was reviewed by several faculty in international business, IS, and IT programs. This input resulted in revisions to the survey instrument. A pilot test was administered to a representative group of both faculty and students at the author’s home institution, which helped to further refine the instrument.

Population and Sample The author prepared a list of IS, business information systems, management information systems, or computer systems programs at many major universities and colleges accredited by either Accreditation

26 Journal of Applied Business and Economics vol. 14(3) 2013 Council for Business Schools and Programs, AACSB, Middle States Association of Colleges and Schools, New England Association of Schools and Colleges, North Central Association of Colleges and Schools, Western Association of Schools and Colleges, or Computing Sciences Accreditation Board to identify potential participants. A total of over 90 student and faculty questionnaires were sent to the Chairs of relevant departments with a request that the questionnaire be forwarded to appropriate faculty members and students. The response rate was 23% for faculty and 29% for students. The study probed deeper with an in-depth survey of: 1) 15 students who had spent at least two weeks in an Asian host country in a IS-related program, and received academic credit for that experience, and 2) 15 faculty who had spent at least two months teaching/conducting research/consulting related to IS or IT applications in business at an educational institution or a company in Asia. The students and faculty were selected based on their exposure to multiple international events and activities, even at their home institution.

DATA COLLECTION AND ANALYSIS

There were four phases in this study; three paper-and-pencil questionnaires and one teleconference followed by online discussion in which data was collected in response to a series of questions. The surveys for the two groups of participants – students and faculty – were collected and analyzed independently of each other.

Phase I The first questionnaire, which appears in Figure 1, asked the participants to identify the international programs in which they participated and rate them on a scale of 1 to 5 regarding their effectiveness in enhancing global education. Next, the respondents were asked to identify the international programs that had proven effective in enhancing their global education and rate each program on a scale of 1 to 5 regarding its chances for successful implementation on a wide-scale at their home institution. They also had the option to provide additional comments.

Phase II The second questionnaire listed: a) the top ten (by frequency) most effective international programs in general identified in the first phase, and b) the top ten (by frequency) international programs identified in the first phase that had proven effective and had the most chances for successful implementation at the participants’ home institution. The participants were asked to rank order the results of the first round. Also, they had the option to provide comments.

Phase III The data from Phase II were analyzed to determine various statistics such as mean, median, mode, and standard deviation using SPSS software. A hierarchical cluster analysis was used to identify common sets of rating pairs to propose categories for reaching consensus. The phase III questionnaire used the same statements as the second round, except that only the top five programs (based on statistical evaluation) were identified in each category. The participants were asked if they would like to modify their answers based on the responses of the other participants. Each participant was given the analysis of the data collected in phase II and a list of comments made in the first and second rounds. After conducting statistical analysis, a trend towards consensus was documented at the conclusion of the third phase.

Phase IV The participants were provided with the results of the third phase after which a one-hour teleconference was held in the fourth and final phase of the study for each of the two groups — faculty and students. This was followed by an online discussion, an open forum that brought together the entire panel so that both faculty and student groups could together discuss an implementation plan to support the strategies. At the end of a four-day online discussion, the group outlined a process for putting in place the

Journal of Applied Business and Economics vol. 14(3) 2013 27 essential elements of a framework to launch international programs that were expected to benefit both faculty and students.

FINDINGS

Data analysis for this study is primarily descriptive. Frequency distributions are used to determine response percentages for various respondent categories. Mean scores and standard deviations provide relative agreement for items and a measure of the degree of consensus for each response. Tables 1 and 2 give insight into student and faculty perspectives on the effectiveness of various international programs and their perceived chances for successful implementation. Values are means ± standard deviation of evaluations, with 1 being the lowest and 5 being the highest rating. Programs that received mean ratings above 2.5 are reported.

TABLE 1 STUDENT FEEDBACK ON EFFECTIVENESS OF INTERNATIONAL PROGRAMS AND THEIR CHANCES FOR SUCCESSFUL IMPLEMENTATION

Chances for International Program Effectiveness Successful Implementation Internship in a foreign country (minimum 2 4.8 ± 0.12 2.6 ± 0.3 weeks) Study abroad (minimum 2 weeks) 4.5 ± 0.24 4.3 ± 0.14 Joint research with international students outside the US 3.1 ± 0.11 2.5 ± 0.22 (communication over internet) Joint research with international students at home 3.0 ± 0.19 4.6 ± 0.19 institution Seminars at home institution by an international 2.8 ± 0.17 4.8 ± 0.11 scholar Take class taught by an international visiting 2.6 ± 0.28 2.9 ± 0.23 faculty

TABLE 2 FACULTY FEEDBACK ON EFFECTIVENESS OF INTERNATIONAL PROGRAMS AND THEIR CHANCES FOR SUCCESSFUL IMPLEMENTATION

Chances for International Program Effectiveness Successful Implementation Internship in a foreign country (minimum 2 4.8 ± 0.17 2.8 ± 0.27 weeks) Joint research with international faculty (travel 4.7 ± 0.23 2.8 ± 0.11 involved) Teach abroad (minimum 2 weeks) 4.8 ± 0.18 2.7 ± 0.21 Seminars at home institution by an international 3.7 ± 0.13 4.6 ± 0.12 scholar Joint research with international faculty at home 3.1 ± 0.1 4.3 ± 0.14 institution

28 Journal of Applied Business and Economics vol. 14(3) 2013 Comments by Faculty Faculty comments mostly addressed the benefits of international exposure and the value added to teaching and research at their home and host institutions. Sample comments are given below. "I think it is hard to describe how much I got from going to Sri Lanka. It is so much more than just a teaching experience. Immersing oneself in a different culture makes all the difference!" "Given the tightening of resources, it is difficult to travel abroad and do research or teach. One might go for a conference for a few days and stay over longer at one's expense but then that does not count as being an 'academic visit'. Overall, traveling abroad for research or teaching is the most effective way for faculty to internationalize their perspectives."

Comments by Students Students comments were documented and categorized so representative samples could be identified. Most of the student comments related to non-discipline-based learning, or in other words, learning that occurred from exposure to culture and customs of people outside the U.S. The value of international experiences was cited as being important to career growth opportunities. "I learned that the United States is not the centre of the universe. Even after interacting with international students here and attending multiple seminars, I hadn't learned half as much as I learned in a week upon arriving in China." "I learned that my way of life is not necessarily the best. Other cultures although very different have their merits. There is no substitute for self-experience. No amount of classes will help." "Visiting a foreign country opened up opportunities for me and I will be working in a foreign country."

DISCUSSION

Overall, both faculty and students were excited to share their perspectives about the value of their international experience. The data show that for both students and faculty, being in a foreign country for studying or teaching, working, or doing research is most beneficial and has tremendous educational value. Based on the participants’ open-ended comments, a major impact of an international experience, even if it is only for two weeks, is increased awareness and appreciation of both the vast and subtle differences in ideas and value systems. Learning about self and others and the development of empathy were cited most often. There was discussion about greater awareness of intercultural sensitivity and diversity. From faculty perspective, international exposure broadened their perspectives and integrating these experiences into their teaching resulted in a more meaningful discussion of global IS practices. Faculty noted an increased self-confidence and self-efficacy with respect to discussions about business case studies outside the US. Many faculty mentioned that involvement in collaborative international research and scholarship has proven valuable for career advancement. When compared to students, faculty were more positive about having international scholars on campus, they opined that visits by international scholars provide greater opportunities for collaboration. The faculty observations noted above are supported by literature. A study examined the cross- cultural differences between South Korea and the U.S. in user behavior towards protective information technologies (Dinev et al., 2009). The authors report that cultural factors should be considered when teaching about information security policies, practices, and technologies in global networks where multiple cultures coexist. Classroom discussions by faculty about cross-cultural factors can lead to improved student understanding and help mitigate the risks associated with cultural differences among diverse project teams in the workplace (Krishna, et al., 2004; David, et al., 2008). From students' perspective 'study abroad' is both highly effective and has the highest chances for successful implementation. Students talked frequently about raised global mindedness. Students mentioned that opportunities for study in international settings may be limited by lack of language proficiency, financial resources, maturity and self-confidence, or the unavailability of comparable academic programs. Internship in a foreign country was identified as being highly effective; however,

Journal of Applied Business and Economics vol. 14(3) 2013 29 students thought that it was very difficult to get paid internships in foreign countries and therefore this mechanism has very low chances for successful implementation. Students noted an increase in internship offers following a significant international experience like study abroad, and attributed those increased employment opportunities to their enhanced cultural-sensitivity and understanding of IS practices in a foreign country. Students observed that when foreign students study in the US, it is typically because of a shortage of university places at home, because foreign students and their families perceive that they can access more prestigious and career beneficial programs abroad, or because higher degrees or highly specialized subject areas are not available in their home country. These students adapt to the US culture and there is little reason for US students to adapt to their foreign culture. As a result, studying with foreign students residing in the US does not help internationalize US students. Literature supports the above viewpoints expressed by the students. Clarke et al. (2009) investigated several of the potential intercultural influences of a semester abroad for U.S. students in IS-related majors. The authors found that students who study abroad may have greater intercultural proficiency, increased openness to cultural diversity, and become more global-minded than those students remaining in a traditional campus setting. Additionally, students with international experiences perceive themselves as being more proficient, approachable, and open to intercultural communication. These are critical traits for IS recruiters who seek students who are open-minded, creative, and have willingly explored international opportunities, especially internships (Fuller, 2005). Several studies provide empirical evidence that cross- cultural learning and adaptation from both client and vendor staff leads to compromises and innovations in IT offshore outsourcing project teams (Brannen et al., 2000; Weber, 2004). Governments strive to achieve a workforce that can function at the cutting edge of the knowledge economy in order to provide sustainable prosperity for all their citizens. Students and faculty need to be prepared to operate in a global environment. For students, the primary personal driver for international learning is career enhancement and broader personal enrichment. Research by the American Council on Education in 2001 suggested that 88% of US students believed that international education would give them a competitive advantage in the workforce (Turlington et al., 2002). A report for the UK’s Higher Education Policy Institute goes even further, stating: international higher education is increasingly seen as a route to good employment (Marshall, 2004). The European Commission funds large schemes to promote collaborative education and research for its member nations.

CONCLUSION

Knowledge has become a critical global commodity; hence, a university needs to measure its educational and research programs by benchmarking itself against leading universities worldwide. Today’s graduates must not only have advanced skills but also a global perspective to be successful in life and work. Information systems and technology is the basic building block for corporate systems everywhere. The career opportunities in IS require students to know both the technology and the business and environment in which they will work. Given the nature of IT — free flow of information that completely blurs national boundaries — the discipline is inherently international but its effective application depends on an understanding of the local culture in which it is being used; students and faculty need to be prepared to operate in this global environment. Organizations need employees who will maintain and enhance business competitiveness in a global, rather than purely local, market. Economic globalization has obligated quality higher education programs to educate students about issues and practices in other parts of the world. Expanded study / teach abroad programs, international internships, events with global themes, and presence of international scholars on campus are key activities that enhance campus internationalization. Studying abroad in an ethnically diverse environment is an academic, cultural, intellectual and emotional journey that facilitates the acquisition of intercultural competencies as well as personal growth. Universities have much to gain from approaching internationalization and ethnic diversity in an integrated fashion with regards to faculty, students, and curriculum. Faculty who have international experiences tend to weave communication standards, business

30 Journal of Applied Business and Economics vol. 14(3) 2013 practices, telecommunications laws, and ethics prevalent in different countries or continents into the IS curriculum. Student and faculty visits to foreign countries promote cultural understanding, build a learning society, globalize the information society, develop entrepreneurs and help make society more socially responsible.

REFERENCES

Avgerou, C. (2008). Information systems in developing countries: a critical research review. Journal of Information Technology, 23, 133–146.

Buckley, P. J., & Ghauri, P. (2004). Globalization, economic geography, and the strategy of multinational enterprises. Journal of International Business Studies, 35(2), 81–98.

Brannen, M. Y., & Salk, J. (2000). Partnering across borders: Negotiating organizational culture in a German-Japanese joint venture. Human Relations, 53(4), 451–487.

Ceccucci, W., & White, B. A. (2008). IS faculty perceptions of ABET accreditation. Information Systems Education Journal, 6(54), 1–8.

Clarke, I., Flaherty, T. B., Wright, N. D., & McMillen, R. M. (2009). Student intercultural proficiency from study abroad programs. Journal of Marketing Education, 31(2), 173–181.

Czinkota, M. R., & Ronkainen, I. A. (2005). A forecast of globalization, international business and trade: report from a Delphi study. Journal of World Business, 40(2), 111–123.

Dar, M. J., Ekstrom, J. J., & Lunt, B. M. (2006). Integrating information assurance and security into it education: a look at the model curriculum and emerging practice. Journal of Information Technology Education, 5, 389–403.

David, C. G., Chand, D., Newell, S., & Resende-Santos, J. (2008). Integrated Collaboration Across Distributed Sites: The Perils of Process and the Promise of Practice. Journal of Information Technology, 23(1), 44–54.

Dewey, P., & Duff, S. (2009). Reason before passion: faculty views on internationalization in higher education. Higher Education, 58, 491–504.

Dillman, D. A. (2000). Mail and internet surveys: the tailored design method, New York: John Wiley & Sons, Inc.

Dinev, T., Goo, J., Hu, Q., & Nam, K. (2009). User behavior towards protective information technologies: the role of national cultural differences. Information Systems Journal, 19, 391–412.

Festervand, T. A. (2002). Delivering a graduate information systems course in France: observations, experiences, and professional development opportunities. Journal of Information Systems Education, 13(2), 89–94.

Fuller, U., Amillo, J., & Laxer, C. (2005). Facilitating student learning through study abroad and international projects. ACM SIGCSE Bulletin, 37(4), 139–151.

Gokhale, A. (2007). Offshore outsourcing: a Delphi study. Journal of Information Technology Case and Application Research, 9(3), 6–18.

Journal of Applied Business and Economics vol. 14(3) 2013 31

Gokhale, A. (2001). Environmental initiative prioritization with a Delphi approach: A case study. Journal of Environmental Management, 28(2), 187–193.

Gorgone, J. T., Gray, P., Stohr, E. A., Valacich, J. S., & Wigand, R. T. (2006). Reprint: MSIS 2006: Model curriculum and guidelines for graduate degree programs in information systems. ACM SIGCSE Bulletin, 38(2).

Hatakenaka, S. (2004). Internationalism in higher education: a review. Journal of Marketing Education, 22, 35–44.

Joseph, M., & Lunt, B. (2006). IT in the Middle East: an overview. Proceedings of the 7th Conference on Information Technology Education SIGITE, 25–30.

King, W. R. (2005). Outsourcing and offshoring: The new is paradigm? Journal of Global Information Technology Management, 8(2), 1–4.

Krishna, S., Sahay, S., & Walsham, G. (2004). Managing cross-cultural issues in global software outsourcing. Communications of the ACM, 47(4), 62–66.

Kvasny, L. (2006). Let the sisters speak: understanding information technology from the standpoint of the 'other'. The Database for Advances in Information Systems, 37(4), 13–25.

Linstone, H., & M. Turoff. (1995). The Delphi method: techniques and applications. Reading, MA: Addison-Wesley.

Marshall, P. (2004). 21st century Britain and the Commonwealth. The Round Table, 93(376), 571–581.

Reichgelt, H., Lunt, B., Ashford, T., et al., (2004). A comparison of baccalaureate programs in information technology with baccalaureate programs in computer science and information systems. Journal of Information Technology Education, 3, 19–34.

Reichgelt, H., Zhang A., & Price, B. (2002). Designing an information technology curriculum: The Georgia Southern University experience. Journal of Information Technology Education, 1(4), 213–221.

Stanek, M. B. (2000). The need for global managers: a business necessity. Management Decision, 38(4), 232–242.

Stephens, C. S., & O’Hara, M. T. (2001). The core information technology course at AACSB-accredited schools: consistency or chaos? Journal of Education for Business, 76(4), 181–84.

Stier, J. (2003). Internationalization, ethnic diversity and the acquisition of intercultural competencies. Intercultural Education, 14(1), 77–91.

Stronkhorst, R. (2005). Learning outcomes of international mobility at two Dutch institutions of higher education. Journal of Studies in International Education, 9(4), 292–315.

Turlington, B., Collins, N. F., & Porcelli, M. (2002). Where credit is due: approaches to course and credit recognition across borders in U.S. higher education institutions. Washington, DC: American Council on Education.

32 Journal of Applied Business and Economics vol. 14(3) 2013 Von Konsky, B. R., Loh, A., Robey, A. Gribble, S., Ivins, J., & Cooper, D. (2006). The benefit of information technology in managing outcomes focused curriculum development across related degree programs. Proceedings of the 8th Australian Conference on Computing Education, 52, 235–242.

Voogt, J., & Knezek, G. (2008). International Handbook of Information Technology in Primary and Secondary. New York, NY: Springer Science Business Media, LLC.

Weber, R. (2004). Some implications of the year-2000 era, don’t-com era, and offshoring for information systems pedagogy. MIS Quarterly, 28(2), 9–17.

Wergin, J. (2005). Higher education waking up to the importance of accreditation. Change, 35(1), 11–19.

Wolk, R. M. (2008). How important is student computing ability? The role of information technology competence in business school accreditation. Information Systems Education Journal, 6(39), 1–16.

Young, T. L. (2006). Prioritized qualities, attributes, and skills of successful home educators: a Delphi study. Ann Arbor, MI: ProQuest Information and Learning Company.

Zammuto, R. F. (2008). Accreditation and the globalization of business. The Academy of Management Learning and Education, 7(2), 256–268.

Journal of Applied Business and Economics vol. 14(3) 2013 33 FIGURE 1 PHASE I QUESTIONNAIRE

Dear Participant:

Please use the following format to identify the international programs in which you have participated and rate them on a scale of 1 to 5 regarding their effectiveness in enhancing global education.

Participation in International Programs Effectiveness in Enhancing Global Education Least Most

1. 1 2 3 4 5 2. 1 2 3 4 5 3. 1 2 3 4 5 4. 1 2 3 4 5 5. 1 2 3 4 5 6. 1 2 3 4 5 7. 1 2 3 4 5 8. 1 2 3 4 5 9. 1 2 3 4 5 10. 1 2 3 4 5

Use the following format to identify the international programs that have proven effective in enhancing global education and rate each program on a scale of 1 to 5 regarding its chances for successful implementation at your institution. Please provide comments on the back of this page.

International Programs Chances for Successful Implementation Least Most

1. 1 2 3 4 5 2. 1 2 3 4 5 3. 1 2 3 4 5 4. 1 2 3 4 5 5. 1 2 3 4 5 6. 1 2 3 4 5 7. 1 2 3 4 5 8. 1 2 3 4 5 9. 1 2 3 4 5 10. 1 2 3 4 5

Please provide comments on the back of this page.

34 Journal of Applied Business and Economics vol. 14(3) 2013

Virtues of Investment Fund Recommendations: Consumer Reports and Money Magazine

C. Edward Chang Missouri State University

Thomas M. Krueger Texas A&M University-Kingsville

INTRODUCTION

Consumer Reports and Money magazine are two of the most popular magazines published in the United States. Consumer Reports claims to be one of the ten leading magazines in the country, with 4.3 million subscribers, and 16 million readers (Center for Advancing Health, 2009, p. 39). Money magazine has a circulation of 1.9 million copies and readership of 8.0 million, and, like Consumer Reports, is published twelve times a year (Nationwide Newspapers, 2012). According to the Audit Bureau of Circulation (2012), this circulation level ranks Money in 36th place. For comparison purposes, People magazine, which ranked #10 in the United States in 2011 had a circulation of 3.6 million copies in June 2011 (Kantor, 2012). Magazines, be they financial or general, rely upon eye-catching stories to sell issues. One of the best ways to attract new subscribers is with claims of having some sort of unique insight from which readers can benefit. Finance publications often rely upon listing of stocks for investors to hone in on, as illustrated by Fortune magazine’s Investor’s Guide 2012 special issue devoted to the top selections for 2012. Within that edition, readers can find stories on purchases of investments including stocks, bonds, metals, and real estate, in articles with such catchy titles as “Where Do I Put My Money Now?” (Sloan, 2011) and “Advice from the Expert Roundtable” (Covin, 2011). The problem is that all publications dealing with prediction of the future have only a conjecture regarding future condition. They have no prior knowledge. While there is a place for strategic investment and investment analysis, the ability to accurately predict the best investments would be an anomaly, running counter to the semi-strong efficient market hypothesis. In 2007, Consumer Reports and Money magazine ran articles claiming to have identified the best investment funds for the future. After reviewing

Journal of Applied Business and Economics vol. 14(3) 2013 35 a variety of studies regarding investment prediction by finance publications, the remainder of this report compares these two investment recommendations.

LITERATURE REVIEW

Stock Recommendations The finance literature is full of examples of finance publications that attempted to predict the future and failed. There are also individual examples of supposed sleuths who had examined investments and came out with the best. A widely-reported example (i.e., Barker (2007) and Orcutt (2012)) is Fortune magazine’s 2000 article titled “10 Stocks to Last the Decade.” Of the ten stocks listed only one finished the decade with a higher price. Twice as many went out of business, and the remaining seven finished the decade with a loss! Instead of drawing conclusions from one published prediction, several researchers have taken a more rigorous approach. In fact, the rigorous study of the success of financial service agencies goes back many decades. In 1933, in the midst of what came to be known as the “Great Depression,” Cowles studied the forecasting ability of leading financial information service agencies. In the first volume of Econometrica, Cowles (1933) reports that, on average, securities purchased on the basis of forecasters’ suggestions underperformed the market. Over the ensuing eighty years, a multitude of research has analyzed the performance of investment gurus. Brody and Rees (1996) divide the information set into the complicated (as they characterize as The Wall Street Journal Transcript) and simplistic (which they characterize as Money magazine). They go on to highlight the success of the simplistic Value Line Investment Survey, and support given to the “Value Line Enigma” by Copeland and Mayers (1982) and Pawlukiewicz and Preece (1991). Brody and Rees’ (1996) contribution to the investor advice literature is an early direct assessment of the performance of stock purchase and sale recommendations published in two popular investment magazines. They analyze the relative performance of 130 recommendations by Money magazine and 109 by Changing Times in 1990, in total there were 214 buy recommendations and 25 sell recommendations. Over the ensuing year, the median cumulative return on the buy recommendations was significantly worse than the market return. Although the “sell” portfolio outperformed the market, results were driven by a single investment. Comparison of Money magazine and Changing Times led to the conclusion that only Changing Times recommendations can provide abnormal profits. However, the market-excess returns evaporated after a short period of time. By comparison, our study examines mutual fund recommendations, over a more recent period of time, is based on a longer holding period, and compares the performance of Money magazine to Consumer Reports.

Mutual Fund Recommendations The mutual fund industry has been one of the extraordinary growth stories in the history of U.S. financial markets. In 1984, net fund assets totaled $370 million; in 2010 they were $10.4 trillion (Haslem, 2010, page xvii). In 2010, 96 million individual investors in 55 million households owned mutual funds. The essential force of this growth has been the effectiveness of mutual funds as vehicles for providing investors with retirement incomes and financial wealth. Performances of various types of mutual funds have been documented in numerous studies; many of which are summarized in Haslem’s masterfully written books (Haslem, 2003; Haslem, 2010). Jensen’s (1968) classic study of mutual fund performance found that the average mutual fund produced disappointing returns. Throughout the remainder of the 20th century and into the current epoch the value of active mutual fund management continues to be hotly debated. In their comprehensive analysis of active versus passive management across investment categories over time, Fortin and Michelson (2002) found significant advantages to indexing. Only in the small company equity market and international stock market were mutual fund managers able to seek out inefficiencies and outperform the benchmark index. In all other instances, the index return was significantly greater than the average fund return. Using a technique that controls for situations where mutual funds ended up with significant alphas

36 Journal of Applied Business and Economics vol. 14(3) 2013 by luck alone, Barras, Scaillet, and Wermers (2010) find that virtually no funds exhibit persistent positive performance. Furthermore, the very small proportion of funds that beat market index, has shrunk over time. Moreover, Comer, Larrymore, and Rodriguez (2008) examined the success of the Wall Street Journal’s predictions reported in its “Smart Money Fund Screen” during 2005. Across 389 mutual funds in the sample, the funds generate positive alphas during the year before publication. However, post publication there is an average decline of over two percent! An important aspect of their findings is that the worst publication performance was experienced by domestic equity funds, which make up a plurality of the investment funds recommended by CR and MM. In an attempt to explain their results, Comer, Larrymore, and Rodriguez (2008) note the findings of Brown, Harlow, and Starks (1996) and Busse (2001) who observed that equity fund managers strategically shift the risk of their portfolio in the opposite direction of performance. Comer, Larrymore, and Rodriguez find that a majority of funds did not alter their benchmark risk levels. Funds chosen by CR and MM tend to have lower risk levels than found within Morningstar benchmark categories, which may provide a partial explanation for the success of CR and MM portfolios.

Performance of Consumer Reports' Recommended Mutual Funds We were unable to find any studies of specific investment fund recommendations by Money magazine. Only two articles have evaluated the performance of Consumer Reports’ recommended mutual funds. The Index Investor critically commented that the Consumer Reports’ (2005) multi-attribute utility analysis “runs the risk of producing answers that are at best misleading and sometimes flat-out wrong” (The Index Investor). After adjusting for matching categories and index fund benchmarks, the Index Investor found that many of the 52 Consumer Reports’ funds actually had negative alphas between 1995 and 2004, with an average alpha of only 0.17%. The Index Investor concluded that Consumer Reports recommended “actively managed mutual funds, whose expense levels and tax costs are demonstrably higher than those on comparable index funds, without any statistically significant evidence that the active funds’ risk/return trade-off is superior.” The Index Investor considered Consumer Reports’ recommendation of mutual funds in 2005 a big mistake by an organization that is generally extremely reliable. Moreover, Index Investor’s alpha calculations are based on a period prior to mutual fund recommendation. The positive 0.17% alpha may be a result of fund selection based on historical performance. However, historical performance is not available to investors. Our analysis of mutual funds subsequent to Consumer Reports’ publication focuses on whether their analysts have any investment selection skills. Chen (2011) examines 60 mutual funds presented in the February 2007 issue of Consumer Reports. Data for the same group of mutual funds were obtained as of September 30, 2008 to evaluate the recommendations provided in Consumer Reports. In order to exclude the major declines that occurred in the stock market in October 2008, the time periods of November 30, 2006 and September 30, 2008 were analyzed in her study. She examines the relationship between mutual funds’ net assets, share prices, manager tenures, expense ratio, tax–cost ratio and annualized returns to see whether Consumer Reports is a reliable source for investors seeking to purchase mutual funds. Chen concludes that Consumer Reports may not be a reliable source for investors seeking to improve their investment decisions. We believe our research is an improvement over that of Chen for several reasons. We deal with a more recent time period and do not exclude any period of abnormal stock market performance. This research investigates an extended time period and expands on the number of return and risk measures in prior studies. Furthermore, we compare the performance of Consumer Reports and Money magazine. Hedge funds have been the focus of many recent empirical studies. One of the more notable of these is Ibbotson, Chen, and Zhu’s (2011) assignment of return to fees, risk, and risk-adjusted return. They find that hedge funds provide excess risk-adjusted return in every year during the 1995-2009 period. Even when replacing the traditional beta measure of systematic risk with Fung and Hsieh’s (2004) seven-factor model, Ibbotson, Chen, and Zhu found significant alphas for hedge funds. Our study is not limited to hedge funds, which are atypical and would be inconsistent with the concept of studying the impact of

Journal of Applied Business and Economics vol. 14(3) 2013 37 picking funds on the basis of widely-disseminated investment information. Our examination of post- announcement performance does not suffer from a backfill bias, which Malkiel and Saha (2005) found could add more than five percent to returns for hedge funds.

DATA AND RESEARCH METHOD

Comparison of Fund Recommendations Of course, this comparison would be irrelevant if CR and MM recommended the same funds. Fortunately, that is far from being the case, as shown in Table 1, which is a summary of the recommendations exhibited in Appendix A. CR recommended 84 funds, while MM recommended 67 funds, which are both less than one percent of the mutual funds within the same Morningstar categories chosen, as shown on the top line of Table 1. In total, the sample consists of 146 investment funds, 79 are unique to CR, 62 are unique to MM, and 5 were in both sets of recommendations. The investment funds selected by both magazines and their Morningstar category are American Funds AMCAP A (Large Growth), T. Rowe Price Equity Income (Large Value), Vanguard Windsor II Investors (Large Value), T. Rowe Price New Era (Natural Resources), and Royce Pennsylvania Mutual Investment (Small Blend). Panel A of Table 1 reports the recommendation concentration across Morningstar categories. Both magazines are in agreement with regard to asset allocation, providing the most recommendations in domestic stocks and the least in municipal bonds. This commonality is probably a function of the total number of funds in these Morningstar categories. As shown in the right column of Panel A, 52.0 percent of all funds in the categories chosen by these magazines were domestic stock funds, while only 3.7 percent were municipal bond funds. At 22.6 percent, balanced funds garnered the second highest number of CR recommendations, which MM excluded from its recommendations. However, MM placed 16.4 percent of its recommendations in fixed-income funds, which were not covered by CR. MM was also twice as likely to propose investment in an international stock fund. Panel B of Table 1 reports that there was a notable difference within the individual Morningstar categories. A plurality of CR funds was in the large value Morningstar category, with selections in the moderate allocation category and small blend categories being the second and third most common. None of these Morningstar categories take the top three spots in the MM listing. Instead, MM urges investment in large blend, foreign large blend, and mid-cap blend funds. As reported in the right column, a plurality of mutual funds falls in Morningstar’s large blend category. Large blend funds are the fourth most frequent category in CR’s recommendations, despite being the most frequent in MM’s listing. Panel C of Table 1 contrasts the concentration of the magazines’ recommendations. It reports the percentage of all recommendations in the top four listed Morningstar categories, which were given in Panel B. Both CR (46.4 percent) and MM (38.8 percent) have a higher concentration than found in mutual funds (35.0 percent) overall. Furthermore, it should be noted that the 35.0 percent value is based solely on the investment fund categories covered by CR or MM. Inclusion of other Morningstar categories (e.g., intermediate government and muni national long) would reduce the mutual fund concentration in the top four Morningstar categories. Panel D and Panel E reveal the biggest differences in CR and MM recommendations. Panel D focuses on stock selections, while Panel E exhibits differences in other fund types. In these categories, there is at least a four fund difference in the number of funds chosen by CR or MM versus the other magazine. Mutual funds totals in the right column demonstrate that at least one hundred and forty-eight funds were available in each Morningstar category. CR recommended almost three times as many large value stock funds, as shown in the top row of Panel D. CR also suggested five additional small blend funds and world stock funds; the latter category was totally devoid of a recommendation in the MM listing. Instead, MM recommended five more foreign large blend stock funds and four more diversified emerging market stock funds. CR did not identify any of the 241 funds in this latter Morningstar category as worthy selections.

38 Journal of Applied Business and Economics vol. 14(3) 2013 As shown in Panel E, which is consistent with the results reported in Panel A, large differences exist in the recommendations of non-stock funds. CR recommended balanced moderate allocation funds and balanced world allocation funds a total of fifteen times. However, MM did not recommend one fund from either Morningstar category. At the other extreme, MM selected four intermediate fixed-income funds, while CR did not recommend one of these 869 funds. Obviously, there are significant differences in the recommendations of CR and MM. This report investigates the relative performance of each recommendation over the ensuing five years.

Data All investment fund information comes from Morningstar. Investment funds are assigned to a specified category by Morningstar. If two or more investment funds within a Morningstar category are recommended an average of the funds’ performance on the measure under consideration is computed. An average of the Morningstar category averages are presented for each magazine’s set of recommendations. These are matched with the average of the portfolios consisting of all mutual funds in the same categories. Two sets of numbers are reported; one based solely on stock investments, while the other consists of stock funds, fixed-income funds, and balanced funds. This will be referred to as the average of all categories. Creation of category averages has several benefits. For instance, investors can invest in a variety of securities and Morningstar categories. By computing averages for each category with magazine recommendations, we even out investment across the Morningstar categories. These averages are combined to create averages of all equity categories and averages of all categories recommended by the magazines. Our results are not biased by the frequency of recommendations within a given category by a magazine. Another advantage of computing Morningstar category averages is that it allows us to create portfolios of all mutual funds in a given category. For instance, our results are not biased by the existence of 1,387 mutual funds in the Morningstar’s large blend category, as compared to there being only 86 mutual funds in Morningstar’s natural resources category. Findings are also not biased by the concentration of magazine recommendations in a given category. A benchmark is created by summing together the mean returns of each category. In creation of the benchmark, we only include categories in which a specified magazine made a recommendation. For instance, in the “all” mutual fund benchmark for CR’s recommendation, real estate is not included. However, it is included in the MM benchmark. Standard pairwise t-tests are computed using Excel to estimate the significance of the differences in these means. A direct comparison between CR and MM is also presented. First, Morningstar categories in which CR and MM made a recommendation were identified. Category average results were then computed for each. Averages of these category averages will be presented in the columns on the right side of the following tables along with t-test statistics. These columns only include equity categories, because no agreement existed on investment in any of the other categories.

Capture Ratios Due to their relative newness to the portfolio evaluation process, additional information will be shared on the capture ratios. Much like the Sortino measure’s focus on standard deviation below the mean, capture ratios have gained popularity because of their ability to provide important investment insight and because they are intuitive and easily understood by investors. Upside capture ratios compare an investment performance against a market index during periods when the benchmark’s performance is positive. On the other hand, downside capture ratios compare the performance to a benchmark during periods when the benchmark’s performance is negative. A value of 100% for either capture ratio implies that the investment fully captures, or matches, the market surrogate’s return during the period evaluated. A value exceeding 100% indicates that the investment captured more return than the benchmark of the upward movement. This is good news if the market advanced, but bad news if the market declined. A value less than 100% means the investment

Journal of Applied Business and Economics vol. 14(3) 2013 39 captured less return than its benchmark, which is good in down markets but bad news in up markets. Capture ratios provide insight regarding whether a fund is relatively aggressive or defensive in nature. Investors can also make portfolio allocation decisions based on their expectation of future market performance; for instance, selecting funds with high capture ratios in anticipation of market advances.

FINDINGS

Investment Fund Characteristics Expense Ratio Abnormal fund performance may arise from astute selection of investment funds with low expense ratios. Selection of funds with high expense ratios would raise the return hurdle necessary to match category performance. Consequently, the first aspect studied of the investment funds recommended by CR and MM is their expense ratios. As shown in Panel A of Table 2, the average expense ratio of CR equity recommendations is 0.40 percent (i.e., 1.00% – 1.40%) below the category average. The average expense ratio of all CR recommendations is 0.36 percent (i.e., 0.95% – 1.31%) below the category average. In both instances, CR recommendations have expense ratios that are significantly lower at the 0.01 level. Expense ratios are even lower for the MM recommendations. Considering the equity recommendations, the expense ratios are 0.62 percent (i.e., 0.76% - 1.38%) lower. Across all recommendations, the MM recommendations have an expense ratio that is 0.59 percent (i.e., 0.66% - 1.25%) lower. Both of these are significant at the 0.01 level. When limiting the analysis to only those Morningstar categories in which CR and MM made a recommendation, MM recommendations on average have an expense ratio that is 0.33 percent (i.e., 1.05% – 0.72%) lower, which is significant at 0.01 level.

Annual Turnover Expense ratios may be minimized by reducing portfolio turnover, which could however result in slower reaction to economic events and diminish return. On the other hand, investors may feel more comfortable investing in portfolios whose contents are stable. As shown in Panel B of Table 2, the annual turnover of CR equity recommendations is under forty percent (i.e., 39.85%), and less than half of the category average. When considering all CR recommendations, annual turnover again is approximately half of the benchmark. The differences in annual turnovers are significant at the 0.01 level. Likewise, MM recommendation’s turnover ratios are lower than their category averages at the 0.01 level. In fact, the turnover falls to only 28.81 percent among the equity component of MM’s listing. Though not significantly different, there is a large disparity in the annual turnover of the CR and MM recommendations, with the MM set having an average turnover ratio that is three fifths (i.e., 29.16% ÷ 48.45%) as large.

Five-Year Tax Cost Ratio Morningstar reports a tax cost ratio to measure the percentage of a fund’s annual return that is reduced by the taxes investors pay on distributions. For instance, a tax cost ratio of 2.1 percent reduces portfolio value by this 2.1 percent to cover taxes. Investors must pay taxes on dividends and capital gains distributions in the year they are received. Like expense ratios, tax cost ratios negatively impact investors and are concentrated in the 0.0 to 5.0 percent range. A zero percent tax cost ratio indicates the fund had no taxable distributions. Assuming equal distributions, a higher tax cost ratio indicates that the fund was less tax efficient. Tax cost ratios of CR equity recommendations were higher than their category average. Given the low variation in tax cost ratios, this 0.09 percent (i.e., 0.87% – 0.78%) difference was significant at almost the 0.01 level. When including balanced funds and the single state long-term municipal bond fund, the difference diminishes to an insignificant 0.04 percent. By comparison, tax cost ratios of MM equity recommendations were lower than their category average by 0.29 percent (i.e., 0.74% - 1.03%), which was almost significant at the 0.05 level. Though the

40 Journal of Applied Business and Economics vol. 14(3) 2013 difference diminishes to 0.23 percent (i.e., 0.90% – 1.13%), additional degrees of freedom from adding the non-equity selections to the sample and lower variability result in a difference that is significant at the 0.05 level. Within the Morningstar categories with both CR and MM recommendations, MM have a tax cost ratio that is 0.14 percent (i.e., 0.69% – 0.83%) lower, which is significant at the 0.10 level. Looking back across Table 2, MM recommendations have preferable lower expense ratios and higher tax efficiency.

Return and Risk Average annual rates of return over the March 2007 to February 2012 period are presented in Panel A of Table 3. The portfolio of funds recommended by CR earned a return that was over one percent higher whether considering equities or all proposed investments. The difference is significant at the 0.01 level. Likewise, MM equity recommendations earned a return that was over one percent higher than their category average. The one instance where the recommended funds do not beat the category average by at least one percent is MM recommendations across all categories. However, the 0.88 percent (i.e., 2.63% - 1.75%) difference is still significant at the 0.01 level. With a difference of only 0.08 percent (i.e., 2.40% - 2.32%), and lack of statistical significance, one cannot say that either the CR or MM recommendations earned a higher rate of return. The standard deviation of monthly returns over the five-year holding period is presented in Panel B of Table 3. Compared with category benchmarks, CR recommendations provide the preferable lower level of variation, whether considering equity categories alone or all recommendations. Across the sixty observations, the standard deviation difference reaches a level of 2.07 percent (i.e., 23.29% - 21.22%) within the equity categories. When expanding the sample to all Morningstar categories with a CR recommendation, the smaller 1.70% (i.e., 20.81% - 19.21%) standard deviation difference is still significant at the 0.01 level. In one of the first instances where MM recommendations do not outperform the category average, the standard deviations are virtually identical for the chosen funds and their category averages. Consequently, the standard deviation of the CR selection has less risk than funds chosen by MM. However, the difference falls short of the 0.05 level. Betas, which are exhibited in Panel C of Table 3, follow the same pattern as standard deviations. Betas of funds chosen by CR are consistently lower than category averages at approximately the 0.01 level. MM betas are very similar to category averages. In fact, the equity fund betas are identical to their category average. The 0.08 (i.e., 1.12 – 1.04) difference in betas between MM recommendations and CR recommendations, is significant at the 0.05 level. The fact that this small difference is significant at this level attests to the fact that the chosen investment funds’ betas have little variation. Looking back across the return and risk information exhibited in Table 3, it is obvious that both magazines identified returns that provided higher returns over the ensuing five years, with CR recommendations having the added bonus of picking funds with significantly lower levels of risk.

Risk-Adjusted Return Measures Given the significantly higher returns and lower or similar risk, as reported in Table 3, it is not surprising that the recommended investment funds provided significantly higher risk-adjusted returns. Sharpe ratios measure excess returns on the basis of total variation, using standard deviation which was presented in Panel A of Table 3. Panel A of Table 4 reveals that Sharpe ratios of recommended funds are consistently better than the category averages. The difference is consistently significant at least at the 0.05 level, and only dropping to that level when measuring significance for the MM selections across all categories. The Sharpe ratios of the CR funds and MM funds are virtually identical, at 0.18 and 0.17, respectively. Treynor measures, presented in Panel B of Table 4, measure return in excess of the Treasury rate relative to systematic risk or beta. CR funds’ typical better relative performance is perhaps most clearly identified using this risk-adjusted return measure. While the category average is a negative 0.21% annually, which implies that average returns were less than Treasury yields, CR equity funds’ Treynor

Journal of Applied Business and Economics vol. 14(3) 2013 41 measure was 1.17%. Even the lesser, “all categories” estimate of excess beta-adjusted returns is a Treynor measure that is six times larger than the category average. While the MM selections were also significant at the 0.01 level, their Treynor measure was a smaller 0.36 for equities. However, relative to the -0.59 for the equity category average, this performance is very good. Across all recommendations the MM choices fell short of providing Treynor measures that were twice their benchmark’s average. In those equity categories where CR and MM are making recommendations, the difference is small and insignificant. Alpha measures are exhibited in Panel C of Table 4. They represent the excess return determined by subtracting the required return estimated using the capital asset pricing model from the actual return. Once again, the CR and MM selections consistently dominate the category averages. Across both recommendations, the average annual excess rate of return is almost one percent, which is always significant at the 0.01 level. Perhaps the biggest surprise is the reversal in recommendation dominance. While CR equities have the larger Treynor measure, MM equity funds have the larger alpha measures. However, the difference in those equity categories where CR and MM are making recommendations is not significant at even the 0.10 level. What can be gained from the study of risk-adjusted returns is that both recommendations clearly provided better choices whether considering total risk or systematic risk in computing risk-adjusted returns.

