FDI Nigeria (1960-2010)

ISSN 2229-6891
International Research Journal of

Applied Finance
Volume. III Issue. 7 July 2012
Contents

A Comparison of the Financial Characteristics of Pharmaceutical Firms in the United States and
the Peoples Republic of China
Christopher Coyne, John L. Haverty, & Jing Lin 889 - 902

Global Conversion Factor Irregularities: Can they be Exploited?
G.D. Hancock-Weise 903 - 928

On Collecting Social Security Benefits after Age 66
Charles J. Higgins 929 - 939

Systematic Risk Estimation: OLS v. State-Space Methods
Joseph Callaghan, Liang Fu, & Jing Liu 940 - 953

Operational Risk and Large Internal Frauds at Financial Institutions
Benton E. Gup 954 - 970

Macroeconomic Factors, Market Volatility and the Performance of Large Cap Funds During the
2008-2011 Period
Dr. Abhay Kaushik 971 - 981

Cash Payout and Long Run Performance of Mergers
Wei Zhang, & Alka Bramhandkar 982 - 994

Too Critical to Fail: What Saved the Public Company Accounting Oversight Board?
Stanley A. Leasure, & Jana Ault Phillips 995 - 1009

Exchange Rates and Stock Market Returns In Mexico: A Markov Regime-Switching Approach
Dr. Ricardo Tovar-Silos 1010 - 1019

Determinants of Foreign Direct Investment in Emerging Markets: Evidence from Nigeria
Dr. Vigdis W. Boasson, Dr. Emil Boasson, & Mr. Hameed H. Salihu 1020 - 1038

Fairness Opinions, Expert Independence and Reputation in Mergers and Acquisitions
Nina T. Dorata, & Takeshi Nishikawa 1039 - 1058

www.irjaf.com

International Research Journal of Applied Finance ISSN 2229 6891
Vol III Issue 7 July, 2012
888

Call for Papers
International Research Journal of Applied Finance (IRJAF) is a double blind peer-reviewed open
access online journal. Every month a copy of IRJAF reaches faculty members in the areas of
accounting, finance, and economics working in 80% of AACSB accredited Business Schools
across the world. The journal provides a dedicated forum for Academicians, practitioners, policy
makers and researchers working in the areas of finance, investment, accounting, and economics.
The Editor of the Journal invites papers with theoretical research/conceptual work or applied
research/applications on topics related to research, practice, and teaching in all subject areas of
Finance, Accounting, Investments, Money, Banking and Economics. The original research
papers and articles (not currently under review or published in other publications) will be
considered for publication in International Research Journal of Applied Finance.

Topics covered include:
*Portfolio Management *Shareholder Responsibilities *Investors and corporate Social
Responsibility *Emerging Markets *Financial Forecasting *Equity Analysis *Derivatives
*Currency Markets *International Finance *Behavioral Finance Financial Accounting *Financial
Management *Cost Accounting *Applied Economics *Econometrics *Financial Engineering
All paper will be double blind reviewed on a continual basis. Normally the review process will
be completed in about 4 weeks after the submission and the author(s) will be informed of the
result of the review process. If the paper is accepted, it will be published in the next month issue.
By the final submission for publications, the authors assign all the copyrights to the Kaizen
Publications. The Editorial Board reserves the right to change/alter the final submissions for
IRJAF for editorial purposes.

Copyright: Articles, papers or cases submitted for publication should be original contributions
and should not be under consideration for any other publication at the same time. Authors
submitting articles/papers/cases for publication warrant that the work is not an infringement of
any existing copyright, infringement of proprietary right, invasion of privacy, or libel and will
indemnify, defend, and hold IRJAF or sponsor(s) harmless from any damages, expenses, and
costs against any breach of such warranty. For ease of dissemination and to ensure proper
policing of use, papers/articles/cases and contributions become the legal copyright of the IRJAF
unless otherwise agreed in writing.

Click here to submit your paper: http://irjaf.com/Submission.html

889

A Comparison of the Financial Characteristics of Pharmaceutical Firms in
the United States and the Peoples Republic of China

Christopher Coyne
St. Josephs University
Philadelphia, PA
ccoyne@sju.edu

John L. Haverty
St. Josephs University
Philadelphia, PA
jhaverty@sju.edu

Jing Lin
St. Joseph's University
Philadelphia, PA
jlin@sju.edu
890

Abstract
This study compares the financial characteristics of pharmaceutical firms from the Peoples
Republic of China (PRC) with the financial characteristics of pharmaceutical firms from the
United States. Five groups of ratios are calculated for a sample of PRC pharmaceutical firms:
liquidity, activity, performance, profitability, and leverage. Data for the PRC firms have been
obtained from the China Stock Market & Accounting Research (CSMAR) Financial Statement
Database produced by the China Accounting and Finance Research Centre at the Hong Kong
Polytechnic University and the Shenzhen GTA Information Technology Company Limited.
Similar ratios are calculated for a sample of U.S. pharmaceutical firms using the Compustat

database. PRC pharmaceutical firms are found to exhibit substantial differences in financial
ratios when compared to U.S. pharmaceutical firms. Analysis of the differences highlights the
role of very large accounts receivable balances and high levels of non-productive assets in PRC
pharmaceutical firms.
Key Words: Pharmaceutical Industry, Peoples Republic of China, Ratio Analysis

The Peoples Republic of China (PRC) has emerged as an economic power in Asia and
represents one of the fastest growing economies in the world. Investors and researchers need to
understand the effects of recent changes in the Peoples Republic of Chinas (PRC) economic
system as well as their accounting system so that reported earnings and cash flows can be
assessed. This research intends to investigate the financial characteristics of one segment of
Chinese firms, pharmaceutical companies, by examining selected financial ratios of these firms
and comparing them to U.S. firms in the same industry. Pharmaceutical companies were chosen
for this investigation due to the increasing importance of this segment as a result of the
increasing governmental and political scrutiny given to healthcare firms worldwide as a result of
the increase in healthcare costs. In addition, the domestic market for pharmaceuticals is growing
rapidly in the PRC as a result of the increasing standard of living, and the PRC government is
encouraging the development of research and development centers inside the PRC. These trends
are increasing investor interest in the PRC pharmaceutical segment and multinationals are
actively seeking and evaluating partners to establish a base to exploit PRC domestic demand as
well as to set up research and development centers within the PRC. Investors worldwide need to
understand PRC financial statements in order to make rational investment decisions in this
highly important sector.
The use of financial ratios to compare the characteristics of different groups of firms has been a
common technique in finance. In an early example of such research, Altman (1968) attempted to
predict bankruptcy by comparing bankrupt firms to healthy firms. Internationally, financial
ratios have been used to compare Japanese keiretsu-affiliated firms with independent Japanese
firms (Meric et al. 2000) and to compare U.S. E.U. and Japanese manufacturing firms (Meric et
al. 2004). There have been only limited studies of ratio comparisons involving PRC firms, most
likely due to the lack of available data. Luo and Chen (1995) used financial ratio comparisons to
compare the financial performance of international joint ventures to wholly foreign-owned
enterprises. Fuglister (1997) analyzed differences in financial ratios between matched pairs of
PRC firms and U.S. firms and concluded that Chinese firms have lower debt to net worth ratios
relative to their U.S. counterparts, have significantly higher average collection periods, higher
interest coverage ratios, lower long-term debt to total capitalization and lower asset turnovers
than their U.S. counterparts.

891

I. Background
The PRC is used in this study due to its growing importance in the world economy. It also
provides an example of a nation whose accounting system is very different from the United
States but is supposedly becoming more similar to Western accounting systems due to its desire
to become more integrated into the world economy. In order to understand the accounting
system of the PRC, it is necessary to view it in the larger context of the monumental economic
restructuring that is taking place within that nation. As part of this restructuring, the PRC has
developed an equity market to raise much needed capital for reform and expansion. In addition,
the accounting system of the PRC is being transformed from a largely Soviet-era model designed
to facilitate central planning to a model more compatible with the PRCs growing international
aspirations.
I.A. Economic Restructuring and Opening in the PRC
Since 1978, the PRC has been undergoing a process of opening and restructuring, moving from
isolationism to integration with the world economy, while simultaneously transforming its
economy from a centrally planned economic system to something called a socialist market
economy. This term has become understood to mean a unique blend of socialism and
capitalism, involving preservation of the economys socialist core, state firms and state banks,
while making incremental changes in other areas. These changes include establishing markets,
eliminating central planning, and allowing non-state-owned industry to prosper (Steinfeld 1999).
The cornerstones of the Chinese socialist economic system were and still are the state-owned
enterprises (SOEs). These are often very large businesses financed and controlled by the central
or local governments. They bought from and sold to other SOEs and their debts were largely
guaranteed by the state. SOEs also had an important social role in addition to their economic
role. As large employers, they offered a system of guaranteed employment as well as various
social services: the SOEs operated their own housing units, schools, hospitals, recreation
facilities, and retirement facilities.
I.B. The PRC Pharmaceutical Industry
The PRC pharmaceutical industry is appropriate for a financial comparison with U.S.
pharmaceutical firms because of its growing importance, as well as its expected structural
differences when compared with the U.S. pharmaceutical industry. In addition, multinational
pharmaceutical firms are aggressively seeking acquisitions and partners to gain a foothold in the
huge PRC market, and need to examine and understand the financial structure of the PRC
pharmaceutical industry in order to be effective partners and competitors.
As a result of the liberalization of the PRC economy, the PRC pharmaceutical industry has seen
a period of restructuring and huge growth. The restructuring of one pharmaceutical firm, Sunve
Pharmaceutical, from SOE to a firm listed on a stock exchange for public investors is described
by Lee (2001). Restructuring involved carving out the social services performed by all SOEs
under the PRCs brand of socialism, and changing the financial and accounting structure of the
firm so that it can gain access to financial markets.
It is currently estimated to be the tenth largest drug market in the world and is expected to
become the largest in the world by 2050 (Economist Intelligence Unit Feb. 3, 2004). Optimistic
forecasts are offset, however, by substantial perceived difficulties in doing business in the PRC,
including linguistic and cultural differences. In addition, in the PRC pharmaceutical industry,
particular concern is expressed over health, safety, environmental and patent issues (Boswell
2004). The PRC pharmaceutical industry is also fiercely competitive and highly fragmented,
with over 5,000 domestic pharmaceutical manufacturers controlling roughly 70% of the market.
892