Performance During Varying Market Conditions The Sortino measure, presented in Panel A of Table 5, is a risk-adjusted return measure similar to the Sharpe measure. Unlike the Sharpe measure’s adjustment of return in excess of the Treasury yield for all price variance as measured by standard deviation, the Sortino measure only includes negative price variation in the denominator. As such, it does not penalize investments for above-average returns. The Sortino measures presented in Table 5 are always greater than the Sharpe measures presented in Table 4, because the denominator is only a portion of the standard deviation. For instance, the equity funds recommended by CR have a Sharpe measure of 0.16 and a Sortino measure of 0.22, both of which are statistically significant at the 0.01 level. Considering only downside risk using the Sortino measure produces a change in the relative confidence one can have in claiming that the CR and MM recommendations provide useful information in one of four instances. However, the statistical significance of MM recommendations, as a whole, drops slightly from 0.05 to 0.10. As with the Sharpe ratios, the CR and MM Sortino ratios for the equity fund categories they have in common is virtually identical. Upside and downside capture ratios are commonly used to determine how much an investment participates in the upward or downward movement of the market. Upside and downside capture ratios are presented in Panel B and Panel C of Table 5, respectively. The importance of reviewing both is immediately evident, when analyzing CR fund performance. Category averages have significantly higher upside capture ratios, in Panel B, which would run counter to the results presented above which indicate that excess returns arise from CR recommendations. Downside capture ratios, given in Panel C, are also significant. However, in this case, significantly higher capture ratios would imply that the categories tend to be populated with funds that magnify market losses. Within all categories, CR funds understate the market decline, resulting in a downside capture ratio less than 100 percent of the decline. This finding is consistent with that reported in Panel A of Table 5, where it is reported that CR funds have a significantly higher Sortino measure. By comparison, there is no significant difference in MM fund capture ratios, in either rising for falling markets. This finding indicates that the abnormally good performance of MM selections is not a function of market conditions. However, the same cannot be said for the CR versus MM comparison exhibited in the right set of columns in Table 5. Here we see that MM choices provide a larger benefit to investors when markets are rising and detriment when markets are falling. The differences, 9.58% (i.e., 115.79% – 106.21%) during market advances and 8.92% (i.e., 111.30% – 102.38%) during market declines come close to offsetting each other. Looking back at the information provided in Table 5, it appears as though CR recommendations should be courted by investors expecting a market decline and MM recommendations should be sought by those anticipating a market increase.

42 Journal of Applied Business and Economics vol. 14(3) 2013 Morningstar Rankings Morningstar rankings in terms of return, risk, and risk-adjusted return (or “star” ranking) are exhibited in Panel A, Panel B, and Panel C of Table 6, respectively. A distribution is created by Morningstar for each category, allowing investors to assess the performance of their fund relative to the other funds in that category. Funds are not ranked across category. Hence, category averages always have a value of three, “3,” which is the middle score on Morningstar’s 5-level rankings system. Ideally, one would like to have a high ranking on the return measure, a low ranking on the risk measure, and a high ranking on the risk-adjusted return measure. In most instances, recommended funds have a Morningstar return ranking that is significantly better than the category average at the 0.01 level. The best ranking is earned by MM equity funds, which have a return ranking of 3.47 on average. The lowest ranking, and one that is significant at only the 0.10 level is that attained by CR equity funds. In categories with both CR and MM recommendations, the average CR recommendation is 3.32 and MM recommendation is 3.55, which are not significantly different. CR recommendations, whether considered on an equity-alone or all-categories basis, have significantly less risk at the 0.01 level. MM risk measures are virtually identical to the category average. As a consequence, it is not surprising that the Morningstar risk ranking of the CR recommendations is lower than that of the MM recommendations in those equity categories where both magazines made recommendations. A lower rating on this measure is preferable. Consistent with the higher returns and lower risk reported in Panel A and Panel B, respectively, Panel C of Table 6 reports that CR choices supply a risk-adjusted return that is significantly higher than the category average at the 0.01 level. The best performance estimate reported on Table 6 is the “star” rating of CR across all categories, which comes in at a ranking of 3.57, or closer to the ranking of “4” than “3.” MM recommendations also beat their category averages on this measure, but have a lesser 0.05 level of significance. Given the higher return ranking of MM recommendations and lower risk ranking of CR recommendations, it is not surprising that the rankings based on risk-adjusted return are similar and not statistically different. Taking the information in Table 6 as a whole, when considering the distribution of returns, risk and risk-adjusted returns it appears as though both CR and MM identified very useful selections for their readerships, with the CR suggestions providing less risk.

CONCLUSION

This investigation examines the investment fund selection ability of Consumer Reports relative to the equally-widely disseminated Money magazine. Although a majority of recommended funds invest in stock in both instances, CR and MM chose vastly different investment funds. Within the equity fund universe, the magazines focused on different Morningstar categories. Furthermore, within the same Morningstar category, the magazines seldom chose the same stock. Nonetheless, the findings of this study consistently laud the performance of both recommendations, finding both recommendations outperform category averages. Furthermore, the difference is frequently significant at the 0.01 level. There is a slight preponderance of CR recommendations over its category average versus the performance of the MM choices relative to its category average. On a head-to-head basis, in those categories where both magazines made recommendations each publication has its own strengths. CR funds’ strength appears to be based on minimizing risk, by picking funds with significantly lower standard deviations, betas, downside capture ratios, and Morningstar risk rankings. On the other hand, MM funds’ strength lies in reducing costs (through lower expense ratios and better tax efficiency) and making selections that will benefit more from market advances. Despite its apparent ability to pick investment funds, CR has only produced one other listing of mutual fund recommendations. Future research could analyze how its February 2005 recommendations performed relative to the recommendation made by MM near the same point in time. Alternatively, one could expand the study by including recommendations made by Fortune magazine or another investment- oriented publication. Such research would provide insight to the robustness of these results which found

Journal of Applied Business and Economics vol. 14(3) 2013 43 startling market inefficiencies relative to recommendations made by both Consumer Reports and Money magazine.

REFERENCES

Audit Bureau of Circulation. (2012). “Circulation averages for six months ended 12/31/2011," downloaded on May 17, 2012 at http://abcas3.accessabc.com/ecirc/magtitlesearch.asp.

Barker, B. (2007). “10 Stocks to Last the Decade, Revisited,” downloaded on June 8, 2012 at http://www.fool.com/investing/general/2007/02/02/10-stocks-to-last-the-decade-revisited.aspx.

Barras, L., Scaillet, O. & Wermers, R. (2010). False Discoveries in Mutual Fund Performance: Measuring Luck in Estimated Alphas. Journal of Finance, 65, (4), February, 179-216.

Brody, R.G. & Rees, L. (1996). The Performance of Popular Investment Magazine Stock Analysts. Journal of Applied Business Research, 12, (1), 42-46.

Brown, K., Harlow, W.N. & Starks, L. (1996). Of Tournaments and Temptations: An Analysis of Managerial Incentives in the Mutual Fund Industry. Journal of Finance, 51, (1), June, 85-110.

Busse, J. (2001). Another Look at Mutual Fund Tournaments. Journal of Financial and Quantitative Analysis, 36, (1), March, 53-73.

Center for Advancing Health. (2009). “Consumer Reports: Car Buying Guide: A Case Report For Getting Tools Used,” downloaded on June 12, 2012 at http://www.cfah.org/activities/getting_tools_used/consumer.pdf.

Chen, H. (2011). Analysis of Consumer Reports' Recommended Mutual Funds Compared to Actual Performance. Journal of Financial Services Marketing, 16, (1), June, 42-49.

Comer, G., Larrymore, N. & Rodriguez, J. (2008). “Do Mutual Fund Media Recommendations Hold Value? An Empirical Analysis of The Wall Street Journal’s Smart Money Fund Screen,” working paper, Georgetown University, downloaded June 5, 2012 from uark.edu and papers.ssrn.com

Consumer Reports. (2005). Stock Mutual Funds: 70 All-weather Winners. Consumer Reports, March, 28- 33.

Consumer Reports. (2007). 60 Funds You Can Count On. Consumer Reports, February, 18-22.

Copeland, T.E. & Mayers, D. (1982). The Value Line Enigma (1965–1978): A Case Study of Performance Issues. Journal of Financial Economics, 10, (3), November, 289-321.

Cowles, A., III. (1933). Can Stock Market Forecasters Forecast? Econometrica, 1, (3), July, 309-324.

Covin, G. (2011). Advice from the Expert Roundtable. Fortune, 164, (10), December 26, 88-98.

Fortin, R. & Michelson, S. (2002). Indexing versus Active Mutual Fund Management. Journal of Financial Planning, 15, (9), September, 82-94.

Fung, W. & Hsieh, D.A. (2004). Hedge Fund Benchmarks: A Risk-Based Approach. Financial Analysts Journal, 60, (5), September/October, 65–80.

44 Journal of Applied Business and Economics vol. 14(3) 2013

Haslem, J. (2003). Mutual Funds: Risk and Performance Analysis for Decision Making, Malden, MA: Blackwell Publishing.

Haslem, J. (2010). Mutual Funds: Portfolio Structures, Analysis, Management, and Stewardship. Hoboken, NJ: Wiley.

Ibbotson, R.G., Chen, P. & Zhu, K.X. (2011). The ABCs of Hedge Funds: Alphas, Betas, and Costs. Financial Analysts Journal, 67, (1), January/February, 15-25.

The Index Investor. (undated). “Consumer Reports: Mutual Fund Review - Critique”

Jensen, M. (1968). The Performance of Mutual Funds in the Period 1945-1964. Journal of Finance, 23, (2), May, 389-416.

Kantor, S. (2012). “The Top 25 U.S. Consumer Magazines from June 2011,” downloaded on May 17, 2012 at http://accessabc.wordpress.com/2011/08/09/the-top-25-consumer-magazines-from-june-2011-fas- fax/.

Malkiel, B.G. & Saha, A. (2005). Hedge Funds: Risk and Return. Financial Analysts Journal, 61, (6), November/December, 80–88.

Money. (2007). The Money 70. Money, 36, (2), February, 64-66.

Nationwide Newspapers. (2012). “Advertise in Money Magazine,” downloaded on June 6 at http://www.nationwideadvertising.com/adinmoma.html.

Orcutt, J.R. (2012). “Common Sense Investing: Long-Term Predictions Sure to Go Wrong,” downloaded on June 6, 2012 at http://allocationforlife.blogspot.com/2012/01/long-term-predictions-sure-to-go- wrong.html.

Pawlukiewicz, J.E. & Preece, D.C. (1991). The Impact of Value Line Special Situation Recommendations on Stock Prices: Evidence from the Over-the-Counter Market. The Financial Review, 26, (4), November, 547-568.

Sloan, A. (2011). Where Do I Put My Money Now? Fortune, 164, (10), December 26, 85-87.

Journal of Applied Business and Economics vol. 14(3) 2013 45 TABLE 1 COMPARISON OF CONSUMER REPORTS AND MONEY MAGAZINE SELECTIONS

Consumer Reports Money Magazine All Mutual Funds Number of funds 84 67 12,847

Panel A: Allocation Across Major Morningstar Categories Domestic Stock 65.5% 61.2% 52.0% International 10.7% 20.9% 15.3% Balanced 22.6% 0.0% 12.8% Fixed-Income 0.0% 16.4% 16.2% Municipal Bond 1.2% 1.5% 3.7%

Panel B: Morningstar category with Largest Allocations Top Category Large Value Large Blend Large Blend Second Category Moderate Allocation Foreign Large Blend Large Growth Third Category Small Blend Mid-Cap Blend Large Value Fourth Category Large Blend Large Value Intermediate Term bond

Panel C: Percent of Funds from four Largest Morningstar Categories Percent of Funds 46.4% 38.8% 35.0%

Panel D: Biggest Differences in stock allocation ( Categories with at least 4 fund difference) Large Value 14 funds 5 funds 973 funds Small Blend 8 funds 3 funds 493 funds World Stock 5 funds 0 funds 504 funds … … … … Foreign Large Blend 1 fund 6 funds 552 funds Diversified Emerging Markets 0 funds 4 funds 241 funds

Panel E: Examples of Differences in Other Morningstar Categories (at least 4 fund difference) Balanced Moderate Allocation 10 funds 0 funds 749 funds Balanced World Allocation 5 funds 0 funds 148 funds ………. ………. ………. ………. Intermediate Fixed-Income 0 funds 4 funds 869 funds

46 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 2 FUND CHARACTERISTICS: EXPENSE RATIO, ANNUAL TURNOVER, AND TAX COST RATIO

Morningstar Categories Morningstar Categories Categories with CR with CR Selections with MM Selections and MM Overlap Consumer Category Money Category Consumer Money Reports Average Magazine Average Reports Magazine

Panel A. Expense Ratio (%) Equity 1.00 1.40 0.76 1.38 1.05 0.72 Categories 0.000*** 0.000*** 0.000*** All 0.95 1.31 0.66 1.25 Categories 0.000*** 0.000***

Panel B. Annual Turnover (%) Equity 39.85 82.35 28.81 84.24 48.45 29.16 Categories 0.002*** 0.000*** 0.144 All 39.29 78.55 45.47 96.21 Categories 0.000*** 0.000***

Panel C. Five-Year Tax Cost Ratio (%) Equity 0.87 0.78 0.74 1.03 0.83 0.69 Categories 0.017** 0.060* 0.070* All 0.85 0.81 0.90 1.13 Categories 0.194 0.046** ***, **, * reflect significance at the 0.01, 0.05, and 0.10 levels, respectively

Journal of Applied Business and Economics vol. 14(3) 2013 47 TABLE 3 COMPARISON OF RETURN AND RISK

Panel A. Five-Year Average Annual Return (%) Equity 1.89 0.67 1.53 0.38 2.32 2.40 Categories 0.007*** 0.003*** 0.436 All 2.51 1.17 2.63 1.75 Categories 0.001*** 0.002***

Panel B. Five-Year Standard Deviation (%) Equity 21.22 23.29 24.68 24.89 21.79 23.51 Categories 0.007*** 0.355 0.060* All 19.21 20.81 19.29 19.44 Categories 0.008*** 0.358

Panel C. Five-Year Beta Equity 0.99 1.07 1.13 1.13 1.04 1.12 Categories 0.010** 0.481 0.048** All 1.00 1.08 1.07 1.05 Categories 0.008*** 0.307 ***, **, * reflect significance at the 0.01, 0.05, and 0.10 levels, respectively

48 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 4 RISK-ADJUSTED RETURN MEASURES

Panel A. Five-Year Sharpe Ratio Equity 0.16 0.11 0.15 0.10 0.17 0.18 Categories 0.004*** 0.011** 0.427 All 0.21 0.14 0.40 0.32 Categories 0.001*** 0.047**

Panel B. Five-Year Treynor Measure (%) Equity 1.17 -0.21 0.36 -0.59 1.28 1.19 Categories 0.005*** 0.008*** 0.401 All 1.80 0.30 2.06 1.18 Categories 0.001*** 0.009***

Panel C. Five-Year Alpha Measure (%) Equity 1.57 0.77 1.80 0.69 1.50 1.86 Categories 0.008*** 0.004*** 0.207 All 1.23 0.22 1.22 0.43 Categories 0.001*** 0.009*** ***, **, * reflect significance at the 0.01, 0.05, and 0.10 levels, respectively

Journal of Applied Business and Economics vol. 14(3) 2013 49 TABLE 5 ADDITIONAL PERFORMANCE MEASURES

Panel A. Five-Year Sortino Ratio Equity 0.22 0.16 0.21 0.15 0.24 0.25 Categories 0.007*** 0.013** 0.369 All 0.31 0.21 0.73 0.56 Categories 0.002*** 0.098*

Panel B. Upside Capture Ratio (%) Equity 101.64 109.53 115.12 113.74 106.21 115.79 Categories 0.015** 0.307 0.044** All 101.91 107.47 109.91 107.80 Categories 0.028** 0.153

Panel C. Downside Capture Ratio (%) Equity 97.44 109.09 111.58 114.38 102.38 111.30 Categories 0.004*** 0.152 0.046** All 99.15 110.30 115.54 116.29 Categories 0.001*** 0.441 ***, **, * reflect significance at the 0.01, 0.05, and 0.10 levels, respectively

50 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 6 MORNINGSTAR RANKINGS: RETURN, RISK, AND RISK-ADJUSTED RETURN

Panel A. Five-Year Morningstar Return Ranking Equity 3.29 3.00 3.47 3.00 3.32 3.55 Categories 0.060* 0.008*** 0.173 All 3.42 3.00 3.41 3.00 Categories 0.007*** 0.005***

Panel B. Five-Year Morningstar Risk Ranking Equity 2.25 3.00 2.98 3.00 2.34 3.06 Categories 0.002*** 0.459 0.038** All 2.43 3.00 2.97 3.00 Categories 0.004*** 0.408

Panel C. Five-Year Morningstar Rating (“Star”) Ranking Equity 3.49 3.00 3.39 3.00 3.51 3.38 Categories 0.009*** 0.017** 0.260 All 3.57 3.00 3.31 3.00 Categories 0.001*** 0.026** ***, **, * reflect significance at the 0.01, 0.05, and 0.10 levels, respectively

Journal of Applied Business and Economics vol. 14(3) 2013 51 APPENDIX A

Numbers of Available Investment Funds Recommended by Consumer Reports (CR) and Money Magazine (MM) in February 2007 and All Mutual Funds as of March 31, 2012

Morningstar Category CR MFs MM MFs CR MM Domestic Stock Funds Commodities Broad Basket 34 1 34 1 Financial 1 86 86 1 Health 1 118 118 1 Large Blend 7 1,387 9 1,387 7 9 Large Growth 2 1,266 4 1,266 2 4 Large Value 14 973 5 973 14 5 Mid-Cap Blend 6 314 6 314 6 6 Mid-Cap Growth 3 595 3 595 3 3 Mid-Cap Value 5 309 3 309 5 3 Natural Resources 3 86 2 86 3 2 Real Estate 196 3 196 3 Small Blend 8 493 3 493 8 3 Small Growth 1 572 1 572 1 1 Small Value 4 255 1 255 4 1 International Stock Funds Diversified Emerging Mkts 241 4 241 4 Foreign Large Blend 1 552 6 552 1 6 Foreign Large Growth 1 165 165 1 Foreign Large Value 1 264 2 264 1 2 Foreign Small/Mid Growth 106 1 106 1 Foreign Small/Mid Value 1 35 35 1 Global Real Estate 104 1 104 1 World Stock 5 504 504 5 Balanced Funds Aggressive Allocation 2 287 287 2 Conservative Allocation 2 462 462 2 Moderate Allocation 10 749 749 10 World Allocation 5 148 148 5 Fixed-Income Funds High Yield Bond 438 1 438 1 Inflation-Protected Bond 139 2 139 2 Intermediate-Term Bond 869 4 869 4 Short Government 125 1 125 1 Short-Term Bond 324 2 324 2 World Bond 180 1 180 1 Municipal Bond Funds Muni National Interm 182 1 182 1 Muni Single State Long 1 289 289 1 Total 84 12,847 67 12,847 84 67 MFs: all mutual funds in this Morningstar category. Funds in shaded cells are category-matched, and are used for comparisons.

52 Journal of Applied Business and Economics vol. 14(3) 2013

What Drives Employee Stock Options Programs? Safeguarding Human Capital and Recruiting Wanted Skills

Yu Peng Lin University of Detroit Mercy

INTRODUCTION

Employee stock options are contracts that give employees the right to buy a share of a firm’s stock at a pre-specified exercise price and under pre-specified terms. Most employee stock options expire in ten years and are granted an exercise price equal to the market price on the date of the grant (i.e., at-the- money stock options). Typically, a grant of stock options cannot be exercised immediately, but only over time, and most commonly, employees can exercise stock options grants gradually over four to five years. Once a stock option is exercisable, the option becomes vested. Employee stock options are non-tradable, and are typically forfeited if the employee leaves the firm prior to vesting. Since 1980, the use of stock options programs has been growing and is now a widespread practice. Broad-based stock options programs, adopted mainly by firms in the 1990s, can be generally defined as stock options for employees other than the firm’s top five executives. In a 2001 study, the percentage of options held by non-executives was seen as a large share of total options outstanding (Core and Guay 2001). In 2002, it was reported that over 70 percent of options granted went to employees below the top five executives (Murphy 2002). The literature generally supports a positive impact of broad-based stock options programs on firm performance (Sesil and Lin 2011). We do not replicate analyses of past research studies here. However, the results from these studies should not be interpreted causally, i.e., the employment of broad-based stock options programs does not necessarily cause a firm to have better performance or productivity. On the contrary, the results may reflect, at least in part, a selection mechanism by which those firms that are embedded with certain characteristics have chosen to adopt these programs. The results seem to suggest that there are strong economic incentives within some firms to disperse stock options broadly. In other words, the adoption of the program is not likely to be random; rather it is driven by the actors in the system who are most affected. This study examines the determinants of broad-based stock options programs in an attempt to shed some light on the selection mechanism1.

Journal of Applied Business and Economics vol. 14(3) 2013 53 The current literature attempts to justify the adoption decision with several arguments with some success. Among these, a majority of the literature seems to agree on the attraction and retention justifications in which a firm can attract and retain highly motivated, skilled, and optimistic employees by offering them stock options. We extend the findings of this observation by further stating the following: First, stock options serve as a mechanism to safeguard a firm’s investment in firm-specific capital. Firms that have provided on-the-job training for their employees will aim for a low rate of voluntary turnover. Second, in the context of capital-skill complementarity, investment in additional physical capital may lead to adopting a stock options program in an attempt to recruit employees with wanted skills. To the best of our knowledge, the current literature remains silent on these two possible avenues, which is the major task we embark upon in this study. This work is organized as follows. Section ІI covers literature review and introduces our hypotheses. Section III introduces our data set and empirical strategy. Section IV documents the estimation results. Section V concludes this study.

LITERATURE REVIEW AND HYPOTHESES

Literature Review Incentive/Motivation Several studies, employing some version of the standard Principal-agent theory, suggest that the introduction of stock options plans can be justified by the need to align interests between principals (shareholders) and agents (employees). When shareholders are too diffuse to monitor employees, corporate assets can be used for the benefit of employees rather than for maximizing shareholders’ wealth. The provision of ownership rights reduces the incentive for agents’ moral hazard since it makes their compensation dependent on their performance (Jensen 1983). The stock options program is then regarded as one way of attaining this goal. A reasonable case can be made for this action in stock options granted to top executives whose decisions affect the value of the firm. As discussed in Jensen and Murphy (1990), executives are provided with variable compensation and incentives through three primary mechanisms: (1) flow compensations, such as annual salary and bonus; (2) changes in the value of the CEO’s portfolio of stocks and options; and (3) the market’s assessment of the CEO’s human capital. In 1980, CEO annual flow compensation was mainly in the form of salary and bonus (Hall and Liebman 1998) with only 30% of CEOs receiving new option grants. Mean salary and bonus was $655 thousand compared to $155 thousand from new option grants. By 1994, options had become a major component of CEO flow compensation with 70% of CEOs receiving new option grants, and mean option grants amounting to 1.2 million compared to $1.3 million in cash pay. Running through the statistics, one can immediately observe that stock options represent a large and significant proportion in a CEO’s total compensation, which, in theory, leads to the relevance of the incentive effect2. However, as one moves deeper into the organization to employees subordinate to the executives and especially below management level, equity-based incentives take on a relatively less important role. In particular, while the size of the grants of stock options is small compared to total compensation, the incentive effect is probably ambiguous. Further, it runs into more difficulties when applied to stock options granted to employees without significant decision power. Individual actions of rank-and-file employees do not have a discernible effect on the firm’s overall performance (Murphy 2002). On the contrary, an alternative approach stressing a potential mutual monitoring effect of granting broad- based stock options on group rather than individual behavior (Kroumova and Sesil 2006; Lin 2009) emerges. Kandel and Lazear (1992) is an example of a contribution along these lines. They argue that group-based compensation programs (e.g., stock options programs) may actually induce employees to monitor the behavior of co-workers and impose social sanctions (peer pressure) on those employees who shirk from cooperative work group norms. Consequently, one can expect monitoring costs and mutual monitoring to drive the use of stock options programs for non-executive employees especially in large firms3.

54 Journal of Applied Business and Economics vol. 14(3) 2013 Attraction/Selection and Retention The second approach to stock options programs based on the idea of sorting comes from the perception that option grants may attract employees with certain characteristics that are particularly valuable to the firm. Consider, for instance, by offering stock options, firms can hope to attract employees who are optimistic about the firm’s particular prospects since this ties the employees’ compensation with the firm’s future performance. This, in turn, may contribute to a better working environment and more innovative practices. Firms may be able to attract individuals with a willingness to take more risk by offering such programs because stock options are embedded with the risk of stock price fluctuations. Furthermore, the vesting provisions of stock options programs also aid in retention. Firms can retain key employees, who can only exercise their stock options after they are vested in the program. Further, employee turnover becomes costly particularly while such separation constitutes the loss of firm-specific skills (Lin and Sesil 2011). Hence, if innovation and willingness to take risk are considered as important characteristics and if firms greatly value firm-specific human capital, we can expect firms to invest in attracting and retaining better motivated and better skilled employees by introducing the stock options program.

Existing Empirical Evidence on Alternative Determinants According to the National Center for Employee Ownership (NCEO) data (Weeden, Carberry, and Rodrick, 1998), 91% of the firms surveyed initiated a broad-based stock options program as a means for improving employee attraction (selection) and retention. Ittner, Lambert, and Larcker (2003) summarize the relative importance of self-reported objectives in a sample of 194 new economy firms. Employee retention is the most often cited objective for stock options plans. In a selection model, Lazear (1999) concludes that many facts regarding the prevalence of stock options programs are more consistent with selection than with providing incentives. Oyer and Schaefer (2004) reject an incentive-based explanation for broad-based stock options plans, and conclude that selection and retention explanations appear to be consistent with the data gathered from three distinct sources. Kroumova and Sesil (2006) conduct a cross-sectional and longitudinal analysis on a panel of firms and provide support to the claim that higher monitoring costs prompt firms to adopt and maintain employee stock options programs. Oyer and Schaefer (2003) document that if firms’ option-granting decisions are driven by economic profit maximization, then the observed broad-based stock options grants are most consistent with explanations involving retention and attraction of employees. Running through these academic discussions and evidence, one can immediately observe the lack of further detail available in the area of attraction and retention. Under what circumstances do adoption decisions emerge in an attempt to attract and retain valuable employees? To the best of our knowledge, the current literature remains largely silent on this critical question. Although it is implied that the major reason firms care about retention is to avoid the loss of firm-specific skills, stock options are a costly way of reducing turnover (Lazear and Gibbs 2008) since options introduce substantial risk into the compensation package. Given the rank-and-file employees’ risk aversion, there will be a hefty risk premium the firm needs to pay. This may not be the most desirable way to accomplish the purpose of attraction and retention. Nonetheless, the program may carry its own significance given its widespread usage. We extend the current findings by examining alternative thoughts underlying the two justifications. In what follows, we introduce two hypotheses that constitute the broad-based stock options employment decisions under the notion of attraction and retention. To the best of our knowledge, the lines of reasoning of the hypotheses have not been researched, and this is the major task we undertake in this study.

Hypotheses Hypothesis A: Stock Options Are Used to Safeguard a Firm’s Investments in Firm-Specific Training Although it is true that firms value retention since a separation may lead to a loss in firm-specific human capital, such human capital is firm-specific in the notion that it is only valuable to the current employer and is thus non-transferable. According to Robinson and Zhang (2005), an investment in firm-

Journal of Applied Business and Economics vol. 14(3) 2013 55 specific human capital is fairly risky from the point of view of individual employees, since the firm (ex post) may threaten not to use the services rendered by the investment as a way to extract a greater share of the surplus value that has resulted from this investment. Similarly, the employees may act much the same way to extract greater returns for themselves. Equilibrium exists when both firms and employees refrain from paying for any investments in specialized human capital. On the other hand, this action may also seriously dilute a firm’s competitive advantage. The human capital theory implies that there exists little incentive for firms to compensate employees for the firm-specific human capital since it is non-separable from the current employer. However, it is important to note that the vast majority of human capital exists somewhere between firm- specific and general (Stevens 1996; Becker 1993). The combination of the two boundaries suggests, therefore, that employee ownership may be used to encourage and safeguard investments in human capital. Employee-owned companies are the ultimate examples of governance structures that empower employees and protect investments in firm-specific human capital (Blair 1995; Inderst and Mueller, 2007). Firms that have provided on-the-job training for employees will aim for a low rate of turnover. Along these lines, granting stock options to non-executive employees, if properly structured, is hypothesized as having the ability to encourage and safeguard investments in firm-specific human capital via reduced voluntary employee turnover.

Hypothesis B: Stock Options Are Used in an Attempt to Improve the Match Between Physical Capital and Skill As well documented in the literature, employee stock options are employed for the purpose of attracting and retaining individuals with specific characteristics such as a special taste for risk and motivation. Yet, another possibility emerges while firms experience production expansion. In the context of capital-skill complementarity (Griliches, 1969; Flug and Hercowitz, 2000), firms benefit from a better match of physical capital and employee skills. Indeed, Kruse (1993) argues that profit sharing firms may hire more capable (or skilled) employees. During expansions, firms are more likely to invest in the latest physical capital, which leads to higher production premiums if matched with the right set of skills. This argument is further implied by the uneven distribution of stock options inside a firm. According to a survey by the National Center for Employee Ownership (NCEO), besides the stock options issued to executives, 52% of options allocated to such employees who could be assumed as valuable employees. Hence, we conjecture that investments in physical capital lead to a positive employment of broadly dispersed stock options as a means to obtain a better match.

DATA AND EMPIRICAL STRATEGY

Our initial data set comes from the National Center for Employee Ownership (NCEO) 4. Using its own resources and knowledge of the field, as well as information obtained from the media and consultants, NCEO identified a total of 563 public and private companies from different industry sectors reported as sponsoring some form of broad-based stock options plans. From this list, the NCEO had information on the start year of broad-based stock options plans for 193 firms. Using the original list of 563 firms, the start dates for another 98 firms were confirmed. This was accomplished through survey data collected in 2001 and early 2002 and by examining SEC 10-K and 8-K forms between the years 1983 and 20025. In total, there are 291 public and private companies with a confirmed adoption year. While broad-based stock options are generally defined as the options grants toward employees below the top 5 executives, adopting such a broad definition may cloud the empirical analysis since the options granted to the executives/managers right below the top 5 level (such as the 6th, 7th… executives) may have quite different implications than the grants toward non- management employees on the program impact and determinants (e.g., Core and Guay ,2001). For the purpose of this study, we follow NCEO’s definition of broad-based stock options plans as the plan with at least 50 percent of non-management employees who received stock options. Conventionally, broad-based stock options programs were largely adopted by firms in the 1990s. This is confirmed on the NCEO list. Of the 291 firms with an identified adoption year, only 79 (27%) adopted the program outside the

56 Journal of Applied Business and Economics vol. 14(3) 2013 1990s. We retain only those firms with adopting years from 1990 to 1999 for the following reasons. First, firms that employed the program in other time periods may be structurally different from those adopted in the 1990s. To reduce the impact of such unobserved heterogeneity, we analyze only those firms that adopted the program in the 1990s. Second, while it remains possible that other compensation schemes may carry similar effects as stock options programs, it is less likely that firms simultaneously adopt these programs for a shorter time period as in the current study (Blasi and Kruse 2006). The information on start years was then combined with 2006 Standard and Poor COMPUSTAT’s full coverage, firm-level data for 1990-1999. After eliminating missing values on such interested variables as sales and employment, we arrived at an unbalanced panel of 86 firms and 676 observations as our adopters. The total number of adopters retained in this study reduced substantially from 291 to 86. The reasons for the drop were twofold: First, among the 291 public and private firms in NCEO, 206 appear to be public firms. Second, within these 206 firms, 120 companies are eliminated due to either missing data (59 firms) or adoption year outside the 1990s (61 firms). Table 1 compiles the industry distribution of the 86 adopting firms. The adopters are drawn from 31 industries and more skewed toward the new economy6 industries. In line with the literature, firms that adopt employee stock options programs are more likely new economy firms.

Journal of Applied Business and Economics vol. 14(3) 2013 57 TABLE 1 INDUSTRY COMPOSITION OF THE ADOPTING FIRMS

Adopters COMPUSTAT firms # of # of SIC Industry Name firms Percentage firms Percentage 2040 Grain Mill Products 1 1.16% 16 0.17% 2621 Paper Mills 1 1.16% 23 0.25% 2810 Industrial inorganic chemicals 1 1.16% 24 0.26% 2820 Plastic, synthetic materials; ex glass 1 1.16% 10 0.11% 2834 Pharmaceutical preparations 2 2.33% 210 2.25% 2836 Biological products, ex diagnostics 1 1.16% 142 1.52% 2911 Petroleum refining 1 1.16% 45 0.48% 3312 Steel works & blast furnaces 1 1.16% 41 0.44% 3452 Bolt, nut, screw, rivets, washers 1 1.16% 6 0.06% 3533 Oil & gas field machy, equip. 1 1.16% 18 0.19% 3559 Special industry machinery, nec. 6 6.98% 61 0.65% 3569 General indl mach &eq., nec. 1 1.16% 17 0.18% 3571 Electronic computers 3 3.49% 26 0.28% 3572 Computer storage devices 4 4.65% 33 0.35% 3576 Computer communications quip. 5 5.81% 68 0.73% 3577 Computer peripheral eqip., nec. 1 1.16% 61 0.65% 3620 electrical industrial apparatus 1 1.16% 20 0.21% 3661 Tele & telegraph apparatus 4 4.65% 72 0.77% 3663 Radio, TV broadcast, comm. eqip. 2 2.33% 101 1.08% 3674 Semiconductor, related device 12 13.95% 140 1.50% 3823 Industrial measurement instr. 1 1.16% 32 0.34% 3825 Elec. meas & test instruments 4 4.65% 42 0.45% 3826 Lab analytical instruments 2 2.33% 44 0.47% 3829 Meas & controlling dev., nec. 1 1.16% 31 0.33% 3841 Surgical, med instr, apparatus 3 3.49% 65 0.70% 3844 X-ray & related apparatus 1 1.16% 14 0.15% 3845 Electro. medical apparatus 4 4.65% 102 1.09% 3861 Photographic equip. & supply 1 1.16% 23 0.25% 7370 Computer programming, data process 2 2.33% 290 3.10% 7372 Prepackaged software 16 18.60% 517 5.53% 7373 Computer integrated system design 1 1.16% 169 1.81% Total 86 100% Total COMPUSTAT population firms: 9343

58 Journal of Applied Business and Economics vol. 14(3) 2013 Table 2 and Figure 1 demonstrate the distribution of the adopting year, while showing that there does not seem to be an adopting year clustering pattern. This helps to rule out a year-trend on adoption.

TABLE 2 ADOPTING YEAR DISTRIBUTION

Adopting Year # of firms 1990 7 1991 10 1992 6 1993 9 1994 11 1995 11 1996 9 1997 13 1998 6 1999 4 Total No. of Firms 86

One potential drawback of the NCEO data set is that it includes only those firms that eventually adopted broad-based stock options programs; therefore, it is not possible to compare them with firms that did not have these programs. Comparing adopting and non-adopting firms, however, is an important part of our analysis, where NCEO data set is added to our own set of “peer” control firms. The control group of non-adopting firms comprises firms that are similar in size and within the same industry as firms in the NCEO sample (adopters). The idea is that firms that operate within the same industry and are similar in size will tap into the same labor market and thus employ human capital of similar quality. These firms may also use similar human resource management practices. To construct the control group, we identified, for every adopter, the next largest or the next smallest (in terms of total employment) or both (if available) non-stock options companies within the same 4-digit industry classification. This was done by first excluding the stock options firms (the NCEO list) from the COMPUSTAT population of firms and then matching each broad-based stock options firm to the non- stock options companies at the adoption year. We arrived at 133 firms with 933 observations as the non- stock options “peer companies”. One advantage of constructing a control group in such a way is that it helps to control much of the industry-specific factors. In order to convert nominal numbers into real terms, the variables to the 1997 dollar are deflated using the GDP deflator.

Journal of Applied Business and Economics vol. 14(3) 2013 59 FIGURE 1 ADOPTING YEAR DISTRIBUTION

# of firms

Adopting Year

To summarize, there are 219 firms with 1609 observations in total. This allows for an in-depth empirical investigation of the determinants of broad-based stock options programs that could not be conducted in the previous literature, including the literature using the original NCEO dataset. We estimate the probability of adopting a broad-based stock options program as a function of a list of firm characteristics. The model is

P BBD x titi −1,, PA +== βα 1 *()1( lnemp ti − + β 21, *lncapitalint ti − + β 31, *lnRandDemp ti −1,

* + β * lnproducitivty − + β * MBratio − + β *lnSGAemp − 4 ti 51, ti 61, ti 1, * + β 7 * lncapitalinvestment ti − + β81, * DAratio ti − + β91, * New Economy ,ti

+ β y *Year dummies)...... a)...... (...... where PA stands for the population-averaged logit model.

BBD ,ti is a dummy variable indicating the appearance of a broad-based stock options program in firm i at year t. lnemp ti −1, is the natural log of total employment of firm i in year t-1 lncapitalint ti −1, is the natural log of total capital stock per employee of firm i in year t-1 lnRandDemp ti −1, is the natural log of research and development expenditure per employee of firm i in year t-1 lnproductivity ti −1, is the natural log of total sales per employee of firm i in year t-1

MBratio ti −1, is the market-to-book ratio of firm i in year t-1

60 Journal of Applied Business and Economics vol. 14(3) 2013 lnSGAemp ti −1, is the natural log of sales, general, and administrative expenditure per employee of firm i in year t-1 lncapitalinvestment ti −1, is the natural log of capital investment per employee of firm i in year t-1

DAratio ti −1, is the debt-to-asset ratio of firm i in year t-1

New Economy ,ti is a dummy variable indicating whether the firm belongs to the new economy industries6.

Logit specification is adopted for a more consistent estimator yielded in panel data (Hsiao 1986). Based on past studies (e.g. Kroumova and Sesil, 2006), total employment is used as the proxy variable for the motivation/mutual monitoring arguments and supplemented by capital stock. The two variables are expected to yield a positive sign revealing a concern of monitoring costs, but carry a negative sign if firms adopt the program as a way to motivate employees, particularly, since the concern of free riding is at a minimum in smaller firms. Research and development expenses are employed as the proxy variable in an effort to examine the general selection and retention justifications. This is then supplemented by the market-to-book ratio. The two variables are anticipated to carry a positive sign if attraction and retention of employees is a major program adopting consideration. We attempt to examine the two alternative avenues underlying attraction and retention mechanisms as outlined in Hypotheses A & B by employing two variables – selling, general, and administration expenses (hereafter: SGA) and capital investment. SGA is utilized because it includes outlays related to employee training as well as to brand promotion, distribution channels, and information systems (Lev and Radhakrishnan 2003). Thereby, SGA is associated with investments in firm-specific human capital. If the adoption decision is in an attempt to secure and encourage such investments, SGA is expected to be positively related to the likelihood of adoption. Lastly, as outlined in Hypothesis B, in an attempt to promote a better match of capital and skill, firms are more likely to disperse stock options broadly when they are on the verge of a growing stage. Accordingly, we employ capital investment to examine this hypothesis. We attempt to provide additional control for firm size by normalizing selected firm characteristics by total employment. Although this way of normalizing will make the independent variables a function of firm size, it is unlikely to contaminate the estimated parameters. The new economy dummy variable aims at providing more control over industry effects. To avoid inconsistency in resulting standard errors due to serially correlated outcomes, all standard errors (and hence z statistics) are clustered by firms. Consequently, all our estimates are calculated in this fashion as a means to obtain more robust inferences (Bertrand, Duflo, and Mullainathan 2004). Endogeneity is a key and serious issue for the kind of analysis conducted in this research. In general, endogeneity refers to the fact that an independent variable included in the model is potentially a choice variable, correlated with observables or unobservables relegated to the error term. Wooldridge (2001) lists three common sources of endogeneity: (1) simultaneity, (2) omitted variables, and (3) measurement error. We believe simultaneity and omitted variables are the most relevant sources that could largely undermine the effort and the results in the current work. Simultaneity arises when at least one of the explanatory variables is determined simultaneously along with the dependent variable. If so, the disturbance and the explanatory variable will be correlated, leading to the endogeneity bias. In the current framework, such bias exists when companies have pre-existing high productivity or when companies have been on an upward growth path in productivity and are more likely to adopt broad-based stock options programs (i.e., reverse causality). Hence, it is important to include a measure of productivity as a predictor. In fact, the pre-adoption levels in all independent variables are used to avoid such a problem. The second source of endogeneity (i.e., omitted variable bias) may arise during some alternative compensation schemes that carry similar attraction and retention effects but are omitted from the estimation. It would be ideal to include the variables that help to control for such a source of bias. However to our knowledge, such a data is not available. Nonetheless, we believe the bias (if there is any)

Journal of Applied Business and Economics vol. 14(3) 2013 61 is at the minimum in our framework for the following two reasons. First, there is little evidence supporting a correlation of the adoption decisions of employee stock options programs and other similar compensation schemes. Even the decisions are related, the shorter time periods (1990-1999) in our estimation help to control such a bias since it is unlikely that firms would adopt compensation programs with similar effects in a short period of time. Second, if the employment decision of employee stock option programs was affected by some pre-existing compensation schemes that carry similar attraction and retention effects, the impact is captured by the constant term but at the expense of assuming the unobservable remains constant over our analyzed periods. Yet, according to Cole (1898), once a human resource practice is adopted, it is unlikely to be discontinued. Consider, for instance, Employee Stock Ownership Plan (ESOP). ESOP is a type of employee benefit plan, similar in some ways to a profit-sharing program. According to NCEO, ESOPs are a very common form of employee ownership in the United States. They have been growing in strength since about 1974. In an ESOP, a company sets up a trust fund, into which it contributes new shares of its own stock or cash to buy existing shares. Shares in the trust are allocated to individual employee accounts. As employees accumulate seniority with the company, they acquire an increasing right to the shares in their account, a process known as vesting. Employees must be 100% vested within 3 to 6 years, depending on whether vesting is all at once or gradual. Hence, similar to a broad-based stock options program, ESOP provides employees a way to participate in firm growth. Thus, we can conjecture that these programs might be considered to be substitutes, perceived to have the same attraction and retention effects (Kim and Ouimet, 2008). The constant term in our statistical model could help to capture this effect. Table 3 summarizes the proxy variables in this study. Debt-to-asset ratio is added as an additional control due to the fact that a broad dispersion of stock options does not constitute a cash outlay, which is more of a concern under cash constraints. Table 4 provides the summary statistics of the variables in equation (a).

TABLE 3 PROXY VARIABLES FOR ADOPTING AND MAINTAINING JUSTIFICATIONS

Adoption Proxy Variables Justifications Motivation/Mutual (1) Total Employment¹ Monitoring (2) Capital Intensity² (1) Research and Development expenses per employee³ Attraction (2) Market-to-Book Ratio4 (1) Research and Development expenses per employee³ Retention (2) Market-to-Book Ratio4 Reward (1) Productivity5 Investment in firm- specific human (1) Sales, General, and Administrative Expenditure (SGA) per employee6 capital Capital-Skill (1) Capital Investment per employee7 Complementarity Cash Constraint (1) Debt to Asset Ratio8 1. Total employment is the total number of employees excluding temporary ones. 2. Capital intensity is measured by net Property, Plant, and Equipment (COMPUSTAT data #8) per employee 3. Research and Development expenses is the R&D expenses a firm incurred in a year. 4. Market to Book ratio is calculated as (Adjusted fiscal year ending stock price/ total common equity).

62 Journal of Applied Business and Economics vol. 14(3) 2013 5. Productivity is calculated as (Total Sales/total employment) 6. The Sales, General, and Administrative Expenditure is obtained from COMPUSTAT data #189. 7. The capital investment figures are obtained from COMPUSTAT data #30. 8. Debt to Asset Ratio is computed as (Total Debt/Total Assets)

TABLE 4 VARIABLE DEFINITIONS AND DESCRIPTIVE STATISTICS

Standard Variable Definitions Means Deviations Dummy variable indicating the presence of BBDt broad-based stock options program in year t 0.273 0.45 lnEmpt-1 Natural log of total employment in year t-1 0.495 1.77 Natural log of total capital stock per employee in lnCapitalintt-1 year t-1 3.478 0.93 Natural log of research and development lnRanddempt-1 expenditure per employee in year t-1 2.586 1.18 lnProductivityt-1 Natural log of total sales per employee in year t-1 5.186 0.57 MBratiot-1 Market to Book ratio in year t-1 16.115 385.56 Natural log of sales, general, and administrative lnSGAempt-1 expenditure per employee in year t-1 4.06 0.738 Natural log of capital investment per employee in lnCapitalinvestmentt-1 year t-1 2.263 0.988 DAratiot-1 Debt to Asset ratio in year t-1 0.11 0.145 Dummy variable indicating New Economy New Economy industries 0.51 0.5 No. of Observations – 1609 No. of Firms – 219

ESTIMATION RESULTS

By the way of constructing a control group, we are able to assign a “virtual” adopting year to control firms corresponding to their adopting peers. Consider, for instance, firm A adopted a broad-based stock options program in 1995. Firm AA is identified as firm A’s controlling peer and would be assigned a virtual adopting year of 1995. This enables us to do a preliminary comparison between adopters and their controlling peers, in particular, during pre-adoption periods. Table 5 shows the comparisons between adopters and their controlling peers, both in overall level and in pre- and post-periods. Generally, the adopting firms show higher levels of employment, capital intensity, research and development expenditure, productivity, market-to-book ratio, SGA expenditure, capital investment, but slightly lower debt-to-asset ratio. These lead to the following observations, at least in terms of basic summary statistics. First, in line with the monitoring cost argument, large firms and firms with higher monitoring difficulties are more likely the adopters. Second, more productive firms tend to employ the broad-based program in which rises the concern of reverse causality. Third, firms with more growth potential are more likely to disperse stock options broadly. Fourth, the program is more likely to be adopted to safeguard the investment in firm-

Journal of Applied Business and Economics vol. 14(3) 2013 63 specific capital as evidenced by the generally higher SGA expenditures experienced by the adopters. Fifth, it is likely that firms on the verge of an expansionary stage engage in attracting better-skilled employees by introducing the stock option program. Sixth, the employment of the employee stock options program does not seem to be driven by the need of reducing cash outlay. Generally, the summary statistics support hypotheses A and B. To gain more insight into the program adopting decision-making, we turn to a formal statistical method – model (a).

TABLE 5 PRE AND POST COMPARISONS

Adopters Non-adopters Variable Overall Pre Post Overall Pre Post 11.069 19.247 6.682 5.993 7.727 5.088 Employment (23.16) (32.29) (14.53) (12.89) (13.57) (12.43) 64.242 73.496 59.278 46.671 52.400 43.681 Capital Intensity (80.16) (100.09) (66.69) (68.94) (95.81) (49.27) R and D expenditure per 30.036 19.076 35.915 18.915 15.246 20.831 employee (26.77) (18.52) (28.62) (19.53) (16.27) (20.79) 250.898 235.008 259.421 195.121 195.902 194.713 Productivity (144.54) (157.20) (136.69) (149.23) (181.49) (129.40) 25.223 3.044 37.118 2.595 0.533 3.669 Market to Book Ratio (540.10) (8.47) (669.38) (21.53) (23.61) (20.30) SGA expenditure per 87.310 67.597 97.883 67.455 59.119 71.806 employment (50.68) (42.81) (51.46) (57.47) (40.19) (64.28) Capital Investment per 18.981 18.900 19.025 12.782 12.959 12.69 employment (18.88) (19.94) (18.31) (16.473) (16.57) (16.44) 0.101 0.118 0.091 0.121 0.112 0.126 Debt to Asset Ratio (0.136) (0.14) (0.13) (0.16) (0.15) (0.17) No. of Observations 676 236 440 933 320 613 Standard deviations are in the parentheses

To test the attraction and retention justifications in our data, the first column of Table 6 reports the estimation results for model (a) without the entry of the determinants proposed in Hypotheses A and B. Agrees with the literature, research and development expenses and market-to-book ratio are both significant and positive predictors of the program which supports the above two justifications. We now turn to the second column. It indicates that the addition of SGA expenses and capital investments in the model renders research and development expenses insignificant but not the market-to-book ratio. This suggests that at least part of the attraction and retention justifications is attributable to the consideration of safeguarding firm- specific capital investments and improving the capital-skill match.

64 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 6 POPULATION AVERAGE LOGIT ESTIMATES Prob(BBD=1)

Variables Estimates Estimates Estimates 0.088 0.151 0.103 lnEmpt-1 (0.92) (1.35) (1.05) 0.009 -0.262 lnCapitalintt-1 - (0.06) (-0.21) 0.218 -0.115 -0.134 lnRandDempt-1 (1.93)* (-0.73) (-0.87) 0.407 0.072 0.0002 lnProductivityt-1 (1.54) (0.29) (0.001) 0.00002 0.00003 0.00003 MBratiot-1 (1.73)* (2.72)*** (3.39)*** 0.899 0.849 lnSGAempt-1 - (2.31)** (2.43)** 0.295 0.184 lnCapitalinvestmentt-1 - (2.11)** (1.64)* -0.648 -0.659 - 0. 699 DAratiot-1 (-1.01) (-0.97) (-1.03) -0.101 -0.407 New Economy - (-0.28) (-0.96) -5.380 -6.279 -6.344 Constant (-3.72)*** (-4.10)*** (-4.36)*** Year Dummies Yes Yes Yes Wald Chi² 68.03 65.64 65.06 No. of firms 219 219 219 No. of Observations 1609 1609 1609 z statistics are in parentheses. z statistics are adjusted for clustering on firms * Statistically significant at the 0.10 level ** Statistically significant at the 0.05 level *** Statistically significant at the 0.01 level

More specifically, the model estimates that the average firm has a 35.6% probability of having a broad- based program. Of the motivation/mutual monitoring related factors, firm size is positively associated with the likelihood of adopting broad-based stock options program in our sample. Yet it is not statistically significant. Neither is capital intensity. The sample studied does not yield conclusive evidence supporting a grant of stock options as a motivational means. As far as reverse causality is concerned, productivity is not a significant predictor, suggesting the program does not seem to be adopted as a reward for better performance. Nor do we find evidence supporting conserving cash as a program determinant. Although market-to-book ratio is predictive of the adoption of a broad-based stock options program, its magnitude is not economically

Journal of Applied Business and Economics vol. 14(3) 2013 65 meaningful. In line with Hypothesis A, we find the associated factor, i.e., SGA expenditures, is positively and significantly associated with the likelihood of dispersing stock options broadly. More specifically, the probability of having a broad-based plan increases by approximately 2% as SGA expenditures increase by 10% from their average value, controlling for firm size. Firms that provide more on-the-job training are more likely to adopt broad-based programs in an attempt to secure investments in firm-specific human capital. Significantly, in the context of capital-skill complementarity, capital investment is positively associated with the probability of adoption. A 10% increase in capital investment is associated with a 0.6% increase in the probability of adopting. While on the verge of expansion, firms disperse stock options broadly to obtain a better match of capital and skill. Note that the constant term is negatively significant, indicating that some pre-existing and time-constant firm characteristics reduce the likelihood of adoption. One characteristic that we believe to be relevant to the current study is the alternative compensation scheme, which has similar effects on attraction and retention. If so, firms that already had similar plans in place would indeed be in less need of a stock options program, which might explain this finding. Taken together, our results provide support to the attraction and retention explanations, with an emphasis on safeguarding investments in firm-specific capital and improving match of capital and skill. While intangible capital is evidenced in the literature as a major adopting predictor, our findings narrow this capital to investments in firm-specific capital. Since such investments become costly with high employee turnover, dispersing stock options broadly is one of the means that firms could employ to safeguard their investments. However, the existence of alternative compensation schemes with similar retention effect could discourage firms from adopting employee stock options programs. Moreover, our empirical evidence provides support in that a stock options program is adopted in an attempt to improve match of capital and skill during the verge of expansion phase of a company. Finally, new economy industries do not seem to be a significant determinant. This could be attributed to the possibility that new economy firms made tremendous capital investments and were positively evaluated by the market during the 1990-1999 time period. Also, the impact of capital intensity could be absorbed by capital investment. As a way to further test the robustness of our results, we dropped capital intensity and new economy variables from model (a). The estimation results, which do not yield significant difference, are shown in the third column in Table 6.

CONCLUSION

By using a unique data set containing the start year of a broad-based stock options program, we contribute to the current literature in the following ways. First, we extend the existing research findings in which attraction and retention is evidenced as the major determinants by adding more insights. While investment in firm-specific human capital is costly, stock options serve as a means to encourage and secure such investment. Our longitudinal results support a positive association between investments in firm-specific human capital and the employment of stock options programs. Second, within the context of capital-skill complementarity, our empirical evidence supports that investments in physical capital are positively related to the likelihood of adopting broad-based programs. Meanwhile, on the verge of expansion, firms disperse stock options broadly to obtain a better match of capital and skill. Third, we do not find conclusive evidence on reverse causality according to which stock options grants are a reward for better performance. Overall, our results suggest that broad-based stock options do not appear to be a reward for better performance, but carry their own significance. They create value by encouraging investments in firm-specific human capital and improving the match of physical capital and skill in the context of attraction and retention. This research has a number of limitations that need to be considered before broader and more generalizable conclusions could be drawn from the analysis. First, our sample size is rather small and may not be representative of the population of public firms. However, given the fact that our sample firms are skewed more toward the new economy industries which is in parallel with the literature, we feel there is some level of validity in our results. Second, the SGA expenditures may contain some outlays not clearly related to employee training. More detailed firm level and worker level data are needed to address this concern.

66 Journal of Applied Business and Economics vol. 14(3) 2013 ENDNOTES

1. Justifying the reasons for why firms choose a certain compensation scheme against another is an ambitious project, beyond the scope of the present study. We focus on the determinants of broad-based stock options programs. It remains true that alternative compensation schemes may carry similar effects as stock options programs. Contrasting different programs would be a good way of providing much detail into the mechanisms underlying firms’ adoption decisions on compensation methods. To our knowledge, there exists no suitable data set for answering these questions. 2. Indeed, in addition to being an important component of CEO compensation, stock options are also an important element of CEO equity incentives (i.e., sensitivity of a CEO’s portfolio value to stock price). Hall and Liebman (1998) reported that in 1980, 57% of CEOs held some amount of options, and by 1994, this figure had reached nearly 90%. In Core and Guay’s (1999) sample of CEOs from the period 1992-1996, options contributed approximately one-third to the value of the median CEO’s equity portfolio and contributed roughly half of the median CEO’s total equity incentives. In defining incentives as the sensitivity of the CEO’s wealth to stock price changes, for most CEOs, the assumption that the majority of incentives are driven by variation in the value of equity holdings is realistic. Jensen and Murphy (1990), and Hall and Liebman (1998) show that the vast majority of a typical CEO’s incentives to increase stock price are driven by variation in the value of his/her stock and option portfolio (not by flow compensation). Core, Guay, and Verrecchia (2000) show that for a typical CEO, non-price incentives provided by flow compensation are not economically large in comparison to the price-based incentives provided by the CEO’s equity portfolio. 3. However, two reasons have been used to support forceful predictions that mutual monitoring will not occur in large firms. First, to the extent that it is costly to monitor and sanction co-workers, there is an incentive to free ride on the monitoring and sanction efforts of other co-workers. Hence, the likelihood of effective mutual monitoring is decreasing in both the size of the group and the cost of the mutual monitoring. Second, while a firm’s employees are more dispersed, it will be more difficult to have direct interactions and observe other employees’ effort. Similarly, less interdependency among a firm’s business (operating) units leads to more difficulty in promoting mutual monitoring. By utilizing the data on Continental Airlines, Knez and Simester (2001) show that group incentive programs have positive impact on firm performance if the firm’s operating units are more interdependent. Consequently, while it is commonly argued that group compensation programs help to reduce monitoring costs in large firms, the effect may be conversely more pronounced in small firms since their employees are presumably less dispersed. 4. The National Center for Employee Ownership (NCEO) is a private, nonprofit membership and research organization that serves as the leading source of accurate, unbiased information on employee stock ownership plans (ESOPs), equity compensation plans such as stock options and ownership culture. 5. This was accomplished by NCEO and by researchers at Rutgers University. 6. New Economy industries include (SIC codes): Computer related industries (3570, 3571, 3572, 3576, 3577, 5045), Electronics and Semiconductor (3661, 3674), Communications (4812, 4813), Retail (5961), and Business Services (7370, 7371, 7372, 7373).

REFERENCES

Blair, M. (1995). Ownership and Control: Rethinking Corporate Governance for the Twentieth Century. The Brookings Institute, Washington, D.C.

Blasi, J. and Kruse D. L. (2006). U.S. High-Performance Work Practices at Century’s End. Industrial Relations, 45;547-578.

Becker, G. S. (1993). Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education. 3rd ed. The University of Chicago Press.

Bertrand, M., Duflo E., and Mullainathan, S. (2004). How Much Should We Trust Difference-in- Differences Estimates? Quarterly Journal of Economics, 119(1); 249–275.

Journal of Applied Business and Economics vol. 14(3) 2013 67 Cooke, W. N. (1994). Employee Participation Programs, Group-Based Incentives, and Company Performance: A Union-Nonunion Comparison. Industrial and Labor Relations Review, 47(4);594–609.

Cole, R. E. (1989). Strategies for Learning. Berkley, Calf. University of California Press.

Core, J. and Guay, W. (1999). The Use of Equity Grants to Manage Optimal Equity Incentive Levels. Journal of Accounting and Economics, 28;151-184.

——— (2001). Stock Option Plans for Non-executive Employees. Journal of Financial Economics, 61(4);253–287.

Core, J., Guay, W., and Verrecchia, R. (2000). Are Performance Measures Other Than Price Important to CEO Incentives? Working paper. University of Pennsylvania.

Fitzroy, F., and Kraft, K. (1987). Profitability and Profit Sharing. Journal of Industrial Economics 35;113–130.

Flug, K. and Hercowitz, Z. (2000). Equipment Investment and the Relative Demand for Skilled Labor: International Evidence. Review of Economics Dynamics 3(3);461–485.

Griliches, Z. (1969). Capital-Skill Complementarity. Review of Economics and Statistics, 51; 465–468.

Hall, B. J. and Liebman, J. (1998). Are CEOs Really Paid Like Bureaucrats? The Quarterly Journal of Economics 113;653–691.

Hsiao, C. (1986). Analysis of Panel Data. Cambridge University Press: Cambridge.

Ittner, C. D., Lambert, R. A. Lambert, and Larcker, D. (2003). The Structure and Performance Consequences of Equity Grants to Employees of New Economy Firms. Journal of Accounting and Economics 34;89–127.

Inderst, R. and Mueller, H. (2007). Safeguarding Specific Human Capital Investments. Working paper, 2007.

Jensen, M. C. (1983). Organization Theory and Methodology. The Accounting Review, LVII; 319–399.

Jensen, M. C., and Meckling, W. (1976). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics 3(4); 305–360.

Jensen, M. C., and Murphy, K. (1990). Performance Pay and Top-Management Incentives. Journal of Political Economy 98;225–264.

Kandel, E. and Lazear, E. (1992). Peer Pressure and Partnerships. Journal of Political Economy 100(4);801–817.

Kim, H. E. and Ouimet, P. (2007). Employee Capitalism or Corporate Socialism? Broad-based Employee Stock Ownership. Working paper, 2008.

Knez, M. and Simester, D. (2001). Firm-Wide Incentives and Mutual Monitoring at Continental Airlines. Journal of Labor Economics 19(4);743–772.

68 Journal of Applied Business and Economics vol. 14(3) 2013 Kroumova, M. K., and Sesil, J. C. (2006). Intellectual Capital, Monitoring, and Risk: What Predicts the Adoption of Employee Stock Options? Industrial Relations 45(4);734–752.

Kruse, D. L. (1993) Does Profit Sharing Affect Productivity? NBER working paper.

Lazear, E. P. (1999). Output-Based Pay: Incentives or Sorting? NBER working paper #7419.

Lazear, E. P. (2000). Performance Pay and Productivity. American Economic Review 90(5); 1346–1361.

Lazear, E. P., and Gibbs, M. (2008). Personnel Economics in Practice, 2nd ed. Wiley.

Lev, B., and Radhakrishnan, S. (2003). The Measurement of Firm-Specific Organization Capital.” NBER working paper #9581.

Lin, Y. P. (2009). Broad-Based Stock Options Program: Adopters vs. Non-adopters. Journal of Applied Business and Economics 9(3); 50–66.

Lin, Y. P, and Sesil, J. C. (2011). Do Broad-Based Stock Options Promote Organization Capital? British Journal of Industrial Relations, 49(S2).

Murphy, K. J. (2002). Stock-Based Pay in New Economy Firms. Journal of Accounting and Economics 34; 129–147.

Oyer, P. (2004). Why Do Firms Use Incentives That Have No Incentive Effects? Journal of Finance. 59; 1619–1649.

Oyer, P. and Schaefer, S. (2003). Compensating Employees Below the Executive Ranks: A Comparison of Options, Restricted Stock, and Cash. Working paper.

——— (2004). Why Do Some Firms Give Stock Options to All Employees? An Empirical Examination of Alternative Theories.” Working paper.

Robinson, A. M., and Zhang, H. (2005). Employee Share Ownership: Safeguarding Investments in Human Capital. British Journal of Industrial Relations 43(3); 469–488.

Sesil, J. and Lin, Y. P. (2011). The Impact of Employee Stock Option Adoption and Incidence on Productivity: Evidence from U.S Panel Data. Industrial Relations 50(3), 514~534.

Stevens, M. Transferrable Training and Poaching Externalities. (1996). In Acquiring Skills: Market Failures, Their Symptoms, and Policy Responds, edited by Alison L. Booth and Dennis J. Snower, 19–40. Cambridge: Center for Economic Policy Research, Cambridge University Press. 1996.

Weeden, R, E. C., and Rodrick, S. (1998). Current Practices in Stock Option Plan Design. National Center for Employee Ownership Press.

Welbourne, T. M., Balkin, D. B., and Gomez-Kejia, L. R. (1993). Gainsharing and Mutual Monitoring: A Combined Agency-Procedural Justice Interpretation. Center for Advanced Human Resource Studies (CAHRS) working paper series. Cornell University.

Wooldridge, J. M. (2001). Econometric Analysis of Cross Section and Panel Data. The MIT press.

Journal of Applied Business and Economics vol. 14(3) 2013 69

Empirical Analysis of Bicultural Border College Students’ Attitudes Toward Money

Yeong Nain Chi The University of Texas at Brownsville

Gaurango Banerjee The University of Texas at Brownsville

INTRODUCTION

Money is not only an instrument of commerce; it also has a multidimensional psychological meaning. Money also can be treated as a symbol of status, prestige, power and value [Yamauchi and Templer, 1982]. The attitudes towards use of money may be influenced by demographic factors such as age, gender, income, education, cultural background, etc. Since 2001, according to the U.S. Census Bureau, the Hispanic American population in the U.S. has grown at four times the rate of the general population. By 2020 their population is projected to be 60 million, accounting for 20 percent of the U.S. population. According to the 2010 Census, 308.7 million people resided in the United States, out of which 50.5 million (or 16 percent) were of Hispanic or Latino origin. The Mexican origin population increased by 54 percent between 2000 and 2010, and had the largest numeric change among the Hispanic groups, increasing from 20.6 million in 2000 to 31.8 million in 2010. According to the 2010 Census summary, the border city in south Texas where our study was conducted had a total population of 175,023 out of which 93.2 percent was Hispanic or Latino population. The Hispanic American population was much younger with a median age of 27.2 years compared to the median age of the overall population of 36.2 years in 2005. Hispanic Americans are also the youngest population group in the United States: approximately one-third of the entire Hispanic American population is under 18 years of age, compared with one-fourth of the total population. Hispanic American buying power in the United States continues to energize the nation’s consumer market. Hispanic Americans controlled $978 billion in buying power in 2009 according to Selig Center projections [Humphreys, 2009]. In 2009, Hispanic Americans accounted for 9.1 percent of all U.S. buying power, up from 6.8 percent in 2000 and from 5 percent in 1990. According to market research publisher

70 Journal of Applied Business and Economics vol. 14(3) 2013 Packaged Facts, "The Hispanic Market in the U.S.: A Generational View", Hispanic Americans annual buying power totals more than $980 billion [Brown & Washton, 2009]. This report [Brown & Washton, 2009] also mentioned that Hispanic Americans from ages 18-44 are particularly influential, because they control more than 60% of all Hispanic American buying power. Considering projections by the U.S. Census Bureau that place Hispanics as the largest minority group in the United States (about 60 million by 2020), understanding money attitudes among Hispanic consumers and Mexican-Americans in particular, should continue to be the focus of study by financial counseling and education providers, money managers, and professional financial planners. The main motivation of this study was to obtain a better understanding of bicultural, US-Mexico border college students’ attitudes toward money, and to educate students to appreciate the link between their money attitudes and financial behavior patterns. The conceptual framework of examining money attitudes designed by Yamauchi and Templer [1982] was employed to guide this study. Specifically, the objectives of this study were to employ the Money Attitude Scale (MAS); to measure attitudes toward money among bicultural border college business students; to identify specific clusters among the college business students that exhibit common patterns of responses; and to examine the differences among money attitude dimensions of the college business students for each variable specified, namely gender and student classification. The results of the money attitudes dimensions of bicultural border college students could be very useful in educating students about the possible impact of these attitudes on their financial behavior. This study has both theoretical and practical implications. Employing the well-developed conceptual framework of the MAS among young bicultural border Hispanic American students, this study contributes to existing money attitude research literature. The results could provide more evidence of validity and robustness of this framework or provide suggestions for modifying this framework to understand consumer groups across different cultural backgrounds. The study may provide practical financial strategies and implications for both students and business organizations by proposing effective ways to understand and address this consumer segment based on their money attitudes. The findings in our paper may be able to bridge an important gap in the financial education process of students by introducing the need to educate students to develop smart money attitudes. These attitudes may influence their financial behaviors during their working life after graduation.

LITERATURE REVIEW

Although spending and other money related aspects have been studied extensively [Wallace et al., 2005], the study of money attitudes is relatively new. Wiseman [1974] observed that psychological aspects of money suffered from lack of standardized assessment instruments. In 1982, Yamauchi and Templar developed and quantified specific money attitude scales. There have been a number of psychometrically based attempts to measure money attitudes among people in general. Yamauchi and Templer [1982] constructed the Money Attitude Scale (MAS) from an original set of 62 items, of which 34 emerged, defining five factors. More precisely, items loading on the factor for Power-Prestige pointed to the use of money as a symbol of success to impress and influence others. Items loading on the factor for Retention-Time correspond to careful spending behavior and meticulous planning of monetary resources to get a sense of security. Gutter et al. [2010] obtained results that showed significant relationships existed between financial behaviors and social learning opportunities and attitudes. Students who budgeted and saved scored higher on the social learning opportunities index score. Items loading on two of the remaining factors pertain more clearly to emotion laden aspects. The factor titled Distrust was interpreted as reflecting suspicion and doubt in situations involving money, and the other factor entitled Anxiety reflected distress and worry over money matters. Roberts and Jones [2001] found that money attitude factors, namely, power-prestige, distrust, and anxiety had a strong relation with compulsive buying and overspending. Hanley and Wilhelm [1992] used the Rosenberg’s Self-Esteem scale to support their model that compulsive spenders had lower self esteem, thereby displaying more anxiety, than ‘normal’ consumers. The fifth factor related to the consumer attitude of

Journal of Applied Business and Economics vol. 14(3) 2013 71 paying a higher price to get the desired quality. As Yamauchi and Templer dropped the latter factor, since all the items in the ‘quality’ factor had been on the original Power-Prestige dimension that was already reflected in the factor for Power-Prestige (factor 1), the final scale consisted of 29 items. Their scale has been employed in several papers [Gresham and Fontenot, 1989; Medina, et al., 1996; Roberts and Sepulveda, 1999; Yang and Lester, 2002] and has been found to have acceptable reliability. Although Furnham’s [1984] money beliefs and behavior scale (MBBS) appears more comprehensive, problems with psychometric attributes and cross-cultural issues persist [Bailey, et al., 1993; Yang and Lester, 2002]. Additionally, Tang’s [1992] money ethic scale (MES) does not include an “anxiety” dimension identified in Yamauchi and Templer’s [1982] work. As noted previously, reliability and validity of the Yamauchi and Templer [1982] instrument suggests a psychometrically sound measure. Attitudes will predict behavior when there is a high correspondence between the attitude, object and the behavioral option [Ajzen and Fishbein, 1977]. A person’s attitudes about money are influenced by culture and individual differences [Mitchell and Mickel, 1999] including personal values [Medina et al., 1996; Gbadamosi and Joubert, 2005]. Demographic factors such as family life cycle [Tang, 1993], age [Furnham, 1984; Tang, 1993; Tang and Gilbert, 1995; Roberts and Sepulveda, 1999], gender [Hanashiro et al., 2004; Masuo, et al., 2004; Roberts and Sepulveda, 1999], income [Roberts and Sepulveda, 1999], educational level [Roberts and Sepulveda, 1999], and occupation [Roberts and Sepulveda, 1999] are also important determinants of money attitudes. Roberts and Sepulveda [1999] used the MAS of Yamauchi and Templer [1982] to measure the effects of demographic factors on the money attitudes among young adults in Mexico. Using five separate regression, they showed that each of the demographic measures, mentioned above, had a significant relation to one or more of the five identified money attitude factors. Cultural background is another source of difference in people’s attitudes towards money [Medina et al., 1996; Roberts and Sepulveda, 1999; Hanashiro, et al., 2004; Masuo, et al., 2004; Burgess, et al., 2005; Özgen and Bayoğlu, 2005; Engelberg and Sjöberg, 2006; Bonsu, 2008; Fünfgeld and Wang, 2009]. Medina et al. [1996] conducted a study employing the MAS factors of Yamauchi and Templer [1982] to compare money attitudes between Mexican-American and Anglo-American consumers. Mexican- American consumers were found to have lower scores on the Retention-time factor and the Quality dimension relating to the power-prestige factor compared to the Anglo-American consumers. Peñaloza and Gilly [1986] noted the influence of cultural traits on Hispanic-American consumers by suggesting that “there may be cultural differences in the symbolic nature and perceived value of goods and services”.

METHODS

Using a relatively homogeneous group such as undergraduate and graduate students, we minimize random error that might occur by using a heterogeneous sample such as the general public [Calder, et al., 1981]. In most of the studies conducted and published in leading journals on MAS to date, non- probability sampling technique was used. Although the results of the studies conducted on such samples cannot be generalized, since the common aim of these studies were to test the transferability of the inventory to different environments, therefore, non-probability sampling techniques were found to be appropriate. This study was used to describe bicultural border Hispanic American college business students’ attitudes toward money as well as to provide an understanding of differences in money attitudes due to gender, student classifications based on year of study, and clusters based on students’ spending behavior. The 29-item MAS [Yamauchi and Templer, 1982] was chosen because the subscales on the survey represent attitudinal factors that appropriately reflect students’ attitudes towards money. Also, MAS has been used in previous research and its reliability and validity indices have been empirically documented. The questionnaire was constructed in a Likert-type scale ranging from 1 to 5 (1= strongly disagree, 3 = neither disagree nor agree, 5 = strongly agree). The survey (as shown in the Appendix) was conducted among students who enrolled in an introductory statistics, macroeconomics, and microeconomics course at a university with predominantly (more than 90 percent) Hispanic American students in South Texas during the fall semester 2009 and the

72 Journal of Applied Business and Economics vol. 14(3) 2013 spring semester 2010. The analysis examined the psychometric properties of the original 29-item MAS. First, the dimensionality of MAS was assessed by examining the factor solution followed by Yamauchi and Templer [1982]. Then, the t-test and one-way ANOVA were employed to compare gender difference and other variables specified in this study among the factors identified. Descriptive statistics of the 29- item MAS in this sample are shown in Table 1.

TABLE 1 DESCRIPTIVE STATISTICS OF HISPANIC AMERICAN COLLEGE BUSINESS STUDENTS’ MONEY ATTITUDES

Money Attitude Item Mean S.D. Communalities Power-Prestige I use money to influence other people to do things for me. 2.20 1.23 0.55 I must admit that I purchase things because I know they will impress others. 2.38 1.20 0.78 In all honesty, I own nice things in order to impress others. 2.16 1.14 0.71 I behave as if money were the ultimate symbol of success. 2.23 1.14 0.61 I must admit that I sometimes boast about how much money I make. 1.81 0.98 0.56 People I know tell me that I place too much emphasis on the amount of 1.79 1.00 0.56 money a person has as a sign of his success. I seem to find that I show more respect to people with more money than I 1.94 1.06 0.58 have. Although I should judge the success of people by their deeds, I am more 1.86 0.98 0.54 influenced by the amount of money they have. I often try to find out if other people make more money than I do. 2.21 1.16 0.35 Retention-Time I do financial planning for the future. 3.71 1.13 0.53 I put money aside on a regular basis for the future. 3.37 1.19 0.67 I save now to prepare for my old age. 2.91 1.26 0.46 I keep track of my money. 3.92 1.02 0.60 I follow a careful financial budget. 3.12 1.15 0.69 I am very prudent with money. 3.10 1.02 0.60 I have money available in the event of another economic depression. 2.76 1.19 0.51 Distrust I argue or complain about the cost of things I buy. 3.02 1.15 0.62 It bothers me when I discover I could have got something for less elsewhere. 3.82 1.09 0.72 After buying something, I wonder if I could have got something for less 3.46 1.05 0.67 elsewhere. I automatically say, “I can’t afford it” whether I can or not. 2.65 1.12 0.52 When I buy something, I complain about the price I paid. 2.46 0.99 0.67 I hesitate to spend money, even on necessities. 2.45 1.12 0.58 When I make a major purchase, I have the suspicion that I have been taken 2.63 1.13 0.65 advantage of. Anxiety It’s hard for me to pass up a bargain. 3.21 1.17 0.78 I am bothered when I have to pass up a sale. 2.89 1.19 0.77 I spend money to make myself feel better. 2.64 1.26 0.49 I show signs of nervousness when I don’t have enough money. 3.03 1.24 0.73 I show worrisome behavior when it comes to money. 2.90 1.13 0.70 I worry I will not be financially secure. 3.41 1.25 0.69

Journal of Applied Business and Economics vol. 14(3) 2013 73 RESULTS AND IMPLICATIONS

The sample consisted of 224 bicultural border Hispanic American college students majoring in Business Administration. Of the total sample, 113 (50.4%) were female and 111 were male (49.6%). The majority of respondents were Junior (n = 98, 43.8%), followed by Sophomore (n = 79, 35.3%), Senior (n = 27, 12.1%), Freshman (n = 13, 5.8%), and Graduate students (n = 7, 3.1%). Approximately 28% of the total participants reported that they had shopped at department stores at least once a week, 29% shopped once every two weeks, 26% shopped once a month, 13% shopped once every three months, and 4% shopped once a year. Approximately 39% of the total participants reported that they had shopped online once a year, 30% once every three months, 19% once a month, 5% once every two weeks, 6% at least once a week, and 1% never shopped online. Approximately 67% and 92% of the total participants reported that they had owned credit cards and debit cards, respectively. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.79 which met the fundamental requirements for factor analysis. The Bartlett’s test of Sphericity showed that nonzero correlations exist at the significance level of 0.001. Reliability of the MAS, as measured by coefficient alpha, was reported as 0.77 which was acceptable. Reliability coefficients for the four subscales of the final MAS: Power- Prestige, Retention-Time, Distrust, and Anxiety were reported as 0.84, 0.85, 0.79, and 0.75 respectively (Table 2).

TABLE 2 DESCRIPTIVE STATISTICS AND RELIABILITY OF THE FOUR IDENTIFIED SUBSCALES

Mean S.D. Sample Original Dimension Dimension Cronbach’s Cronbach’s Items Range Alpha Alpha Overall 0.77 0.77 29 Power-Prestige 18.58 6.55 0.84 0.81 9 9 ~ 45 Retention-Time 22.89 5.78 0.85 0.78 7 7 ~35 Distrust 20.47 5.08 0.79 0.73 7 7 ~ 35 Anxiety 18.07 4.86 0.75 0.69 6 6 ~ 30

In order to classify respondents through their attitudinal dimensions, cluster analysis was utilized. Cluster analysis techniques assign objects to groups so that there is as much similarity within groups, and difference between groups, as possible [Churchill, Jr., & Iacobucci, 2005]. It has been referred to as typology construction. Factor scores of the money attitude dimensions were used to cluster bicultural border Hispanic American college business students. As a cluster analysis technique, K-means cluster analysis was performed. A three-cluster solution was agreed upon. The clusters were labeled as Confident Consumers, Conscious Planners, and Careless Spenders groups (Table 3). Confident Consumers cluster is the largest group, comprising of approximately 40.6% of the respondents, named after the low mean factor scores association with Power-Prestige, Distrust and Anxiety factors, but the high mean factor score with Retention-Time factor. The ‘Confident’ consumers seem to have more self-esteem reflected in their lower scores for power-prestige, anxiety and distrust. Also, they seem to more confident based on their inclination to budget and keep track of their finances as shown on their higher scores for the retention-time factor. The second cluster, classified Conscious Planners cluster with 32.6% of the respondents, is named because of the high factor score associated with Retention-Time factor and Distrust factor among these respondents. The Retention-Time factor included items relating to financial planning for the future, saving on a regular basis, budgeting and keeping track of money. The Retention-Time factor items emphasize careful budgeting and being prudent with spending money. The Distrust factor relates to this cluster as it includes items concerning hesitation to spend

74 Journal of Applied Business and Economics vol. 14(3) 2013 money, even on necessities, looking for bargain purchases, and complaints regarding cost of things purchased. Finally, the Careless Spenders cluster is the smallest group comprising of approximately 26.8% of the respondents. These respondents put more emphasis on Power-Prestige and Anxiety factors with the high mean factor scores. The Power-Prestige factors includes items relating to careless, compulsive spending such as using money for consumption purposes in order to impress or influence other people. The Anxiety factor correlates with items emphasizing impulsive spending such as being bothered at passing up a sale or bargain, and feeling better after spending money.

TABLE 3 MEAN FACTOR SCORES OF THE FOUR IDENTIFIED SUBSCALES FOR CLUSTER

Confident Conscious Careless Median Factor Score Consumers Planners Spenders Power-Prestige 13 21 24 25 Retention-Time 25 26 16 20 Distrust 18 24 19 20 Anxiety 15 20 21 17.5 Number of Cases 91 73 60 (N = 224) Percentage 40.6% 32.6% 26.8%

Results of the cluster analysis were tested for accuracy using the multiple discriminant analysis. This can be employed as a useful complement to cluster analysis and is used primarily to predict membership in two or more mutually exclusive groups. In this case, the null hypothesis of equal population covariance matrices was rejected significantly (Box’s M = 35.738; F = 1.739; p = 0.021), and Wilk’s Lambda scores were 0.158 (χ2 = 405.111; df = 8; p = 0.000) and 0.512 (χ2 = 147.153; df = 3; p = 0.000) for both discriminant functions, respectively, indicating that group means were significantly different. The canonical correlation results were 0.831 and 0.699, supporting that there were strong relationships between the discriminant score and the cluster membership. The Table 4 presents the correlation matrix. It also analyzes the multicollinearity of the constructs. It means that constructs with correlation above ± 0.85 [Kline, 1985] can be considered the same. As shown in Table 4, no correlation above this value was found. The strongest correlation found was between Power-Prestige and Anxiety. Both the Power-Prestige and Anxiety factors include items in the Money- Attitude Scale relating to careless and impulsive spending by the college students. Lower self-esteem and impulsive spending of consumers is reflected in both power-prestige and anxiety dimensions of money attitudes. Power-Prestige was correlated positively with Anxiety and negatively with Retention-Time. This is consistent the MAS items classified under each factor. The Power-Prestige factor emphasizes more conspicuous spending in order to impress others, while the Retention-Time factor correlates with financial planning and budgeting. Power-Prestige has a low correlation with Distrust factor, since this factor is more consistent with consumers searching for bargain purchases. Similarly, Retention-Time was negatively correlated with Anxiety, while having a low positive correlation with Distrust. Retention-time emphasizes budgeting, while Anxiety factor correlates with impulsive spending habits. Both Retention- Time and Distrust factors discourage overspending, but Distrust factor does not include financial planning and budgeting which is a major characteristic of the Retention-Time factor. The Distrust factor was positively correlated to the Anxiety factor. The items in the Distrust factor such as hesitancy to spend, complaints about cost of things bought correlate positively with items in the Anxiety factor such as signs of nervousness and worrisome behavior with money matters and financial insecurity.

Journal of Applied Business and Economics vol. 14(3) 2013 75 TABLE 4 CORRELATION AMONG THE FOUR IDENTIFIED SUBSCALES

Power-Prestige Retention-Time Distrust Anxiety Power-Prestige 1 Retention-Time -0.224** 1 Distrust 0.095 0.110 1 Anxiety 0.373** -0.302** 0.332** 1 ** Correlation is significant at the 0.01 level (2-tailed)

Since one of the purposes of the study is to compare the differences in money attitudes between female and male students, the factor score of the four subscales was saved for further statistical analysis. In order to test the significant difference between the two samples, t-test is performed with the four- subscale scores. Gender only had significant differences in Anxiety at the 0.10 level, but no significant differences in Power-Prestige, Retention-Time, and Distrust. The results showed that female students felt more worrisome and more anxious with money in terms of financial insecurity than their male counterparts. Also, female students may find it harder to pass up a bargain sale compared to the male students. (t = -1.88, p = 0.061). Males had no significant differences in money attitudes than females in Power-Prestige (t = 1.37, p = 0.172); Retention-Time (t = -0.279, p = 0.780); and Distrust (t = -0.777, p = 0.439) as shown in (Table 5).