Some of the fierce competition might be fueled by the relative lack of patent protection in the
PRC market. The top ten companies only control one-fifth of the market, compared with up to
one-half in western countries. This fragmentation could be viewed as a difficulty, since finding a
reliable partner can be quite challenging, but it could also be seen as a strategic opportunity to
obtain a large share of the market by competing with smaller domestic players. Overall, the PRC
pharmaceutical market is expected to be a place of opportunity for foreign producers as Chinas
accession to the World Trade Organization forces additional PRC market reforms.
I.C. Accounting in the PRC
As the Chinese economic system has been undergoing transformation, Chinese financial
accounting has been transformed as well. The role of accounting information in a centrally
planned economy is far different from its role in a socialist market economy. In a centrally
planned economy, the users of accounting information are primarily the central planners
responsible for making broad macroeconomic decisions as well as decisions on allocating
government funds to various state-owned-enterprises (SOEs). In order to serve this group of
users, the accounting information from the SOEs needs to emphasize uniformity and rigidity and
provide information to the central planners to show that the enterprises are meeting state-
mandated production quotas, and efficiency standards. In a market economy, financial
accounting serves external users, such as banks and equity investors, who provide capital and
need high quality information to assess the loan or investment potential of an individual
enterprise. As the users of an accounting system change, the accounting system must also change
to provide relevant information to the appropriate users. International lenders and investors need
credible information to compare a Chinese investment opportunity with similar investments in
their own nation as well as other nations.
The Chinese government has developed a set of accounting standards for China that incorporates
IFRS but reflects the unique aspects of the Chinese environment, including the fact that China
was a large, developing country with wide regional development disparities, and the concept of a
socialist market economy was quite new to China. Thus, Chinese accounting standards are a
rapidly evolving set of standards based largely on IFRS but with unique, evolving aspects, and
heavy government involvement. Recently, the China Accounting Standards Committee and the
Chairman of the International Accounting Standards Board announced in a joint statement (IASB
2005) that they have agreed to work to eliminate existing differences between Chinese
accounting standards and IFRS. As of April, 2006, the Ministry of Finance announced the
issuance of a new Accounting Standards for Business Enterprises including 38 specific
accounting standards. These standards have become effective January 2007 for all listed Chinese
enterprises.
During the period of this study (1997-2002), therefore, the PRC accounting system was in the
process of gradual convergence with IFRS, but with many still unresolved differences. Some of
these differences have been noted by Ball and Wu (2004) and arose because financial accounting
in the PRC is governed by the Ministry of Finance rules that are used to determine taxable
income. In an effort to protect its tax revenue, the Ministry of Finance had several rules in place
to that end. Lower of cost or market inventory valuation was not allowed. The provision of bad
debts was limited to 3% of receivables. The PRC government establishes fixed depreciation
schedules. Impairment of long term assets has generally not been recognized. Up to 20% of
profits must not be distributed to shareholders. In addition, there are many more difficult aspects
of PRC accounting that make analysis of PRC financial statements quite challenging. Ball and
Wu (2004) note the existence of obligations known as guanxi, in which one party has obligations
893

to another party due to past relationships, and these are not shown on financial statements. Lee
(2001) illustrates how SOEs often disguise obligations among themselves due to a process
known as triangular debt, in which many transactions among SOEs are made with no intention
to settle.
In the face of these difficulties, however, the desire to attract foreign investment and become
integrated into the worldwide economy has resulted in a process of gradual change, resulting in
an accounting system that is becoming more similar to IFRS and to U.S. GAAP. Cultural
differences such as the existence of guanxi obligations and triangular debt, however, are
expected to change more slowly.
I.D. Financial Ratio Comparisons and International Accounting Diversity
One difficult fact of international business is that the accounting systems of every nation are
different, and any comparison of firms between two countries must take this consideration into
account. In an early study of this issue Choi et al. (1983) noted even when financial ratios of two
countries are based on U.S. GAAP they can still be misinterpreted because the U.S. investor may
not understand a particular foreign environment that influences all financial ratios in that
environment. The PRC accounting system is certainly different from the U.S. system, so
comparisons of financial ratios between the two systems may be artifacts of the difference
between the two accounting systems, as opposed to real structural differences. There are several
factors that mitigate this difference, however. As noted above, the economic pressures of
globalization are forcing the PRC accounting system to become closer to a western accounting
system, so the differences are becoming less significant. In addition, Lainez and Callao (2000)
in a study of accounting harmonization within the European Union, have noted the difficulties in
inter-country financial ratio comparisons due to differences in accounting systems, but they did
note that the most significant cause of any such difference was the use of different methods to
value tangible fixed assets. The United States values fixed assets on a strict historical cost
method, while certain other countries, notably Great Britain (and Hong Kong as a result of its
British colonial heritage) periodically assess the current value of their plant and equipment and
revalue it to approximate current value. International Financial Reporting Standards (IFRS)
permit either method but with appropriate disclosure. Differences in the valuation method of
fixed assets result in significantly different depreciation charges and cause significantly different
reported net incomes. Since the PRC companies included in our sample are drawn from the PRC
A-share market, they all report under PRC GAAP, and value their fixed assets on an historic cost
basis, so at least that significant difficulty in comparing financial ratios of U.S. companies to
PRC companies is mitigated.

II. Data and Methodology
In order to compare the financial ratios of U.S. pharmaceutical firms to PRC pharmaceutical
firms we obtained a sample of U.S. pharmaceutical firms, calculated some composite ratios
based on that sample, and compared them to a similar set of financial ratios obtained from a
sample of PRC pharmaceutical firms.
II.A. The U.S. Sample
The U.S. sample was drawn from the Compustat
database, using the Standard Industrial

Classification (SIC) for pharmaceutical firms (2834). We removed all firms that were not
headquartered in the United States. Other exclusions included any firm having either less than
$100 million in total assets or a minimum level of sales of $70 million in 1995. The purpose of
this control is to remove firms from the sample that might have a patent on a pharmaceutical
894

product as their only asset. Additionally all the firms chosen were required to have an end of
year fiscal year. This is to ensure that the reporting periods are comparable to the PRC firms
which by law all have a fiscal year ending December 31. The last consideration was that any
firm that had come through these screenings had to have a complete history of data for the time
period studied, 1997 2002. The sample that results from these screenings produced 16 firms.
We excluded one firm from the analysis as an outlier due to a series of negative ratios, resulting
in a final sample of 15 firms for the U.S. firms for the entire period. Table I shows the U.S.
pharmaceutical firms in our sample. The total assets of each of these firms in the year 2003 are
also shown.
Table I: U.S sample of pharmaceutical firms
Total assets 2003

II.B. The PRC Sample
The sample of PRC pharmaceutical firms was drawn from a database called the China Stock
Market Accounting Research Database, produced by the Shenzhen Guo TaiAn Information
Technology Company Ltd. This database is designed to be compatible with the constructs used
in Compustat

and has similar data for the PRC firms listed in the various PRC exchanges. This
database uses word descriptions instead of SIC, and the industry we selected was medicine
manufacturing. All firms in China report their financial results on a calendar year, and all the
firms met the other criteria used to select the U.S. firms. Since there were only 13 firms with
data in 1997 we did not screen for size, and kept all firms in our sample. As a result, the sample
size for the Chinese firms increased each year of the comparison increasing from 13 firms in
1997 to 60 firms in 2002. The PRC pharmaceutical firms in our sample are shown in Table II.
The total assets of each of these firms are also shown.
A comparison of Table II with Table I shows that the PRC pharmaceutical firms in our sample
are considerably smaller than the pharmaceutical firms in the U.S. sample in terms of total assets,
since the official exchange rate during the time of this study was approximately 8.29RMB per
U.S.$. This illustrates the fragmented nature of the PRC pharmaceutical market in comparison
with the U.S. pharmaceutical market.