TABLE 5 GENDER DIFFERENCE WITH THE FOUR IDENTIFIED SUBSCALES

Gender N = 224 Mean S.D. P-value (2-tailed) Power-Prestige Male 111 19.19 6.60 0.172 Female 113 17.99 6.48 Retention-Time Male 111 22.78 5.92 0.780 Female 113 23.00 5.67 Distrust Male 111 20.21 5.48 0.439 Female 113 20.73 4.66 Anxiety Male 111 17.46 4.70 0.061 Female 113 18.67 4.95

In addition, a one-way ANOVA test was performed to examine the effects of student classification on the four subscales identified (as seen in Table 6). Those of significant difference were Distrust (F(4, 219) = 2.644, p = 0.035), and Anxiety (F(4, 219) = 2.115, p = 0.080); but no significant difference on Power- Prestige (F(4,219) = 0.431, p = 0.786), and Retention-Time (F(4,219) = 0.674, p = 0.611). This shows that hesitancy to spend and concerns about cost of purchases changes as the students mature from their first year of college studies to their final year of studies. In terms of conspicuous consumption to impress others (Power-Prestige factor), and financial planning and budgeting (Retention-Time factor), there were no significant differences in money attitudes among students in different student classifications based on the years of study. Also, there were no significant differences between credit card and debit card ownerships on money attitudes under the four subscales identified.

76 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 6 STUDENT CLASSIFICATION DIFFERENCES WITH THE FOUR IDENTIFIED SUBSCALES

df F P-value Power-Prestige Between Groups 4 0.431 0.786 Within Groups 219 Retention-Time Between Groups 4 0.674 0.611 Within Groups 219 Distrust Between Groups 4 2.644 0.035 Within Groups 219 Anxiety Between Groups 4 2.115 0.080 Within Groups 219

The results also showed that significant differences with department store shopping behavior were found among the four subscales (shown in Table 7). Those of significant difference was Anxiety (F(4, 219) = 2.601, p = 0.037); but no significant difference on Power-Prestige (F(4, 219) = 0.423, p = 0.792), Retention-Time (F(4,219) = 0.878, p = 0.478); and Distrust (F(4,219) = 0.638, p = 0.636). Department store shopping may be influenced strongly by anxiety in passing up a bargain or sale vis-a-vis other shoppers in the same store.

TABLE 7 DEPARTMENT STORE SHOPPING BEHAVIOR WITH THE FOUR IDENTIFIED SUBSCALES

df F P-value Power-Prestige Between Groups 4 0.423 0.792 Within Groups 219 Retention-Time Between Groups 4 0.878 0.478 Within Groups 219 Distrust Between Groups 4 0.638 0.636 Within Groups 219 Anxiety Between Groups 4 2.601 0.037 Within Groups 219

Similarly, significant differences with online shopping behavior were found among the four dimensions. Those of significant difference was Power-Prestige (F(5, 218) = 2.048, p = 0.073); but no significant difference on Retention-Time (F(5,218) = 0.845, p = 0.519); Distrust (F(5,218) = 1.562, p = 0.172), and Anxiety (F(5, 218) = 0.627, p = 0.680). Online shopping (as seen in Table 8) may be correlated strongly with the Power-Prestige dimension, as online purchases may reduce search time and allow the students to purchase items relatively faster in order to impress others. By examining how independent variables influence some patterning of response on the dependent variables, a multivariate analysis of variance (MANOVA) was employed (shown in Table 9). The independent variables studied were the identified cluster, gender, student classification, frequencies of department store and online shopping behaviors, ownership of credit and debit cards. The dependent variables considered in this study were: Power-Prestige, Retention-Time, Distrust, and Anxiety.

Journal of Applied Business and Economics vol. 14(3) 2013 77 TABLE 8 ONLINE SHOPPING BEHAVIOR WITH THE FOUR IDENTIFIED SUBSCALES

df F P-value Power-Prestige Between Groups 5 2.048 0.073 Within Groups 218 Retention-Time Between Groups 5 0.845 0.519 Within Groups 218 Distrust Between Groups 5 1.562 0.172 Within Groups 218 Anxiety Between Groups 5 0.627 0.680 Within Groups 218

Preliminary assumption testing was conducted to check for normality, linearity, univariate and multivariate outliers, homogeneity of variance-covariance matrices, and multicollinearity, with no serious violations noted. There was a statistically significant difference among three respondent cluster groups on the combined dependent variables: Pillai’s Trace =1.18, F(8, 438) = 78.745, p = 0.000; Wilk’s Lambda = 0.158; partial eta squared = 0.590. When the results for the dependent variables were considered separately, all the dependent variables reached statistical significance among these three groups. Similarly, for gender, Pillai’s Trace=0.04, F(4, 219) = 2.255, p = 0.064; Wilks’ Lambda = 0.960; partial eta squared = 0.040, for student classification, F(16, 876) = 1.159, p = 0.296; Wilks’ Lambda = 0.919; partial eta squared = 0.021, for the frequency of department store shopping behavior, Pillai’s Trace=0.111, F(16, 876) = 1.557, p = 0.074; Wilks’ Lambda = 0.8929; partial eta squared = 0.028, for the frequency of online shopping behavior, Pillai’s Trace= 0.111, F(20, 872) = 1.245, p = 0.209; Wilks’ Lambda = 0.892; partial eta squared = 0.028, for credit card ownership, Pillai’s Trace=0.006, F(4, 219) = 0.327, p = 0.859; Wilks’ Lambda = 0.994; partial eta squared = 0.006, and for debit card ownership, Pillai’s Trace=0.012, F(4, 219) = 0.648, p = 0.629; Wilks’ Lambda = 0.988; partial eta squared = 0.012. Cluster, Gender and the frequency of Department Store Shopping behavior showed statistically significant differences on the combined dependent variables, respectively, at the 0.10 level.

TABLE 9 SIGNIFICANT MULTIVARIATE EFFECTS ON THE FOUR IDENTIFIED SUBSCALES (MANOVA)

Variable Pillai’s Trace F df Error df P-value Cluster 1.180 78.745 8 438 0.000 Gender 0.040 2.255 4 219 0.064 Student Classification 0.083 1.159 16 876 0.296 Department Store Shopping 0.111 1.557 16 876 0.074 Online Shopping 0.111 1.245 20 872 0.209 Credit Card Ownership 0.006 0.327 4 219 0.859 Debit Card Ownership 0.012 0.648 4 219 0.629

CONCLUSION

As the Hispanic American population grows and matures, its structure is changing in almost every way, from educational levels and labor force composition to household characteristics and accumulation of wealth. It is these evolving factors that are driving the increasing influence of Hispanic Americans in

78 Journal of Applied Business and Economics vol. 14(3) 2013 U.S. consumer markets. Multiple studies have compared the generalizability of the money attitudes across different cultural populations. However, no systematic study has been conducted on understanding young Hispanic American consumers from a money attitude perspective and specifically on addressing these groups of consumers using the MAS approach of Yamauchi and Templer [1982]. In this regard, the factor analysis using the four subscales explained 48.24 % of the total variance in our sample, considerably more than that explained in earlier studies. The reliability coefficients (alpha) for each of the subscales in our sample was improved compared to the original Cronbach’s alpha scores. As seen in Table 2, the alphas for the four factors, namely power-prestige, retention-time, distrust, and anxiety were 0.84, 0.85, 0.79, and 0.75 respectively. It was encouraging to find that the MAS scale that was developed in the United States would retain much of its original structure when administered in a bicultural, border setting as in our paper. Three consumer segments were formed using cluster analysis on the four identified subscales in the MAS scale. The clusters were classified as Confident Consumers, Conscious Planners, and Careless Spenders groups. Statistical analyses revealed that there were significant differences among the money attitude dimensions of the bicultural, border college business students with respect to gender, student classification, identified clusters, department store and online shopping behaviors. Gender only had significant difference in the Anxiety subscale at the 0.10 level, as shown in Table 5. Female students felt more worrisome and more anxious with money in terms of financial insecurity than their male counterparts. Also, female students may find it harder to pass up a bargain sale. The result could be used to educate female students about the pitfalls of their anxiety dimension, and how such an attitude may be influencing their spending patterns. Student classification had significant differences in the Distrust and Anxiety dimensions as seen in Table 6. This indicates that inclinations to spend and concerns about cost of purchases change as the students mature from their first year of college studies to their final year of studies. Students in different levels (i.e. freshman, juniors, etc) could be enlightened about the distrust and anxiety dimensions of money attitudes, and how these dimensions may affect their financial behaviors, in terms of their inclinations to spend and the effects on their savings and net cash flows. In terms of conspicuous consumption to impress others (Power-Prestige factor), and financial planning and budgeting (Retention-Time factor), there were no significant differences in money attitudes among students based on their years of study. As shown in Table 7, Department store shopping seems to be influenced strongly only by the Anxiety factor that may arise as a result of passing up a bargain or sale to other shoppers in the same store. Students could be informed how the anxiety dimension may affect their department store shopping behavior, and have negative impacts on their net cash flows. Online shopping, as seen in Table 8, is correlated strongly with the Power-Prestige dimension, as online purchases may reduce search time and allow the students to purchase items relatively faster in order to impress others, Students could be educated on the effects of being overly influenced by the power- prestige dimension resulting in more online shopping behavior. By examining how independent variables (gender, student classification, identified clusters, department store and online shopping behaviors, credit card and debit card ownership) influence some patterning of response on the dependent variables, a multivariate analysis of variance (MANOVA) was employed. According to the results, as illustrated in Table 9, gender, identified clusters and the department store shopping behavior showed statistically significant differences on the combined dependent variables, i.e., power-prestige, retention time, distrust, and anxiety, at the 0.10 level. Future studies on money attitudes of Hispanic Americans should take into account the adult market. Furthermore, the differentiation among the dimensions could be evaluated through additional demographic variables such as age, income, educational level and occupation.

REFERENCES

Ajzen, I., & Fishbein, M. (1997). Attitude-Behavior Relations: A Theoretical Analysis and Review of Empirical Research. Psychological Bulletin, 84, 888-918.

Journal of Applied Business and Economics vol. 14(3) 2013 79 Bailey, W. C., & Lown, J. M. (1993). A Cross-Cultural Examination of the Aetiology of Attitudes towards Money. Journal of Consumer Studies and Home Economies, 17, 391-402.

Bonsu, S. K. (2008). Ghanaian Attitudes towards Money in Consumer Culture. International Journal of Consumer Studies, 32, 171-178.

Brown, R., & Washton, R. (2009). The Hispanic Market in the U.S.: A Generational View. Seventh Edition, Packaged Facts, Retrieved from http://www.packetdfacts.com

Burgess, S. M., Battersby, N. B., Gebhardt, L., & Steven, A. (2005). Money Attitudes and Innovative Consumer Behavior: Hedge Funds in South Africa. Advances in Consumer Research, 32, 315-321.

Calder, B. J., L. W. Phillips, & A. M. Tybout. (1981). Designing Research for Application. Journal of Consumer Research, 8, 197-207.

Churchill, G.A. Jr, & Iacobucci, D. (2005). Marketing Research: Methodological Foundations (Ninth Edition). Mason, OH: Thomson/South-Western.

Engelberg, E., & Sjöberg, L. (2006). Money Attitudes and Emotional Intelligence. Journal of Applied Social Psychology, 36, 2027-2047.

Fünfgeld, B., & Wang, M. (2009) Attitudes and Behavior in Everyday Finance: Evidence from Switzerland. International Journal of Bank Marketing, 27, 108-128.

Furnham, A. (1984). Many Sides of the Coin: The Psychology of Money Usage. Personality and Individual Differences, 5, 501-509.

Gbadamosi, G., & Joubert, P. (2005). Money Ethic, Moral Conduct and Work-related Attitudes: Field Study from the Public Sector in Swaziland. Journal of Management Development, 24, 754-763.

Gresham, A. B., & Fontenot, G. F. (1989) The Differing Attitudes of the Sexes toward Money: An Application of the Money Attitude Scale. Journal of Economic Psychology, 13, 5-18.

Gutter, M. S., Garrison, S., & Copur Z. (2010). Social Learning Opportunities and the Financial Behaviors of College Students. Family & Consumer Sciences Research Journal, 38, 387-404.

Hanashiro, R., Masuo, D. M., Kim, J. H., & Malroutu, Y. L. (2004). Money Attitudes and Gender Comparison Between Japanese Students and Asian American Students. Okinawan Journal of American Studies, 1, 38-45.

Hanley, A., & Wilhelm, M. S. (1992). Compulsive Buying: An Exploration into Self-esteem and Money Attitudes. Journal of Economic Psychology, 13, 5-18.

Humphreys, J. M. (2009). The Multicultural Economy 2009. Georgia Business and Economic Conditions, 69, 1-15.

Kline, M. (1985). Mathematics and the Search for Knowledge. New York: Oxford University Press.

80 Journal of Applied Business and Economics vol. 14(3) 2013 Masuo, D. M., Malroutu, Y. L., Hanashiro, R., & Kim, J. H. (2004). Do Men and Women Perceive Money Differently? A study of Asian and Asian American College Students. Papers of the Western Family Economics Association, 19, 12-23.

Medina, J. F., Saegrert, J., & Gresham, A. (1996). Comparison of Mexican American and Anglo- American Attitudes toward Money. The Journal of Consumer Affairs, 30, 124-145.

Mitchell, T. R., & Mickel, A. E. (1999). The Meanings of Money: An Individual Difference Perspective. Academy of Management Review, 24, 568-578.

Özgen, Ö., & Bayoğlu, A. S. (2005) Turkish College Students’ Attitudes towards Money. International Journal of Consumer Studies, 6, 493-501.

Peñaloza, Lisa, & Gilly, M. C. (1986) The Hispanic Family-Consumer Research Issues. Psychology and Marketing, 3, 291-304.

Roberts J. A., and Jones, E. (2001). Money Attitudes, Credit Card Use, and Compulsive Buying among American College Students. Journal of Consumer Affairs, 35, 213-240.

Roberts, J. A., & Sepulveda, M, C. J. (1999). Demographics and Money Attitudes: A Test of Yamauchi & Templer’s (1982) Money Attitude Scale in Mexico. Personality and Individual Differences, 27, 19-35.

Tang, T. L. P. (1993). The Meaning of Money: Extension and Explanation of the Money Ethic Scale in a Sample of University Students in Taiwan. Journal of Organizational Behavior, 14, 93-99.

Wallace, D. S., Paulson, R. M., & Lord, C. G. (2005). Which Behaviors Do Money Attitudes Predict? Meta-Analyzing the Effects of Social Pressure and Perceived Difficulty. Review of General Psychology, 9, 214-227.

Wiseman, J. (1974). The Money Motive. New York: Random House.

Yamauchi, K. T., & Templer, D. I.. (1982). The Development of a Money Attitude Scale. Journal of Personality Assessment, 46: 522-528.

Yang, B., & Lester, D. (2002). Internal Consistency of the Yamauchi/Templer Money Attitude Scale. Psychological Reports, 91, 994.

Journal of Applied Business and Economics vol. 14(3) 2013 81 APPENDIX (Money Attitude Scale)

This data will be used to evaluate money attitudes. Please answer each question by selecting the answer you think BEST describes your honest attitudes, beliefs, and practices. Statement Strongly Disagree Neutral Agree Strongly Disagree Agree I use money to influence other people to do things for me. I must admit that I purchase things because I know they will impress others. In all honesty, I own nice things in order to impress others. I behave as if money were the ultimate symbol of success. I must admit that I sometimes boast about how much money I make. People I know tell me that I place too much emphasis on the amount of money a person has as a sign of his success. I seem to find that I show more respect to people with more money than I have. Although I should judge the success of people by their deeds, I am more influenced by the amount of money they have. I often try to find out if other people make more money than I do. I do financial planning for the future. I put money aside on a regular basis for the future. I save now to prepare for my old age. I keep track of my money. I follow a careful financial budget. I am very prudent with money. I have money available in the event of another economic depression. I argue or complain about the cost of things I buy. It bothers me when I discover I could have gotten something for less elsewhere. After buying something, I wonder if I could have gotten something for less elsewhere. I automatically say, “I can’t afford it” whether I can or not. When I buy something, I complain about the price I paid. I hesitate to spend money, even on necessities. When I make a major purchase, I have the suspicion that I have been taken advantage of. It’s hard for me to pass up a bargain. I am bothered when I have to pass up a sale. I spend money to make myself feel better. I show signs of nervousness when I don’t have enough money. I show worrisome behavior when it comes to money. I worry I will not be financially secure. On average how often do you shop at department stores?  at least once a week  once every two weeks  once a month  once three month  once a year On average how often do you shop online?  at least once a week  once every two weeks  once a month  once three month  once a year Do you have your own credit card(s) (that is not used primarily as a debit card)?  Y e s  N o Do you have your own debit card(s)?  Y e s  N o Your gender is  Male  Female You are a  Freshman  Sophomore  Junior  Senior  Graduate Student

82 Journal of Applied Business and Economics vol. 14(3) 2013

The New Economic Reality and the Unemployment Rate: Will It Ever Get Below 5% Again?

Robert L. Howard North Carolina A&T State University

Belinda P. Shipps North Carolina A&T State University

INTRODUCTION

The unemployment rate has been at 8% or higher during most months since June 2009, when the last recession ended. According to the National Bureau of Economic Research, the U.S. experienced a recession beginning in 2007 (Isidore, 2008). This recession has been a leading cause of unemployment, which has continued to rise. During the first ten months of 2008, employers reduced jobs by 1.2 million (Isidore, 2008). In the past, the unemployment rate rose during recessions, continued to rise after the official end of the recession, but then declined substantially. According to the Bureau of Economic Research, a recession starts at the peak of a business cycle and ends at the bottom of a cycle (Business Cycle, 2012). Figure 1 shows the unemployment rate for persons aged 16 and over for 1979 – 2012. The unemployment rate at the end of the July 1981 – November 1982 recession was 10.8%. It remained at 10.8 % in December 1982, and then declined during the 1980s, reaching a low of 5.0% in March 1989 (US Business Cycle, 2012; Unemployment Rates, 2012). The unemployment rate at the end of the July 1990 – March 1991 recession was 6.8%. It then rose to a high of 7.8% in June 1992, and declined during the 1990s, reaching a low of 3.9% in September 2000. Our next recession was the period March 2001 through November 2001, which ended with an unemployment rate of 5.5%. The rate then rose to a high of 6.3% in June 2003, and declined, reaching a low of 4.4% in May 2007. Our most recent recession was the period December 2007 through June 2009, which ended with an unemployment rate of 9.5%. The rate reached a high of 10.0% in October 2009, but has declined very slowly since that time (Databases, 2012). On Labor Day, 2011 the unemployment rate was 9.0% and there were approximately 14 million unemployed persons in the country; of that number 6 million were

Journal of Applied Business and Economics vol. 14(3) 2013 83 unemployed for at least 27 weeks, and 4 million of those were jobless for over a year (Jobs in America, 2011; Nearly 1 in 3, 2011).

FIGURE 1 UNEMPLOYMENT RATE OVER TIME, PERSONS 16 YEARS AND OVER

Economists have identified four types of unemployment: frictional, seasonal, structural, and cyclical (Bade and Parkin, 2011). Frictional unemployment arises from individuals changing jobs and new workers entering the labor force. Seasonal unemployment arises because of persons not being able to work because of changes in the season. Persons who work outside and those who work in seasonal industries are affected by seasonal unemployment. Structural unemployment arises from workers’ lack of skills needed to perform jobs that are available. Cyclical unemployment arises because of changes in the business cycle; this unemployment rises during recessions and declines during expansionary periods. Unemployment rates differ significantly by industry and class of worker. Figure 2 shows the number of unemployed persons and the unemployment rates for several industries and class of workers for February 2011 and February 2012 (Unemployed Persons, 2012). The highest unemployment rates were for construction and agricultural workers, while the lowest unemployment rates were for government workers and employees in financial services, education, and health services.

84 Journal of Applied Business and Economics vol. 14(3) 2013 FIGURE 2 UNEMPLOYED PERSONS BY INDUSTRY AND CLASS OF WORKER

Number of unemployed persons Unemployment (in thousands) rates Feb. Feb. Feb. Feb. Industry and class of worker 2011 2012 2011 2012

Total, 16 years and over(1) 14,542 13,430 9.5 8.7

Nonagricultural private wage and salary workers 11,641 10,517 9.9 8.8 Mining, quarrying, and oil and gas extraction 41 66 5.6 7.2 Construction 1,883 1,404 21.8 17.1 Manufacturing 1,492 1,294 9.9 8.4 Durable goods 989 787 10.5 8.1 Non durable goods 503 506 8.9 9.0 Wholesale and retail trade 1,889 1,824 9.2 8.9 Transportation and utilities 499 440 9.0 7.6 Information 205 247 6.7 8.4 Financial activities 636 498 6.9 5.3 Professional and business services 1,469 1,590 10.1 10.3 Education and health services 1,198 1,197 5.6 5.4 Leisure and hospitality 1,783 1,496 13.8 11.6 Other services 546 461 8.9 7.1 Agriculture and related private wage and salary workers 270 290 18.4 19.5 Government workers 927 848 4.2 3.9 Self-employed workers, unincorporated, and unpaid family workers 601 579 5.9 5.9

Footnotes (1) Persons with no previous work experience and persons whose last job was in the U.S. Armed Forces are included in the unemployed total.

NOTE: Updated population controls are introduced annually with the release of January data.

Educational achievement is also highly correlated with unemployment rates. Figure 3 shows, for 2011, the unemployment rate and the median weekly earnings by highest level of education obtained, from less than high school to professional and doctoral degrees (Education Pays, 2012). As might be expected, the unemployment rate decreased and the median weekly earnings increased as more education was obtained. There was much discussion during the 2012 presidential campaign about what measures should be taken to decrease the number of jobless workers. Several changes have occurred in our economy, however, which may prevent the normal decline in the unemployment rate after our most recent recession; it appears likely that we will see high levels of unemployment for many months into the future. Our research addresses the question of what measures should be taken to decrease the number of jobless workers. We address this question by first identifying and evaluating some of the changes that are contributing to unemployment. We then provide conclusions and suggestions for future research.

Journal of Applied Business and Economics vol. 14(3) 2013 85 FIGURE 3 UNEMPLOYMENT RATE AND MEDIAN WEEKLY EARNINGS BY EDUCATION OBTAINED

This discussion is valuable to researchers as well as practitioners. In order for the country to continue to grow and prosper, there is a need for people to be employed. Prolonged high unemployment has the potential to impact many people and industries and to hamper the ability of the economy to provide products, services and jobs. There is a need to study the factors involved in the current unemployment environment as well as possible alternatives and solutions which may lower unemployment. This research is unique in addressing the new and changing nature of unemployment and the need for creative solutions to the new jobless economy. In this paper we discuss some causes of unemployment that cannot be neatly classified into the categories of frictional, seasonal, structural, or cyclical unemployment. There has not been much discussion of these issues in the literature.

FACTORS KEEPING UNEMPLOYMENT HIGH

Over the years, the rise in unemployment has been attributed to many factors and has been associated with short and long term effects on individuals, businesses and economies of one or many countries. In earlier times, severe recessions with high unemployment have shown similar patterns of unemployment recovery: after a recession ended, unemployment rates steadily dropped to more pre-recession rates and the average time frame for job recovery was ten months (Knotek and Terry, 2009). In reviewing more recent time periods and industries, the job recovery patterns for unemployment may be changing.

Companies Still Struggling from the Recession Although the Great Recession officially ended in June 2009, many firms have not fully recovered from the effects of the recession. In December 2011, Texas Instruments and Altera Corporation reduced their forecasts for the fourth quarter because of weaker demand from their customers who were concerned about the broader economy (Tibken, 2011). A survey of small businesses by American Express revealed that many firms, because of cash-flow problems and concern about the economy, reduced their year-end raises and eliminated the traditional holiday party (Maltby, 2011). CEOs of large companies are stepping

86 Journal of Applied Business and Economics vol. 14(3) 2013 gingerly because of uncertainty about the economy. The Business Roundtable, an association of CEOs of the largest U.S. corporations, reported that two-thirds of its members planned to keep their staffs steady or reduce employees during the first six months of 2012; only 1/3 were planning to add additional workers (Rugaber, 2011). The group expected a 2% growth in the economy for 2012, a rate not large enough to increase hiring substantially.

Large Firms Announcing Job Cuts The lingering effects of the recession have prompted many large firms to announce substantial job cuts for 2012 and beyond. Many of the planned cuts are at finance related firms. HSBC Holdings, a London based financial organization, is planning to cut 30,000 jobs from its international work force; some of those cuts will be in this country (Muñoz and Patrick, 2011). Bank of America also plans to eliminate 30,000 positions in a company-wide cost cutting program (BofA Layoffs, 2011). Citigroup eliminated thousands of jobs during the financial crisis, and has announced a cut of an additional 3000 jobs (Kapner, Citi Poised, 2011). Credit Suisse, Wells Fargo, Goldman Sachs, and Morgan Stanley have announced job cuts of 2,000, 1,900, 1,000, and 580, respectively (Muñoz and Patrick, 2011; Morgan Stanley, 2011). MF Global, the New York securities firm that filed for bankruptcy, recently fired all of its 1,066 employees, although some of the former employees were rehired to assist with the legal aspects of the bankruptcy (Failed Trading Company, 2011). Small banks have not been immune from the recent job cutting. In the third quarter of 2011, small and midsize banks (defined as banks with assets of $10 billion or less) eliminated 2,714 jobs; 605 more of these banks cut jobs than added jobs (Fitzpatrick and Barry, 2011). Disappointing 2011 holiday sales led Sears Holdings Corporation, which owns Sears and K-Mart, to announce that it would close as many as 120 stores in the near future. The company did not indicate the number of employees that would lose their jobs, but the number will probably be in the thousands. The number of store closings represents about 5% of its full-size stores, and there are approximately 250,000 employees (Bustillo and Zimmerman, 2011). Another firm announcing the closing of stores is Lowes, the home improvement giant. It is closing 20 stores and will open only half of the new stores it had planned for the coming years. There will be approximately 1,950 job terminations (Lowe’s Closing, 2011). Gap, Inc., the large specialty retail chain, will close 189 of its 889 stores by the end of 2013; the number of job losses was not announced (Mattioli and Hudson; 2011). Whirlpool, the world’s largest appliance manufacturer, is closing two factories and eliminating 5,000 jobs. The move is expected to reduce annual costs by about $400 million. Electrolux, the world’s second largest appliance manufacturer, is also reducing production capacity. Both firms have suffered from weak consumer demand as individuals are reluctant to spend on big-ticket items (Hagerty and Tita, 2011). To save more than $200 million in operating costs in 2012, Advanced Micro Devices announced that it would cut about 1,400 jobs and place the savings in new business areas. High-level managers as well as lower level employees were affected (Tibken and Clark, 2011). Altria, the country’s largest cigarette maker, announced that it would cut 15% of its cigarette- related salaried work force, although it did not specify the number of jobs that would be lost. The job cuts are in response to a steady decline in cigarette sales which followed an increase in April 2009 of the federal excise tax (Companies Flourish, 2011). Other large manufacturers that have announced job cuts include Novartis, the Swiss pharmaceutical company (2,000 jobs), Nokia Siemens, the Finnish-German wireless equipment maker (17,000 jobs), and Gannett, which publishes USA Today (700 jobs) (Mijuk and Whalen, 2011; Nokia, 2011; Adams, 2011). In contrast to the grim statistics on job cuts, there is one major area that is experiencing significant expansion: the automobile industry. Ford, General Motors, and Chrysler are all profitable, auto sales have been increasing annually since 2009, and additional employees have been added to meet the increased demand. The government bailouts of General Motors and Chrysler in 2009 saved the domestic automobile industry, and it is now outperforming most other areas of the economy (Auto Industry, 2011).

Journal of Applied Business and Economics vol. 14(3) 2013 87 Low Rate of Job Creation There must be a high level of job creation if the unemployment rate is to be reduced substantially. The economy lost more than 5 million jobs in 2009. In 2010, 940,000 jobs were created, and there were 1.6 million new jobs in 2011 (Wiseman and Rugaber, 2012). Figures from the Bureau of Labor Statistics at the end of 2011 indicated that there were approximately 14 million unemployed workers, but only 3.3 million job openings. Thus, the number of unemployed workers to job openings was greater that 4-to-1, indicating that there were no jobs available for more than three out of four unemployed workers. This ratio has remained over 4-to-1 for almost three years (Shierholz and Finio, 2011); before the recession it was only 3-to-2 (Wessel, 2011). It has been estimated that, considering the growth in the working age population, it will require adding 275,000 new jobs each month for 5 years to bring the unemployment rate down to where it was when the Great recession began (Shierholz, 2011). Considering the issues discussed in this paper, this level of job growth is not likely.

Companies with Cash, but Not Expanding There is an assumption among many politicians that a drop in the corporate tax rate would spur employment: a tax reduction would give firms more money to spend, they could expand their operations, and then hire more workers to produce more products. The reality today is that many firms already have excess cash and rising profits, but are not spending that cash to expand operations and increase hiring. Citigroup, for example, reported earnings of $3.8 billion in the third quarter of 2011, a rise of 73% from the $2.2 billion reported in the third quarter of 2010 (Kapner, Citi Shines, 2011). This large profit increase, however, has not increased Citigroup’s hiring. Actually, Citigroup announced plans to cut 4,500 jobs over the next few quarters because of concerns about worldwide financial markets and new regulations (Kapner and Matthias, 2011). Honeywell International also had outstanding third quarter, 2011, results: its profits increased from $596 million in the third quarter of 2010 to $863 million in the third quarter of 2011, a 45% increase. However, in an interview with The Wall Street Journal, Honeywell’s CEO David Cote indicated that although the company is generating cash, he is cautious about bringing on additional employees in the prevailing economic atmosphere (Linebaugh, 2011). What do firms do with their excess cash when it is not used to expand capacity? They can use it to increase dividends to stockholders and buy back shares of the firm’s common stock. Pfizer, Inc., the large pharmaceutical firm, had about $3.7 billion in cash and cash equivalents plus $25.3 billion in short-term investments, which can readily be converted into cash, in October, 2011. It was announced that this excess cash would be used to finance a 10% dividend increase to stockholders and a stock repurchase plan of up to $10 billion. This new repurchase program is in addition to the $6.5 billion of shares repurchased in 2011 (Loftus, 2011). But Pfizer also announced plans to lay off 16,500 employees because of an expected drop in sales of its best-selling cholesterol product, Lipitor (Edwards, 2011) because the patent for Lipitor expired on November 30, 2011 (Countdown, 2011). McGraw-Hill Companies, Inc. also announced that it will use some of its excess cash to implement a share buyback program, and it will cut jobs. About 550 positions, including both executive and lower level personnel, will be cut, and $1.5 billion will be used to repurchase shares of the firm’s outstanding common stock (McGraw-Hill, 2011). Campbell Soup Company and Best Buy Company, Inc. are two other well-known companies that have recently announced share repurchase programs ($1 billion and $5 billion, respectively) (Rougemont, 2011). Using excess cash to repurchase outstanding shares, rather than expanding operations, has become a common practice for companies today. Companies in the SandP 500 Index spent a total of $109.2 billion on stock buybacks during the second quarter of 2011 and $118.4 billion during the third quarter; the expectation is that over $120 will be spent during the fourth quarter (SandP Indices, 2011). It is apparent that companies are skeptical of putting additional funds into productive facilities in the current economic climate. Until their perception of the economic outlook changes, excess cash will not lead to additional hiring.

88 Journal of Applied Business and Economics vol. 14(3) 2013 Companies Expanding But Not Hiring Some companies have found it profitable to expand operations in the current economy but many of them have not hired additional workers. The wireless industry, for example, has shown rapid growth over the past 5 years as more consumers use smartphones, wireless applications, and network technology. Revenue in the industry has grown 28% since 2006 when employment in the industry peaked at 207,000 employees, but productivity gains, consolidation, and outsourcing have led to a decline of 20% of workers in the industry over the past 5 years. Sprint Nextel Corporation has decreased its number of call centers from 74 in 2007 to 44 in 2010, with a corresponding drop in workers from 60,000 to 40,000. ATandT, Inc. and Verizon Communications, Inc. have kept their number of employees relatively constant over the past few years, but their revenues increased from $100 billion in 2008 to $122 billion in 2010 Troianovski, (2011). Some jobs have been created in other industries as a result of the wireless expansion, but those numbers do not match wireless job losses. Exxon Mobil Corporation, the world’s most profitable company, reported third quarter income of $10.33 billion in 2011, an increase of 41% from 2010; revenue increased by 32% (Ordonez, 2011). The firm has reported huge profits in other quarters as shown in Figure 4 (Why Tax Cuts, 2009):

FIGURE 4 TOP 10 QUARTERLY EARNINGS OF ALL TIME (PRIOR TO 2011)

1. Exxon Mobil Corp: 2008, 2Q $11.68 billion 2. Exxon Mobil Corp: 2007, 4Q $11.66 billion 3. Exxon Mobil Corp: 2008, 1Q $10.89 billion 4. Exxon Mobil Corp: 2005, 4Q $10.71 billion 5. Exxon Mobil Corp: 2006, 3Q $10.49 billion 6. Exxon Mobil Corp: 2006, 2Q $10.36 billion 7. Exxon Mobil Corp: 2007, 2Q $10.26 billion 8. Exxon Mobil Corp: 2006, 4Q $10.25 billion 9. Exxon Mobil Corp: 2005, 3Q $9.92 billion 10. Exxon Mobil Corp: 2007, 3Q $9.41 billion

However, job growth over the years has not accompanied these profits, as shown in Figure 5:

FIGURE 5 NUMBER OF EMPLOYEES AT EXXON MOBIL

Year Employees at Exxon Mobil 1999 106,900 2000 99,600 2001 97,900 2002 92,500 2003 88,300 2004 85,900 2005 83,700 2006 82,100 2007 80,800

Thus, Exxon has been able to expand with fewer workers. There are many other exceptions to the notion that increased expansion lead to increased hiring in this economy. In North Carolina there is a new toll road in the Raleigh-Durham-Research Triangle Park area. However, there are no toll booths on this

Journal of Applied Business and Economics vol. 14(3) 2013 89 road and thus no toll collectors and minimal job creation; tolls are collected electronically (Free Rides, 2012). Drivers can set up an account with the North Carolina Turnpike Authority and install on the car windshields a N.C. Quick Pass electronic transponder that has a customer ID number. Overhead frequency readers communicate with the transponders and deduct tolls from a prepaid account. Overhead cameras will take photos of the license plates of cars that do not have the transponders, and the owners will be sent a monthly bill (Toll Operations, 2012). In a speech in Kansas, President Obama noted the disconnect between production and employment: “Steel mills that needed 1,000 employees are now able to do the same work with 100 employees, so layoffs too often became permanent, not just a temporary part of the business cycle” (Friedman, 2011). Economists refer to this situation as a jobless recovery: companies are able to expand without hiring additional workers (Knotek and Terry, 2009).

Fundamental Change in Labor Markets Economists have noticed a fundamental change in labor markets since the early 1980’s. In the recession of July 1981 – November 1982, layoffs tended to be temporary – employees who were laid off generally returned to their old jobs when economic conditions improved. Also, in the early stages of the recovery, companies hired permanent workers to meet the expected need for increased production (Knotek and Terry, 2009). In the recession of December 2007 – June 2009, layoffs tended to be permanent in nature – workers were forced to look for new jobs at other companies, a process that resulted in longer periods of unemployment and contributed to long-term unemployment experienced by many workers. Also, during the recovery following the most recent recession, companies that needed more labor tended to hire fewer permanent employees; they met their needs by having their current employees work overtime and hiring part-time and contract workers.

Unemployment Compensation Some economists argue that extending unemployment benefits, which are given to unemployed workers to help them to survive temporarily from loss of income, can prolong unemployment. The argument suggests that people receiving government assistance from unemployment pay are less willing to work as long as the benefits continue (Bade and Parkin, 2011).

ADDITIONAL FACTORS KEEPING UNEMPLOYMENT HIGH

There are several other factors that have changed since the recession of July 1981 – November 1982 that are contributing to a high unemployment rate.

Business Shifted Abroad To take advantage of lower labor costs, companies have been shifting production to Mexico, China, and other low wage countries. Because of weak demand for its appliances, Whirlpool is cutting 5000 jobs and closing a plant in Arkansas, which manufactures refrigerators. Whirlpool is shifting the production of refrigerators to its plant in Mexico (Hagerty and Tita, 2011; Smith, 2011). In recent years, other companies, including Ford, General Motors, General Electric, Coca Cola, and RCA have opened plants in Mexico. General Electric employs 30,000 employees in its 35 Mexican plants. These moves, of course, have cost thousands of American jobs (Ensinger, 2010). Companies do not have to open plants abroad to transfer American jobs. Many companies outsource services such as payroll, call centers, and information technology to other countries. President Obama noted in his Kansas speech that, “Today, even higher-skilled jobs, like accountants and middle management, can be outsourced to countries like China and India” (Friedman, 2011). When our economy improves to the point where firms decide to expand and hire additional workers, many of those new jobs will be located in foreign countries and will have minimal impact on our unemployment rate.

90 Journal of Applied Business and Economics vol. 14(3) 2013 Decline in New Business Formation It has often been noted that major job growth in the economy comes from small firms and new business startups. According to the Small Business Administration, small businesses – firms with fewer than 500 employees – provide jobs for over half of the nation’s workforce. They create more than 50% of the private, non-farm gross domestic product, and they create between 60% and 80% of the nation’s net new jobs (Longley, 2012). A study at the Federal Reserve Bank of Cleveland evaluated several measures of entrepreneurship over the past few years. It found that the number of businesses in the country reached a peak in early 2005 and then began to decline. After the recession began in December 2007, the decline was magnified. Some of the decline was due to business failures, but a larger portion of the decline was because of a decrease in new business formations. It was noted that “68,490 more businesses closed in 2009 than in 2007, an 11.6% increase in the business closure rate. But by 2009, 115,795 fewer employer businesses were founded than in 2007, a 17.3% decline in firm formation” (Rampell, 2011). Since most new businesses are small, and since small business drives job formation, one can conclude that hiring will remain depressed until the rate of new business formation improves.

Accelerated Pace of Automation The history of capitalism has been to use new technologies to improve productivity and replace workers. Farm machines have eliminated thousands of farm jobs, automatic teller machines have reduced the need for bank tellers, and self-service gas stations have reduced the need for service station attendants. In the past, new technologies, new products, and new ideas combined to create new jobs that replaced those that were lost. In recent years, however, the pace of automation has accelerated to the point where the number of new jobs created is less than the number of jobs eliminated. In their book Race Against the Machine, the authors Erik Brynjolfsson and Andrew P. McAfee argue that workers are losing the battle against machines because the pace of automation has increased in recent years. Clever software, faster computers, robotics, voice recognition technologies, computerized inventory control, and online commerce have enabled corporations to substitute technology for workers at a faster pace (Lohr, 2011). The result has been more spending on equipment (since the end of the recession in June 2009, spending on equipment increased 26%), increasing profits (corporate profits as a share of the economy is at a 50- year high), and flat payrolls. The capability of machines will certainly improve in the future, decreasing further the need for workers.

Companies' Unwillingness to Hire Those Already Unemployed A disturbing trend has developed concerning unemployed workers: many firms that are hiring are refusing to consider applications from persons that are currently unemployed. A survey by the National Employment Law project found over 150 postings on employment web sites requiring that applicants must be currently employed. Job applicants from around the country have been told that they have been out of work too long, and would not be considered for a position that they would otherwise be qualified for (Unemployed Seek Help, 2011). This issue has become particularly important because of the rise in the number of persons who have been unemployed for longer that one year. The number of persons unemployed for over a year rose from 645,000 in the second quarter of 2007 (9.5% of the total unemployed) to 4.5 million in the second quarter of 2010 (30.9% of the total unemployed) (Office of Publications, 2010). It is possible that firms’ preference for hiring persons already employed or with a short duration of unemployment may have contributed to this increase. If this trend becomes more widespread, it will become more difficult to bring down the unemployment rate.

Stalemate in the Present U.S. Congress There is massive political gridlock in the current U.S. Congress; the two sides have very different views on what should be done to improve the economy, and they find it difficult to compromise (Austerity, 2011; Von Drehle, 2012). One side believes that taxes should be reduced, government spending should be decreased, health care reform should be repealed, government regulations should be

Journal of Applied Business and Economics vol. 14(3) 2013 91 reduced, entitlement programs (social security, medicare, and Medicaid) should be substantially modified to reduce costs, and the Recovery and Reinvestment Act of 2009 (the stimulus package) did not work since there was no boom in employment. The other side believes that taxes should be increased on the wealthy, a decrease in government spending would hamper economic growth and shrink the safety net the government provides, health care reform should be modified and improved, government regulations should be kept in place to protect consumers and prevent another financial crisis, entitlement programs should be evaluated and revised in a manner that protects their long-term solvency, and the stimulus package was not large enough considering the slump we were in. In the summer of 2011 there was a need to increase the national debt limit so that the government could continue to borrow funds needed to meet its financial obligations. In previous years, under the Bush, Clinton, Reagan, and other administrations, the debt ceiling was lifted with little or no partisan debate. On this occasion, however, there was passionate debate on the national debt, with each side unwilling to yield on what it considered to be fundamental principles. An agreement was reached just hours before the nation would have defaulted on its debt. Because of the difficulty in reaching an agreement, the outstanding debt of the U.S. government was downgraded by the credit rating agency Standard and Poor’s (White House, 2011). One of the terms of the debt agreement was to reduce Medicare payments to health care providers by $3.87 billion, a move likely to result in job losses and poorer care for the nation’s seniors (Korn and Kamp, 2011). Business leaders are concerned about the gridlock in Washington and its potential impact on credit markets. Many of them are uncertain about the desirability of committing investment funds until there is a clearer picture of the direction of government policy. A cutback in government spending will reduce aggregate demand, and it will reduce the revenue of firms that had anticipated selling products or services to the government. These firms will probably curtail their hiring or reduce their staffs.