Company Total Assets
(US$)
Abbott Laboratories $26,715,342,000
Allergan Inc. 1,754,900,000
Alpharma Inc. -Cl A 2,327,801,000
Bausch & Lomb Inc 3,006,400,000
Bristol Myers Squibb 27,471,000,000
Chiron Corp 4,195,169,000
Genentech Inc. 8,736,171,000
Ivax Corp 2,372,934,000
Johnson & Johnson 48,263,000,000
Lilly (Eli) & Co. 24,867,000,000
Merck & Co. 40,587,500,000
Pfizer Inc. 116,775,000,000
Schering-Plough 15,102,000,000
Watson Pharmaceuticals Inc. 3,282.600,000
Wyeth 31,031,922,000
895

Table II: PRC sample of pharmaceutical firms
Total assets 2002
COMPANY TOTAL ASSETS
2002 (RMB)
COMPANY TOTAL
ASSETS 2002
(RMB)
NEPTUNUS 3,429,311,783 TOPFOND 1,141,971,155
XINLI PHARM 684,422,288 HAIZHENGYAOYE 1,575,521,724
E-JIAO 982,280,276 HENGRUIYIYAO 1,204,575,771
LIVZON 1,856,439,162 LINGRUIKEJI 697,745,052
BYS 2,521,578,446 MEILOUYAOYE 925,623,595
YUNNAN BAIYAO 920,651,110 TZXP 3,008,126,000
HENGHE 371,399,885 GPC 4,173,194,491
QQHY 356,985,457 TAITAI PHARM 2,523,199,898
THUGB 724,861,291 SDJT 316,087,173
TJG 1,301,052,537 KUNMING ZHIYAO 763,284,617
NEP. 2,680,988,507 DIKANG 779,445,158
JLAD 1,881,530,079 TJPC 1,419,005,482
GLJQ 1,025,449,209 CNM 808,370,603
SD XINHUA 1,905,456,216 KANGMEI PHARM 629,578,765
GOLDHORSE 1,530,786,036 SDPG 943,593,775
SSP CO., LTD. 785,666,905 TTPC 1,250,698,633
HOIST INC. 573,430,692 BEISHENG PHARM 1,275,588,245
JINLING 1,537,690,857 JRBCT 426,072,253
GUANGJI 819,906,886 HP 5,953,473,459
HD MEDICINE 1,370,947,229 SWP 450,336,536
JIUZHITANG 1,019,085,828 TMSP 373,529,789
SANJIU MED&PHARM

6,850,211,320 NJMC 1,816,297,163
DCPC 3,616,464,655 DFPC 740,781,442
GINWA 1,333,162,162 HACIGUFEN 1,226,280,055
TONGRENTANG 2,459,189,103 LKPC 2,361,076,789
TAIJI 3,195,601,490 NCPC 6,958,987,345
SFIC 3,139,570,098 ZXP 742,911,995
XIZANGYAOYE 733,078,599 SHANGHAIYIYAO 5,402,814,501
ZHEJIANGYIYAO 1,636,671,688 THDB 1,568,532,868
ZHONGSHEN 692,994,762 S & P 339,957,251
Note: The RMB (also known as the yuan). The exchange rate is 8.0168RMB per U.S. Dollars. During the period of
the study the exchange rate was fixed at 8.2865 per U.S. Dollars.

II.C. Methodology
We calculated financial ratios for a six-year period (1997-2002) for each of the firms in our U.S.
sample and our PRC sample. The ratios we calculated were in five categories: liquidity, activity,
performance, profitability, and leverage. The ratios calculated are listed in Table III.
For each financial ratio, the means and medians were calculated for the firms in the U.S. sample
for each year. Similar calculations were performed for the firms in the PRC sample. For each
financial ratio, we compare the means of the U.S. sample to the means. In order to test for the
statistical significance of any observed differences in the means we performed a parametric t-
test. In addition, for each financial ratio, we compared the median values in the U.S. sample to
the median values in the PRC sample. We tested for the statistical significance of any observed
differences in the medians using the z-statistic obtained from the non-parametric Mann-Whitney
U-test.
896

Table III: Financial ratios reported in the study
Type of Ratio Ratio
Liquidity Current Ratio
Quick Ratio
Activity Inventory Turnover
Receivables Turnover
Total Asset Turnover
Average Collection Period (days)
Days to Sell Inventory
Cash Conversion Cycle (days)

Performance Sales / Property Plant and Equipment

Profitability Net Profit Margin %
Return on Assets %
Return on Equity %

Leverage Long-Term Debt / Shareholder Equity
Total Debt / Total Assets
III. Data Analysis and Results
Table IV compares the means and medians for the liquidity ratios calculated for the U.S.
pharmaceutical industry sample to those calculated for the PRC sample.
Table IV: Comparison of U.S. Pharmaceutical Industry Averages and PRC Pharmaceutical
Industry Averages
Liquidity Comparison
Test of Means Test of Medians
U.S.
Sample
Mean
PRC
Sample
Mean
t-score U.S.
Sample
Median
PRC
Sample
Median
Z-score
Current Ratio 2002 2.05 1.70 1.30 1.75 1.50 1.76
2001 2.07 2.11 -0.11 1.93 1.58 0.78
2000 2.22 2.48 -0.58 2.04 1.78 0.73
1999 1.92 2.12 -0.53 1.67 1.52 -1.34
1998 1.98 2.68 -1.16 1.79 2.00 -0.63
1997 1.78 3.08 -1.81 1.63 2.09 -1.37

Quick Ratio 2002 1.40 1.39 0.02 1.12 1.14 -0.34
2001 1.41 1.78 -1.29 1.15 1.28 -0.46
2000 1.50 2.08 -1.39 1.33 1.50 -0.50
1999 1.28 1.75 -1.36 1.09 1.20 -0.60
1998 1.31 2.12 -1.63 1.09 1.67 -1.64
1997 1.13 2.52 -2.10 0.95 1.43 -2.60
*Significant at .05 **Significant at .01

This comparison between the means of the current ratios observed for the U.S. sample and the
means of the current ratios observed for the PRC sample shows no significant differences for all
years of the study. Likewise, the comparison of the medians of the current ratios observed for the
U.S. sample and the medians of the current ratios observed in the PRC sample shows no
significant differences. The comparisons of the means and medians of the quick ratios show
similar results.
897

Table V compares the means and medians of the activity ratios calculated for the U.S.
pharmaceutical industry sample to those calculated for the PRC sample.

Table V: Comparison of U.S. Pharmaceutical Industry Averages and PRC Pharmaceutical
Industry Averages
Activity Comparison
Year U.S.
Sample
Mean
PRC Sample
Mean
t-score U.S.
Sample
Median
PRC
Sample
Median
Z-score
Receivables Turnover
2002 5.87 3.43 4.99** 5.66 2.94 3.78**
2001 6.09 3.58 3.90** 6.03 2.97 3.62**
2000 5.88 2.68 6.80** 5.96 2.26 4.34**
1999 5.85 2.08 6.85** 5.55 1.74 5.19**
1998 6.02 3.39 4.02** 5.94 3.22 3.46**
1997 5.85 2.92 4.79** 5.35 2.92 4.24**
Inventory Turnover
2002 2.59 3.00 -0.78 2.05 3.00 -2.05*
2001 2.58 2.88 -0.61 2.08 2.88 -1.27
2000 1.50 2.64 -4.14 2.03 2.33 -2.89**
1999 2.35 2.49 -0.35 2.31 2.32 -0.26
1998 2.30 2.49 -0.49 2.30 2.18 0.34
1997 2.34 2.90 -1.02 2.18 2.21 -0.48
Total Asset Turnover
2002 0.66 0.56 1.40 0.63 0.44 1.78
2001 0.71 0.51 2.66* 0.77 0.40 2.80**
2000 0.77 0.51 3.04** 0.92 0.43 2.65**
1999 0.78 0.47 3.52** 0.83 0.38 3.24**
1998 0.80 0.49 3.82** 0.80 0.44 3.66**
1997 0.79 0.48 4.35** 0.80 0.42 3.90**
Average Collection Period
(days)

2002 63.7 251.7 -4.11** 63.6 124.3 -3.85**
2001 62.8 245.7 -2.79** 59.7 115.5 -3.31**
2000 64.3 248.6 -5.96** 60.4 161.7 -4.37**
1999 67.1 285.6 -5.90** 64.9 210.8 -5.21**
1998 67.1 158.5 -4.36** 60.6 113.5 -3.46**
1997 68.8 166.8 -4.96** 67.3 125.0 -4.29**
Days to Sell Inventory
2002 182.9 262.6 -0.79 175.2 121.9 1.92
2001 178.2 165.5 0.34 173.4 123.0 1.42
2000 175.0 196.1 -0.82 177.5 156.5 0.66
1999 186.9 217.3 -0.90 155.7 157.5 -0.14
1998 208.1 203.6 0.10 156.1 167.4 0.23
1997 181.9 222.2 -0.87 164.8 164.9 -0.02
Cash Conversion Cycle
(days)

2002 169.1 354.9 -2.67** 181.8 198.3 -1.25
2001 163.1 321.5 -1.71 166.1 179.1 -1.30
2000 159.6 353.1 -4.44** 150.9 241.4 -2.54**
1999 174.5 394.0 -4.47** 185.6 351.9 -3.36**
1998 182.9 269.6 -1.87 133.8 223.9 -1.99*
1997 166.6 309.6 -2.67** 143.9 240.1 -2.38*
898

This comparison shows that inventory turnover is generally somewhat higher for PRC
pharmaceutical firms than it is for U.S. pharmaceutical firms, but the observed differences are
not significantly different. Days to sell inventory is correspondingly higher for U.S.
pharmaceutical firms than it is for PRC pharmaceutical firms, but again, the results are not
significantly different. On the other hand, receivables turnover is substantially higher for U.S.
pharmaceutical firms and average collection period is substantially lower for U.S.
pharmaceutical firms. These results are statistically significant for all years of the study as
evidenced by both the test of means and the test of medians. Total asset turnover is generally
higher for U.S. pharmaceutical firms, and operating cycle is generally higher for PRC
pharmaceutical firms. These results are statistically significant for 5 out of 6 years of the study
as evidenced by both the test of means and the test of medians. The overall cash conversion
cycle is longer for PRC pharmaceutical firms than it is for U.S. pharmaceutical firms. These
results are statistically significant for 4 out of 6 years of the study as evidenced by both the test
of means and the test of medians. These longer cash conversion cycles in the PRC are driven by
the occurrence of high levels of accounts receivable in the PRC pharmaceutical companies
relative to U.S. pharmaceutical companies. These results are consistent with Fuglister (1997)
who also found relatively high average collection periods among PRC companies. These
relatively large PRC accounts receivable could be manifestations of triangular debt among SOEs
that are never intended to be collectible.
Table VI compares one performance ratio, sales to net property, plant, and equipment, calculated
for the U.S. pharmaceutical industry sample to those calculated for the PRC sample. The mean
ratios and median ratios for each sample are calculated for each year.
Table VI: Comparison of U.S. Pharmaceutical Industry Averages and PRC Pharmaceutical
Industry Averages
Performance Comparison
U.S.
Sample
Mean
PRC
Sample
Mean
t-score U.S.
Sample
Median
PRC
Sample
Median
Z-score
Sales / Net
Property Plant
and Equipment

2002 3.09 2.30 1.67 3.03 1.50 3.70**
2001 3.25 2.01 3.68** 3.15 1.60 3.94**
2000 3.27 2.21 3.01** 3.14 1.83 3.38**
1999 3.26 1.81 2.73** 3.03 1.47 3.89**
1998 3.13 1.94 3.05** 3.07 1.55 3.70**
1997 2.96 0.87 9.37** 3.11 1.46 3.92**


Sales / Net property plant and equipment is generally higher for U.S. pharmaceutical firms than
it is for PRC pharmaceutical firms. These results are statistically significant for 5 out of 6 years
of the study in the tests of means, and in all years of the study in the tests of medians. This may
indicate either poor performance by the PRC pharmaceutical firms, or even more likely, the
existence of obsolete property, plant and equipment due to the strict definition of impairment
utilized by the PRC Ministry of Finance in financial reporting during the period of this study.
899

Table VII compares the profitability ratios calculated for the U.S. pharmaceutical industry
sample to those calculated for the PRC sample. The mean ratios and median ratios for each
sample are calculated for each year.
Table VII: Comparison of U.S. Pharmaceutical Industry Averages and PRC Pharmaceutical
Industry Averages
Profitability Comparison
U.S.
Sample
Mean
PRC
Sample
Mean
t-score U.S.
Sample
Median
PRC
Sample
Median
Z-score
Net Profit Margin (%)
2002 13.65 -14.35 2.34* 14.37 6.33 2.81**
2001 13.53 -4.50 1.63 15.26 7.37 2.20*
2000 13.22 10.35 0.78 16.47 8.62 0.18
1999 8.33 12.23 -0.50 18.01 8.44 0.55
1998 13.18 8.56 0.64 17.13 12.20 0.17
1997 9.71 8.64 0.13 14.39 13.69 0.46

Return on Assets (%)
2002 9.17 0.60 3.59** 8.18 3.12 3.33**
2001 9.54 1.53 3.57** 9.95 3.65 3.41**
2000 10.74 4.03 2.83** 11.10 4.57 0.97
1999 10.43 3.97 2.12* 14.05 4.46 2.83**
1998 9.80 5.01 1.72 11.29 6.90 1.74
1997 8.09 6.37 0.50 9.81 7.03 1.16

Return on Equity (%)
2002 21.16 -0.40 2.85** 22.68 6.42 3.44**
2001 23.25 4.21 3.34** 23.19 6.30 3.18**
2000 20.63 8.50 1.84 24.62 7.75 0.97
1999 22.29 -1.78 2.29* 25.70 8.22 3.06**
1998 21.21 43.73 -0.66 22.13 11.00 1.52
1997 16.77 9.24 1.12 15.96 10.58 1.40


All the profitability ratios, net profit margin, return on assets, and return on equity are all higher
for U.S. pharmaceutical firms as compared to PRC pharmaceutical firms, illustrating the fiercely
competitive nature of the PRC pharmaceutical industry. The net profit margins were only
statistically different from one another in 1 year of the study in the tests of means and two years
of the study in the tests of medians. Return on assets and return on equity only displayed
statistically significant differences in about half of the years of the study, but the observed
statistically significant differences were clustered in the most recent years of the study.

Table VIII compares the leverage ratios calculated for the U.S. pharmaceutical industry sample
to those calculated for the PRC sample.
Long-term debt / shareholder equity are all higher for U.S. firms than for PRC pharmaceutical
firms. The total debt / total assets ratio is also higher for U.S. firms than for PRC pharmaceutical
firms. These results are statistically significant for all years of the study as evidenced by the tests
of means, and in 5 out of 6 years of the study as evidenced by both the test of medians. This is
again consistent with findings by Fuglister (1997).
900

Table VIII: Comparison of U.S. Pharmaceutical Industry Averages and PRC Pharmaceutical Industry
Averages
Leverage Comparison
U.S.
Sample
Mean
PRC
Sample
Mean
t-score U.S.
Sample
Median
PRC
Sample
Median
Z-score
Long-Term Debt /
Shareholder Equity (%)

2002 45.79 5.53 3.77** 40.08 4.01 3.84**
2001 54.02 8.57 3.37** 44.09 4.22 4.06**
2000 32.39 13.52 2.04* 24.28 3.85 0.04
1999 35.23 4.77 2.23* 18.00 3.76 3.10**
1998 38.53 10.11 2.16* 21.88 5.10 2.86**
1997 28.46 10.08 2.48* 17.72 6.84 2.41*

Total Debt / Total
Assets (%)

2002 23.70 7.13 4.59** 25.75 2.81 4.39**
2001 23.73 8.25 3.67** 22.67 3.46 4.37**
2000 16.62 6.79 3.16** 16.15 2.45 1.34
1999 19.33 5.90 4.11** 16.85 2.89 4.09**
1998 19.17 6.16 3.78** 17.79 3.97 4.13**
1997 17.83 6.02 3.78** 15.94 2.99 3.52**


IV. Summary and Conclusions
This study found that PRC pharmaceutical does not display significantly different current ratios
or quick ratios than their U.S. counterparts. The most striking difference in the study is that PRC
accounts receivable turnover is generally substantially lower than U.S. pharmaceutical firms.
PRC pharmaceutical firms have a lower total asset turnover than U.S. pharmaceutical firms
largely due to the high accounts receivable demonstrated by the PRC pharmaceutical firms. The
high level of accounts receivable may illustrate the existence of triangular debt, which is never
intended to be repaid. In addition, the leverage ratio comparison shows PRC pharmaceutical
firms use significantly less debt in their capital structures than U.S. pharmaceutical firms. This is
most likely due to the relatively underdeveloped debt market in the PRC. Since PRC firms
apparently cannot make use of an efficient market mechanism to borrow long-term funds, PRC
pharmaceutical firms stretch their accounts receivables in order to compensate.
In addition, PRC pharmaceutical firms demonstrate a lower sales to net property, plant and
equipment ratio than U.S. firms, most likely due to the existence of obsolete property, plant and
equipment on the books of PRC pharmaceutical firms.
Finally, PRC pharmaceutical firms are less profitable than U.S. pharmaceutical firms, although
these results do not show a pattern of statistical significance.
These observed differences might be viewed as either artifacts of the different accounting
systems or as structural or cultural differences between the PRC pharmaceutical industry and its
U.S. counterpart. The fact that the PRC accounting system was continually evolving during the
901

period of this study makes it somewhat difficult to separate these influences. Since the PRC
accounting system has been gradually converging with IFRS, however, it should be expected that
any observed differences due to accounting differences should be decreasing over the period of
the study. The data do not support his possibility, however. Where there are significant
differences observed, they either persist throughout the period of the study (Tables V, VI, and
VIII) or they arise in the latter years of the study (Table VII), lending support to the conclusion
that these observed ratio differences represent are caused by either structural and/or cultural
differences between the PRC pharmaceutical industry and the U.S. pharmaceutical industry.
Two important structural and cultural differences are the existence of triangular debt, and the
existence of obsolete property, plant and equipment. Triangular debt is manifested in relatively
large amounts of accounts receivable balances in PRC firms. This situation is most likely driven
by the underdeveloped market for debt within the PRC, and a preference for short-term debt
since most of the financing in the PRC comes from banks. Financial statement users need to be
alert to this fact of life in PRC businesses, since much of this debt may never be collected.
Obsolete property, plant and equipment is manifested in the relatively low sales to net property,
plant and equipment ratios exhibited by the PRC pharmaceutical firms. Users of PRC financial
statements and financial statement ratios need to take these two issues into consideration in their
decisions.
This research is limited to pharmaceutical firms in both the U.S. and the PRC. In addition, the
structure of the PRC pharmaceutical industry is significantly more fragmented than the U.S.
pharmaceutical industry. The PRC pharmaceutical firms in this study are substantially smaller
than the firms in the U.S. sample, and the results of this study may be an artifact of that
difference.