Expenditure Cutbacks at the Federal, State, and Local Levels According to a survey released by the National League of Cities, 2011 was the fifth consecutive year of declining revenues at cities around the country (More Cities, 2011). Second quarter revenues at the state level were higher than those during the same period in 2010, but still lower than state revenues collected in the second quarter of 2008 (Dougherty, 2011). In most cases, these revenues have dropped because the main sources of revenue for states and local governments, income taxes, sales taxes, and property taxes, have fallen. Income taxes are lower because of the high unemployment rate, sales taxes are lower because of the drop in consumer spending, and property taxes are lower because of the drop in housing prices. In other cases, revenues are down because state and local legislatures have deliberately cut taxes. Many state and local governments around the country have adopted the “cut taxes, cut spending” policy. The result has been significant job losses and cuts in programs and services that had previously been considered desirable. Almost a third of cities in the country laid off workers in 2011 and more than half cancelled or postponed infrastructure projects (More Cities, 2011). Persons who would have taken jobs to work in these projects were not hired. At the state level, over 200,000 state employees lost their jobs in 2011 (Dougherty, 2011). In the state of North Carolina, a temporary sales tax increase was scheduled to expire in 2011. It had been implemented during the Great Recession and it had produced more than $1 billion annually during each of the previous two years. The state legislature could have extended the tax, but is chose to let it expire. Instead, it passed a budget that required state agencies to cut their budgets by $1.3 billion from levels the agencies considered necessary to continue their operations at current levels (Democrats Begin, 2011). According to the state Department of Commerce, these budget cuts resulted in a loss of 11,200 government jobs in 2011 (Behind on Jobs, 2011). According to the state Department of Public Instruction the school systems in North Carolina cut more than 2,400 positions (Binker, 2011). At the university level, the budget of the University of North Carolina system was cut by $414 million, resulting in a loss of 3,000 jobs (Davis, 2011).

92 Journal of Applied Business and Economics vol. 14(3) 2013 These job losses at the state and local levels are likely to continue and will hamper the nation’s ability to decrease its unemployment rate.

Defense Cutbacks and Returning Veterans from Iraq and Afghanistan The U.S. has withdrawn its troops from Iraq and is decreasing its presence in Afghanistan. Secretary of the Navy Ray Mabus recently indicated that the Navy and Marine Corps will reduce their numbers as these wars end. He did not specify an exact number of the reductions, but as these returning veterans enter the labor force, they will add to the number of persons looking for work in the civilian marketplace and may increase the unemployment rate (Official Says, 2011). Much more ominous are proposals being considered by a Congress that is predisposed to budget cuts rather than revenue increases or debt increases. The largest component of the federal budget is national defense and is a prime target for budget cutting. Throughout the years our number one priority has been to maintain a military presence capable of protecting our interests around the world. Some feel that this mission will be seriously compromised if substantial military cutbacks take place (McKeon, 2011). A bipartisan subcommittee (consisting of three Senate Democrats, three Senate Republicans, three House Democrats, and three House Republicans) was appointed to evaluate the growing federal deficit. The subcommittee failed to reach an agreement on how to reduce the government deficit by at least $1.2 trillion over the next decade. As a result automatic across the board spending cuts were scheduled to take place beginning in 2013 (Bendavid and Hook, 2011). The plan was to cut $500 - $600 billion from the defense budget and $500 – $600 billion from other areas of the federal government over the next decade. It has been estimated that these budget cuts and their secondary effect would destroy up to 800,000 active duty, civilian, and industrial jobs (McKeon, 2011). Actually, the job cutting has already begun, since the cuts will have to take place gradually over a period of time. The Army recently announced plans to cut about 8,700 positions from its civilian workforce by the end of September 2012. It hopes to meet its goal through a combination of early retirement offers, job buyouts, and attrition; however, layoffs will be used if necessary (O’Keefe, 2011). To implement the congressionally mandated cuts, it is the responsibility of the administration to draw up the specific plans. President Obama recently announced the initial components of the plan, which call for a delay in the Pentagon’s most expensive weapons, a 14% reduction in the Army’s troops, a reduction in our nuclear arsenal, and a total cut in military spending of $487 billion over 10 years. Full details of the reductions were presented in the Pentagon’s annual budget (Barnes and Hodge, 2012).

Spillover from Europe's Financial Problems The European Union (EU) is the largest trading partner of the United States. The U.S. and Europe exchanged $475 billion in goods during the first nine months of 2011 (Rugaber et al., 2011). Revenue for the 500 largest U.S. companies, totaling approximately 14% (1.3 trillion) comes from Europe. Therefore, the financial crisis in Europe is closely linked to the U.S. The U.S. is watching Europe very cautiously because Europe’s high level of government debt and collapsing financial institutions can impact the decision making and the profit and resources of U.S. firms, government, consumers and individual investors (Rugaber, 2011). The European Crisis is affecting the U.S. in the following ways (Rugaber, 2011):

•Stock-market fluctuations affect consumer spending •Exports to Europe impact sales, prices and profits •U.S. Banks worldwide are reducing lending; many have outstanding loans with Europe •Financial crisis uncertainty in Europe leads to cautious reductions in hiring and investing

Many companies such as General Motors are feeling the pain of the crisis in Europe. The general manager for General Motors stated that the European financial crisis was a more serious problem for General Motors than the financial crisis of 2007 (Trudell, 2011).

Journal of Applied Business and Economics vol. 14(3) 2013 93 General Motors’ Challenges Relating to the European Financial Crisis (Trudell, 2011): • No annual profits in Europe for more than ten years • Decline in New York trading • European operations lost $292 million before interest and taxes in the quarter ending 9/30/2011 • 2.5 percent drop in third-quarter, 2011, net income • GM had $900 million in restructuring and early-retirement costs in Europe and cut 5,800 jobs in 2010 • Additional layoffs took place in 2011 in Europe

Other American companies are having similar types of problems in terms of higher operational costs and lower sales and profits. The impact of the European financial crisis is causing American firms to resist hiring new workers and to cut their current labor supply by laying workers off. This impact on the US economy contributes to the jobless recovery from the recession.

Banking Crisis The banking crisis of 2007 has contributed significantly to unemployment in the U.S (Uchitelli et al, 2008). There are many factors that have been linked to the financial crisis and unemployment such as excessive and risky consumer and corporate lending and borrowing, and excessive sub-prime mortgage lending. Morgan Stanley, Citigroup Inc., and Bank of America Corporation are good examples of lending institutions that financed large numbers of risky loans before 2007. By 2008, when the housing bubble burst, these same companies took billions of dollars in emergency funds from the U.S. Federal Reserve to get them out of financial trouble (Keoun, 2011; Trudell, 2011). Other companies that received federal help to save their companies and alleviate unemployment were Chrysler and General Motors. In many situations, the result of bad debt and unhealthy consumer and corporate spending led to increased bankruptcies and massive job layoffs. The relation between banking crises, recession, and unemployment were evaluated in a study by the Federal Reserve Bank of Kansas City (Knotek and Terry, 2009). The conclusion was that recessions coupled with banking crises are associated with high levels of unemployment and slow declines in the unemployment rate following the recessions.

Real Estate Bubble The unprecedented real estate bubble has had a major impact on unemployment. Falling housing prices and mortgage foreclosures have made it impossible for many families to sell one property and purchase another, resulting in a drop in existing home sales and new home sales. Many individuals employed in the housing industry have lost their jobs as a result of the drop in sales and drop in new construction. There are millions of individuals who work in the housing industry; occupations include real estate agent, appraiser, mortgage lender, property insurance agent, title insurance agent, mortgage lender, pest inspector, home inspector, contractor, builder, carpenter, brick layer, plumber, air conditioning installer, painter, drywall installer, carpet and tile installer, graders, and landscape architect. The employment of persons in these occupations is dependent on a vibrant housing market. One indication of the severity of the problem is the number of housing starts during and following the recession. The number of housing starts during and following the July 1981-November 1982 recession was: 1981 – 1,084,200; 1982 – 1,062,200; 1983 – 1,703,000; 1984 – 1,749,500. In contrast, the number of housing starts during and following the December 2007 – June 2009 recession were: 2007 – 1,355,000; 2008 – 905,500; 2009 – 554,000; 2010 – 586,900; 2011 – 609,200.

Job Losses Because of the Internet Perhaps the greatest factor keeping the unemployment rate high is job losses because of the Internet. As the Internet continues to grow, it provides many e-business opportunities for startups as well as existing businesses. However, it also provides many opportunities for job loss as the Internet becomes a

94 Journal of Applied Business and Economics vol. 14(3) 2013 more robust player in job loss and unemployment. In the following sections we present evidence of several instances of job loss because of the Internet.

U. S. Postal Service The U.S. Postal Service is an agency in financial disarray; it has had large losses in recent years primarily because of drops in revenue. The Internet is primarily to blame: individuals pay bills on line, taxes are largely paid on line, individuals send email and text messages rather than send letters, social security checks and other payroll checks not sent by mail - direct deposit is used. It has been estimated that the amount of first class mail declined by 19% over the last decade, and it is expected to fall by another 37% over the next ten years (Garvin, 2011). Even those who use the mail can purchase their stamps online. Since about 80% of the postal service’s budget goes to employees in the form of salaries and benefits, any cost cutting to reduce the agency’s losses will require significant job cuts (Garvin, 2011). Over the last four years 110,000 jobs have been cut, 7,500 administrative staff positions have been eliminated, and the agency is seeking congressional approval to cut an additional 120,000 positions and close 3,653 post offices (Schmid, 2011; Levitz, 2011).

Book Stores Borders, the second largest bookstore in the nation, filed for bankruptcy and closed 237 of their 642 stores, and 11,000 jobs were lost in February 2011 (Barror, 2011). All of their stores were closed by the end of 2011. One factor blamed on their closing is the increased use of downloadable e-books (Barror, 2011). Bookstores are experiencing major challenges as Internet downloading of digital books to e- readers and electronic books (E-books) continue to increase in popularity. According to the Association of American Publishers, e-book sales have doubled since 2010 and make up 9% of total consumer book sales. E-readers, such as the Apple Ipad; Sony Reader; Amazon Kindle and Barnes and Noble Nook, are devices that allow Internet access to e-books and other reading materials quickly and easily. With e-readers, customers do not need to go to the bookstores to purchase their books or get help from the sales staff. In many cases, they are buying at reduced costs by cutting out the middleman. The use of the e-reader and the Internet to access digital books is eliminating the need for many employees working in the bookstores (Barror, 2011). As the number of e-book downloads are increasing, disintermediation is leading to job losses for employees.

Record Stores The Internet and the use of digitized music have had a major impact on the basic operating principles of the record industry. Increased use of the Internet for digital music has contributed to online music piracy, the decline of purchasing music CDs and DVDs and the decline of the traditional record label business. As digital music downloads to digital electronic devices continue to escalate, CD in-store sales continue to decline (Bachman, 2007). According to the U.S. census data there was a net loss of 1,900 record stores from 2002 to 2005. In other words, for every four record stores, one record store closed (Tozzi, 2008). As music devices and digitized music increased, the Internet provided easy access to the music through downloads. Since 1996 when illegal downloading of music gained major attention, the retail music industry has been steadily declining (Ryan and Hadida, 2010). Users were able to use different websites such as Napster to access music for free. The music was accessed and delivered directly over the Internet, which eliminated the need for the intermediaries in the physical record stores. Customers were able to select which music they wanted rather than being limited by the selections on the shelf. Thus, convenience, cost and accessibility helped to diminish the value of the record store employees. Physical distribution of CDs dropped to 21 percent in 2009 (Ryan and Hadida, 2010). Many employees in the music industry are no longer needed when customers are able to obtain their music without the help of music employees. This is another case of disintermediation leading to job loss.

Journal of Applied Business and Economics vol. 14(3) 2013 95 Attorneys Attorneys are also in danger of losing some of their business because of the Internet. Attorneys answer legal questions for their clients, provide legal advice, prepare documents, negotiate settlements, and represent their clients in court. Some of these functions can be performed online, without a face-to- face meeting with an attorney. Legal questions can be answered, advice can be given, and simple forms can be completed and sent by email to the client. There is a website, www.justanswer.com, which provides experts to answer legal questions that one might have. For a nominal fee, one can ask a legal question; an expert (an attorney with several years experience) will obtain as much information from the client as possible concerning the issue, and provide an answer and give advice when needed. The experts are rated by individuals who submit questions, and the ratings generally range from 95% to 100% satisfaction. There will still need for attorneys to negotiate, prepare complex documents, and represent their clients in court, although there is a company that provides a “Complete Case-Winning 24-hour Self- Help Course” that promises to teach one how to win in court without a lawyer (Graves, 2011). Access to the course is delivered by email for a cost of $249.00. There will be some persons who will use this course to avoid hiring a lawyer, and others who will get their questions answered through the Internet. Legalzoom.com also provides some services previously provided only by attorneys. Thus, there will likely be some decrease in the need for attorneys in the future because of the Internet.

Banking The Internet is having a substantial impact on employment in the commercial banking industry. The need for bank branches at traditional banks will be decreased because more and more deposits are being made online, payments are being made online, and cash can be obtained from automated teller machines. Brokerage firms Charles Schwab and E*Trade offer online banking services without brick and mortar banking locations, and there are several Internet only banks; ING Direct, Ally Bank, FNBO Direct, and HSBC Advance are several examples. As branch banking declines, there will be declines in the need for tellers, head tellers, branch managers, janitors, security guards, and other branch related personnel.

College and University Faculty Positions There is a cost squeeze in college education today, and the public anxiety is becoming more evident. David Shi, a former president of Furman University, noted that college costs have increased 50% over the last decade, whereas family incomes actually fell between 2000 and 2009 (Fischer, 2011). In a survey conducted by the Pew Research Center, 75% of those polled felt that college education was out of reach of most families because of the cost. In this year of cost cutting, state legislatures around the country have not been much help; they have been cutting higher education budgets at a time when enrollments are increasing. The Internet has an answer to this problem: cut costs by eliminating faculty positions. A continuation of events occurring today will lead to a reduction of thousands of faculty positions. What are some of these events? An online course on artificial intelligence is being offered by the computer science department at Stanford University by two leading experts in artificial intelligence. There is no cost for the course, and over 58,000 students from around the world have registered for it. This course is one of three that Stanford is offering on an experimental basis (Markoff, 2011). Technical assistance from Google and Amazon will provide for grading and online chat sessions related to the course (Chu, 2011). Consider the cost savings (and job reductions) if other courses in the computer science department were prepared by experts and offered in a similar fashion; video tutorials could be prepared and students could review them at their leisure. Students from other universities around the country could be directed by their schools to take the courses and thus eliminate the need for those schools to offer them. Courses in other disciplines could also be prepared by experts in those fields for online delivery, and thousands of faculty positions could be eliminated. Of course some faculty members in each department at each university would be retained to teach those courses that are not amenable to the online process and to

96 Journal of Applied Business and Economics vol. 14(3) 2013 answer questions raised from the online courses. The loss of faculty positions would also likely result in the loss of ancillary positions and consolidation of some departments. Education online is rapidly growing field; most universities offer some courses online, entire degrees at some traditional universities can be earned online, and there are several online universities. Students can order, pay for, and track transcript requests online, and at least one university, the University of Hawaii, has held a virtual graduation (Gutierrez, 2011). Job losses will be huge as the use of the Internet increases on the college campus. The quality of course offerings may decline, but state governments will save billions of dollars.

Secondary School Teachers At the secondary school level, use of the Internet will also lead to job losses. A forecast of things to come occurred in Guilford County, North Carolina during the summer of 2011. Most school districts offer summer school classes for students who failed a course during the regular school year and wish to make it up, or who wish to get ahead. For the first time ever, no summer school classes were offered in the typical classroom setting with a teacher; instead, all of the classes were offered on-line. Thus students could complete their work at home, at the library, or anywhere else where they could access the Internet. This Internet-only approach enabled the schools to reduce the summer school budget by over $350,000; of course this savings meant fewer jobs for classroom teachers and other school personnel who might be needed to assist the students (Glover, 2011). Online learning has been expanded substantially beyond summer school offerings. In an article in The Wall Street Journal, it was noted that 39 states have established virtual schools that allow students statewide to enroll, providing advanced placement, remedial, and other courses that might not be available at the local level (Moe, 2011). The Florida Virtual School offers a full academic curriculum, with more than 220,000 annual course enrollments, and virtual charter schools operate in 27 states with a full-time enrollment of over 200,000 students (Moe, 2011). Online teaching will certainly increase, thus reducing teaching jobs. In each case mentioned above, the direct use of the Internet by an individual eliminated the need for a middle person. Organizations are looking for ways to reduce costs and improve efficiencies and allow customers more direct, 24/7 access to information. As illustrated, the Internet provides many opportunities to perform this task. New ways of cutting out the middle man and using the Internet are rapidly emerging where more jobs will be eliminated. This form of unemployment can be devastating but it also provides an opportunity for employees to re-tool and re-skill for other positions that may be in demand and are not obsolete. There has been much discussion and debate about the causes of unemployment. Different views place different levels of importance on the various factors. Some factors associated with unemployment include: stock market crashes, monetary policy; housing and real estate bubbles; labor market changes, and the European financial crisis. This next section discusses some of the factors and solution strategies that have been associated with unemployment.

REMEDIES FOR UNEMPLOYMENT

Over the years many different approaches have been adopted to help reduce unemployment. As the length of long-term unemployment continues to rise, there will be a need to address the jobless recovery after a recession. The approaches vary based upon the economic situations of the time. For example, during the mid- 1800s when land was undeveloped and there were a huge number of farmers and a great deal of undeveloped land, the government instituted the Homestead Act of 1862. Any person would receive up to 160 acres of public land if they agreed to farm and cultivate the land for five years (Fitzpatrick, 2009). This Act helped to stimulate the economy by providing work opportunities for any person willing and able to take advantage of the potential that farming offered.

Journal of Applied Business and Economics vol. 14(3) 2013 97 In the 1900’s, economists from the monetarist school of thought argued that a more rapid growth in the nation’s money supply would bring down unemployment. The expanded money supply would be spent in acquiring goods and services, leading to an increase in the demand for labor and a decrease in unemployment (Rose, 2000). Economists from the Keynesian school of thought argued that both fiscal and monetary policies were needed to combat unemployment. On the fiscal policy side, taxes should be reduced and government spending should be increased; these would increase aggregate demand and increase the need for more employees. On the monetary policy side, interest rates should be reduced, thus making credit cheaper and investing by businesses more desirable. Keynesian economists claim that when there is not adequate demand in relation to supply, this situation can lead to high unemployment (Blinder, 2002). Furthermore, during times of high unemployment, government policies can be used to help increase demand and stimulate the economy by encouraging investment and reducing unemployment. The Keynesian approach was used extensively in the New Deal programs that were initiated under President Franklin Roosevelt. These programs were instituted in the 1930’s during the period of the Great depression when the unemployment rate reached a high of 25%. Programs relating to wages, social security and farmer relief helped improve the economy and reduce unemployment, but a complete recovery did not take place until the 1940’s (Isidore, 2008; Delaney, 2011). Monetarists and Keynesians attack the unemployment from the demand side; supply side economists argue that the problem should be attacked from the supply side. They argue that changes should be made to increase supply by making it easier for businesses to produce: reduce taxes, reduce regulations, and reduce government spending. A smaller government would give more room for private sector to innovate, create jobs, and reduce unemployment (Rose, 2000). Monetarists, Keynesians, and supply side economists provide approaches to fighting cyclical unemployment; these approaches do not help much with frictional, seasonal, or structural unemployment. Drucker (1994) recognized the changing nature of society and the economy in the 1980’s and 1990’s. He introduced the idea of the importance of the knowledge worker, skill obsolescence, and the need for continuous learning. He argued that as the economy changed and we become more global there is a need for individuals and firms to continually re-evaluate their strategies, skills and knowledge with new knowledge and skills in order to stay employed and marketable. More widespread knowledge about job openings and requirements can reduce frictional unemployment, while additional training and re-training can reduce structural unemployment. Georgia’s Works Jobs Training Program, Mississippi’s Subsidized Transitional Employment Program and Services, and Minnesota’s Emergency Employment Development program are examples of programs that are designed to provide workers with the skills necessary for current job openings (Georgia Works, 2011; Mississippi’s STEPS2, 2011; Kane and Smith, 2011). The problem still remains, however; the traditional remedies to fight unemployment will have minimal impact on the factors keeping unemployment high discussed in this paper. There is a need for policymakers to consider new approaches to bring these new types of unemployment under control.

CONSEQUENCES AND CONCLUSIONS

There are a number of consequences of the high unemployment rate and expenditure cutbacks, both for society and for the unemployed individuals. According to the Census Bureau, 46.2 million Americans (15.1% of the population) were living in poverty in 2010, and 146.4 million (48% of the population) were classified as low income or in poverty. The poverty level was the highest since 1993. The head of the Census Bureau Poverty Statistics Branch indicated that the single most important factor leading to the increase in poverty was the increase in the number of people who did not work. Many middle-class Americans have dropped into the low income or poverty levels because of pay cuts, reduced hours, or loss of a job (1 in 2 People, 2011; Stanglin, 2011; Yen, 2011). Joblessness also leads to foreclosures, homelessness, and personal bankruptcies. Unemployed homeowners are doing all they can to hold on to their homes. Some measures include using unemploy-

98 Journal of Applied Business and Economics vol. 14(3) 2013 ment benefits, exhausting savings and retirement accounts, borrowing from credit cards and insurance policies, pawning valuables, and renting part of the home to others. If unemployment lasts long enough and these measures don’t work, the home is lost to foreclosure. Homelessness is another consequence of joblessness. Unemployed homeowners and renters who can no longer pay the mortgage or the rent will be foreclosed upon or evicted, and will be forced to live with relatives or friends, in homeless shelters, or on the streets. Schools have seen an increase in children from homeless families, and programs have been established by the Salvation Army and others to provide some assistance to these families (Program Guides, 2012). Each year over a million individuals file for personal bankruptcy. In many cases unemployment is a major cause of the problem. To cover living expenses, unemployed individuals will often exhaust their savings and retirement plan accounts. When they begin working again, often at jobs that pay less than what they were earning before, every penny will be spent on current expenses and bills that mounted during the period of joblessness. Thus, for many years, there will be nothing available for retirement savings, and many of these persons will enter their retirement years with little or no retirement savings. Unemployed persons need money; one source that many have found is the Social Security system. Older unemployed persons may file for early Social Security benefits. One can receive a reduced benefit as early as age 62, rather than waiting until the normal Social Security retirement age of 66. Payment of these early benefits, however, will have a negative impact on the financial health of the Social Security system. Another part of the Social Security system, the disability fund, has been misused by some unemployed individuals. Two studies have shown a correlation between when persons’ unemployment compensation is exhausted and when they file for unemployment benefits. Of course disability benefits are designed for persons who are truly disabled. There may have been some persons who were disabled who had not applied for disability benefits while they were receiving unemployment compensation, and then decided to apply when those benefits ended. But it is likely that many individuals were not really disabled, and saw the disability fund as another potential source of cash. The fragile health of the Social Security system is worsened by payments made to persons illegally (Paletta and Searcey, 2011). Underemployment is rising due to long-term unemployment. After a period of unemployment many individuals will take any job that comes along (Fleck, 2011). This problem is especially acute among many recent college graduates who have not been able to find employment related their major course of study. They have resorted to jobs such as taxi drivers, waiters, and pizza delivery drivers (Vennochi, 2011). The high unemployment rate is certainly a major problem facing our country today. It is our conclusion that a high rate of unemployment will be with us for a long time; we are not likely to see unemployment in the 5% - 6% range. Progress cannot be stopped; jobs will continue to be lost because of technology and the Internet. Jobs are also lost when politicians cut expenditures but are not willing to entertain tax increases or the usage of additional debt to allow for spending increases. The proposed drop in government spending will have a negative impact on economic growth, and other sectors of the economy are not growing fast enough to create the jobs necessary to decrease the unemployment rate. Many politicians argue that taxes were decreased early in the administration of President Ronald Reagan, and the economy expanded substantially thereafter. But they do not point out that spending also increased, not decreased, during the Reagan years. In fact, during each of the eight years of the Reagan administration, spending was greater than tax revenues, producing annual deficits, and this increased spending helped spur economic growth. The government had to borrow to cover the deficits; the total amount of debt taken on during the 8 years of the Reagan administration was greater than the entire national debt that existed when President Reagan took office. Similarly, during the administration of President George W. Bush, taxes were decreased and the economy expanded. But again government expenditures also increased (producing annual deficits), which helped economic expansion. There was no hesitancy to use borrowed funds to finance the wars in Iraq and Afghanistan, and to initiate a prescription drug benefit for seniors, and this spending contributed to

Journal of Applied Business and Economics vol. 14(3) 2013 99 the economic expansion that took place during the Bush years. Leaders in Congress today, however, are not willing to pursue similar borrowing strategies, and thus expenditures will be cut. There are three questions that we pose for further discussion: 1. How do we as a society handle the situation when, on a permanent basis, the economy is not producing enough job growth to provide employment to all of its citizens who want to work? It appears that there are not enough jobs available today for our unemployed workers, and the imbalance is likely to get worse. 2. When agencies are forced to cut their budgets, the cuts can take place in one of two ways: the number of workers can be reduced or the salaries of each worker can be cut to achieve the total amount of reduction needed. Should one approach be encouraged over the other? Managers are often told to “do more with less.” It may be possible to do more with less when the less money is distributed among the same number of workers, with each getting somewhat less. But it is difficult to do more with less money when less money means fewer workers to perform the needed work. The unemployment rate will certainly be higher if the number of workers is reduced. 3. Is it the responsibility of firms to create jobs? It there any responsibility by businesses to help reduce unemployment? In finance and management classes, students are taught that the goal of the firm is to maximize shareholder wealth or stock prices. If share price maximization means substituting machines for workers or moving production facilities to Mexico, then those actions must be taken. But is there any moral obligation to consider the welfare of employees?

REFERENCES

1 in 2 People Low-income or Poor, Census Reveals. (2011). News and Record, December 15, 2011, p. A8.

Adams, R. (2011). Newspapers Edit Down Outlooks The Wall Street Journal, August 22, 2011 , p. B6.

Austerity Worsened Economy in 2011. (2011). News and Record, December 31, 2011, p. A13.

Auto Industry on Hiring Spree (2011, July 7). News and Record, p. B6.

Bachman, J. (2007). The Big Record Labels' Not-so-Big-Future. Bloomberg BusinessWeek, 10(1). Retrieved from http://www.businessweek.com/bwdaily/dnflash/content/oct2007/db2007109_120106.htm

Bade, R., and Parkin, M. (2011). Fundamentals of Economics (5th ed.). Boston: Addison-Wesley.

Barnes, J. E., and Hodge, N. (2012). Military Faces Historic Shift The Wall Street Journal, January 6, 2012, pp. A1, A2.

Barror, F. (2011). Bookstore-Closing-Reflects-Rise-in-Ereaders. Independent Collegian, August 29. Retrieved from http://www.independentcollegian.com/bookstore-closing-reflects-rise-in-ereaders- 1.2615356

Behind on Jobs (2011). News and Record, December 27, 2011, p. A1.

Bendavid, N., and Hook, J. (2011). Deficit Effort Nears Collapse. The Wall Street Journal November 21, A1-A4.

Binker, M. (2011). Numbers do Little to Settle Argument Over State Layoffs. News and Record, November 17, 2011, p. A1, A5

100 Journal of Applied Business and Economics vol. 14(3) 2013

Blinder, A. S. (2002). Keynesian Economics. The Concise Encyclopedia of Economics. Retrieved from http://www.econlib.org/library/Enc/KeynesianEconomics.html

B of A Layoffs are Likely to Hit N.C. Employees. (2011). News and Record, November 17, 2011, p. A6.

Brooks, D. (2011). U.S. Economy in Midlife Crisis. News and Record, December 28, 2011, p. A13.

Builders, N. A. O. H. (Producer). (2011) NAHB: Annual Housing Starts (1978 - 2011). retrieved from http://www.nahb.org/generic.aspx?genericContentID=554

Business Cycle Dating Committee, National Bureau of Economic Research (2012). The National Bureau of Economic Research; January 5, 2012. Retrieved from http://www.nber.org/cycles/sept2010.html

Bussey, J. (2011). Are Companies Responsible for Creating Jobs? The Wall Street Journal, October 28(1), B1-B2.

Bustillo, M., and Zimmerman, A. (2011). Holiday Sales Woes Cast Cloud Over Sears. The Wall Street Journal, December 28(1), A1-A2.

Casselman, B. (2011). Economy Adds Jobs, but Not Enough. The Wall Street Journal, October 8-9(1), A1-A4.

Chu, E. (2011). Stanford Attracts 58,000 Students for Virtual Artificial Course. Wiredacademic, August 16. Retrieved from http://www.wiredacademic.com/2011/08/stanford-attracts-58000-students-for-virtual- artificial-course/

Companies Flourish but Not Workers. (2011). News and Record, November 26, 2011, p. B6.

Countdown to Expiry: Lipitor Goes Generic on 11/30/11. (2011). August 15. Retrieved from http://blogs.wsj.com/health/2008/06/18/countdown-to-expiry-lipitor-goes-generic-on-113011/

Databases, Tables and Calculators by Subject. (2012). Retrieved from http://data.bls.gov/timeseries/lns14000000

Davis, J. (2011). UNCG Chancellor Says Budget Cuts Are Hurting. News and Record, September 9, 2011, p. A6.

Degree of Difficulty. (2011). News and Record, November 20, 2011, p. H2.

Delaney, A. (2011). Long-Term Unemployment: States Search For Strategies To Counter Joblessness The Huffington Post. Retrieved from http://www.huffingtonpost.com/2011/09/05/long-term- unemployment_n_946620.html

Democrats Begin Blitz on State Budget. (2011). News and Record, September 20, 2011, p. A2.

Dougherty, C. (2011). Tax Revenue Snaps Back – States Book More Inflows but Face Tab for Higher Spending, End of Federal Funds. The Wall Street Journal, March 30(1) , A1. Retrieved from http://0-search.proquest.com.sheba.ncat.edu/wallstreetjournal/printviewfile?accountid=12711

Drucker, P. (1994). The Age of Social Transformation. The Atlantic Monthly (November), 53-80.

Journal of Applied Business and Economics vol. 14(3) 2013 101

Education Pays. (2012). Retrieved from www.bls.gov/emp/ep_chart_001.htm.

Edwards, J. (2011). Pfizer Plans 16,300 Layo+ffs Amid Health Benefit Cuts for Retirees. cbsnews.com, August 15(1). Retrieved from http://www.cbsnews.com/8301-505123_162-42849477/pfizer-plans-16300- layoffs-amid-health-benefit-cuts-for-retirees/

Employment Outlook Weak but Stable, Survey Reveals. (2011). News and Record, September 18, 2011, p. B6.

Employment Situation Summary. (2010). Retrieved from http://www.bls.gov/news.release/empsit.nr0.htm.

Ensinger, D. (2010). Whirlpool Moving Jobs to Mexico. economyincrisis.org, February 22. Retrieved from http://economyincrisis.org/content/whirlpool-moving-jobs-mexico

Failed Trading Company Fires all 1,066 Workers. (2011). News and Record, November 21, 2011, p. B8.

Fischer, K. (2011). Crisis of Confidence Threatens Colleges. The Chronicle of Higher Education, May 15(1). Retrieved from http://chronicle.com/article/Higher-Education-in-America-a/127530/

Fitch's Ratings on McGraw-Hill Unaffected by Accelerated Share Buyback. (2011). December 8(1). Retrieved from http://0proquest.unit.com.sheba.ncat.edu/pqdlink?did=2530975711andFmt=27clientId=15100andRQT=3 09andVName=PQD

Fitzpatrick, D., and Barry, R. (2011). Ax Falls at Smaller Banks. The Wall Street Journal, November 30(1), C1-C2.

Fitzpatrick, L. (2009). A Brief History of Unemployment. Time, November 23, 2009. Retrieved from http://www.time.com/time/business/article/0,8599,1942606,00.html

Fleck, C. (2011). When Any Job Must Do. AARP Bulletin, June(1), 16-18.

Free Rides on Triangle Toll Road End Today. (2012). News and Record, January 3, 2012, p. A6.

Friedman, T. (2011). First 100 Days of 2013 is Next Chance To Fix Economy. News and Record, December 13, p. A11.

Garvin, G. (2011). Return to Sender – Postal Service Bailout. News and Record, June 19, 2011, p. 1.

Georgia Works Jobs Training Program: a Peach or the Pits? (2011). PBS News Hour, October 12(1). Retrieved from http://www.pbs.org/newshour/bb/business/july-dec11/georgiaworks_10-12.html

Glover, M. (2011). Summer School Classes Move Online. News and Record, July 3, 2011, pp. B1-B3.

Graves, Frederick. “How to Win…Without a Lawyer.” Jurisdictionary, 2011. Retrieved from http://www.jurisdictionary.com/

102 Journal of Applied Business and Economics vol. 14(3) 2013 Gutierrez, B. (2011). No Caps and Gowns Needed for First-ever Virtual Graduation at UH. Hawaii News Now, May 7. Retrieved from http://www.hawaiinewsnow.com/story/14591601/no-caps-and-gowns- needed-for-virtual-graduation-at-uh

Hagerty, J. R., and Tita, B. (2011). Appliance Sales Tumble. The Wall Street Journal, October 29-30(1), B1-B2.

Isidore, C. (2008). It's Official: Recession since Dec. ‘07. CNN money, December 1, 2008.

Izzo, P. (2011). Bleak News for Americans’ Income. The Wall Street Journal, October 14(1), A6.

Jobs in America (2011). News and Record, September 5, 2011, p. A1.

Kane, M., and Smith, D. (2011). From MEED to SEED, a Minnesota Idea.” Twin Cities Daily Planet. daileyplanet.net, February 8(1). Retrieved from http://www.tcdailyplanet.net/blog/growth-and- justice/meed-seed-minnesota-idea.

Kapner, S. (2011). Citi Poised to Slash More Jobs. The Wall Street Journal, November 16(1), C3.

Kapner, S. (2011). Citi Shines, but Investors Shrug. The Wall Street Journal, October 18, A1-A2.

Kapner, S. and Matthias R. (2011). Citigroup to Cut 4,500 Jobs, Take $400 Million Charge. The Wall Street Journal, December 7(1), C3.

Kell, J., and Warner, M. (2011). Store Closings Slam Lowe’s Profit. The Wall Street Journal, November 15(1), B3.

Keoun, B. (2011). Wall Street Aristocracy Got $1.2 Trillion in Secret Loans. Businessweek, 2(1), 1.

Knotek, E. S., and Terry, S. (2009). How Will Unemployment Fare Following the Recession? Retrieved from http://www.kc.frb.org/PUBLICAT/ECONREV/PDF/09q3Knotek.pdf

Korn, M., and Kamp, J. (2011). Health-Care Firms Come Down With Severe Case of Cutback Jitters. The Wall Street Journal, August 2(1), B1-B2.

Krugman, P. (2011). Phony Fear Factor. News and Record, October 1, 2011, p. A15.

Langley, R. (2011). Small Business Drives U.S. Economy. About.com Retrieved from http://usgovinfo.about.com/od/smallbusiness/a/sbadrives.htm

Levitz, J. (2011). Small Town Post Offices Threatened. The Wall Street Journal, July 26, 2011, p. A3.

Linebaugh, K. (2011). Honeywell is Generating Cash but Still Wary on Hiring. The Wall Street Journal, October 31(1), B8.

Loftus, P. (2011). Pfizer Lifts Payout, Buyback. The Wall Street Journal, December 13(1), B5.

Lohr, S. (2011). Machines in a Better Position To Find Jobs Than Unemployed. News and Record, October 30, 2011, p. H3.

Journal of Applied Business and Economics vol. 14(3) 2013 103 Longley, R. (2012). Small Business Drives U.S. Economy Retrieved from http://usgovinfo.about.com/od/smallbusiness/a/sbadrives.htm

Lowe’s Closing 20 Stores. (2011). News and Record, October 18, 2011, p. B11.

Maltby, E. (2011). No End-of-Year Cheer for Employees. The Wall Street Journal, November 10(1).

Markoff, J. (2011). Virtual and Artificial, but 58,000 Want Course. New York Times, August 15(1). Retrieved from http://www.nytimes.com/2011/08/16/science/16stanford.html.

Mattioli, D., and Hudson, K. (2011). Gap to Slash its Store Count. The Wall Street Journal, October 14(1), B1-B2.

McGraw-Hill Announces Job Cuts, Accelerated Stock Buyback – Update. (2011). RTT News. December 7(1). Retrieved from http://www.rttnews.com/1776907/mcgraw-hill-announces-job-cuts-accelerated- stock-buyback-update.aspx

McGraw-Hill Cutting 10 Pct of Education Unit Work Force, Freezing Pension; Eyes $100M Savings. (2011). Washingtonpost.com, December 7(1). Retrieved from http://www.washingtonpost.com/business/mcgraw-hill-cutting-10-pct-of-education-unit-work-force

McKeon, B. (2011). Why Defense Cuts Don’t Make Sense. The Wall Street Journal, October 14(1), A15.

Mijuk, G., and Whalen, J. (2011). Novartis to Cut 2,000 Jobs. The Wall Street Journal, October 26(1), B5.

Mississippi’s STEPS 2, Program Overview. (2011). Retrieved from http://mdes.ms.gov/Home/docs/STEPS2Web_v1/STEPS2_web_v1.html.

Moe, T. (2011). The Internet Will Reduce Teachers Union Power. The Wall Street Journal, July 18(1), A13.

Moen, A., and Gustafsson, K. (2011). Nokia Cuts 3,500 Manufacturing Jobs. The Wall Street Journal, September 30(1), B5.

More Cities Reporting Layoffs, Lower Revenue. (2011). News and Record, September 27, 2011, p. A10.

Morgan Stanley to cut 580 Jobs in New York. (2011). December 28(1). Retrieved from http://finance.yahoo.com/news/morgan-stanley-cut-580-jobs-125048870.html

Muñoz, S. S., and Patrick, M. (2011). HSBC Strategy Shift Cuts Jobs, Branches. The Wall Street Journal, August 2(1), C1-C2.

Murphy, M. (2011). Buying Shares, Buying Trouble. The Wall Street Journal, October 12(1), B1-B4.

NBER Business Cycle Dating Committee - March 1991. (2012). The National Bureau of Economic Research, January 5, 2012. Retrieved from http://www.nber.org/March91.html.

Nearly 1 in 3 Unemployed Out of Work More Than a Year. (2011). The Wall Street Journal, June 3(1). Retrieved from http://blogs.wsj.com/economics/2011/06/03/nearly-1-in-3-unemployed-out-of-work- more-than-a-year/

104 Journal of Applied Business and Economics vol. 14(3) 2013

Nokia Siemens to Cut 17,000 Jobs to Rein in Its Costs. (2011). News and Record, November 24, 2011, p. B7.

Office of Publications and Special Studies, Ranks of Those Unemployed for a Year or More Up Sharply. (2010). Issues in Labor Statistics Summary, 10-10, October 2010. Retrieved from http://www.bls.gov/opub/ils/summary_10_10/ranks_unemployed_year.htm

Official Says Marines, Navy to Reduce Numbers. (2011). News and Record, December 18, 2011, p. A6.

O’Keefe, E. (2011). Army Slashing 8,700 Jobs as Budget Cuts Begin. Washingtonpost.com, December 9(1). Retrieved from http://www.washingtonpost.com/blogs/federal-eye/post/army-slashing-8700-jobs-as- budget-cuts-begin/2011/12/08/gIQAOWRxhO_blog.html

Ordonez, I., and Stynes, T. (2011). Exxon Profit Soars 41%. The Wall Street Journal, October 28(1), B7.

Paletta, D., and Searcey, D. (2011). Jobless Tap Disability Fund. The Wall Street Journal, December 28(1), A1.

Program Guides Homeless Mothers. (2012). News and Record, January 3, 2012, p. A3.

Rampell, C. (2011). A Decline in American Entrepreneurship: Economix Explaining the Science of Everyday Life, Economix, March 31(1). Retrieved from http://economix.blogs.nytimes.com/2011/03/31/a-decline-in-american-entrepreneurship/

Ranks of Those Unemployed for a Year or More Up Sharply (2010). Retrieved from http://www.bls.gov/opub/ils/summary_10_10/ranks_unemployed_year.htm

Ranks of Those Unemployed for a Year or More Up Sharply. (2010). Retrieved from http://www.bls.gov/opub/ils/summary_10_10/ranks_unemployed_year.htm

Red Kettle Donations Down, Need Is Up. (2011). News and Record, December 18, 2011, p. B7.

Reddy, S. (2010). U.S. Job Market loses Steam. The Wall Street Journal, August 7(1), 1.

Rose, P.S. (2000). Money and Capital Markets (7th Ed.). Boston: McGraw-Hill Higher Education.