References
Altman, Edward I., 1968, Financial ratios, discriminant analysis and the prediction of corporate
bankruptcy, Journal of Finance 23, 589-609.
Ball, Ray, and Joanna S. Wu, 2004, Jinan Quingqi Motorcycle Co. Limited, Journal of
Accounting Education 22, 325-344.
Boswell, Clay, 2004, Chemical Market Reporter, 265, (23), FR4-6.
Choi, Frederick D., Hisaki Hino, Sang Kee Mi, Sang Oh Nam, Juniche Ujiie, and Arthur I.
Stonehill, 1983, Analyzing foreign financial statements: The use and misuse of
international ratio analysis. Journal of International Business Studies 14 (1), 113-131.
Economist Intelligence Unit, 2004, EIU Views Wire, Feb. 3 2004.
Fuglister, Jane, 1997, A comparative ratio analysis between Chinese and U.S. firms, Advances in
International Accounting 10, 185-206.
International Accounting Standards Board, 2005, Joint Statement of the Secretary-General of the
China Accounting Standards Committee and the Chairman of the International
Accounting Standards Board, www.iasplus.com/pressrel/0511china jointstatement.pdf ,
accessed September 20, 2006.
Lainez, Jose A., and Susana Callao, 2000, The effect of accounting diversity on international
financial analysis: Empirical evidence, The International Journal of Accounting 35 (1),
65-83.
Lee, Chi Wan Jevons, 2001, Financial restructuring of state-owned enterprises in China: The
case of the Sunve Pharmaceutical Corporation, Accounting Organizations and Society 26
(7-8), 673-689.
902

Luo, Yadong, and Min Chen, 1995, Financial performance comparison between international
joint ventures and wholly foreign-owned enterprises in China, The International
Executive 37 (6), 599-613.
Mehric, Guiser, Larissa Kyj, Carol Welsh, and Ilhan Mehric, 2000, A comparison of the
financial characteristics of Japanese Keiretsu-affiliated and Independent firms,
Multinational Business Review 8 (2), 26-30.
Meric, Ilhan, Stephanie M. Weidman, Carol N. Welsh, and Guiser Meric, 2002, A comparison of
the financial characteristics of U.S., E.U. and Japanese manufacturing firms, American
Business Review 20 (2), 119-126.
Steinfeld, Edward S., 1999, Beyond the Transition: Chinas Economy at Centurys End, Current
History 98, (629): 271-275.

903

Global Conversion Factor Irregularities: Can they be Exploited?

G.D. Hancock-Weise
Associate Professor of Finance
College of Business Administration
University of Missouri-St. Louis
One University Boulevard
St. Louis, MO 63131
gdweise@umsl.edu

I am grateful for helpful comments from Professor Hung-Gay Fung, Mr. Paul Weise and Mr. Jerry Gorum
904

Abstract
Conversion factors are an integral part of pricing long-term interest rate futures contracts.
Approximate conversion factor methodologies can result in the identification of a cheapest-to-
deliver bond which is different from the cheapest-to-deliver bond identified by a conventional
model or result in the over or underpricing of the futures relative to the underlying bond. This
study finds that the conversion factor methodologies used around the globe are consistently
imprecise relative to the pricing of the underlying cash bond market. Consequently, both the long
and the short are regularly faced with information which can influence the benefits of reversing
versus holding a futures contract until delivery. In addition, the short may be motivated to deliver
a bond other than the CTD identified by the exchanges conversion factor. This paper examines
the conversion factors used by the three largest global derivative exchanges: the CME Group,
Eurex and NYSE-Euronext to determine potential advantages which may be exploited due to the
difference between a conventional approach to the conversion factor and the approximation
approaches used by the exchanges studied in this paper.
Key Words: Conversion factors, long-term interest rate futures contracts, global derivative
exchanges.

The purpose of this paper is to apply a conventional approach to calculating global conversion
factors in order to obtain a delivery price which is based on the fair value of the futures and the
underlying cash bond. Conversion factors represent the price of delivering a bond other than one
with a coupon rate equal to the futures contract rate and are used to adjust the contract price of
interest rate futures. The majority, but not all, credit futures offered around the globe require a
6% coupon bond to be delivered to fulfill the terms of the contract. When the short selects a
deliverable bond which does not have the contracted coupon rate the conversion factor methods,
used by the three largest global derivative exchanges, introduce prejudices which provide
incentives for the selection of some bonds as opposed to others. This paper demonstrates that
when an approximation method is used unintended benefits can accrue to the long or short when
the same bond is identified as the cheapest-to-deliver (CTD) by the conventional approach. In
addition, when different bonds are identified as CTD by the exchange and the conventional
method, the short can sometimes gain more by delivering a bond other than the CTD identified
by the exchange approach. Minor changes can be made to conversion factor methodologies
which remove the unintended prejudices.

I. Introduction
The first long-term interest rate futures contract was introduced by the Chicago Board of Trade
(CBOT) on September 11
th
, 1975 on the Government National Mortgage Association (GNMA)
bond. However, after a short period of success the contracts failed due to design flaws. Two
years after the introduction of GNMA futures, the CBOT introduced futures on the AAA-rated
U.S. T-bond.
1
The contracts quickly became popular in the mortgage industry for hedging
against adverse changes in long-term interest rates and rapidly spread to the hedging of other
long-term rate exposures. The original futures contract required the delivery of an 8% coupon
bond with a maturity of no less than 15 years until maturity, or the first call date, counting from

1
The U.S. T-bond was downgraded in late 2011 by Standard & Poors to AA-rated debt but Moodys and Fitch still maintain their AAA-rating
with a negative outlook.

905

the first day of the delivery month. In 2000, the CBOT reduced the required coupon rate from
8% to 6% and in 2011 the maximum maturity/call date allowed for delivery against the U.S. T-
bond futures was stated as 25 years.
2
Other exchanges followed, changing their coupon rates to
6% as well.
Approximately ten years after the introduction of the T-bond futures contract, in 1987, the
London International Financial Futures Exchange (LIFFE) introduced futures on long-term
Japanese Government Bonds (JGB).
3
Four years after the formation of the Eurex Exchange,
long-term futures were offered on Euro-denominated bonds and in 2002 offerings were expanded
to include the German Bundt Bond and Swiss bond.
The underlying bond for all long-term interest rate futures contract is hypothetical; however, an
actual cash bond is used to settle the contract. This introduces the opportunity for the short to
select a bond for delivery which offers a higher reward than another. This opportunity has been
labeled in the literature as the quality option and has been studied since shortly after the contracts
were first offered (e.g. Hegde, 1990, Kane and Marcus, 1984, Klemkosky and Lasser, 1985 and
Resnick and Hennigan, 1983). Due to the contract design, the CBOT developed a system of
conversion factors to be applied when the delivered bond has a coupon which differed from the
contracted rate.
Long-term credit futures contracts are now traded on exchanges around the globe on bonds and
notes denominated in a wide variety of currencies issued by different governments.
4

Interestingly, each contract is surprisingly similar. Table 1 provides a brief summary of some of
the more popular interest rate futures contracts now available. Note that with only two
exceptions, all of the major contracts state a coupon rate of 6% and the majority are based on a
100,000 par denominated in the domestic currency.
The main distinguishing features between different credit futures contracts are: the underlying
currency, the term to maturity of the underlying instrument, the risk of the issuing government
and the frequency of coupon payments. Although not the largest global derivative exchange,
Eurex offers the worlds most heavily traded bond futures contracts: the Euro-Bund, the Euro-
Bobl and Euro-Schatz futures contracts.
A considerable body of literature has developed since the introduction of T-bond futures
concerning biases introduced by the delivery process (e.g. Arak, Goodman and Ross, 1986,
Meisner and Labuszewski, 1989, Ritchken and Sankarasubramanian, 1992, and Livingston,
1987). The delivery option allows the short the alternative of delivering anytime during the
delivery month depending on conditions which benefit the short. The delivery process creates
circumstances which result in a cheapest-to-deliver (CTD) cash bond. The CTD phenomenon
can be traced to: (1) biases inherent in the conversion factor invoice system and (2) cash market
peculiarities (Benninga and Wiener, 1999, and Kane and Marcus, 1984).
This paper first introduces the conversion factor methodologies used by the CME Group, Eurex
and NYSE-Euronext. Following the exchange approaches to the conversion factor, a
conventional model is outlined and then all approaches are applied to two separate and
independent data sets in order to determine the implications.