Rougemont, I. (2011). 6 Stocks With Major New Share Buybacks. July 1, 2011. Retrieved from http://seekingalpha.com/article/277624-6-stocks-with-major-new-share-buybacks

Rugaber, C. S. (2011). CEOs of Big Companies Foresee no Pickup in Hiring, December 14, 2011. Retrieved from http://www.salon.com/2011/12/14/ceos_of_big_companies_foresee_no_pickup_in_hiring_2/

Rugaber, C.S., Fahey, J., and Liedtke, M. (2011). US Companies are Feeling Impact of European Crisis. AP Business Writers, November 13, 2011. Retrieved from http://news.yahoo.com/us-companies-feeling- impact-european-crisis-160229813.html

Ryan, J., and Hadida, A.L. (2010). One Way to Save the Music Industry. Bloomberg Businessweek, July 29, 2010. Retrieved from http://www.businessweek.com/technology/content/jul2010/tc20100729_564014.htm

Journal of Applied Business and Economics vol. 14(3) 2013 105

SandP Indices Says SandP 500 Stock Buybacks Up 48.8% in Q3 - Quick Facts. (2011). December 21(1). Retrieved from http://www.rttnews.com/1785867/s-p-indices-says-s-p-500-stock-buybacks-up-48-8-in- q3-quick-facts.aspx

Schmid, Randolph. “U.S. Postal Service Proposes Cutting 120,000 Jobs, Pulling out of Federal Retirement, Health Care Plans.” Huffington Post, August 11, 2011. Retrieved from http://www.huffingtonpost.com/2011/08/11/us-postal-service-jobs-benefits-layoffs_n_924927.html

Schmidt, E. M. (2011). McGraw-Hill Plans Job Cuts. Telegraph Herald, December 8, 2011. Retrieved from http://0-www.lexisnexis.com.sheba.ncat.edu/lnacui2api/frame.do?reloadEntirePage=trueandra

Sears to Cut More Jobs After Losses Widen. (2011). August 18(1). Retrieved from http://www.thestreet.com/story/11224747/1/sears-to-cut-more-jobs-after-losses-widen.html

Shierholz, H., and Finio, N. (2011). Job-seekers Ratio Remains Above 4-1 for 34th Straight Month. December 13, 2011. Retrieved from http://www.epi.org/publication/job-seekers-ratio-remains-4-1-34th- straight/

Shierholz, H., Finio, N., Sabadish, N., and Wething, H. (2011). Job Growth Strengthens but Insufficient to Cure Sick Labor Market, December 2, 2011. Retrieved from ttp://www.epi.org/publication/job-growth- strengthens-insufficient-cure-2/

Smith, A. (2011). Whirlpool Cutting Jobs and Closing a Plant. CNN Money, October 28(1). Retrieved from http://money.cnn.com/2011/10/28/news/companies/whirlpool_jobs/index.htm

Stainburn, S. (2011). Business Roundtable Survey: No Hiring Push in Early 2012. December 14, 2011. Retrieved from http://www.globalpost.com/dispatch/news/regions/americas/united- states/111214/business-roundtable-survey-ceos-not-hiring

Stanglin, D. (2011). Census Bureau: U.S. Poverty Rises to 15.1%, Highest Since 1993. USA Today, September 14, 2011. Retrieved from http://content.usatoday.com/communities/ondeadline/post/2011/09/census-bureau-us-poverty-rises-to- 151-highest-since-1983/1

Tibken, S. (2011). Intel Cuts Its Outlook. The Wall Street Journal, December 13(1), B4.

Tibken, S., and Clark, D. (2011). AMD to Cut 1,400 Jobs as New Chief Revamps. The Wall Street Journal, November 4(1), B5.

Toll Operations. (2012) Retrieved from http://www.ncdot.gov/turnpike/tolls/default.html

Tozzi, J. (2008). From Bricks and Mortar to Digital Music Master. Bloomberg Businessweek, April 21, 2008. Retrieved from http://www.businessweek.com/stories/2008-04-21/from-bricks-and-mortar-to- digital-music-masterbusinessweek-business-news-stock-market-and-financial-advice

Troianovski, A. (2011). Wireless Jobs Vanish. The Wall Street Journal, July 18(1), B1-B2.

Troubled BofA to Eliminate 30,000 Jobs (2011). News and Record, September 13, 2011, p. A2.

106 Journal of Applied Business and Economics vol. 14(3) 2013 Trudell, C. (2011). GM Sees Europe Crisis ‘More Serious’ Than 2008 Credit Bubble. BusinessWeek.com, November 22, 2011,, 1. Retrieved from http://www.businessweek.com/news/2011-11-22/gm-sees- europe-crisis-more-serious-than-2008-credit-bubble.html

Two-thirds of CEOs Say They Plan No Hiring (2011). News and Record, December 15, 2011, p. B5.

Uchitelle, L. A., Edmund L., Labaton, S. (2008). U.S. Loses 533,000 Jobs in Biggest Drop Since 1974. New York Times, December 6, 2008. Retrieved from http://www.nytimes.com/2008/12/06/business/economy/06jobs.html

Unemployed Persons by Industry and Class of Worker. (2012). Retrieved from http://www.bls.gov/news.release/empsit.t14.htm.

Unemployed Seek Help Against Bias. (2011). News and Record, October 11, A8.

Unemployment Rates (2012). Retrieved from http://data.bls.gov/pdq/SurveyOutputServlet.

United States Unemployment Rate (2011). December 4(1). Retrieved from http://www.tradingeconomics.com/united-states/unemployment-rate

US Business Cycle Expansions and Contractions. (2012). Retrieved from http://www.nber.org/cycles.html

Vennochi, J. (2011). Graduates: Young, Hip, Educated and Jobless, News and Record, December 29, 2011, p. A8.

Veterans’ College Costs To Rise in N.C. (2011). News and Record, December 28, 2011, p. A2. Retrieved from http://www.grra.org/announcements/2011/12/28/mikes-desk-122811-guilford-schools- rise-and-greensboros-recycling-needs-gro

Von Drehle, D. (2012). The Prophet. Time, December 26, 2012, pp. 113-116.

Wessel, D. (2011). Untangling the Long-Term-Unemployment Crisis. The Wall Street Journal, October 20(1), A6.

White House: U.S. Needs Higher Borrowing Limit. (2011). News and Record, December 28, 2011, ), p. A10.

Why ‘Tax cuts = Jobs’ Is a Lie: Exxon Mobil (2009). February 6. Retrieved from http://www.democraticunderground.com/discuss/duboard.php?az=printer_friendlyandforum=132andtopic _id=8173972

Wiseman, P., and Rugaber, C. (2012). 200,000 Jobs Added Last Month. January 6(1). Retrieved from http://www.dailycamera.com/national-international-business/ci_19693829

Yen, H. (2011). Census shows 1 in 2 People Low-income or Poor. Yahoo News, 12, 1. Retrieved from http://news.yahoo.com/census-shows-1-2-people-poor-low-income-054325860.html

Journal of Applied Business and Economics vol. 14(3) 2013 107

Foreign Direct Investment Flows: An Examination of Its Distribution Among Middle- and Low-Income Countries

Gabriel Manrique Winona State University Fastenal Company

Val Vlad Pennsylvania State University-Erie

Yanping Chong Winona State University

INTRODUCTION

In the last 20 years the world economy has witnessed a tremendous increase in the importance of Foreign Direct Investment (FDI) as a source of economic development. FDI has become particularly important in financing investment and growth for developing countries. The status and importance of FDI as a vehicle of economic development has become even more obvious in recent years. Many scholars have concluded that FDI has positive effects on the balance of payments, promotes exports, creates employment and radiates an array of externalities – financial, technological, institutional, and managerial, into the host countries. The recipient or host countries have preferred FDI to other forms of capital inflows because in addition to being a major source of financing (critical for countries suffering from chronic financial market failures), it stimulates the formation of human capital, promotes successful corporate governance practices, and is often accompanied by technological and institutional transfers. FDI is also more stable than other forms of investment. Once made, FDI tends to be harder to move out of a country, compared to other forms of foreign investment such as portfolio investments. Some studies go even further and point at FDI as a source of economic and political stability for the regions that receive it and for host countries. Given its growing importance, a large pool of countries has entered into keen competition to attract FDI. Nevertheless, the data on FDI location describes a very uneven distribution of FDI. The lion’s share

108 Journal of Applied Business and Economics vol. 14(3) 2013 (75%) goes to the developed world - the group of countries where people make on the average more than $20,000 a year. Out of the remaining 25%, 66% (or another 16.5% of the total) is absorbed by the group of the twelve largest developing economies (with China and India leading the “emerging” group of countries). In spite of being one of the most researched topics in international economics, FDI continues to receive a large amount of research attention. There are still contradicting opinions on what are the ingredients for a successful recipe to attract FDI. A large segment of theoretical literature on FDI location is based on the conventional model of comparative advantage derived from cross-country differences in factor endowments. According to this framework, we should see (FDI) flowing mainly to capital-scarce, poor countries. In reality, the least developed countries received a negligible share of the total world FDI during the 1990s. For example, in 1994, the least developed countries that accounted for more than 4% of world production and 11% of world population received less than 1% of total world FDI (Zhang and Markusen, 1999). We show later on that a similar pattern of uneven FDI distribution prevailed during the 2000s with countries at the bottom of the wealth hierarchy receiving a very small fraction of the world’s total FDI. Logically, some questions arise for both the multinational enterprises (MNEs) seeking to place their investment in the most advantageous locations and for those countries entering the competition to attract more FDI. For MNEs the relevant questions pertain to what makes a successful location choice for their investment. For potential recipient countries the questions pertain to how MNEs decide to locate their investments. In short, what are the most important determinants of FDI location?

REVIEW OF THE LITERATURE

On the question of the location determinants of FDI, Chakrabarti (2001) found that market size (as a proxy for market potential) and market openness are positively associated with the volume of FDI inflows. Asiedu (2002) showed that among the more important factors in attracting large amounts of FDI are the country’s openness to trade, its infrastructure availability, and the potential return to capital. She did not find conclusive evidence, in the case of the African economies in her study, that economic and political variables such as growth rates, government consumption, inflation rate, money growth, and political risk are significant factors. In particular, she investigated Africa’s poor performance in attracting FDI in the late 1990s. For example, sub-Saharan Africa’s share of the world’s FDI had fallen steadily to just below 3% in 1995-99. When using a dummy variable for “Africa” in a cross-country model that included several locational choice variables, the “dummy” coefficient turned out to be significantly negative. She concluded that the African continent has a negative image in the investors’ view. This is different from earlier decades. For example, Agodo (1978) found that US FDI in Africa was determined by market size (market- seeking behavior), the presence of raw materials (resource-seeking behavior), and a sufficiently developed infrastructure along with political stability. In that earlier study, tax concessions and tariff protection did not have a significant influence on foreign investment decisions. Firms also tend to be sensitive to the existing infrastructure of potential FDI locations. The relevant measures of infrastructure sophistication may vary by firm and its needs. The transportation network has been a key measure of infrastructure sophistication particularly where global supply management plays a key role in the ability of firms to compete. Communication infrastructure, measured in the past as telephone lines, has also been used in studies of FDI. More recently, relevant proxies for infrastructure have included network readiness, which has attracted more attention from MNEs as they rely more and more on the internet for vital communications and other functions. In other studies of FDI, Gastanaga et al. (1998) found that the expected rate of growth was a highly significant determinant of FDI while exchange rate volatility did not play an important role. As may be expected, high levels of corruption in the host country can repel FDI while high corporate tax rates can play a negative role in attracting FDI, especially corporate tax rates of 25% or higher. Along the same line of research, attracting FDI has been found to be more likely with the combined presence of an open

Journal of Applied Business and Economics vol. 14(3) 2013 109 economy, smaller government, political stability, market competition, and a higher rate of saving and investment in the host economy. From a different perspective, one can analyze the FDI location factors and how a country is affected by FDI. One may argue that the extent and direction in which FDI benefits the host country influences - endogenously – most of the variables that, in turn, would attract FDI into that country. This argument is one of the bases for our analysis that there has been growing competition among developing countries for attracting FDI. The literature on how FDI can benefit the recipient economies is varied in its conclusions. De Mello (1999) argued that the benefits of FDI depend on spillovers, profitability, value-added content of FDI related production, capital formation, employment, exports, and technology in the recipient economy. Further, he showed that the rate of FDI growth alone would not be sufficient to account for a country’s economic growth. Grossman and Helpman (1991) pointed out that FDI in developing countries may specialize in producing less technology-intensive goods or simply engage in exploiting their natural resource. In both cases spillover effects and learning by doing opportunities are very limited. A slightly different opinion is provided by Borensztein et al. (1998). They argued that FDI is an important determinant for technology transfers and contributes more to growth than domestic investment. They analyzed a group of 69 developing countries over a period of 20 years and concluded that FDI has strong positive impacts on economic growth. However, they argue that the strength of this impact depends on human capital - countries with low levels of human capital hardly benefit from FDI. It may be concluded that human capital is the major ingredient that enables recipient countries to benefit from FDI and from the externalities created by FDI. In addition, Alfaro (2010) found that capturing FDI’s beneficial effects depends on the financial sophistication of a country. In other words, countries lacking modern financial infrastructures are unlikely to reap the benefits of FDI because it is access to finance that allows domestic entrepreneurs to take advantage of backward and forward linkages created by FDI. In looking at locational factors determining FDI flows, we can begin by looking at what distinguishes MNEs from domestic firms. Markusen (1995) enumerated four factors that tend to distinguish between MNEs and local firms: MNEs tend to have higher levels of R&D; have greater need for human capital; are more likely to introduce new and complex products; and are more likely to engage in advertising and product differentiation. Thus, when MNEs are searching for the best destinations for their capital, the availability of human capital is among the highest ranked location variables. A more recent trend in FDI location shows that MNEs invest in skilled-labor countries to outsource white-collar workers. Non-tradable sectors such as bank, insurance, credit-card, accounting, investment banking, and high-tech engineering, companies extend their activities in skilled-labor-abundant developing countries. MNEs have direct requirements such as engineers, technicians, and accountants and indirect requirements such as electrical and water supplies, telecommunications, transport links, and legal institutions. The availability of these elements helps determine FDI location. They indicate the presence of the intangible and firm-specific assets which form the foundation of knowledge capital that is associated with much of modern FDI. Deichmann et al. (2003) considered the role of social and human capital in determining the flow of FDI into Eurasian countries. They found that the volume of FDI that flows into a country, relative to the country’s population, depends heavily on social capital - measured by the infrastructure level; and human capital - measured by the level of professional skills. They also found the depth of financial markets - another dimension of social capital - to be an important determinant of FDI. In further examining the FDI decisions by individual firms, we should differentiate among different reasons for FDI location. Much of early FDI was in extractive industries and such FDI continues to be significant particularly in emerging economies. Efficiency-seeking is another strong motivation for FDI and explains a significant portion of FDI that took place in many emerging economies such as China and Mexico. Market expansion (market-seeking behavior), is a third explanation for FDI. MNEs seeking growth opportunities look to countries with large market potential for their products and services. As average incomes rise in emerging markets - a tendency that seems irreversible - they become increasingly attractive for market seeking MNEs. In addition, a MNE’s ability and willingness to differentiate its

110 Journal of Applied Business and Economics vol. 14(3) 2013 products explain their continuous international expansion, an interpretation of MNE behavior consistent with Hymer’s early explanation of MNE behavior (Hymer, 1976). Firms can also base their FDI decisions on the presence of other MNEs in a particular market. This is known as the agglomeration effect. Such agglomeration effects can explain why FDI tends to cluster in certain countries. Firms may be motivated by the already existing market provided by other upstream MNEs with whom they have had relationships in the past particularly in their home country. The presence of other MNEs can also provide signals to potential entrants regarding the desirability of FDI in a particular location. According to the Capital Markets Consultative Group’s survey of large corporate investors (IMF, 2003) because of significant vertical linkages in production networks, some types of FDI flowing into emerging market economies (EMCs) are motivated by following existing clients into a new market. This generates agglomeration effects and creates FDI clusters. A large majority of the survey’s respondents emphasized also that they invest in EMCs primarily to meet domestic demand rather than to reduce global manufacturing costs. Evidence of the agglomeration effect is provided in a study by Barrell and Pain (1999). They focus on the flow of US FDI into the European Union and discovered that, even with few barriers to trade, FDI still tend to agglomerate for a number of reasons. By locating FDI near other foreign firms, investors can realize beneficial spillover effects including R&D related advantages and linkages to intermediate goods. Shaver (1997) found that there are country-specific and industry-specific information spillover effects for corporations when they consider FDI. Furthermore, FDI tends to be a sequential process whereby a firm’s prior country-specific knowledge positively influences the FDI decision and that furthermore, industry- specific knowledge gleaned from the experiences of other MNEs also influences the decision to invest positively. Lall and Streeten (1977) suggest that this “herd mentality” is even more relevant in relation to host countries where investor risk is high and information is deficient. Similarly, Kindleberger’s (1969) oligopolistic reaction theory suggests that following the leader strategies prevail in making FDI decisions. More evidence on agglomeration as a factor in FDI location decisions is provided by Chang and Park (2005). In their study of Korean firms locating in China, network externalities were deemed important and that such network externalities were sensitive to the sameness of nationality of the origin of FDI and the similarity of industries. This gives us more reason to account for previous FDI in a country as a factor explaining current and future FDI. According to early work on MNEs by Hymer (1976) and Kindleberger (1969), the advantages of setting up FDI must exceed the disadvantages of operating outside the home environment in order for the FDI to take place. The MNE must have location-specific advantages based on markets, resource availability, labor and infrastructure. Furthermore Dunning (1993) pointed out that the importance of each location-specific factor varies according to a firm’s own inclination toward natural resources, markets, efficiency, strategic assets and other considerations. We also cannot discount the importance of cultural connections and political factors in the FDI decision. Bandelj (2002) reported that country dyads, political, migration and cultural relations had a strong positive effect on FDI flows. In an earlier study of US FDI in Latin America, Nigh (1986) found that the political environment also influences firm decisions on FDI – a country’s hostility to US foreign policy is interpreted at the firm level as hostility to US firms thereby affecting the FDI decision. Henisz and Macher (2004) studied the FDI location decisions of semi-conductor firms between 1994 and 2002. Not surprisingly, they found that such companies looked for countries that had high levels of sophistication – indicating the need to complement the production and design needs of the firm. They also found that companies tended to stay away from politically unstable or hazardous countries where the prospects of expropriation were greater. In addition, they found that companies with less advanced technologies were less sensitive to political hazard when making FDI decisions. However, we note that it has not always been assumed that political instability was very important in determining FDI decisions by firms. For example, in a 1972 study of FDI decisions in Latin America made by multinational corporations, Bennett and Green (1972) found that political instability was not very important in deterring FDI.

Journal of Applied Business and Economics vol. 14(3) 2013 111 Cultural and historical factors may be an even more important location factor for developing countries. In their study of the FDI location decisions of MNEs from Spain, Galan et al. (2007) found that when investing in less developed countries in Latin America, historical and cultural factors played a more important role than in the developed countries of the European Union. However, Mitra and Golder (2002) found that the cultural gap between the domestic and foreign markets is not significant in explaining foreign market entry. We may then posit that where markets and institutions are already developed, home and host country historical ties become less important in the FDI decision. As such the firm can focus more on strategic asset-seeking behavior. This may also imply that historical and cultural ties compensate for weaker institutional development. The level of corruption is another factor that may influence FDI decisions of MNEs. Habib and Zurawicki (2002) found evidence that MNEs prefer not to tolerate corruption and provided examples of firms that have successfully instituted zero tolerance for corruption in their FDI. Just as importantly, they found that cultural differences (with respect to the level of corruption tolerated within a country) also influence firm decisions about FDI. Firms from cultures where corruption is less pervasive are less likely to invest in countries where corruption is more pervasive, a finding reinforced by Cuervo-Cazzura (2006). In a world where the developed countries remain the major attraction for FDI, globalization has given developing countries (DCs) more access to the international capital markets. According to the World Bank’s 2009 World Development Indicators, private capital flows into developing countries increased from $208 billion, in 2003, to $961 billion, in 2007. The main source of these flows has been FDI. Approximately 55% of the total private capital flows into developing countries in 2007, was FDI. This has been especially helpful for low income countries for which the net inflows of FDI as a percentage of GDP increased from 1.7% in 2000 to 4.2% (2009 World Development Indicators, WB). For some small countries, FDI inflows can be equivalent to as much as 20% of their GDP. It is not surprising then that countries would establish more Investment Promotion Agencies (IPAs). Harding and Javorcik (2007) reported that by 2005, 68 out of 81 developing countries had investment promotion agencies. Overall more than 160 countries already had IPAs at the national level and that there were 250 sub-national IPAs. The survey of works done by Lim (2008) on the effectiveness of such investment promotion activities reveals that such activities are not misplaced; investment promotion activities are positively linked to increases in FDI. The widespread belief among DCs that FDI is an important way of accelerating economic growth has led to increasing competition to attract FDI. On the grounds that becoming “business friendly” leads to greater FDI, many DCs have tried to change their business climate. In one measure from the Doing Business studies of the World Bank, we found that in a sample of 111 DCs, the cost of starting a business (measured as a percent of per capita income) declined by an average of 53% between 2004 and 2010. In fact, in only four countries did the relative cost of starting a business increase. In this paper, we seek to examine the factors associated with success in attracting FDI inflows from 2000 to 2009. We used panel data of 75 out of 151 countries designated by the World Bank as low- or middle-income countries as of 2000. Many countries were excluded due to serious data shortage. The rest of the paper is organized as follows: data description and summary statistics of the FDI inflow; a discussion of the empirical models used and the implications of our findings; and concluding remarks and directions for future research.

DATA AND SUMMARY STATISTICS

To study the distribution of FDI, we first took the total FDI inflows in our set of 75 low- or middle- income countries. Appendix 1 lists the sample countries by income group. We then calculated the percentage share of each country in this total or “global” FDI inflow. We calculated the similar percentage share of each country in the “global” GDP of this set of countries. This was done for each year from 2000 to 2009. As merely a starting point for comparisons, we posit that absent any differentiation among countries with regard to each one’s attractiveness for FDI, each country’s share of FDI would be proportional to its size relative to others. Hence a country’s share of “global” FDI would reflect its share of

112 Journal of Applied Business and Economics vol. 14(3) 2013 the “global” GDP. We will show momentarily that FDI is in fact not distributed proportionately to GDP and it is the deviations from the proportional distribution of FDI that we seek to examine. For each country, by taking the ratio of its percentage share of “global” FDI to its percentage share of “global” GDP, we can illustrate the proportionality of FDI inflows. A ratio of 1 would indicate that a country’s FDI inflow for the year is proportional to the size of its economy, relative to this set of countries. A ratio greater than 1 would indicate FDI inflow is relatively greater than would be suggested by the size of its economy. As expected, Table 1 shows that FDI inflows are far from proportional to an economy’s size. Some countries received negligible FDI inflows, as shown by the minimum values for the different years while others received FDI as much as 8 times of the proportionate size of their economy. The distribution of this share ratio skews slightly to the right (skewness >0) and is less peaked than a normal distribution. Right skewness indicates more countries in our sample have FDI share to GDP share closer to the minimum than the maximum. The box plot displayed in Figure 1 presents a visual summary of our panel data. Each of the box-and-whiskers plots describes succinctly the distribution of the FDI flow for a particular year, and the series of boxes in chronological order reveals the change in distribution over time. Specifically, the mean and median are quite persistent while the distribution varies over time. It is clear that with regard to attracting FDI inflows, the countries differ significantly. We seek to identify the factors that dictate the distribution in a given year as well as its time variations.

TABLE 1 DESCRIPTIVE STATISTICS FOR SHARE OF FDI TO SHARE OF GDP, 2000-2009

Year Obs. Mean Std.Dev Min Max Skewness Kurtosis 2000 75 1.22 0.95 -0.04 5.17 1.25 2.85 2001 75 1.34 1.15 -1.88 4.15 0.43 0.27 2002 75 1.15 0.86 -0.04 3.94 0.99 0.90 2003 75 1.28 1.06 -1.55 3.84 0.43 0.15 2004 75 1.36 1.01 0.00 4.27 0.96 0.29 2005 75 1.26 1.07 -1.08 5.67 1.17 3.25 2006 75 1.71 1.47 -0.10 8.35 2.01 5.61 2007 75 1.74 1.26 0.02 5.78 1.08 0.83 2008 75 1.69 1.23 0.00 4.98 0.75 0.33 2009 75 1.70 1.38 -0.45 6.36 1.17 1.16 2000- 2009 1.44 1.15 -0.51 5.25 1.02 1.50

Journal of Applied Business and Economics vol. 14(3) 2013 113 FIGURE 1 BOX PLOT FOR SHARE OF FDI TO SHARE OF GDP, 2000 -2009

10.0

7.5

5.0

2.5 FDI share to GDP share

-2.5

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year

Note that the length of the box represents the interquartile range (the distance between the 25th and 75th percentiles). The symbol in the box interior represents the group mean. The horizontal line in the box interior represents the group median. The vertical lines issuing from the box extend to the group minimum and maximum values.

EMPIRICAL METHODS AND RESULTS

In our continuing study of FDI among low- and middle-income countries, we will test for the significance of several factors in explaining for the differences in FDI inflows. Among the factors we will be looking at are: agglomeration, market size, infrastructure, business environment, corruption, human capital, and trade orientation. If agglomeration among MNEs is indeed important, then we would expect that future FDI inflows would be positively related to existing FDI stock. The likelihood that a company would invest in a foreign country would be greater if it were inclined to follow other companies already in that country. Existing FDI stock per capita is one such measure of the presence of foreign companies in a country If market seeking and market expansion are powerful motivators for MNE behavior, then market size would be one consideration for FDI decisions. Markets have different dimensions but the measurement should include sheer size (population) and effective purchasing power (income per capita). We would expect market size to be positively correlated with FDI inflows. We expect FDI flows to be affected by the level of infrastructure in a country. Infrastructure can impact production cost. But just as importantly in today’s competitive environment, the quality of infrastructure affects firms’ global supply chain management. To be able to move products, parts, and services efficiently and promptly, a high level of infrastructure is necessary. The different types of infrastructure need to be considered including road and rail quality, port capacity, reliability of electricity, and in today’s world, network readiness. Institutional factors can impact the attractiveness of the business environment. The popularity of the Doing Business series of the World Bank and the proliferation of other indices that attempt to capture the business environment in different countries should not be surprising. Companies considering going global are sensitive to conditions in countries with which they lack familiarity. But rather than arguing that company decisions are driven by these indices and measures, we expect instead that contained in these

114 Journal of Applied Business and Economics vol. 14(3) 2013 measures are factors such as investment protection and bureaucratic intervention which are important in the FDI decision of firms. Hence, while we expect these indices to be correlated with FDI flows, the indices are not the causes of FDI flows. The level of corruption in a country can serve as a deterrent to FDI. In addition to directly increasing the cost of doing business should one choose to take part in this behavior, corruption may be indicative of other inefficiencies in an economy that will indirectly affect the cost of doing business. It may also be indicative of the business uncertainties one may face including such crucial elements as timely deliveries, capricious inspections, and dispute resolution. While directly measuring corruption is almost impossible for obvious reasons, organizations like Transparency International have developed indirect measures of corruption that may be used to study its impact on FDI. The availability of labor and the quality of human capital are elements that many consider to be critical for FDI. While for simple production processes the availability of cheap, unskilled labor may be the most important determinant of FDI, for more sophisticated FDI, the quality of human capital is what matters most. We expect that FDI flows will be partially explained by both labor size and the level of education of the labor force. A country’s trade orientation and its global readiness are also factors that can be expected to explain FDI flows. As MNEs seek to expand markets and to develop a truly global supply chain, it cannot be assumed that MNEs will engage in FDI only to serve the host market or to export back to the home country. It is increasingly more likely that MNEs engage in FDI to establish a global supply chain that connects production and sales in a number of countries. If that is so, then countries with broad trade orientation, open trade policies and with supporting trade infrastructure are more likely to attract FDI today. To summarize, we use existing FDI stock per capita as a measure of the presence of foreign companies in a country. The market’s size and its growth rate are evaluated along two dimensions: sheer size (population) and effective purchasing power (real GDP per capita and the growth rate of real GDP per capita). Electricity consumption per capita is a proxy for a country’s infrastructure conditions. Business environment is captured by total tax rate, investment protection index, cost of starting business and corruption. We assess labor quality using the rate of secondary school enrollment. Finally, trade volume as a percentage of GDP is regarded as a measure of a country’s trade orientation. We provide further details on the data sources and the construction of the variables in the Data Appendix.

Single Equation Model and Estimation Results We first assess the effects of the aforementioned factors using a simple linear regression model. In practice, there is typically a lag of at least one year between the decision to invest in a country and when the investment is actually made and recorded in the statistics of a country. To account for this time lag, we run regressions using lagged one year values of the explanatory variables. The model is thus specified as follows:

y = + X + , (1)

it it it−1 it where y is theα ratioβ of FDI shareε to GDP share for country i in year t, is the time and country specific intercept term, X is a vector of the explanatory variables for country i in year t, and is the error term. it it Specifically, X consists of log of population, log of real GDP per capita,α growth rate of GDP per capita, electricity consumptionit per capita, total tax rate as a percentage of commercial εprofits,it investment protection index, cost of business start-up procedures as a percentage of GNI per capita, corruption perceptions index, secondary school enrollment as percentage of gross eligible population and trade as percentage of GDP. First, we use F tests to check whether we can pool data across countries and/or over time. In other words, we test the null hypothesis that the intercepts are equal across countries and/or over time. If we fail to reject the null of the poolability, the data are pooled. Otherwise, we apply a Hausman test to test if the random effects estimator is appropriate. The test resultsα푖푡 are in Table 2. The Hausman specification test

Journal of Applied Business and Economics vol. 14(3) 2013 115 results indicate rejection of the null of random effect models at the 5% significance level. The F tests suggest that country fixed effects are significant at the 5% level while time effects are borderline significant at the 10% level. As a robustness check, we estimate the model with and without time effects.

TABLE 2 SPECIFICATION TESTS

Test Specification Tests statistic P-value F test for country effects 5.48** <0.0001 F test for time effects 1.66 0.10 Hausman test for fixed effects 38.18** <0.0001 Notes: *and** indicate the rejection at the 10% and 5% levels, respectively.

Table 3 shows the estimation results for the sample countries. In the model with country fixed effects only, at the 5% significance level, a one percent increase in population contributes to a 2.2526 increase in the FDI share to GDP share, a one percent increase in real GDP per capita leads to a 0.8743 increase in the FDI share to GDP share, a one unit increase in the corruption index (the higher the index, the lower the corruption level) raises the FDI share to GDP share by 0.2060, and a one percent increase in trade as percentage of GDP leads to a 0.0078 increase. The estimated coefficients of the cost of business start-up procedures as a percentage of GNI per capita are also statistically significant. However, they do not have the hypothesized signs. Adding time effects leads to some changes in the estimated coefficients. First, the log of FDI stock per capita becomes significant at the 5% level, with a one percent increase resulting in a 0.2797 rise in the FDI share to GDP share. Second, the log of real GDP per capita is no longer statistically significant. Third, the coefficients of cost of business start-up procedures, the corruption index and trade remain significant albeit with small changes in magnitude. To sum up, results in Table 3 suggest that agglomeration, market size, business environment, corruption and trade orientation are the main determinants of FDI allocation, while electricity consumption and the rate of secondary school enrollment do not matter.

Simultaneous Equation Models and Estimation Results In the previous section, we estimated the effects of the set of variables on the FDI flow using a single linear equation. The implied assumption is that only one such relation as specified in equation (1) exists among the variables. One apparent omission is the contribution of the FDI flow to the FDI stock. On one hand, larger FDI stock can lead to more FDI flow via “agglomeration effects”, and on the other hand, bigger FDI flow adds to higher FDI stock. Out of this consideration, we treat the two variables, the FDI stock and FDI flow, as endogenous variables, determined by a set of exogenous determinants. We estimate the following two Simultaneous Equations (SE) models. Equation (1.1) of SE Model 1 is exactly the same as equation (1). The only difference is that we express X as [S ,M ], where M includes all explanatory variables other than the FDI stock per capita. Equation (1.2) states that the it−1 it−1 it−1 it−1 current FDI stock is determined by previous FDI stock and current inflows. Note that S and F are not the raw series of the FDI stock and flow, and therefore we cannot specify this accumulation process as an identity. Alternatively, in SE model 2, we specify the FDI stock as determined by exogenousit it factors grouped under M in equation (2.2), while the agglomeration effect is arranged formally as equation (2.1). it−1 SE Model 1 = + + + (1.1)

퐹푖푡 α0 α1푆푖푡−1 α2푀푖푡−1 ε푖푡

116 Journal of Applied Business and Economics vol. 14(3) 2013 = + + + (1.2)

푆SE푖푡 Model훽0 2β 1푆푖푡−1 β2퐹푖푡 µ푖푡 = + + (2.1)

퐹푖푡 = α0 +α1푆푖푡−1 +ε푖푡 + (2.2)

푆푖푡 훽0 훽1푆푖푡−1 훽2푀푖푡−1 µ푖푡 TABLE 3 SINGLE EQUATION ESTIMATIONS OF DETERMINANTS OF FDI INFLOWS TO 75 MIDDLE- AND LOW-INCOME COUNTRIES, 2001-2009

Dependent variable Share of FDI to share of GDP Country Fixed Effects Country Fixed Effects and Time Fixed Effects Determinants Estimate t-Value Estimate t-Value

Lag ln(FDI stock per capita) 0.1973 1.61 0.2797** 1.99 Lag ln(Population) 2.2526** 2.05 2.7639* 1.65 Lag ln(Real GDP per capita) 0.8743** 1.79 0.9002 1.47 Lag Real GDP growth rate -0.0069 -0.59 -0.0108 -0.88 Lag ln(Electric power consumption per capita) -0.3698 -1.12 -0.3439 -1.04 Lag Strength of investor protection index (0 to 10) -0.0022 -.02 0.0426 0.38 Lag Total tax rate (% of commercial profits) -0.0048 -1.32 -0.0048 -1.33 Lag Cost of business start-up (% of GNI per capita) 0.0019** 2.02 0.0020** 2.08 Lag Corruption Perceptions Index (0 to 10) 0.2060** 2.31 0.2012** 2.27 Lag Secondary school enrollment (% gross) -0.0001 -0.01 0.0005 0.06 Lag Trade (% of GDP) .0078** 2.18 0.0071** 1.99 Notes: *and** indicate the rejection at the 10% and 5% levels, respectively. Number of cross sections is 75 and number of time periods is 9.

Table 4 presents the estimation results for SE model 1. Compared with the single equation estimation, the agglomeration effect is larger and more significant in the SE model 1. A one percent increase in the FDI stock gives rise to a 0.6421 increase to the FDI share to GDP share in question (1.1) compared to a 0.1973-0.2797 increase in the single equation (1). In other words, the agglomeration effect is more pronounced when we account for the dynamic interactions between the FDI stock and FDI flow. The coefficients of real GDP growth rate, total tax rate, the cost of business start-up, corruption index and secondary school enrollment all have hypothesized signs and are significant at the 5% level. To be specific, a one percent increase in real GDP growth rate contributes to a 0.0512 increase in the FDI share to GDP share, a one percent increase in total tax rate leads to a 0.0034 decline in the FDI share to GDP share, a one percent increase in the cost of business start-up is associated with a 0.0009 drop in the FDI share to GDP share, a one unit increase in the corruption index (the higher the index, the lower the corruption level) raises the FDI share to GDP share by 0.1527, and a one percent increase in the secondary school enrollment leads to a 0.0059 increase. Population, real GDP per capita and trade yield

Journal of Applied Business and Economics vol. 14(3) 2013 117 significant coefficients, but not the anticipated signs. The estimated coefficients in the accumulation equation (1.2) are as expected. The lagged FDI stock has a coefficient of 0.9827 and the share of FDI to share of GDP has a coefficient of 0.0412, and both are significant at the 5% level.

TABLE 4 SE MODEL 1 ESTIMATIONS OF DETERMINANTS OF FDI INFLOWS TO 75 MIDDLE- AND LOW-INCOME COUNTRIES, 2001-2009

Dependent variable Share of FDI to share of GDP FDI stock per capita (1.1) (1.2) Determinants Estimate t-Value Estimate t-Value Lag ln(FDI stock per capita) 0.6421** 12.11 0.9827** 163.82 Lag share of FDI to share of GDP ( not lagged for 1.2) - - 0.0412** 5.85 Lag ln(Population) -0.1693** -6.49 - - Lag ln(Real GDP per capita) -0.9243** -10.80 - - Lag Real GDP growth rate 0.0512** 4.73 - - Lag ln(Electric power consumption per capita) 0.0723 0.98 - - Lag Strength of investor protection index (0 to 10) 0.0120 0.42 - - Lag Total tax rate (% of commercial profits) -0.0034** -2.01 - - Lag Cost of business start-up (% of GNI per capita) -0.0009* -1.69 - - Lag Corruption Perceptions Index (0 to 10) 0.1527** 3.70 - - Lag Secondary school enrollment (% gross) 0.0059** 2.18 - - Lag Trade (% of GDP) -0.0025** -2.11 - - Notes: *and** indicate the rejection at the 10% and 5% levels, respectively. Number of cross sections is 75 and number of time periods is 9.

Table 5 shows the results for the alternative specification, SE model 2. The agglomeration effect is still significant in equation (2.1). A one percent increase in the log FDI stock per capita raises the FDI share to GDP share by 0.3652, smaller than the estimate from the SE model 1 while larger than that from the single equation model. The estimation results for equation (2.2) show that the FDI stock per capita appears to be very persistent and the estimated coefficient of the lagged FDI stock per capita is about 0.94. In addition, a one percent rise in real GDP per capita and real GDP growth rate increase the FDI stock per capita by 0.0792 and 0.0063 percent respectively. The cost of business start up also has a significant coefficient. However, it does not have the hypothesized sign. The rest of the factors are not significantly correlated with the current FDI stock per capita once we include the lagged value of FDI stock per capita. This is not surprising as the FDI stock per capita is a stock variable and its current level is the result of the influence of these factors over time and not just in a single period. In summary, among the factors we examined, the agglomeration effect is robust against different model specifications, suggesting that the MNEs have the tendency to invest in countries that are already heavily invested by foreign companies. Our analysis also suggests that countries can take certain

118 Journal of Applied Business and Economics vol. 14(3) 2013 measures to attract foreign direct investment, though we come to slightly different conclusions as to what factors are most influential depending on model specifications.

TABLE 5 SE MODEL 2 ESTIMATIONS OF DETERMINANTS OF FDI INFLOWS TO 75 MIDDLE- AND LOW-INCOME COUNTRIES, 2001-2009

Dependent variable Share of FDI to share of FDI stock per capita GDP (2.1) (2.2) Determinants Estimate t-Value Estimate t-Value Lag ln(FDI stock per capita) 0.3652** 12.29 0.9385** 82.50 Lag share of FDI to share of GDP - - - - Lag ln(Population) - - 0.0012 0.21 Lag ln(Real GDP per capita) - - 0.0792** 4.35 Lag Real GDP growth rate - - 0.0063** 2.73 Lag ln(Electric power consumption per capita) - - 0.0193 1.24 Lag Strength of investor protection index (0 to 10) - - -0.0049 -0.80 Lag Total tax rate (% of commercial profits) - - -0.0004 -1.04 Lag Cost of business start-up (% of GNI per capita) - - 0.0002** 2.07 Lag Corruption Perceptions Index (0 to 10) - - -0.0051 -0.58 Lag Secondary school enrollment (% gross) - - 0.0005 0.91 Lag Trade (% of GDP) - - 0.0003 1.11 Notes: *and** indicate the rejection at the 10% and 5% levels, respectively. Number of cross sections is 75 and number of time periods is 9.

CONCLUSION

Foreign Direct Investment (FDI) as a source of economic development plays an important role in financing investment and growth for developing countries. We find that the distribution of FDI among developing countries is very uneven, suggesting that some countries are much better in attracting FDI than others. This paper aims to find out influential factors in determining the FDI flows. We look at factors including agglomeration, market size, infrastructure, business environment, corruption, human capital, and trade orientation. We first estimated a single equation model of the FDI share to GDP share onto measures of the aforementioned factors. The empirical results show that the existing FDI stock, population, real GDP per capita, corruption level, and linkage with the global market through trade are important factors in determining FDI inflow to the developing countries. To obtain more accurate estimation results, we used a simultaneous equations model to account for the bidirectional determination between the FDI stock and FDI flow. This more comprehensive model reveals larger and significant agglomeration effects and at the same time, suggests countries with higher GDP growth rate, lower tax rate, smaller cost of business start- up, less corruption and higher secondary school enrollment are more successful in attracting FDI. The empirical findings in this paper are relevant for both MNEs and potential recipient countries. Our findings provide answers to what makes a successful location choice for MNEs’ investment, and point to

Journal of Applied Business and Economics vol. 14(3) 2013 119 the significant factors that potential recipient countries should focus on to attract more FDI. There has been a growing literature in studying FDI flows in a bilateral setup. It would be interesting to expand our analysis framework to bilateral FDI data. We leave that to future research.

REFERENCES

Agodo, O. (1978). The Determinants of U. S. Private Manufacturing Investments in Africa. Journal of International Business Studies, 9 (3), 95-107.

Alfaro, L. et al. (2010). Does Foreign Direct Investment Promote Growth? Exploring the Role of Financial Markets on Linkages. Journal of Development Economics, 91 (2), 242-256.

Asiedu, E. (2002). On the Determinants of Foreign Direct Investment to Developing Countries: Is Africa Different? World Development, 30 (1), 107-119.

Bandelj, N. (2002). Embedded Economies: Social Relations as Determinants of Foreign Direct Investment in Central and Eastern Europe. Social Forces, 81 (2), 411-444.

Barrell, R. & Pain, N. (1999). Domestic Institutions, Agglomerations and Foreign Direct Investment in Europe. European Economic Review, 43, 925-934.