2
This change was made in response to the introduction of the Ultra T-bond futures contract which requires the delivery of an underlying bond
with 25 years or more until maturity.

3
LIFFE is now part of NYSE Group.

4
Futures are not, at this time, traded on corporate bonds on any exchange.
906

Table 1: Long-Term Interest Rate Futures
Issuer CR Exchange Par
10-year euro Notional 6% March Terme de France 1,000,000
3-year Commonwealth T-
bonds 6%
Sydney Futures Exchange
(SFE) A$100,000
5-year German euro-Bobl 6% EUREX 100,000
10-year German euro-Bund. 6% EUREX 100,000
2-year German euro-Schatz 6% EUREX 100,000
10-year JGB 6% NYSE-Euronext/TSE 1,000,000
Long Gilt 6% LIFFE 50,000
5-year Malaysia Govt. 6% Bursa Malaysia RM100,000
5-year Thai Govt. 6% Thailand Futures Ex (TFEX) 1,000,000
3-year Hong Kong Note 5% Hong Kong Futures Exchange
HK$1,000,00
0
10-year Italian Govt. 6% EUREX 100,000
30-year Euro-Buxl
(German) 4% EUREX 100,000
15-year U.S. T-bond 6% CME Group $100,000
10-year CONF (Swiss) 6% EUREX CHF100,000
10-year Canadian Govt. 6% Montreal Exchange C$100,000

II. Conversion Factor Methodologies
Derivative exchanges employ various conversion factor (CF) techniques to price their long-term
interest rate futures contracts with varying degrees of precision relative to the underlying cash
bond. Some of the differences between exchange approaches to conversion factor calculations
can be explained by the different features of the underlying bond. The conversion factor
techniques currently used are reviewed below.
All approaches to the conversion factor follow the same simple format b(a + d + e) f,
where:
b = (1+(c/m))
-t1/T
the present value interest factor which discounts the cash flows
from the next coupon payment date (NCD) back to the delivery date (DD). The
variable m is the number of coupon payments per year.

a = (C/c)(1-(1+c/m)
-n
) the discounted value of the delivered coupon stream,
C,from maturity back to the NCD at the contracted rate of c.

d = (1+(c/m))
-n
the discounted value of par from maturity back to the NCD.

e = (C/m) the coupon paid on the delivered bond on the NCD. This is not part of
the annuity stream unless delivery takes place on a coupon payment date.

f = (C/m)(t
o
/T) the accrued interest due on the DD.

CME Group
The CME Group, Inc. is based in Chicago and is, at this time, the worlds second largest
derivatives exchange with 2011 trading volume reported by the Futures Industry Association to
be more than 2.91 billion.
5
On July 12
th
, 2007 the Chicago Mercantile Exchange (CME) merged

5
The largest global derivatives exchange is the Korea Exchange with trading volume of 3.356 billion in 2011.
www.futuresindustry.org/fimagazine Jan/Feb 2012.
907

with the CBOT, the oldest and, at one time, the largest futures exchange, expanding its offerings
of interest sensitive futures contracts. After more than two years of negotiations, the CBOT and
the CME became officially wholly owned subsidiaries of The Group.
The CME Group offers long-term credit futures contracts on the U.S. T-bond/T-Note (UST)
spanning the yield curve from 2-years to 25+ years and uses the conversion factor in equation 1
to price the delivery of a bond other than the 6% coupon described in the futures contract:
6

CF
UST
= (1+(c/2))
-v/6
*[(C/c)(1-(1+c/2)
-N*2
)

+ (1+(c/2))
-N*2
+ (C/2)] (C/2)(6-v)/6 1
The variables are defined as:
CF
UST
= the dollar-denominated conversion factor;
C = the coupon rate on the delivered bond;
c = the contracted coupon rate (usually 6%);
N = the number of whole years from the first day of the delivery month until the first
day of the month of maturity of the delivered bond, when z 7 then 1 should be
added to n;
z = the number of whole months between the last full integer year and maturity;
v = round z down to the nearest quarter, z. If z < 7, then v = z. If z 7, then v = 3.
7

The CF
UST
components are as follows:
b = (1+(c/2))
-v/6

a = (C/c)(1-(1+c/2)
-N*2
)
d = (1+(c/2))
-N*2

e = (C/2)
f = (C/2)(6-v)/6
The CBOT's procedure includes quarters of a year counting from the first day of the delivery
month to the first day of the delivered bonds maturity.
8
The number of full years until the cash
bond matures (or the first call date) is first determined, and then the numbers of months
remaining until maturity, after the full years, are counted. The remaining months are
subsequently rounded down to the nearest quarter.
Several notable deviations exist between the cash flow pattern of a deliverable cash UST and the
CME Group conversion factor used to adjust the price:
(i) The conversion factor counts years starting and ending with the first day
of the relevant months while the price of the underlying deliverable bonds
is based on an actual ABF day-count cash flow pattern.
9

(ii) Rounding rules are applied to the number of months used in the
calculation of conversion factors while the price of a deliverable bond is
based on actual months.
(iii)Whole months are used to obtain accrued interest for the CF
UST
while day-
count is the basis for accrued interest on the deliverable bond.

6
Interest Rate Resource Center. 2008. Calculating U.S. Treasury Futures Conversion Factors. Chicago Mercantile Exchange Group, Brochure.

7
This holds for the Ultra, the 15 year T-bond and the 10 year T-note. However, for the three year or five year T-notes z should be rounded down
to the nearest month, z, and if z 7 for the two year, then v = z - 6.

8
The CME approach to T-Notes is to round down to the nearest month. Note that the exchanges approximation holds regardless of the actual
delivery day selected by the short.

9
Day count conventions and definitions are in the appendix.
908

(iv) The CF
UST
is rounded to four decimal places but cash T-bonds are rounded
to seven decimal places.
Eurex
Eurex Group offers long-term interest rate futures which represent all major points along the
yield curve and are denominated in either Euros or Swiss francs (CONF). Eurex Group is the
worlds third largest derivatives exchange, and Europes largest, with 2011 trading volume in
excess of 2.43 billion contracts. Eurex was created in 1996 with the merger between the Deutshe
Brse AG and SWX Swiss Exchange, originally offering derivatives only on issues originating
in Europe. However, in February 2004, Eurex U.S. opened for trade in dollar-dominated
products. In its first year, Eurex U.S. achieved a record volume of 1.066 billion. From 1999-
2006, prior to the CME-CBOT and the NYSE-Euronext mergers, Eurex was the undisputed
largest derivatives exchange in the world with trading volume of 1.526 billion contracts.
Eurex Group offers credit futures contracts covering the German yield curve from the one month
to 30 years in length and Swiss Confederation bond futures covering the yield curve from 8 to 13
years. The Euro Bund (German) futures is considered the benchmark for the European yield
curve. Specifically, Eurex offers futures on the medium and long-term notional debt presented in
Table 2:

Table 2: Eurex Credit FuturesContracts
Contract ID Term CR Currency
P
Euro-Schatz Futures FGBS 1.75 to 2.25 6% EUR
Euro-Bobl Futures FGBM 4.5 to 5.5 6% EUR
Euro-Bund Futures FGBL 8.5 to 10.5 6% EUR
Euro-Buxl Futures FGBX 24.0 to 35.0 4% EUR
Short-Term Euro-BTP
Futures
FBTS 2 to 3.25 6% EUR
Long-Term Euro-BTP
Futures
FBTP 8.5 to 11 6% EUR
CONF Futures CONF 8.0 to 13.0 6% CONF
Source: www.Eurexchange.com

Three different methods are employed by Eurex Group for determining the conversion factor
depending on the features of the underlying cash bond. The exchange separates the futures into
three different categories due to differences in the cash flow pattern of the underlying bond: the
Euro-BTP bond futures, Euro-denominated bond futures and CONF bond futures.
10
All
conversion factors are rounded to six decimal places.
BTP Futures
The Euro-BTP (Buoni del Tesoro Poliennnali) futures first began trading on Eurex on Sept 14
th
,
2009. The BTP is an Italian government issued bond which serves to expand the sovereign euro-
denominated futures offerings by providing a higher risk hedging alternative (AA-rating). The
coupon payments on the BTP are treated in much the same way as the UST bond in that equal
payments, at equal intervals, are made twice per year but the ISDA day-count is used rather than
the AFB day-count. Likewise, each time the bond trades, accrued interest is considered in the
invoice pricing. While the pricing and accrued interest characteristics of the Euro-BTP are

10
The Euro-denominated bond futures include: Euro-Schatz, Euro-Bobl, Euro-Bund and Euro-Buxl. The CONF is the only Eurex bond futures
offered denominated in Swiss francs (CHF).

909

similar to the UST bond, the conversion factor is slightly different as shown in equation 2
below:
11

CF
BTP
= (1 + c)
-x/2
*[(C/c)(1 -(1+c)
-n/2
) + (1+c)
-n/2
+ (C/2) + (C/2)(t
o
/T)] - (C/2)((t
o
-t
1
)/T) 2

All variables are as previously defined and:

CF
BTP
= the euro-denominated BTP conversion factor;
n = the number of coupon payments; and,
x = (1 + t
1
/365).
The BTP bond pays interest semi-annually, however, the conversion factor inputs several of the
variables as if the bond is paid annually and others as semi-annual.
b = (1 + c)
-x/2

a = (C/c)(1-(1+c)
-n/2
)
d = (1+c)
-n/2

e = (C/2) + (C/2) (t
o
/T)
f = (C/2)((t
o
-t
1
)/T)
Several notable deviations exist between the cash flow pattern of a deliverable Euro-BTP bond
and the Eurex Group conversion factor used to adjust the price:
(i) The power of the discount factor, b, used in the CF
BTP
approach adds 1 to
the proportion of time (t
1
/T) and then divides by 2 while the price of the
underlying BTP cash bonds conforms to a (t
1
/T) power.
(ii) The annuity component, a, in the CF
BTP
, approach discounts at a rate of c
while the market discounts at a rate of (c/2). Likewise, the CF
BTP
approach
uses whole years rather than the coupon periods priced in the cash bond.
(iii)The par value is discounted at (c), instead of (c/2), over the life of the
bond rather than over the relevant number of coupon periods.
(iv) The CF
BTP
approach results in net accrued interest of (t
1
/T)(C/2), however,
the portion reflecting the accrued interest at t
o
is discounted in the BTP
model. This results in a deviation from the cash market accrued interest
equal to: (C/2)(t
o
/T)(b-1).

Euro-Denominated Bond Futures
Germany and France are the main issuers of the benchmark, risk-free (AAA-rated), Euro
denominated bonds on which futures contracts are written. The interest payments and accrued
interest calculations for Euro-denominated bonds are based on annual coupons and an AFB day-
count. Eurex offers four different interest sensitive futures contracts in this category ranging in
maturity from 1.75 years to 35 years until maturity: the Euro-Buxl, the Euro-Bund, the Euro-
Bobl and the Euro-Schatz. Only one of the four Euro-bond futures requires a coupon rate which
differs from the current global standard of 6%; the Euro-Buxl contract with a 4% coupon.

11
Eurex. Conversion Factor Calculation for Euro-BTP Futures. http://www.Eurexchange.com/market/clearing/statistics/calculator_en.html.

910

Equation 3 below is the conversion factor methodology applied by Eurex to Euro-denominated
bonds:
12

CF
= (1+c)
-x
[(C/c){(1+c) (1+c)
-N
} + (1+c)
-N
+ C(t
o
/365)] C((t
o
-t
1
)/365) 3
The variables are as previously defined and CF

denotes the conversion factor for euro-
denominated bonds.
b = (1 + c)
-x

a+ C = (C/c)((1+c)-(1+c)
-N
)
d = (1+c)
-N

e = C(t
1
/T) see also a
f = C((t
o
-t
1
)/T)
Three notable deviations exist between the cash flow pattern of deliverable cash Euro-bonds and
the Eurex Group conversion factor used to adjust the price:
(i) The power of the discount factor, b, used in the CF
BTP
approach adds 1 to
the proportion of time (t
1
/T) while the price of the underlying BTP cash
bonds conforms to a (t
1
/T) power.
(ii) The CF
BTP
approach results in net accrued interest as (t
1
/T)(C/2), however,
the portion reflecting the accrued interest at t
o
is discounted in the BTP
model. This results in a deviation from the cash market accrued interest of:
C(t
o
/365)(b-1).
(iii)The CF
BTP
uses a combination of an annual and semi-annual valuation
approach whereas the cash bonds are priced as strictly semi-annual
CONF- Bond Futures
The Swiss Confederation (CONF) is the only issuer of the CHF-denominated AAA-rated bonds.
The underlying instruments are non-vers issues and, like the BTP bond, pay interest only once
per year. Non-vers issues are partly paid such that when the flat flag is set no accrued interest is
included in the calculation. All CONF-bond futures contracts require a 6% coupon rate and a
maturity in the range of 8-13 years. Equation 4 below shows the general formula applied to
CF
CHF
calculations. Adjustments are made as needed to consider partly paid issues and accrued
interest conditions.
13

CF
CHF
= (1+c)
-z/12
* [(C/c)((1+c) (1+c)
-N
) + (1+c)
-N
] C(1-(z/12)) 4
The variables are as previously defined and CF
CHF
denotes the conversion factor for the CONF-
bond futures.
b = (1 + c)
-z/12

a+C = (C/c)((1+c)-(1+c)
-N
)
d = (1+c)
-N

e = included in a
f = C(1-(z/12))
There are two notable deviations between the cash flow pattern of a deliverable cash CONF bond
and the Eurex Group conversion factor used to adjust the price:

12
Eurex. Conversion Factor Calculation for Euro-denominated Bonds. http://www.Eurexchange.com/market clearing/statistics/calculator
_en.html.

13
Eurex. Conversion Factor Calculation for CONF-Bonds. http://www.Eurexchange.com/market/clearing/statistics/calculator _en.html.

911

(i) The CONF uses a month-count, rather than the day-count, employed by
the market when discounting from the NCD back to the DD.
(ii) The CONF conversion approach to accrued interest uses a month-count
instead of a day-count. This results in a larger difference as time passes.

NYSE-Euronext
The last CF evaluated is the NYSE-Euronext method applied to the Japanese Government
Bonds (JGB) which was launched in 1987 on the London International Financial Futures
Exchange (LIFFE).
14
An agreement was forged between the Tokyo Stock Exchange (TSE)
and LIFFE which established a cooperative network linkage. Although LIFFE is now a
subsidiary of the NYSE Group, the cooperative relationship with the TSE has continued to
solidify. As recently as March 1, 2011 the exchanges signed a letter of intent establishing
alternatives to provide for greater customer accessibility.
The most actively traded NYSE-Euronext credit futures are written on the U.K. gilt bonds. Gilts
are predominantly registered investments making it necessary to establish the identity of the
recipients of each coupon payment ahead of the coupon date. Consequently, there is a period
prior to each coupon date when Gilts trade without entitlement to the upcoming coupon payment
(i.e. it is traded ex-coupon). This complicates the applied conversion factor beyond those used
by other exchanges making it a special case. As such the Gilt conversion factor is not presented
here. Instead, NYSE-Euronexts AAA-rated Japanese Government bond (JGB) futures
conversion factor is reviewed. The JGB pays interest semi-annually and the CF
JGB
is rounded to
seven decimal places, as are the underlying cash bonds, using equation 5 below.
15

CF
= (1+c/2)
-nm/6
[(C/c)((1+c/2)
np
1) + 1] C((6-t
m
)/12) 5
C and c are as previously defined and:

CF
= the yen-denominated conversion factor;

n
m
= the number of months from the delivery day until the maturity date;
n
p
= the number of payments from the delivery day until the maturity date; and,
t
m
= the number of months from the delivery day until the next coupon payment date.
The CF

components are:
b = (1+c/2)
-nm/6
the present value of the future value lump sum from the
maturity date back to the delivery date;

a = (C/c)((1+c/2)
np
1) the future value of the delivered coupon stream upon
maturity;
d = 1. the future value of par;
e = 0. included in factor a; and,
f = C((6-t
m
)/12) accrued interest.
The JGB conversion factor approach is unique in that the cash flows are compounded to maturity
and then discounted as a lump-sum back to the delivery day. This is theoretically equivalent to

14
The TSE also trades the JGB futures using an identical conversion factor methodology.

15
Japanese Government Bond Conversion Factor Formula. Tokyo Stock Exchange. http://www.tse.or.jp/english/rules/derivatives/jgbf/b7/
gje60000003zw2-att/cfformulae.pf.

912

the other approaches which obtain the discounted value of the cash flows. Therefore, this fact
alone will not cause notable differences, however, the following source of deviation from the
cash market is present:
(i) The JGB conversion factor approach to calculating accrued interest is
based on a month-count rather than a day-count.