Bennett, P. & Green, R. (1972). Political Instability as a Determinant of Direct Foreign Investment in Marketing. Journal of Marketing Research, 9 (2), 182-186

Borensztein, E., DeGregorio, J. & Lee, J.W. (1998). How Does Foreign Direct Investment Affect Economic Growth? Journal of International Economics, 45, 115–35.

Chakrabarti, A. (2001). The Determinants of Foreign Direct Investments: Sensitivity Analyses of Cross- Country Regressions. Kyklos, 54 (1), 89-114.

Chang, J. & Park, S. (2005). Types of Firms Generating Network Externalities and MNCs' Co-Location Decisions. Strategic Management Journal, 26 (7), 595-615.

Cuervo-Cazurra, A. (2006). Who Cares about Corruption? Journal of International Business Studies, 37 (6), 807-822.

De Mello, L. (1999). Foreign Direct Investment-Led Growth: Evidence from Time Series and Panel Data. Oxford Economic Papers, 51 (1), 133-151.

Deichmann, J. et al. (2003). Foreign Direct Investment in the Eurasian Transition States. Eastern European Economics, 41(1) 5-34.

Dunning, J. (1993). Multinational Enterprises and the Global Economy. Harlow: Addison-Wesley.

Galan, J., Gonzalez-Benito, J. & Zuniga-Vincente, J. (2007). Factors Determining the Location Decisions of Spanish MNEs: An Analysis Based on the Investment Development Path. Journal of International Business Studies, 38 (6), 975-997.

Gastanaga, V. et al. (1998). Host Country Reforms and FDI Inflows: How Much Difference Do They Make? World Development, 26 (7), 1299-1314.

120 Journal of Applied Business and Economics vol. 14(3) 2013 Grossman, G. & Helpman, E. (1991). Innovation and Growth in the Global Economy. Cambridge, MA: MIT Press.

Habib, M. & Zurawicki. L. (2002). Corruption and Foreign Direct Investment. Journal of International Business Studies, 33(2), 291-307.

Harding, T. & Javorcikr, B. (2007). Developing Economies and International Investors: Do Investment Promotion Agencies Bring Them Together? Policy Research Working Paper 4339, Washington, D.C.: The World Bank Group.

Henisz, W. & Macher, J. (2004). Firm- and Country-Level Trade-Offs and Contingencies in the Evaluation of Foreign Investment: The Semiconductor Industry, 1994-2002. Organization Science, 15 (5), 537-554.

Hymer, S. (1976). The International Operations of National Firms: A Study of Direct Investment. Cambridge: Massachusetts Institute of Technology Press.

IMF (2003). Foreign Direct Investment in Emerging Markets. Report of the Working Group of the Capital Markets Consultative Group. Washington, D.C.: International Monetary Fund.

Kindleberger, C. (1969). American Business Abroad. New Haven: John Wiley Co.

Lall, S. & Streeten, P. (1977). Foreign Investment, Transnationals and Developing Countries. Boulder, CO: Westview Press.

Lim, E. (2001). Determinants of, and the Relation Between, Foreign Direct Investment and Growth: A Summary of the Recent Literature. IMF Working Paper 01/175, Washington, D.C.: International Monetary Fund.

Markusen, J. (1995). The Boundaries of Multinational Enterprises and the Theory of International Trade. The Journal of Economic Perspectives, 9 (2), 169-189.

Mitra, D. & Golder, P. (2002). Whose Culture Matters? Near-Market Knowledge and Its Impact on Foreign Market Entry Timing. Journal of Marketing Research, 39 (3) 350-365.

Nigh, D. (1986). Political Events and the Foreign Direct Investment Decision: An Empirical Examination. Managerial and Decision Economics, 7 (2), 99-106.

Shaver, J., Mitchell, W. & Yeung, B. (1997). The Effect of Own-Firm and Other-Firm Experience on Foreign Direct Investment Survival in the USA, 1987-92. Strategic Management Journal, 18 (10), 811- 824.

The World Bank Group and International Finance Corporations. Doing Business (various editions). Washington, D.C.: The World Bank.

Zhang, K. & Markusen, J. (1999). Vertical Multinationals and Host-Country Characteristics. Journal of Development Economics, 59, 233-252.

Journal of Applied Business and Economics vol. 14(3) 2013 121 APPENDIX 1

Data on foreign direct investment (FDI) inflows and stocks are from UNCTAD, Division on Investment and Enterprise. FDI inflows comprise capital received by a foreign direct investor from a FDI enterprise. FDI includes the three following components: equity capital, reinvested earnings and intra- company loans. Data on FDI flows are presented on net bases (capital transactions' credits less debits between direct investors and their foreign affiliates). Net decreases in assets or net increases in liabilities are recorded as credits, while net increases in assets or net decreases in liabilities are recorded as debits. Hence, FDI flows with a negative sign indicate that at least one of the three components of FDI is negative and not offset by positive amounts of the remaining components. These are called reverse investment or disinvestment. The Corruption Perceptions Index (CPI) data are extracted from publications by Transparency International (TI). Countries are ranked on a scale from 10 (very clean) to 0 (highly corrupt) by their perceived levels of corruption, as determined by expert assessments and opinion surveys. Data on population, nominal and real GDP, electric power consumption per capita, strength of investor protection index, total tax rate, cost of business start-up, secondary school enrollment and trade are from the World Bank database. Note that real GDP per capital is measured in constant 2000 dollars. Missing data are interpolated using the Last observation carried forward (LOCF) method. For each series, missing values are replaced by the last observed value of that variable. Table A.1 presents the list of 75 countries in our sample based on the World Bank’s income classification as of 2000. A few countries migrated from one income group to another during our sample period.

TABLE A1 LIST OF 75 COUNTRIES BY WORLD BANK INCOME CLASSIFICATION 2000

Income Group Countries

Bangladesh, Benin, Cambodia, Ethiopia, Guinea-Bissau, Kenya, Low Kyrgyzstan, Mozambique, Nepal, Tanzania, Togo

Armenia, Belize, Bolivia, Cameroon, Cape, Verde, Cote d'Ivoire, Egypt, El Salvador, Georgia, Ghana, Guatemala, Guyana, Honduras, India, Indonesia, Lesotho, Moldova, Mongolia, Morocco, Nicaragua, Nigeria, Pakistan, Lower Middle Paraguay, Philippines, Senegal, Sri Lanka, Swaziland, Syria, Ukraine, Vanuatu, Vietnam, Yemen

Albania, Algeria, Argentina, Belarus, Bosnia and Herzegovina, Botswana, Brazil, Chile, China, Colombia, Costa Rica, Croatia, Dominican Republic, Ecuador, Gabon, Iran, Jamaica, Jordan, Kazakhstan, Malaysia, Mauritius, Upper Middle Mexico, Namibia, Panama, Peru, South Africa, Thailand, Trinidad and Tobago, Tunisia, Turkey, Uruguay, Venezuela

122 Journal of Applied Business and Economics vol. 14(3) 2013

Risk, Return, and Income Mix at Commercial Banks: Cross-Country Evidence

Rifat Gorener Roosevelt University

Sungho Choi Chonnam National University

INTRODUCTION

In the traditional sense, banks are intermediaries that add value and earn income based on the spread between the interest paid on deposits and the interest received on loans. Market analysis of bank performance, prudential regulation of banks, and theoretical models of bank “uniqueness” are all based primarily on this framework of intermediation and spread income. However, non-interest or “fee-based” activities account for a substantial and growing portion of the earnings in the banking sector. For example, non-interest income comprised about 47 percent of the net revenue on average at the 25 largest US banking companies in 2004, a rise from about 40 percent in the mid-1990s and about 35 percent in the mid-1980s. Banking companies in other developed economies have experienced similar trends.1 The surge in non-interest income at commercial banks has come from two sources: the non-traditional financial services that banks have only recently begun to provide and the traditional activities that commercial banks have always provided but that are now produced and priced differently.2 During the 1990s, industry deregulation permitted commercial banks in many countries to expand into non- traditional product offerings, such as investment banking, merchant banking, insurance agency, and

Journal of Applied Business and Economics vol. 14(3) 2013 123 securities brokerage. Unlike traditional intermediation business, which generates an interest margin between depositors and borrowers, these new banking activities primarily generate fee income.3 At around the same time, the advances in information, communications, and financial technologies altered the production and distribution of many traditional banking products; in the process, these products became sources of non-interest income. For example, credit scoring and asset securitization have transformed the production of consumer credit and home mortgages from a traditional portfolio lending process in which banks earn mostly interest income into a transactions lending process in which banks earn mostly non-interest income from loan origination and servicing fees. Similarly, some deposit customers have demonstrated a willingness to pay higher fees (or accept lower interest rates on their balances) for the convenience associated with widespread networks of branches, ATMs, and/or Internet banking facilities.4 Initially, there was conventional wisdom among bankers, regulators, and industry analysts that increased non-interest income would reduce the risk profile of commercial banks: less exposure to interest rate movements and credit risk would stabilize bank revenues and diversification gains from a broader business mix of fee-based activities would stabilize bank profits. However, the empirical literature has not systematically confirmed these initial beliefs. While a handful of studies have found evidence linking non-interest income to lower bank risk, a growing majority of studies are finding contradictory evidence. In practice, the diversification gains appear to be limited (Stiroh, 2004) and/or tend to be consumed by increased risk-taking in other areas (Demsetz & Strahan, 1997). Streams of non-interest revenue from some activities appear to be more, rather than less, volatile than traditional loan-based revenue streams and expose the bank to risk from increased operating leverage (DeYoung & Roland, 2001). Additionally, at the average bank, expansion into non-interest activities appears to offer a poor risk–return trade-off (DeYoung & Rice, 2004). In this study, we examine whether and how increased reliance on non-interest income affects the financial performance of banks, as measured by the stock market return data for 877 publicly traded commercial banking companies in 42 countries between 1995 and 2002. By examining these relationships in this broad, multi-national, financial markets framework, we advance the literature in at least three important ways. First, we provide a global robustness test for previous studies, which have focused on banks in single countries or small groups of similar countries. Second, most (though not all) previous studies have been based on performance data from banks’ financial statements, and as such may reflect the myriad economic misrepresentations embedded in accounting documents. Third, we can exploit the substantial cross-country heterogeneity in our data to test how regulatory practices, economic conditions, and social institutions affect these relationships. Our main results are derived from a two-stage risk–return regression framework, which we estimate multiple times using a variety of risk measures derived from a two-factor market model. In general, we find that non-interest income is associated with riskier stock returns at commercial banking companies, due primarily to increased market, or systematic, risk. The magnitude of this relationship is non-trivial: a one-standard-deviation increase in cross-sectional non-interest income is associated with about a 6 percent increase in the market beta for the average bank in our sample. These findings are new to the literature, and they suggest that fee-based banking activities (e.g., merger financing, loan origination fees, brokerage commissions) increase banks’ exposure to the business cycle. In contrast, we find almost no evidence linking non-interest income to changes in total risk (the variability of stock returns), interest rate risk (a second factor in a market model), or idiosyncratic risk (the residual from a two-factor market model). Our estimation results suggest that, on average, the stock markets in these 42 countries efficiently price the increased risk associated with the non-interest income market. That is, after controlling for cross-sectional differences in risk, the market returns do not fluctuate with the mix of bank income. However, when we re-estimate our models using accounting returns (ROA and ROE) rather than market returns, we find evidence of a premium in accounting returns that is positively related to non-interest income. This set of results is also new to the literature, and it offers a potential explanation for the initial

124 Journal of Applied Business and Economics vol. 14(3) 2013 conventional wisdom among industry participants that expansion into non-interest activities would result in an improved risk–return trade-off at commercial banks. Finally, we find that cross-country differences in regulatory practices, economic conditions, and social institutions influence our main results in important ways, but on average our risk–return results appear to be robust across countries. Because deregulation was important in removing the constraints that prevented banks from participating in certain fee-based activities, reliance on non-interest income will vary across banks based on country-specific regulations. In contrast, increased reliance on non-interest income due to innovations in financial markets and information flows is less likely to vary across countries, because technology allowing more efficient production of traditional banking products spreads quickly. In general, we find that tight controls of the banking system designed to reduce risk (e.g., entry restrictions, explicit deposit insurance, activity restrictions, ownership restrictions, competition policy) are associated with less risky non-interest activities. Our findings have implications for regulation. Activities that generate non-interest income can be less transparent than traditional banking activities, and thus more difficult to monitor for safety and soundness. For example, reductions in equity capital caused by credit losses are obvious and easily identifiable (i.e., the progression of loan delinquency, loan classification, loan provisioning, and loan charge-off), but shocks to non-interest income affect equity capital more subtly through ex post reductions in retained earnings, which for most established firms are the primary source of capital. Activities that generate volatile earnings streams make this source of capital riskier, and this volatility is exacerbated by financial leverage. Under the current regulatory capital rules, banks are not required to hold capital against most fee-generating activities.5 Our findings invite a discussion about whether the risk associated with certain non-interest activities is large enough and systematic enough to merit a required capital charge in future versions of these capital regulations. Our findings also have implications for investors. If the additional risk associated with non-interest income is predominantly systematic risk (as we find here), then investors will require higher expected returns to accept this non-diversifiable risk in their portfolios. The rest of paper is set out as follows. In Section 2, we discuss the literature that investigates the relation between non-interest income and bank profitability and risk. Section 3 describes our data set and presents the methodology employed in our paper. In Section 4, we report our empirical results. Section 5 examines the effect of cross-country institutional difference on banks. Then, in Section 6, we conclude our analysis with a brief summary of our main findings and an assessment of their implications.

LITERATURE REVIEW

The impact of non-interest income and non-traditional activities on commercial bank performance has been the subject of academic and regulatory study for three decades. Most of these studies were performed prior to the large expansion of permissible commercial bank activities made possible by the major deregulatory acts, e.g., the Gramm–Leach–Bliley (GLB) Act of 1999, allowing US financial holding companies to participate in both banking and non-banking activities, and the Second Banking Co- ordination Directive of 1989 (implemented in 1993–94), permitting universal banks to expand anywhere within the EU regardless of the host country restrictions on product powers. These early studies were limited to examining the small number of non-interest activities permissible at the time, and they produced a mixture of interesting although somewhat contradictory results. In general, the studies found that diversifying into non-traditional financial activities could potentially reduce risk, but in practice the post-diversification changes in risk followed no clear patterns, either increasing, decreasing, or remaining unchanged across the various studies. Moreover, the risk reductions that were achievable in practice tended to diminish quickly, and in some instances were reversed as banks ramped up their risk-taking in other areas, such as greater financial leverage. More recent studies—performed after the implementation of the major deregulatory acts, and also in the wake of the financial and technical innovations that changed the production processes for generating both interest and non-interest income—have been able to examine a much broader set of non-interest activities. These later studies have generally concluded that expansion into fee-based financial activities has increased the riskiness of banking companies.

Journal of Applied Business and Economics vol. 14(3) 2013 125 The earliest group of studies provided suggestive evidence that banks could potentially reduce their riskiness by diversifying into non-banking activities. Heggestad (1975), Johnson and Meinster (1974), Litan (1985), and Wall and Eisenbeis (1984) used industry-level IRS data from the 1950s, 1960s, and 1970s to compare the aggregate earnings stream of the banking industry with the aggregate earnings streams of other financial industries (e.g., securities firms, insurance companies, real estate brokers, leasing companies, thrift institutions). While the results of these studies did not always agree across industries, a common thread ran through the studies: over long periods of time, banking industry earnings and non-banking financial industry earnings were quite uncorrelated with each other, and in extreme cases these correlations were close to zero or even negative. This basic result suggested that if banks were allowed to add some non-banking financial products to their traditional mix of banking services, the resulting portfolio diversification effects could potentially increase banks’ expected returns without increasing their riskiness (or, equivalently, reduce banks’ riskiness without reducing their expected returns). Most studies have used firm-level data rather than industry averages, and have found mixed results. One set of these studies concluded that diversification into non-banking activities increased the riskiness of banks. Boyd and Graham (1986) examined large bank holding companies (BHCs) that diversified into non-banking activities during the 1970s and concluded that, in the absence of strict regulatory oversight and control, expansion into these activities can increase the risk of failure. Sinkey and Nash (1993) found that commercial banks that specialized in credit card lending (an often-securitized type of lending that generates substantial fee income) generated higher and more volatile accounting returns, and had higher probabilities of insolvency, than commercial banks with traditional product mixes during the 1980s. Demsetz and Strahan (1997) studied the stock returns of BHCs and found that greater diversification across product lines did not necessarily reduce risk because the diversifying BHCs tended to shift to riskier mixes of activities and hold less equity. Roland (1997) discovered that abnormal returns from fee- based activities were less persistent (more short-lived or volatile) than abnormal returns from lending and deposit-taking at large BHCs. Kwan (1998) found that the accounting returns of Section 20 securities affiliates tended to be more volatile, but not necessarily higher, than the accounting returns of their commercial banking affiliates. On the contrary, other firm-level studies found that diversifying into non-banking activities reduced the bank risk, although these gains tended to be limited in size, scope, or practice. Boyd et al. (1980) measured the correlations between accounting returns at the bank and non-bank affiliates of BHCs during the 1970s, and found that the potential for risk reduction was exhausted at relatively low levels of non- banking activities. Eisenbeis, Harris, and Lakonishok (1984) found positive abnormal stock returns associated with the formation of one-bank holding companies between 1968 and 1970, a brief time period during which these firms were permitted to engage in a wide variety of non-banking activities. Kwast (1989) examined the accounting returns of the securities and non-securities activities of commercial banking companies between 1976 and 1985, and found limited potential for risk reduction by diversifying into securities activities. Brewer (1989) reached similar results using the market returns of US bank holding companies. Gallo, Apilado, and Kolari (1996) ascertained that high levels of mutual fund activity were associated with increased profitability, but only slightly moderated risk levels, at large BHCs between 1987 and 1994. Rogers and Sinkey (1999) found that non-traditional activities were associated with larger size, smaller interest margins, less core deposit funding, and less risk. Because the set of permissible non-banking activities was substantially constrained prior to the late 1990s and early 2000s, an alternative research approach examined the return streams of unrelated banking firms and non-banking financial firms, and then calculated the hypothetical reduction in earnings variability based on the covariances of those earnings streams. Rosen et al. (1989) found minimal financial benefits from the hypothetical diversification of banking companies into real estate activities. Wall et al. (1993) constructed synthetic portfolios based on the accounting returns of banks and non- banking financial firms, and concluded that banks would have experienced higher returns and lower risk had they been able to diversify into small amounts of insurance, mutual fund, securities brokerage, or real estate activities during the 1980s. Using both accounting data and market data from the 1970s and 1980s,

126 Journal of Applied Business and Economics vol. 14(3) 2013 Boyd, Graham, and Hewitt (1993) concluded that BHCs could have reduced their riskiness by merging with life insurance or property/casualty insurance firms, but would likely have increased their riskiness by merging with securities or real estate firms. Laderman (1998) applied a similar synthetic merger approach to data from the 1980s and 1990s, and concluded that BHCs could have reduced the volatility of their accounting returns by offering “modest to relatively substantial amounts” of life insurance or casualty insurance underwriting. Allen and Jagtiani (2000) used stock market data to construct return streams for synthetic “universal banks” consisting of a commercial banking company, a securities firm, and an insurance company, and found that the exposure to market risk increased with the addition of these non- banking activities. While these studies yielded provocative results, this research approach is limited because (by construction and by necessity) it cannot capture the positive and/or negative synergies in production, marketing, or organizational control from combining banking activities with non-banking activities. Despite the very mixed evidence produced by the academic research outlined above, by the late 1990s there was still conventional wisdom among industry participants that fee-based activities had a stabilizing effect on bank income and that diversifying into non-banking activities reduced bank risk. This conventional wisdom was documented by DeYoung and Roland (2001), who also provided three new conceptual arguments for why non-interest income may be less stable than income from traditional banking activities. First, the revenue from traditional lending activities may be relatively stable over time because switching costs and information costs make it costly for either borrowers or lenders to walk away from a lending relationship; in contrast, the revenue from some fee-based activities may be relatively unstable because banks face a high level of competitive rivalry, low information costs, and fluctuating demand in a number of these product markets (e.g., investment banking, retail brokerage, securitized mortgage refinance). Second, fee-based services can require the bank to increase its ratio of fixed-to- variable expenses (e.g., skilled labor, retail sales space); the higher operating leverage that results increases the sensitivity of the bank earnings to fluctuations in the bank revenues. Third, although banks internally allocate some capital to their fee-based lines of business, there is no regulatory capital requirement for most of these activities, which suggests a higher degree of financial leverage—and thus higher earnings volatility—for these activities. Using quarterly data from US commercial banks between 1988 and 1995, DeYoung and Roland (2001) showed that non-interest income (from non-deposit-related activities) was associated with more volatile revenue streams, a higher degree of total leverage, and more volatile earnings streams, when compared with interest-based lending activities. Similarly, the studies of US banking companies that followed also contradicted the conventional industry wisdom. Cooper, Jackson, and Patterson (2003) found that non-interest income was associated with lower risk-adjusted stock returns at bank holding companies. Stiroh (2004a) found little risk- reduction benefit for small banking companies that engaged in both traditional and non-traditional banking activities, although he found substantial diversification benefits for small banks that diversified within traditional or non-traditional areas. In another study that included data from both large and small banks, Stiroh (2004b) found reductions in earnings volatility at banks with large amounts of non-interest income, but concluded that these were due to the reduced volatility of net interest income at those banks, and were not due to diversification gains. DeYoung and Rice (2004) concluded that expansion into non- interest activities appeared to offer a poor risk–return trade-off. Studies using non-US data remain rare, and to date have found mixed results. Esho, Kofman, and Sharpe (2005) found that increased reliance on fee-based activities was associated with increased risk at Australian credit unions. Like the US studies, Smith, Staikouras, and Wood (2003) found that non-interest income exhibited more volatility than net interest income over time; the negative correlations between these two income streams led them to conclude that diversification effects exist. Overall, these more recent studies have benefited from the availability of more detailed information about the composition of non-interest-based activities and, since they used more recent data either wholly or partially drawn from post-deregulation markets, as such reflected less constrained behavior and performance of banking companies.

Journal of Applied Business and Economics vol. 14(3) 2013 127 METHODOLOGY

Our primary objective is to determine whether and how non-interest income affects the riskiness of banking companies, and whether the market returns of these companies are sensitive to any such changes in risk. To answer these questions, we first need to generate bank-level measures of risk (RISK) and return (RETURN). We use the following two-factor market model to generate bank-level measures of RISK:

Rit = αi + βi MRit + γi INTit + εit (1) where Rit is the return on the stock of bank i during period t, MRit is the return on a stock market index in bank i’s country during period t, INTit is the change in the market interest rate on a benchmark long-term government bond in bank i’s country during period t, and the error term εit is assumed to be distributed symmetrically with mean zero. We estimate (1) separately for each bank i in each year of our 1995–2002 database, using weekly data (i.e., t = one week) and ordinary least squares techniques with robust errors. We use four different RISK measures in our tests, each of which is derived from equation (1). Systematic or market risk (MKTRISK) is the estimated value of βi. Interest rate risk (IRRISK) is the estimated value of γi. Idiosyncratic risk (IDIORISK) is the standard deviation of the estimated residual terms εit. Total risk (TOTRISK) is the standard deviation of Rit. Our primary RETURN measure is the annualized average of the weekly returns Rit. In alternative tests, we use two accounting return measures, the annual return on assets (ROAi) and the annual return on equity (ROEi). We estimate the relationships between non-interest income and banking company performance in the following system of RISK–RETURN equations:

RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit (2)

RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit (3) which we estimate using two-stage least squares estimation techniques and an unbalanced panel of annual data from 1995 through 2002.6 NII equals the ratio of non-interest income to total assets at bank i in year t. We use alternative definitions of NII in robustness tests. Thus, the main test statistics are the estimated coefficients λ, φ, and δ. Yit and Zit are vectors of control variables (described below). T is a vector of time fixed-effects dummies. We also use bank fixed effects and country fixed effects in robustness tests. The error terms πit and ηit are assumed to be distributed symmetrically with mean zero. The coefficient λ has a straightforward interpretation: it is the sensitivity of a bank’s riskiness to increases in its non-interest income. As the previous literature has found evidence of both positive and negative associations between non-interest income and risk, we have no a priori expectations about the sign of λ. However, the interpretation of λ will depend on how RISK is defined. If RISK = TOTRISK, then we are testing whether non-interest income increases the risk exposure of undiversified stakeholders, such as bank employees, bank managers, or bank regulators. However, if RISK = IDIORISK, then we are testing whether the risk associated with non-interest income is idiosyncratic to bank i, and hence is of no concern to a diversified investor. Conversely, if RISK = MKTRISK, then we are testing whether non- interest income increases banks’ exposure to the ups and downs of the stock market and the economy in general, creating risk that cannot be avoided by diversified shareholders. Finally, if RISK = IRRISK, then we are testing whether non-interest income reduces banks’ exposure to movements in interest rates, part of the conventional wisdom initially posited by bankers and industry analysts. The coefficient δ also has a straightforward interpretation: it measures the marginal risk–return trade- off required by investors in banking company stock. Obviously, we expect this effect to be positive in an efficient market. Because this trade-off may not be linear, we also estimate alternative regression specifications that include both linear and quadratic RISK terms. The coefficient φ has a more subtle

128 Journal of Applied Business and Economics vol. 14(3) 2013 interpretation: it is the sensitivity of stock returns to increases in non-interest income after controlling for any change in the risk–return trade-off (i.e., δRISK) caused by the increase in non-interest income. A zero estimated value for φ would be consistent with an efficient stock market in which the risks associated with non-interest activities are priced just as efficiently as the risks associated with more traditional, interest- based banking activities. However, it is possible that the risks associated with relatively new banking activities are not yet well understood by investors, and hence may be poorly priced.7 If the market systematically underprices (overprices) the risk associated with non-interest activities, then we expect a positive (negative) φ coefficient. Given the large number (42) of nations in our data, it is not possible to observe the various sources of non-interest income and how they may differ across banks. Our non-interest income variable NII combines all the non-interest income earned by a bank—fees, commissions, trading gains/losses, etc.— from all its lines of business. Because demand schedules and production functions can vary greatly across these different lines of business, the unobserved composition of NII may drive some of our findings. For example, if we find that NII is positively associated with MKTRISK, then we might infer that NII is weighted toward activities that are especially sensitive to the business cycle, such as investment banking, merger finance, or retail brokerage. Similarly, if we find that NII is positively associated with IRRISK, then we might infer that NII is weighted toward activities that are especially sensitive to interest rate fluctuations, such as the fee income earned via loan origination, securitization, and servicing. Not having access to the line-of-business composition of non-interest income also limits our ability to test for diversification effects. At best, we can draw inferences about whether combining non-interest income with all the other banking activities creates diversification gains. For instance, a finding that NII is positively associated with MKTRISK, but negatively associated or neutral with respect to TOTRISK, would imply that an increase in non-interest income generates gains from diversification within the average bank. These gains could be due to cost synergies (input sharing), revenue synergies (e.g., cross- selling), or less than perfect covariation of the cash flows between interest-based and non-interest-based activities. The variable definitions, data sources, and summary statistics for all of the regression variables used to estimate (1), (2), and (3) are displayed in Table 1. Our primary data source was the Fitch-IBCA Bankscope database, from which we drew our initial data set of 14,404 financial institutions and banks from 46 different countries observed annually from 1995 through 2002. From this large initial sample, we retained only publicly traded bank holding companies (BHCs), commercial banks, and savings banks for which we had full information.8 This resulted in a final data set of 877 banking companies from 42 countries. Table 1 displays some descriptive information for our bank data set. The accounting data for banks come from the Fitch-IBCA Bankscope database and the market data, including individual stock return data, interest rate data, and market index, for each country come from Datastream. The mean NII is 1.85 percent with a range from -6.27 percent to 39.26 percent, which indicates the wide variation among our sample banks. For the US studies, DeYoung and Rice (2004) reported that the aggregate non-interest income was 2.39 percent for the year 2001 and differed greatly between larger banks and smaller banks. Smith et al. (2003) reported that in EU countries, the average non-interest income increased from 0.88 percent of the total assets in 1994 to 1.09 percent in 1998. The difference among studies is mainly caused by the different samples under study. The mean TOTRISK is 2.28 percent, with a range from 0.08 to 39.45 percent. Our sample banks on average are less risky than the overall market of their country: the mean market risk is 0.45, which is substantially smaller than the market beta of 1. The mean of the interest rate risk and bank-specific risk is -0.3268 and 0.0207, respectively. Finally, the average stock return of banks is 3.46 percent per annum, with wide variation among banks.

Control Variables We include vectors of control variables in equations (2) and (3), denoted as Y in the RISK regression and Z in the RETURN regression. These variables are included to specify each equation better and to identify the system of equations. In addition, we include a vector of fixed time effects T in each regression to absorb the average year-to-year variation in risk and return in the banking sector.

Journal of Applied Business and Economics vol. 14(3) 2013 129 Six of the control variables appear in both the RISK equation (2) and the RETURN equation (3). lnASSETS is the natural logarithm of assets for bank i. Size may be risk-reducing because it increases the potential for diversification and gives banks greater access to sophisticated risk-management tools; however, the very largest banks may have greater systematic risk since their earnings are likely to be closely aligned with the general economy. Holding risk constant (recall that RISK appears on the right- hand side of the RETURN regression), the returns will increase (decrease) with the bank size if the average bank experiences scale economies (diseconomies); if the bank size is associated with a high- volume, low-margin, transactions banking strategy, then the returns may decline with the bank size (DeYoung, Hunter, & Udell, 2004); and there may be higher returns if the market prices-in a “too big to fail” premium for the very largest banks. LOANS is the ratio of total loans to total assets for bank i. A high loan-to-asset ratio may be risk-increasing because loans expose the bank to greater credit risk and greater market risk than other assets do; however, this ratio may signal an efficiently run bank, and hence less risk due to effective risk management. Holding risk constant, a high loan-to-asset ratio is likely to generate higher returns. LIQUID is the ratio of liquid assets to total assets for bank i. A store of liquid assets may represent a risk cushion that allows a bank to take on more risk, or it could signal risk-averse management, in which case it will be associated with lower levels of risk. Holding risk constant, a large store of low-yielding liquid assets is an opportunity cost and hence will be likely to be associated with lower returns. LOSSPROV is the ratio of provisions for loan losses to total assets for bank i. Assuming that this ratio indicates a bank’s expected credit risk, it should be associated with higher total risk, market risk, and idiosyncratic risk. Holding risk constant, high provisioning should reduce the returns. EXPLICIT is a dummy variable indicating that bank i operates in a country with explicit deposit insurance. If explicit deposit insurance schemes increase the moral hazard incentives for bank managers, banks may take more risks on average; however, if market investors interpret this policy as an implicit government guarantee, the market risk may decline. Depositors can discipline banks by demanding higher interest rates or withdrawing their deposit. However, the explicit deposit insurance reduces depositors’ incentives to monitor a bank. Therefore, the explicit deposit insurance lowers banks’ interest expenses and makes interest payments less sensitive to bank risk. Therefore, we expect that EXPLICIT has a positive impact on returns. CR3 is the three-bank concentration ratio in bank i’s country. In the absence of competitive rivalry (high CR3), managers may choose the “quiet life,” in which case the risk will decline, and/or they may slacken off and run the bank poorly, in which case the earnings could grow more volatile. Holding risk constant, high market concentration will be associated with higher returns due to less competitive rivalry. There are five control variables that appear only in the RISK equation (2). CAPITAL is the ratio of market-value equity to book-value assets for bank i. A large store of market-value equity (relative to the size of the bank) indicates a vote of confidence from investors that allows the bank to take on more risk without immediate penalty. RESTRICT is an index constructed by Barth, Caprio, and Levine (2001) that increases with the restrictions on permissible bank activities (e.g., securities, insurance, real estate) in bank i’s country. To the extent that these restrictions prevent banks from achieving their most efficient (or most preferred) risk–return trade-off, this index could be either positively or negatively related to risk. CORPGOV is an index constructed by Kaufman, Kraay, and Zoido-Lobaton (the KKZ index,9 1999 and 2005) that increases as the governance environment in bank i’s country improves. We expect this index to be negatively related to risk, because poor corporate governance is typically associated with higher risk- taking by managers. BANKFREE10 is an index that increases as restrictions on the banking and finance decline (easier entry, less state ownership, less government credit allocation, etc.) in bank i’s country. The sign of this variable is ambiguous: while greater bank freedom allows banks to engage in risky activities, it also removes the impediments to achieving an efficient risk–return trade-off. MULTSUPS is a dummy variable that indicates that bank i’s country has more than one bank supervisory body. If coordination between supervisors is difficult, or if multiple chartering authorities precipitate a “race to the bottom,” then this variable will be associated with greater risk. There are two control variables that appear only in the RETURN equation (3). The coefficients for these variables should be interpreted holding risk constant. STATE and FOREIGN are dummy variables

130 Journal of Applied Business and Economics vol. 14(3) 2013 and measure the ownership of bank i by either the state or foreign, respectively. The expected coefficient sign for STATE would depend on whether bank i is a private bank expecting greater returns because state-owned rival banks provide poor competition or a state bank expecting lower returns. A similar story holds for foreign ownership.

BASIC RESULTS

The results from the full-sample estimations of equations (2) and (3) are displayed, respectively, in Tables 2 and 3. There are three main findings in these regressions. First, non-interest income alters the composition of risk at the average bank in our sample, but neither increases nor decreases the overall risk. Second, we find positive associations in the data between the various risk measures and market returns, an indication that investors are pricing bank risk in these 42 countries. Third, after controlling for risk, we find no statistical relationship between non-interest income and market returns—that is, there is no premium or discount associated with non-interest activities on average. In Table 2, we find a positive and statistically significant association between non-interest income and bank risk only when risk is defined as MKTRISK. Based on our estimates, a one-standard-deviation increase in NII is associated with a 6 percent increase in market risk at the means of the data. There is no evidence linking non-interest income to total risk, interest rate risk, or idiosyncratic risk. These results imply the existence of gains from diversification: greater reliance on non-interest activities makes the average bank in our data more sensitive to general market volatility (systematic risk) without increasing its overall return volatility. In Table 3, we find positive and statistically significant risk–return relationships, evidence that investors require higher returns in exchange for accepting higher (diversifiable and non-diversifiable) risks. Based on our estimates, one-standard-deviation increases in TOTRISK, MKTRISK, and IDIORISK are associated with 82 percent, 94 percent, and 80 percent increases in RETURN, respectively, at the means of the data. Given this robust evidence of risk pricing in the data, we conclude that the lack of statistical significance for IRRISK in column [c] is likely to be due to the difficulties associated with measuring interest rate risk. The coefficient for NII is statistically non-significant in all four regressions, consistent with efficient markets that bid away any excess risk-adjusted returns associated with the product mix.

Control Variables The bank’s total assets might be the most obvious factor related to the level of bank risk. In Table 1, the sign of LogAssets is negative and significant in three regression models, except the market risk model, which indicates that larger banks are relatively more diversified, which lowers their total, bank-specific, and interest risk. However, as banks’ size grows, they involve more non-interest income activities, which have more comovements with the market, and thus their systematic market risk increases. The bank size has a negative effect on the stock return controlling for the risk–return trade-off. This result indicates that the average bank experiences scale diseconomies or is associated with a high-volume, low-margin, transactions banking strategy. LOANS is negative and significantly associated with all four risk measures. The results indicate that a high loan-to-asset ratio may signal an efficiently run bank, and hence less risk due to effective risk management. Holding risk constant, however, we find that a high loan-to-asset ratio is likely to generate lower returns. The measure of financial leverage (CAPITAL) is positive and significantly associated with total risk as well as market risk measures. The result is consistent with our expectation: a large store of market-value equity indicates a vote of confidence from investors that allows the bank to take on more risk without immediate penalty. The measure of asset liquidity (LIQUID) is positively and significantly related to most risk measures and this effect might be due to the fact that liquid assets are shifty assets that are more difficult to monitor and can be moved quickly into other investments, meaning that as banks increase their liquid assets, the risk increases. The liquidity reduces the profitability significantly. The ratio of a provision for loan losses to total assets (LOSSPROV) in our regression to control for the credit

Journal of Applied Business and Economics vol. 14(3) 2013 131 risk is positively related to all the risk measures. Obviously, LOSSPROV reduces the performance of a bank significantly. The country-specific variables are significantly associated with the risk of banks as well as the performance of banks. The dummy variable for a market-based economy (MARKET) is positive and weakly significant, indicating that banks in a market-based country have higher risk than banks in a bank- based country. This is mainly due to the limited positive role of banks in a market-based country. The existence of explicit deposit insurance (EXPLICIT) reduces the risk of banks. This result is not consistent with the moral hazard of bank managers. The result indicates that market investors interpret this policy as an implicit government guarantee, resulting in lower risk. However, we do not find any significant effect on the stock return. CORPGOV is highly significant and negative in risk regressions. This suggests that better governance reduces the risk of banks significantly. Better governance is associated with lower risk- taking behavior of managers in general. BANKFREE is negative and significantly associated with bank risk. This indicates that if a bank is more free and open, then it can achieve an efficient risk–return trade- off. The banking concentration increases the bank’s market risk significantly, while it also increases its profitability. This is consistent with the “too-big-to-fail” argument. Multiple existences of bank supervisory bodies increase banks’ risk in general. The state ownership of banks significantly reduces the performance of banks.

Non-Linear Risk–Return Trade-Off It is unlikely that the risk–return trade-offs for non-interest banking activities and interest-based banking activities will be the same. Non-interest income derived from traditional banking activities, such as fees charged to core depositors, is likely to be relatively stable, while non-interest income derived from less traditional activities, such as investment banking, securities brokerage, or mortgage origination/securitization, may fluctuate substantially with the local, regional, or international business conditions (DeYoung & Roland, 2001). We cannot test this proposition directly because, as stated above, our non-interest income data are not reported by lines of business; however, if we reasonably assume that banks with especially large volumes of non-interest income tend to have less traditional business mixes, then we can test this proposition indirectly. Table 4 shows the partial results from alternative specifications of equation (3) that include the interaction variable RISK*NII, and reports the partial derivatives with respect to both RISK and NII. The coefficients of the RISK*NII term offer weak evidence that the market requires a higher return by banks with especially high levels of non-interest income: these coefficients are positive (negative for IRRISK) but not statistically significant. We conduct two additional tests in Table 4. First, we calculate a partial derivative of return with respective to NII and RISK and test whether the partial derivatives are statistically different from zero. Generally, we find that the partial derivatives are significantly different from zero, which suggests that both NII and Risk have a significant effect on the returns of banks. Second, we conduct the joint significance test to determine the joint significance of NII and RISK*NII as well as the joint significance of RISK and RISK*NII. While we find that NII and RISK*NII are weakly jointly significant, RISK and RISK*NII are always jointly significant. Overall, the findings suggest weak evidence that the market requires a higher return from banks with especially high levels of non-interest income.

Accounting Returns Strategic investment and product mix decisions at all firms, including banking companies, are often made and evaluated based on accounting information rather than market returns. Perhaps the main reason for this is that, while it is difficult to isolate the impact of a given business decision or investment project on a firm’s stock price, the project’s accounting-based rate of return can be calculated by tracing the impact of that project on the revenue and expense lines in the firm’s income statement. It is also likely that managers make investment and product mix decisions based on the probable impact of those decisions on their firms’ financial statements, because their promotions and compensation (and this is especially true for middle-level and quasi-upper-level managers) are heavily influenced by easily quantifiable information in financial statements. Thus, having found no statistical relationship between

132 Journal of Applied Business and Economics vol. 14(3) 2013 non-interest income and risk-adjusted market returns, we are interested in whether non-interest income is related to accounting rates of return. Table 5 displays the results of equation (3) re-estimated using accounting returns (ROA and ROE) in place of RETURNS as the dependent variable. We find a strong statistical relationship between NII and ROA (first panel) and also between NII and ROE (second panel). Based on our column [a] estimates, a one-standard-deviation increase in NII is associated with a 76 percent increase in ROA (84 basis points) and a 41 percent increase in ROE (347 basis points) at the means of the data. Indeed, the evidence here suggests that had an accounting-based financial analysis been performed during our sample period, there would have appeared to be a risk-adjusted return premium associated with non-interest income. This is consistent with the conventional wisdom at the time among industry participants that expansion into non- interest activities resulted in improved risk–return trade-offs at commercial banks.

Subsamples by Time Non-interest income has been increasing as a percentage of commercial bank income in the US since the early 1980s (e.g., DeYoung & Rice, 2004; Kaufman & Mote, 1994,). Figure 1 shows that non-interest income increased for both US and non-US commercial banks during our 1995–2002 sample period, although the increase moderated in the later years. This change in the income composition has required banks to conduct many things differently from in the past, including using new production processes, new financial products and services, and new pricing methods. As discussed above, it is likely that these changes affected banks’ risk profiles through changes in their revenue streams, operating and/or financial leverage, exposure to interest rate risk, etc. It is also likely that these continuous innovations made it difficult at first for market investors to price commercial bank equity shares accurately; risk-averse investors with little knowledge of the true risk profiles of these new business methods may have initially priced-in a large risk premium, and relaxed this stance after several years of financial performance data observations made the risks associated with these new methods more transparent.11 We re-estimated our system of equations (2, 3) after splitting the data into an early time period (1995–1998) and a later time period (1999–2001). The results are shown in Tables 6 and 7. These subsample regressions are important tests of robustness, as they are largely consistent with our main full- sample results (see Tables 2 and 3) in both the early and the later subsamples. In the equation (2) regressions displayed in Table 6, MKTRISK is positively and significantly associated with NII in both time periods, while TOTRISK and IDIORISK continue to be statistically unrelated to NII in both time periods. In the equation (3) regressions displayed in Table 7, there is a positive estimated trade-off between RETURN and TOTRISK, MKTRISK, and IDIORISK in both time periods, and no statistical relationships between RETURN and NII in either time period. Moreover, these regressions are consistent with (though by no means constitute strong proof of) our above conjecture that investors reduced their assessments of the riskiness of non-interest income as time passed. For example, in Table 6, the positive coefficient for NII declines in size, suggesting that the market scaled back its perception of the riskiness of non-interest income as time passed. Similarly, the market associated non-interest income with decreased interest rate risk early in the sample period, but this belief disappeared in the later period. The data in Table 7 tell a similar story. The positive risk–return trade-offs declined by about 50 percent for TOTRISK and IDIORISK, and by about 25 percent for MKTRISK, as time elapsed.12

Additional Robustness Tests We performed a number of additional robustness tests of the main results displayed in Tables 2 and 3. When we imposed bank fixed effects on the model, the signs and statistical significance levels for all the main test coefficients, NII in equation (2), and NII and RISK in equation (3), were unchanged. The results are not reported here, but they are available from the authors upon request.

Journal of Applied Business and Economics vol. 14(3) 2013 133 CROSS-COUNTRY RESULTS

To investigate further the relationship between the non-traditional activities of a bank and its risk and return, in this section we examine the extent to which the institutional environment affects the relationship. Many economists have argued that bank-based systems are better at mobilizing funds, managing risk, and exerting sound corporate control, especially in weak institutional environments (Levine, 1997). Calomiris (1994) showed that the historical evidence indicates that German universal banks (with a relatively high level of non-traditional activities) were more efficient (lower cost of capital) than US banks and suffered fewer systemic problems than the US banking system. On the other hand, if a bank is operating in a market-based economy, its positive role is relatively limited and results in low demand for non-traditional activities. Therefore, banks are not good at managing risk due to non-interest income activities. Tables 8A and 8B show the results. In a market-based country, NII has a significant and positive association with the market risk. We do not find any impact of NII on risk in a bank-based country. In Table 8B, we find a more interesting result. In a market-based country, there is a significantly positive risk–return trade-off. On the other hand, we do not find a significant risk–return trade-off in a bank-based country. The result, however, shows that NII has a significantly positive relationship with stock returns. This suggests that the market efficiently prices the risk–return trade-off in a market-based economy, but not in a bank-based economy. The risk of non-traditional activities can be contained through effective and prudential regulation and supervision. Depositors frequently lack the incentives and capabilities to monitor and discipline banks. Hence, governmental regulation and supervision may reduce the information asymmetries and are often essential to ensure the solvency of the whole banking system. The stronger supervisory power and strict regulation will improve the governance of banks by direct monitoring and discipline and will increase the likelihood that banks will allocate resources efficiently based on risk–return trade-offs. The enhanced governance of banks through strengthened supervisory power and stringent regulation may reduce the moral hazard of banks engaging in non-traditional activities. Therefore, tougher supervision and regulation lead to risk reduction. On the other hand, supervisors may maximize their private welfare instead of maximizing more social welfare, which preserves the safety and soundness of the banking system. In addition, tighter regulation and more supervision may reduce the competition, efficiency, and competitiveness of the banking system. This view, therefore, suggests that the risk profile due to non- traditional activities may differ among the levels of supervision and regulation. We use the bank freedom index as our main variable to test the potential different effects of NII on risk and return. Tables 9A and 9B document the findings. We find that NII has a significant and positive association with the total and market risk if the bank freedom is high. In contrast, there is a negative association between NII and risk in a low-bank-freedom country. The results, therefore, suggest that stronger supervisory power and strict regulation will improve governance and lead to risk reduction. In Table 9B, we find a significant risk– return trade-off in both countries. The deposit insurance, especially explicit deposit insurance (EDI), reduces the losses that depositors incur in the case of bank failure. However, having an explicit deposit insurance scheme may lead to greater moral hazard for bank managers, who may take advantage of the deposit insurance program by engaging in more activities that may increase risk. The banking literature suggests that the more generous deposit insurance is, the greater are the risk-taking incentives for banks. Deposit insurance may make depositors less likely to enforce market discipline on banks and may induce banks to take additional risks.13 The empirical evidence also shows that deposit insurance increases the probability of banking crisis (Demirguc-Kunt & Detragiache, 2002; Demirguc-Kunt & Huizinga, 2004). A more generous deposit insurance scheme may therefore lead to greater moral hazard for the bank managers. Bank managers may take advantage of the explicit deposit insurance by greatly engaging in non-traditional activities, which may be riskier than traditional lending activities. Tables 10A and 10B show the results based on the existence of EDI. Consistent with the literature, we find a strong and positive association between NII and market risk in a country with EDI, while we find a negative relationship in a country

134 Journal of Applied Business and Economics vol. 14(3) 2013 without EDI. The finding suggests that explicit deposit insurance increases the moral hazard and thereby induces the risk-taking incentives for banks, resulting in a great risk to banks. Finally, concentration (CONCEN) is the share of assets of the three largest banks in a country. If the banking sector is concentrated due to regulation, few large banks may enjoy rents. Bank concentration may lead to a “too-big-to-fail” policy and depositors do not have to care about bank failure. Bank concentration may also lead banks to engage in more non-traditional activities by exploiting the implicit guarantee from the government. We include a measure of banking concentration of each country from the World Bank Database (2004). This is the measure of the fraction of assets in the five largest banks that is owned by commercial banks and/or financial conglomerates. If there are fewer than five banks, it uses that number to calculate the index. Bank concentration may lead to the TBTF policy and depositors not having to care about bank failure. Bank concentration may also lead banks to engage in more risky activities by exploiting the implicit guarantee from the government. Tables 11A and 11B show the results based on the concentration. We find a strong and positive association between NII and market risk in a less concentrated market. The finding suggests that concentration leads banks to engage in more non- traditional activities by exploiting the implicit guarantee from the government and it results in a great risk to banks.

CONCLUSION

Even though there is evidence of decreasing traditional activities among banks, the empirical studies on banks have mainly analyzed the role of the banks in terms of the traditional activities generating interest income. Kaufman and Mote (1994) argued that the nature of banking activities changed steadily in the 1990s. DeYoung and Roland (2001) showed that non-interest income at FDIC-insured commercial banks increased from 25% to over 40% of their aggregate income over the period 1984 to 2001. This is not only a trend for US banks, but also a trend for banks worldwide. The growth of non-interest income seems to have a positive effect on bank profitability and few studies have found consistent evidence on this trend in European countries (Smith et al., 2003). Although there is a general belief that fee-based earnings are more stable than loan-based income, DeYoung and Roland (2001) provided three potential observations that explain that fee-based income may not be more stable than income from traditional banking activities: the low switching and information costs, the need to hire an additional fixed labor input (more expense ratio), and the no regulatory capital requirement that banks do not need to hold capital against fee-based activities (higher degree of leverage). The empirical results of the effects of non-traditional banking activities on the riskiness of bank are mixed. It is therefore interesting and important to know whether and how these new activities affect the nature of riskiness and the performance of banks. This paper investigates the effect of non-interest income on the riskiness and profitability of banks. Specifically, using the data of 877 banks in 42 countries, we investigate whether non-traditional activities of banking institutions affect the risk and returns of banks. This is the first paper to examine this issue in a global context using capital market measures of risk as well as market and book value measures for performance. In general, we find that non-interest income is associated with riskier stock returns at commercial banking companies, due primarily to increased market, or systematic, risk. These findings are new to the literature, and they suggest that fee-based banking increases banks’ exposure to the business cycle. In contrast, we find almost no evidence linking non-interest income to changes in other risk measures, such as total risk, interest rate risk, or idiosyncratic risk. In addition, our findings suggest that, on average, the stock markets efficiently price the increased risk associated with the non-interest income market. That is, after controlling for cross-sectional differences in risk, market returns do not fluctuate with the mix of bank income. However, we find evidence of a premium in accounting returns that is positively related to non-interest income. This set of results is also new to the literature, and offers a potential explanation for the initial conventional wisdom among industry participants that expansion into non-interest activities would result in an improved risk–return trade-off at commercial banks. Finally, we find that cross-country differences in regulatory practices, economic conditions, and social institutions

Journal of Applied Business and Economics vol. 14(3) 2013 135 influence our main results in important ways, but on average our risk–return results appear to be robust across countries. Our findings have several policy implications for regulation. Non-traditional activities can be less transparent than traditional banking activities, thus they are more difficult to monitor for safety and soundness. For example, reductions in equity capital caused by credit losses are obvious and easily identifiable (i.e., the progression of loan delinquency, loan classification, loan provisioning, and loan charge-off), but shocks to non-interest income affect equity capital more subtly through ex post reductions in retained earnings, which for most established firms are the primary source of capital. Activities that generate volatile earnings streams make this source of capital riskier, and this volatility is exacerbated by financial leverage. Under the current regulatory capital rules, banks are not required to hold capital against most fee-generating activities. Our findings invite a discussion about whether the risk associated with certain non-interest activities is large enough and systematic enough to merit a required capital charge in future versions of these capital regulations. Our findings also have implications for investors. If the additional risk associated with non-interest income is predominantly systematic risk (as we find here), then investors will require higher expected returns to accept this non-diversifiable risk in their portfolios.

ENDNOTES

1. Based on Federal Reserve Y-9C filings. We define net revenue as net interest income plus non-interest income. See Table 1 and Figure 1 below for some trends in non-interest income for non-US banking companies. Early studies documenting increases in non-interest income at commercial banks include those by Boyd and Gertler (1994) and Kaufman and Mote (1994). Choi (2005) documents the differences in the growth and determinants of non-interest income across banks in 42 different nations. 2. This organizational dichotomy follows the analysis of DeYoung and Rice (2004, Table 1). 3. In the US, the barriers between banking, securities, and insurance activities were reduced on an ad hoc basis during the late 1980s and early 1990s; the crowning blow was the Gramm–Leach–Bliley (GLB) Act of 1999, which allowed banking and non-banking activities to be affiliated within a single financial holding company. In the EU, the Second Banking Co-ordination Directive of 1989 (implemented in 1993–94) permitted universal banks to expand anywhere within the EU regardless of the host country restrictions on product powers. 4. The elimination of deposit interest-rate ceilings in many countries has also led to increased fee income from depositor services, by allowing banks to price depositor services in a more rational and competitive fashion. 5. The yet-to-be-implemented Basel II capital framework includes capital charges on various fee-generating lines of business, but it uses largely ad hoc risk weights with no rigorous justification for the magnitudes. There is also a capital charge for operational risk, which is determined by the level of total operating (net interest plus non-interest) income. 6. Equation (2) is estimated by itself, and then equation (2) is estimated using the fitted value of RISK from (1) as a right-hand-side instrumental variable. 7. DeLong and DeYoung (2007) provided evidence that stock investors learn with experience how better to price new phenomena such as large, complex bank M&As. 8. We excluded central banks, cooperative banks, investment banks, Islamic banks, medium- and long-term credit banks, non-banking credit institutions, real estate/mortgage banks, and specialized governmental credit institutions. 9. The KKZ index is the aggregate indicators of six dimensions of governance: 1. Voice and Accountability— measuring political, civil, and human rights; 2. Political Instability and Violence—measuring the likelihood of violent threats to, or changes in, government, including terrorism; 3. Government Effectiveness— measuring the competence of the bureaucracy and the quality of public service delivery; 4. Regulatory Burden—measuring the incidence of market-unfriendly policies; 5. Rule of Law—measuring the quality of contract enforcement, the police, and the courts, as well as the likelihood of crime and violence; 6. Control of Corruption—measuring the exercise of public power for private gain, including both petty and grand corruption and state capture. The indicators are constructed using an unobserved components methodology described in detail in the paper. The index is measured ranging from about -2.5 to 2.5, with higher values

136 Journal of Applied Business and Economics vol. 14(3) 2013 corresponding to better governance. References: Kaufmann, D., Kraay, A., & Zoido-Lobaton, P. (1999a). Aggregating Governance Indicators. World Bank Policy Research Department Working Paper No. 2195; Kaufmann, D., Kraay, A., & Zoido-Lobaton, P. (1999a). Governance Matters. World Bank Policy Research Department Working Paper No. 2196. 10. The index is from the Heritage Foundation and the WSJ—the Heritage Foundation/Wall Street Journal Index of Economic Freedom. 11. Recent studies by DeLong and DeYoung (2007) and Pastor and Veronesi (2003) provided empirical evidence linking stock market prices to investor learning. 12. We acknowledge that other phenomena may be partially or wholly responsible for the observed reductions in the perception and pricing of risk associated with non-interest income in Tables 6 and 7. For example, the mix of financial services that generated the non-interest income in the two subsample periods may have been different. 13. See Bhattacharya, Boot, and Thakor (1998) and Demirguc-Kunt and Detragiache (2002).

REFERENCES

Allen, L.G., & Jagtiani, J. (2000). The Risk Effects of Combining Banking, Securities, and Insurance Activities. Journal of Economics and Business, 52, 485–497.

Barth, J.R., Caprio Jr., G., & Levine, R. (2001). The Regulation and Supervision of Bank around World: A New Fatabase. In R.E. Litan, & R. Herring (Eds.), Integrating Emerging Market Countries into the Global Financial System. Washington, DC: Brooking Institution Press.

Bhattacharya, S., Boot, A., & Thakor, A.V. (1998). The Economics of Bank Regulation. Journal of Money, Credit & Banking, 30, 745–770.

Boyd, J.H., & Gertler, M. (1994). Are banks dead? Or are the reports greatly exaggerated? Quarterly Review, Federal Reserve Bank of Minneapolis, 2-23.

Boyd, J. H., Hanweck, G. A., and Pithyachariyakul, P. (1980). Bank Holding Company Diversification, Federal Reserve Bank of Chicago, Proceedings from a Conference on Bank Structure and Competition, 105-120.

Boyd, J.H., & Graham, S.L. (1986). Risk, Regulation, and Bank Holding Company Expansion into Nonbanking. Federal Reserve Bank of Minneapolis Quarterly Review, 10 (2), 2–17.

Boyd, J.H., Graham, S.L., & Hewitt, R.S. (1993). Bank Holding Company Mergers with Nonbank Financial Firms: Effects on the Risk of Failure. Journal of Banking & Finance, 17, 43–63.

Brewer III, E. (1989). Relationship between Bank Holding Company Risk and Nonbank Activity. Journal of Economics and Business, 41 (4), 337–353.

Calomiris, C.W. (1994). The Costs of Rejecting Universal Banking: American Finance in German Mirror, 1870–1914. In N. Lamoreau, & D. Raff (Eds.), The Coordination of Economic Activities Within and Between Firms. Chicago: University of Chicago Press.

Choi, S. (2005). Three Essays in Banking and Finance. Rensselaer Polytechnic Institute, PhD Dissertation.

Cooper, M.J., Jackson III, W.E., & Patterson, G.A. (2003). Evidence of Predictability in the Cross- Section of Bank Stock Returns. Journal of Banking & Finance, 27, 817–850.

Journal of Applied Business and Economics vol. 14(3) 2013 137 DeLong, G. and DeYoung, R. (2007), Learning by Observing: Information Spillovers in the Execution and Valuation of Commercial Bank M&As. The Journal of Finance, 62, 181–216. doi: 10.1111/j.1540- 6261.2007.01205.x

DeYoung, R., Hunter, W., & Udell, G. (2004). The Past, Present, and Probable Future for Community Banks. Journal of Financial Services Research, 25(2), 85-133.

DeYoung, R., & Roland, K.P. (2001). Product Mix and Earnings Volatility at Commercial Banks: Evidence from a Degree of Leverage Model. Journal of Financial Intermediation, 10, 54–84.

DeYoung, R., & Rice T. (2004). Noninterest Income and Financial Performance at U.S. Commercial Banks. Financial Review, 39, 101-127. doi: 10.1111/j.0732-8516.2004.00069.x

Demirguc-Kunt, A., & Detragiache, E. (2002). Does Deposit Insurance Increase Banking System Stability? An Empirical Investigation. Journal of Monetary Economics, 49, 1373–1406.

Demirguc-Kunt, A., & Huizinga, H. (2004). Market Discipline and Deposit Insurance. Journal of Monetary Economics, 51, 375–399.

Demsetz, R.S., & Strahan, P.E. (1997). Diversification, Size, and Risk at Bank Holding Companies. Journal of Money, Credit, and Banking, 29, 300–313.

Eisenbeis, R.A., Harris, R.S., & Lakonishok, J. (1984). Benefits of Bank Diversification: The Evidence from Shareholder Returns. Journal of Finance, 34, 881–892.

Esho, N., Kofman, P., & Sharpe., I. G. (2005). Diversification, Fee Income, and Credit Union Risk. Journal of Financial Services Research, 27(3), 259-281.

Gallo, J.G., Apilado, V.P., & Kolari, J.W. (1996). Commercial Bank Mutual Fund Activities: Implications for Bank Risk and Profitability. Journal of Banking & Finance, 20, 1775–1791.

Heggestad, A. (1975). Riskiness of Investments in Nonbank Activities by Bank Holding Companies. Journal of Economics & Business, 219–223.

Johnson R.N. & Meinster, D.R. (1974). Bank Holding Companies: Diversification Opportunities in Nonbank Activities. Eastern Economic Journal, 1, 316-323.

Kaufmann, D., Kraay, A., & Zoido-Lobaton, P. (1999). Aggregating Governance Indicators. Manuscript, The World Bank.

Kaufman, G.G., & Mote, L.R. (1994). Is Banking a Declining Industry? A Historical Perspective. FBR NY Economic Perspectives, 18, 2–21.

Kwan, S.H. (1997). Securities Activities by Commercial Banking Firms’ Section 20 Subsidiaries: Risk, Return, and Diversification Benefits. Working paper, Federal Reserve Bank of San Francisco.

Kwast, M.L. (1989). The Impact of Underwriting and Dealing on Bank Returns and Risks. Journal of Banking and Finance, 13, 101–125.

Laderman, E. S. (1998). The Potential Diversification and Failure Reduction Benefits of Bank Expansion into Nonbanking Activities. Federal Reserve Bank of San Francisco, Working Paper 2000-01.

138 Journal of Applied Business and Economics vol. 14(3) 2013

Levine, R. (1997). Financial Development and Economic Growth: Views and Agenda. Journal of Economic Literature, 688–726.

Litan, R.E. (1985). Evaluating and Controlling the Risks of Financial Product Deregulation. Yale Journal of Regulation, 3, 1-52.

Pastor, L., & Veronesi, P.. (2003). Stock Valuation and Learning about Profitability. Journal of Finance, 58:1749–89.

Rogers, K.E., & Sinkey, J.F., Jr. (1999). An Analysis of Nontraditional Activities at US Commercial Banks. Review of Financial Economics, 8, 25–39.

Roland, K. P. (1997). Profit Persistence in Large U.S. Bank Holding Companies: An Empirical Investigation. Office of the Comptroller of the Currency, Economics Working Paper, 97-2.

Rosen, R. J., Lloyd-Davies, P. R., Kwast, M.L., & Humphrey, D. B. (1989). New Banking Powers. Journal of Banking and Finance, 13, 355–366.

Sinkey J., & Nash, R. (1993). Assessing the Riskiness and Profitability of Credit Card Banks. Journal of Financial Services Research, 7, 127-150.

Smith, R., Staikouras, C., & Wood, G. (2003). Noninterest Income and Total Income Stability. Working paper, Bank of England.

Stiroh, K.J. (2004a). Diversification in Banking: Is Noninterest Income the Answer? Journal of Money, Credit, and Banking, 36(5), 853-882.

Stiroh, K.J. (2004b). Do Community Banks Benefit from Diversification? Journal of Financial Services Research, 25(2-3), 135-160. doi: 10.1023/B:FINA.0000020657.59334.76

Wall, L.D. & Eisenbeis, R. A.(1984). Risk Considerations in Deregulating Bank Activities. Federal Reserve Bank of Atlanta, Economic Review, 69, 6-18.

Wall, L. D., Reichert, A. K., & Mohanty, S. (1993). Deregulation and the Opportunities for Commercial Bank Diversification. Federal Reserve Bank Atlanta Economic Review, 1-25.

Journal of Applied Business and Economics vol. 14(3) 2013 139 TABLE 1 SUMMARY STATISTICS

Standard Variable Definition Mean Median Minimum Maximum Deviation Variables used to estimate the market model, equation (1)

R Return of an individual bank -0.0001 0.0000 0.0262 -0.3966 0.5576

MR Market return of a country -0.0002 0.0000 0.0175 -0.1941 0.2246

Change in interest rate in a INT -0.0000 0.0000 0.0014 -0.0162 0.0171 country RISK, RETURN, and NII variables used in equations (2) and (3) Total risk estimated from market TOTRISK 0.0228 0.0205 0.0145 0.0008 0.3945 model (1) Market risk estimated from MKTRISK 0.4533 0.3544 0.4624 -1.3422 3.0279 market model (1) Interest rate risk estimated from IRRISK -0.3268 -0.1892 3.4478 -17.7642 28.5523 market model (1) Idiosyncratic risk estimated from IDIORISK 0.0207 0.0186 0.0119 0.0008 0.1730 market model (1) Annual stock return (annualized RETURN 0.0346 0.0232 0.1089 -0.3732 2.0467 average of weekly returns)

ROA Return on book assets 0.0111 0.0101 0.0267 -0.4820 0.5608

ROE Return on book equity 0.0849 0.1049 0.1752 -3.5711 1.2082

Non-interest income divided by NII 0.0185 0.0116 0.0307 -0.0627 0.3926 total assets Non-interest income divided by NII(alt) operating revenue (non-interest 0.1901 0.1538 0.1568 -0.0738 1.0000 income plus net interest income) Control variables used in equations (2) and (3) Total assets, expressed in ASSETS 22,315,504 2,684,594 71,483,148 6,887 1,097,190,000 thousands of US dollars lnASSETS Natural log of ASSETS 14.9887 14.8030 1.9493 8.8374 20.8160 LOANS Total loans divided by total assets 0.6016 0.6297 0.1618 0.0002 0.9704 Liquid assets divided by total LIQUID 0.1605 0.0988 0.1494 0.0002 0.9185 assets Loan loss provisions divided by LOSSPROV 0.0053 0.0029 0.0094 0.0000 0.1811 total assets Dummy = 1 for explicit deposit EXPLICIT 0.8874 1.0000 0.3162 0.0000 1.0000 insurance CR3 Three-bank concentration ratio 0.3652 0.2700 0.2060 0.2000 1.0000 Market value of bank equity CAPITAL 0.1038 0.0822 0.1177 0.0012 0.9071 divided by book value of assets

140 Journal of Applied Business and Economics vol. 14(3) 2013 Increasing index of permissible RESTRICT 10.3558 12.0000 2.5738 3.0000 14.0000 bank activities. Barth et al. (2001) Increasing index of corporate CORPGOV governance. Kaufman, Kraay, 1.1070 1.2900 0.4514 -0.7600 1.7200 and Zoido-Lobaton (KKZ) Index of bank freedom of the BANKFREE 3.7216 4.0000 0.5655 2.0000 5.0000 country Dummy = 1 if multiple banking MULTSUPS 0.5914 1.0000 0.4916 0.0000 1.0000 supervisors Dummy = 1 if economy is MARKET 0.6858 1.0000 0.4642 0.0000 1.0000 market-based (versus bank-based) Dummy = 1 if bank is a foreign- FOREIGN 0.0416 0.0000 0.1998 0.0000 1.0000 owned bank Dummy = 1 if bank is a state- STATE 0.0209 0.0000 0.1432 0.0000 1.0000 owned bank

The Fitch-IBCA Bankscope database is our primary source for annual bank-level financial data. The bank stock returns Rit, stock market indices MRit, and interest rates INTit come from Datastream. Appendix I contains additional details about the data sources for MR. Datastream does not report long-term government bond yields for all countries; in these cases, we use a long-term corporate bond yield to calculate INT. The remainder of the control variables are observed from various sources.

Journal of Applied Business and Economics vol. 14(3) 2013 141 TABLE 2 MAIN REGRESSIONS, EQUATION (2) RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

[a] [b] [c] [d] Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value Constant 0.0424 (.000) -1.205 (.000) 6.4148 (.000) 0.0476 (.000) NII 0.0058 (.396) 1.0072 (.000) 0.4166 (.873) 0.0053 (.400) LOANS -0.0073 (.001) -0.1785 (.000) -1.3412 (.004) -0.0052 (.000) lnASSETS -0.0015 (.000) 0.3008 (.000) -0.4921 (.000) -0.0029 (.000) CAPITAL 0.0023 (.002) 0.5847 (.000) -0.0442 (.936) -0.0012 (.430) LIQUID 0.0061 (.021) 0.1501 (.009) -0.3906 (.504) 0.0034 (.056) LOSSPROV 0.1624 (.000) 1.3610 (.166) 6.4626 (.388) 0.1332 (.000) MARKET 0.0002 (.829) 0.0638 (.004) -0.0684 (.660) 0.0002 (.781) EXPLICIT -0.0019 (.138) -0.1258 (.000) -0.0360 (.866) -0.0006 (.262) RESTRICT 0.0005 (.000) 0.0158 (.000) -0.1626 (.000) 0.0004 (.000) CORPGOV -0.0077 (.000) -0.2173 (.000) -0.1428 (.388) -0.0068 (.000) BNKFREE -0.0020 (.000) -0.1097 (.000) -0.0227 (.853) -0.0010 (.028) CR3 -0.0019 (.195) 0.4446 (.000) -0.8500 (.061) -0.0075 (.000) MULTSUPS 0.0061 (.000) 0.1747 (.000) 0.1291 (.439) 0.0037 (.000) YR95 -0.0014 (.233) -0.0810 (.003) -1.2639 (.000) -0.0009 (.133) YR96 -0.0030 (.000) -0.0257 (.287) -1.1179 (.000) -0.0015 (.005) YR97 0.0009 (.159) 0.0297 (.156) -1.2797 (.000) 0.0014 (.016) YR98 0.0048 (.000) 0.1056 (.000) -0.5676 (.004) 0.0044 (.000) YR99 0.0031 (.000) -0.0547 (.003) -0.7478 (.000) 0.0040 (.000) YR00 0.0048 (.000) -0.0890 (.000) -0.5266 (.014) 0.0060 (.000) YR01 0.0005 (.320) -0.0267 (.145) -0.1371 (.441) 0.0012 (.016) Adj-R2 0.2313 0.4728 0.0380 0.2541 N 4444 4444 4444 4444

142 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 3 MAIN REGRESSIONS, EQUATION (3) RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: RETURN RETURN RETURN RETURN coefficient p-value coefficient p-value coefficient p-value coefficient p-value Constant -0.0665 (.168) 0.3061 (.000) 0.2194 (.000) -0.1090 (.076) NII 0.2486 (.145) 0.0296 (.869) 0.2845 (.107) 0.2622 (.125) LOANS -0.0356 (.098) -0.0099 (.642) -0.0942 (.000) -0.0456 (.038) lnASSETS 0.0010 (.761) -0.0550 (.000) -0.0154 (.002) 0.0090 (.065) TOTRISK 4.6876 (.000) MKTRISK 0.1770 (.000) IRRISK -0.0151 (.206) IDIORISK 5.3747 (.000) LIQUID -0.0993 (.002) -0.0735 (.013) -0.0374 (.138) -0.0914 (.004) LOSSPROV -2.1926 (.000) -1.6851 (.000) -0.9360 (.001) -2.1386 (.000) EXPLICIT -0.0037 (.749) 0.0180 (.111) -0.0382 (.008) -0.0132 (.279) FOREIGN -0.0290 (.032) -0.0308 (.017) -0.0091 (.466) -0.0256 (.057) STATE -0.0672 (.000) -0.0734 (.000) -0.0475 (.000) -0.0647 (.000) CR3 0.1129 (.000) 0.0290 (.016) 0.0327 (.011) 0.1291 (.000) YR95 0.0170 (.001) 0.0253 (.000) -0.0102 (.262) 0.0146 (.005) YR96 0.0361 (.000) 0.0274 (.000) 0.0044 (.613) 0.0295 (.000) YR97 0.0525 (.000) 0.0515 (.000) 0.0383 (.005) 0.0488 (.000) YR98 -0.0205 (.009) -0.0162 (.024) -0.0067 (.413) -0.0221 (.008) YR99 -0.0059 (.382) 0.0186 (.027) -0.0028 (.728) -0.0136 (.068) YR00 -0.0332 (.000) 0.0051 (.402) -0.0183 (.000) -0.0431 (.000) YR01 0.0063 (.326) 0.0133 (.057) 0.0068 (.000) 0.0020 (.750) Adj-R 0.0606 0.0715 0.0460 0.0581 N 4444 4444 4444 4444

Journal of Applied Business and Economics vol. 14(3) 2013 143 TABLE 4 ALTERNATE SPECIFICATION OF EQUATION (3), SELECTED RESULTS RETURNit = ν + φNIIit + δRISKit + φNIIit*RISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d)

Dependent variable: RETURN RETURN RETURN RETURN coefficient p-value coefficient p-value coefficient p-value coefficient p-value NII -0.2036 (.052) -0.8503 (.009) 0.3248 (.046) -0.0215 (.943) TOTRISK 2.3354 (.019) TOTRISK*NII 10.8248 (.473) MKTRISK 0.1252 (.004) MKTRISK*NII 1.1036 (.075) IRRISK -0.0152 (.025) IRRISK*NII -0.1699 (.047) IDIORISK* 2.7180 (.018) IDIORISK*NII 5.4947 (.722)

∂RETURN/∂NII 0.0432 [2.07] -0.3500 [-3.04] 0.0880 [2.26] 0.0923 [0.74] Joint significance of (.770) (.005) (.030) (.7558) NII and RISK*NII ∂RETURN/∂RISK 2.5356 [2.88] 0.1456 [3.55] -0.0184 [-2.84] 2.8196 [2.71] Joint significance of (.005) (.002) (.002) (.014) RISK and RISK*NII

Adj-R 0.0481 0.0686 0.0444 0.0461 N 4444 4444 4444 4444

144 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 5 ALTERNATE SPECIFICATION OF EQUATION (3), SELECTED RESULTS

ROAit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: ROA ROA ROA ROA coefficient p-value coefficient p-value coefficient p-value coefficient p-value NII 0.3120 (.000) 0.2756 (.000) 0.3095 (.000) 0.3123 (.000) TOTRISK 0.1282 (.413) MKTRISK 0.0282 (.000) IRRISK 0.0017 (.209) IDIORISK -0.0964 (.643)

Adj-R 0.2348 0.2566 0.2350 0.2219 N 4444 4444 4444 4375

ROEit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: ROE ROE ROE ROE coefficient p-value coefficient p-value coefficient p-value coefficient p-value NII 1.1119 (.000) 1.0512 (.000) 1.1045 (.000) 1.1144 (.000) TOTRISK 0.9363 (.401) MKTRISK 0.0484 (.086) IRRISK 0.0058 (.464) IDIORISK 0.6479 (.648)

Adj-R 0.2601 0.2160 0.2598 0.2599 N 4444 4444 4444 4444

Journal of Applied Business and Economics vol. 14(3) 2013 145 TABLE 6 EQUATION (2) FOR THE 1995–1998 AND 1999–2002 SUBSAMPLES RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value 1995–1998 NII -0.0093 (0.331) 1.0869 (0.022) -1.4021 (0.595) -0.0038 (0.595) Adj-R2 0.2537 0.4612 0.0444 0.3410 N 1791 1791 1791 1791

1999–2002 NII 0.0124 (0.177) 0.9556 (0.001) 1.4387 (0.701) 0.0092 (0.296) Adj-R2 0.2293 0.4998 0.0307 0.2135 N 2653 2653 2653 2653

TABLE 7 EQUATION (3) FOR THE 1995–1998 AND 1999–2002 SUBSAMPLES RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value 1995–1998 NII 0.1902 (0.247) 0.0210 (0.911) 0.1525 (0.356) 0.1850 (0.262) TOTRISK 2.6354 (0.134) MKTRISK 0.0886 (0.128) IRRISK 0.0138 (0.324) IDIORISK 2.6011 (0.206) Adj-R2 0.0431 0.0459 0.0386 0.0653 N 1791 1791 1791 1443 1999–2002 NII 0.2558 (0.289) 0.0553 (0.828) 0.4233 (0.094) 0.2781 (0.249) TOTRISK 5.7614 (0.000) MKTRISK 0.2169 (0.000) IRRISK -0.0382 (0.000) IDIORISK 7.2757 (0.000) Adj-R2 0.0839 0.0988 0.0718 0.0812 N 2653 2653 2653 2653

146 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 8A EQUATION (2) FOR THE MARKET VS BANK-BASED SUBSAMPLES RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

(a) (b) (c) (d)

Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value Market NII -0.0027 (0.711) 0.7813 (0.001) 0.9233 (0.792) -0.0025 (0.710) Adj-R2 0.2676 0.5490 0.0537 0.2345 N 3057 3057 3057 3057

Bank NII -0.0055 (0.769) 0.9920 (0.111) 3.3694 (0.345) 0.0041 (0.754) Adj-R2 0.2292 0.4460 0.0410 0.3374 N 1387 1387 1387 1387

TABLE 8B EQUATION (3) FOR THE MARKET VS BANK-BASED SUBSAMPLES RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value Market NII 0.2974 (0.228) 0.0463 (0.863) 0.3763 (0.140) 0.3133 (0.203) TOTRISK 9.4899 (0.000) MKTRISK 0.2593 (0.001) IRRISK -0.0330 (0.000) IDIORISK 13.0544 (0.000) Adj-R2 0.1047 0.0875 0.0656 0.1048 N 3057 3057 3057 3057 Bank NII 0.2599 (0.077) 0.2499 (0.089) 0.3249 (0.032) 0.2628 (0.073) TOTRISK 0.3342 (0.656) MKTRISK 0.0007 (0.969) IRRISK -0.0119 (0.067) IDIORISK 0.4761 (0.570) Adj-R2 0.1322 0.1320 0.1340 0.1323 N 1387 1387 1387 1387

Journal of Applied Business and Economics vol. 14(3) 2013 147 TABLE 9A EQUATION (2) FOR THE BANK FREEDOM SUBSAMPLES RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value High bank freedom NII 0.0180 (0.059) 1.7571 (0.000) 1.4354 (0.704) 0.0101 (0.273) Adj-R2 0.2884 0.5560 0.0649 0.2727 N 2995 2995 2995 2995

Low bank freedom NII -0.0360 (0.002) -0.4942 (0.163) 3.5523 (0.200) -0.0204 (0.016) Adj-R2 0.1981 0.3952 0.0348 0.2659 N 1449 1449 1449 1449

TABLE 9B EQUATION (3) FOR THE BANK FREEDOM SUBSAMPLES RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value High bank freedom NII -0.0182 (0.844) -0.0500 (0.654) 0.0927 (0.290) -0.0010 (0.992) TOTRISK 3.9950 (0.012) MKTRISK 0.0606 (0.102) IRRISK -0.0142 (0.084) IDIORISK 5.8796 (0.011) Adj-R2 0.1321 0.1191 0.1147 0.1323 N 2995 2995 2995 2995 Low bank freedom NII 0.7394 (0.072) 0.5809 (0.138) 0.1875 (0.686) 0.6953 (0.094) TOTRISK 4.3039 (0.005) MKTRISK 0.1342 (0.046) IRRISK 0.0567 (0.013) IDIORISK 4.5154 (0.013) Adj-R2 0.1406 0.1465 0.1452 0.1406 N 1449 1449 1449 1449

148 Journal of Applied Business and Economics vol. 14(3) 2013 TABLE 10A EQUATION (2) FOR THE EDI SUBSAMPLES RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

(a) (b) (c) (d)

Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value With EDI NII 0.0111 (0.126) 1.3824 (0.000) 0.2991 (0.916) 0.0047 (0.505) Adj-R2 0.2864 0.5149 0.0405 0.2672 N 3947 3947 3947 3947

No EDI NII -0.0687 (0.002) -3.0716 (0.000) 4.1214 (0.455) -0.0292 (0.016) Adj-R2 0.1700 0.4646 0.1040 0.3793 N 497 497 497 497

TABLE 10B EQUATION (3) FOR THE EDI SUBSAMPLES RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value With EDI NII 0.0684 (0.351) -0.0435 (0.629) 0.0765 (0.310) 0.0889 (0.221) TOTRISK 2.2981 (0.012) MKTRISK 0.0708 (0.007) IRRISK 0.0031 (0.733) IDIORISK 2.6314 (0.020) Adj-R2 0.0876 0.0856 0.0785 0.0861 N 3947 3947 3947 3947 No EDI NII 3.5572 (0.001) 2.6461 (0.007) 2.0042 (0.041) 3.1363 (0.001) TOTRISK 19.0767 (0.000) MKTRISK 0.01274 (0.004) IRRISK 0.0618 (0.027) IDIORISK 34.1633 (0.001) Adj-R2 0.2393 0.2183 0.2173 0.2393 N 497 497 497 497

Journal of Applied Business and Economics vol. 14(3) 2013 149 TABLE 11A EQUATION (2) FOR THE CONCENTRATION SUBSAMPLES RISKit = μ + λNIIit + ψYit + τΣt=1,TTt + πit

Less concentrated NII 0.0039 (0.729) 1.6594 (0.000) -2.5819 (0.468) -0.0039 (0.708) Adj-R2 0.1284 0.5268 0.0550 0.1633 N 2842 2842 2842 2842

TABLE 11B EQUATION (3) FOR THE CONCENTRATION SUBSAMPLES RETURNit = ν + φNIIit + δRISKit + ρZit + τΣt=1,TTt + ηit

(a) (b) (c) (d) Dependent variable: TOTRISK MKTRISK IRRISK IDIORISK coefficient p-value coefficient p-value coefficient p-value coefficient p-value Highly concentrated NII 0.2598 (0.424) 0.2159 (0.503) -0.3570 (0.473) 0.2591 (0.424) TOTRISK 4.1318 (0.001) MKTRISK 0.1635 (0.000) IRRISK 0.1182 (0.035) IDIORISK 4.3674 (0.002) Adj-R2 0.0746 0.0885 0.0711 0.0720 N 1602 1602 1602 1602 Less concentrated NII 0.1465 (0.086) 0.2158 (0.033) 0.1669 (0.054) 0.1746 (0.042) TOTRISK 1.8536 (0.379) MKTRISK -0.0205 (0.500) IRRISK 0.0007 (0.873) IDIORISK 5.7615 (0.038) Adj-R2 0.1736 0.1734 0.1730 0.1748 N 2842 2842 2842 2842

150 Journal of Applied Business and Economics vol. 14(3) 2013 FIGURE 1

Noninterest Income as Percentage of Total Revenues 28% 26% 24% ALL 22% US 20% NON-US 18% 16% 14% 12% 10% 1995 1996 1997 1998 1999 2000 2001 2002

Noninterest Income as Percentage of Total Assets 2.60% 2.40%

2.20% ALL 2.00% US 1.80% NON-US 1.60% 1.40% 1.20% 1.00% 1995 1996 1997 1998 1999 2000 2001 2002

Noninterest Income as Percentage of Net Operating Revenues 38% 36% 34% ALL 32% US 30% NON-US 28% 26% 24% 22% 20% 1995 1996 1997 1998 1999 2000 2001 2002

Journal of Applied Business and Economics vol. 14(3) 2013 151 APPENDIX I

Country Country Code Name of Index Datastream Mnemonic Argentina AR Merval ARGMERV Austria AT ATX ATXINDX Australia AU ASX ALL ORDINARY ASXAORD Belgium BE Bel20 BGBEL20 Brazil BR BOVESPA BRBOVES Canada CA S&P/TSX Composite TTOCOMP Chile CL IGPA General IGPAGEN Czech Republic CZ PX50 CZPX50I Denmark DK KFX index DKKFXIN Estonia EE Talse index ESTALSE Finland FI HEX General Index HEXINDX France FR CAC40 FRCAC40 Germany DE DAX 30 Performance DAXINDX Greece GR ASE General GRAGENL Hong Kong HK Hang Seng Index HNGKNGI Hungary HU Budapest index BUXINDX India IN Bombay SE 200 IBOM200 Indonesia ID Jakarta Composite Index JAKCOMP Ireland IE Overall index ISEQUIT Israel IL Tel Aviv SE Moaf 25 ISTMAOF Italy IT MIB 30 ITMIB30 Japan JP TOPIX TOKYOSE Luxembourg LU Datastream market index TOTMKLX Malaysia MY Kuala Lumpur Composite Index KLPCOMP Mexico MX IPC (Bolsa) MXIPC35 Netherlands NL AEX NLALAEX Norway NO OBX OSLOOBX Pakistan PK Karachi SE 100 PKSE100 Poland PL WIG POLWIGI Portugal PT PSI (BVL) General POPSIGN Singapore SG STI SNGPORI South Korea KR KOSPI KORCOMP South Africa ZA JSE All Share JSEOVER Spain ES IBEX 35 IBEX35I Sri Lanka LK Colombo SE All Share SRALLSH Sweden SE Affarsvarlden General Index AFFGENL Switzerland CH Swiss Market Index SWISSMI Taiwan TW TWSE TAIWGHT Thailand TH SET Index BNGKSET Turkey TR ISE National 100 TRKISTB UK GB FTSE 100 FTSE100 US US S&P 500 S&PCOMP

152 Journal of Applied Business and Economics vol. 14(3) 2013