III. A Conventional Methodology
When conversion factors are not precise, inefficiencies are introduced into the market which can
result in unintended consequences. The conventional long-term interest rate futures conversion
factor approach, CF
c
, replicates the cash flow pattern of the underlying cash bond including the
day-count for which accrued interest is determined. As such the conventional model is the
present value of a 100% par bond on a coupon paying bond as shown in equation 6:

b a d e f
|----------------| |-----------------------| |-------------| |--------| |-------------|

CF
c
= (1+(c/m))
-t1/T
*[(C/c)(1-(1+c/m)
-n
)

+ (1+(c/m))
-n
+ (C/m)] (C/m)(t
o
/T) 6

The variables are redefined below for clarity:
CF
c
= the conventional conversion factor rounded to nine decimal places;
c = the annual percentage coupon rate stated in the futures contract (frequently
6%);
m = the number of coupon payments made per year (m=1 or m=2);
C = the annual percentage coupon rate on the delivered bond;
n = the number of coupon payments remaining counting the NCD
+1
until maturity
(NCD = next coupon date, where +1 indicates the coupon date following the
NCD);
t
1
= the number of days from the DD until the NCD;
t
o
= the number of days from the last coupon date (LCD) until the DD;
T = the number of days in the payment interval.
When the conventional approach is applied during leap years the actual/actual day-count
approach, which is consistent with the underlying bond, is employed. Table 3 summarizes the
conversion factors by component for ease of comparison.

Table 3 Summary of Conversion Factor Components
CF
c
CF
UST
CF
BPT
CF
CF
CHF
CF

b (1+(c/2))
-
t1/T

(1+(c/2))
-v/6
(1 + c)
-x/2
(1 + c)
-x
(1+ c)
-z/12
(1+c/2)
-nm/6

a (C/c)(1-
(1+c/2)
-n
)
(C/c)(1-
(1+c/2)
-N*2
)
(C/c)(1-(1+c)
-
n/2
)
(C/c)((1+c)-
(1+c)
-N
)
(C/c)((1+c)-
(1+c)
-N
)
(C/c)((1+c/2)
n
p
1)
d (1+(c/2))
-n
(1+(c/2))
-N*2
(1+c)
-n/2
(1+c)
-N
(1+c)
-N
1
e (C/2) (1+(c/2))
-N*2
(C/2)+(C/2)
(t
o
/T)
C(t
1
/T)see
also a
included in a included in a
f (C/2)(t
o
/T) (C/2)(6-v)/6 (C/2)((t
o
-t
1
)/T) C((t
o
-t
1
)/T) C(1-(z/12)) C((6-t
m
) /12)
913

The next section present the first of two tests applied to each exchange CF approach and
examines the source and magnitude of the difference between the approximation methods and
the conventional approach.

IV. Evidence of Conversion Factor Discrepancies
In this section, a hypothetical Feb 15 cash bond with 5, 10, 15 or 20 years until maturity with
coupon rates ranging from 2% to 10% is assumed to be delivered. In order to identify possible
patterns resulting from currently used CF methods, all possible delivery dates, from the first day
after the last coupon payment until the day before the next coupon payment date for each coupon
rate selected, is evaluated. For each day, the difference between the conventional CF and the
exchange CF was calculated and the results appear in Tables 4-8.
Casual observation of the results indicates that, even after adjusting for differences in bond
characteristics, the exchange CF approaches show differences from the conventional
methodology. Some approaches show greater differences than others and the differences are
larger for some delivery dates and coupon rates than others. Even in the case when the delivered
bond meets the maturity and coupon requirements of the futures contract rate, the conversion
factors diverge from the expected solution except when the delivery date falls on a coupon
payment date.
The CME Group conversion factor method is first reviewed followed by Eurex and then NYSE-
Euronext. Throughout, delivery is assumed to take place on the last business day of the delivery
month.
UST
Table 4 summarizes the differences between the conventional CF and the CME CF across all
coupons and maturities studied. As expected, deliverable bonds with a coupon rate less than 6%
result in a discount to the contracted futures price while those with coupon rates greater than 6%
result in a premium to the contract price.
The magnitude of the conversion difference between the conventional approach and the CME
Group approach is obtained as: CF
c
- CF
UST
. After testing all possible delivery days, the results
indicate that when the coupon rate on the delivered bond is equal to the contract rate of 6%
virtually no difference exists between the two models regardless of the maturity of the
underlying cash bond or the delivery day selected by the short. Likewise, when delivery is at the
mid-point (May 15
th
) between coupon payment dates the difference between the two conversion
factor methods is indiscernible.
However, when the delivery date is prior to the coupon mid-point, the CME Group CF method
overestimates the price of delivering a low coupon bond and underestimates the price of a high
coupon bond. Similarly, when the delivery date is past the coupon mid-point, the CF
UST
underestimates the price of delivering a low coupon bond and overestimates the price of
delivering a high coupon bond. This pattern holds true regardless of the maturity of the
underlying bond delivered. This implies that if the CTD has a high coupon, the short is better off
in a contract that will deliver prior to the coupon mid-point. Likewise, if the CTD has a low
coupon, the short if better off selecting a futures that will deliver after the coupon mid-point.
The implications of the CME Group results suggest that all else the same, the short is better off
selecting a low-coupon bond when the month of delivery is prior to the coupon mid-point and
initiating delivery as early as possible during the delivery month. However, when the delivery
month falls after the coupon mid-point the short will benefit by selecting a high-coupon bond
and initiating delivery as late in the month as possible. The opportunity to select in such a
914

manner provides the short with a valuable conversion option which is not unlike the delivery
option described by Hedge, 1988, and Helmer, 1990.
The estimation technique used by the CME Group introduces unnecessary rounding to the
conversion price which results in the differences observed in Table 4. Specifically, when the
components of equation 2 are tested separately, a, d and e contribute very little to the observed
differences between the CF
c
approach and the CF
UST
approach. Instead, the major sources of the
differences lie in components b and f and are due to the approximation procedures applied to
accrued interest.

BTP
Applying equations 2 and 6, Table 5 shows the BTP deviations from the CF
c
using the same data
set described above. The Eurex BTP conversion factor utilizes a more conventional day-count
for accrued than does the CME UST approach, however, the discount factor, identified by
component b, introduces large relative differences. Observed deviations are also attributable, to a
lesser degree, to components a and d.

Table 4 CME UST Conversion Factor Results

5 Year
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -0.002446396 -0.001196667 5.30623E-05 0.001302791 0.002552521
4/15/2011 -0.001244915 -0.00061566 1.35955E-05 0.000642851 0.001272106
5/15/2011 -5.98333E-05 -2.98969E-05 3.94983E-08 2.99759E-05 5.99123E-05
6/15/2011 0.004699489 0.002300132 -9.9225E-05 -0.00249858 -0.00489794
7/15/2011 0.005929623 0.002933244 -6.3136E-05 -0.00305951 -0.00605589

10 Year
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -0.001811296 -0.000879117 5.30623E-05 0.000985241 0.00191742
4/15/2011 -0.000924014 -0.000455209 1.35955E-05 0.0004824 0.000951205
5/15/2011 -4.45149E-05 -2.22377E-05 3.94983E-08 2.23167E-05 4.45939E-05
6/15/2011 0.003479933 0.001690354 -9.9225E-05 -0.0018888 -0.00367838
7/15/2011 0.004401425 0.002169145 -6.3136E-05 -0.00229542 -0.0045277

15 Year
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -0.001222118 -0.000584528 5.31E-05 0.000690652 0.001328243
4/15/2011 -0.000626317 -0.000306361 1.36E-05 0.000333552 0.000653508
5/15/2011 -3.03E-05 -1.51E-05 3.95E-08 1.52E-05 3.04E-05
6/15/2011 0.00234856 0.001124667 -9.92E-05 -0.00132312 -0.00254701
7/15/2011 0.002983727 0.001460296 -6.31E-05 -0.00158657
-
0.003109998

20 Year
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -0.000928389 -0.000437663 5.30623E-05 0.000543788 0.001034513
4/15/2011 -0.000477903 -0.000232154 1.35955E-05 0.000259345 0.000505094
5/15/2011 -2.32194E-05 -1.15899E-05 3.94983E-08 1.16689E-05 2.32984E-05
6/15/2011 0.001784524 0.000842649 -9.9225E-05 -0.0010411 -0.00198297
7/15/2011 0.002276947 0.001106906 -6.3136E-05 -0.00123318 -0.00240322

915

Table 5 Eurex BPT Conversion Factor Results

5 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%

3/15/2011 0.007040605 0.004002172 0.00096374 -0.00207469 -0.005113128

4/15/2011 0.009646898 0.006776805 0.00390671 0.00103662 -0.001833474

5/15/2011 0.01217097 0.009460833 0.0067507 0.00404056 0.001330421

6/15/2011 0.014781137 0.012233226 0.00968531 0.0071374 0.004589492

7/15/2011 0.017309045 0.014915155 0.01252126 0.01012737 0.007733483

10 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%

3/15/2011 0.004658066 0.002810902 0.00096374 -0.00088343 -0.002730589

4/15/2011 0.00709763 0.005502171 0.00390671 0.00231125 0.000715794

5/15/2011 0.00945918 0.008104938 0.0067507 0.00539645 0.004042211

6/15/2011 0.011900193 0.010792754 0.00968531 0.00857788 0.007470437

7/15/2011 0.014263218 0.013392241 0.01252126 0.01165029 0.01077931

15 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%

3/15/2011 0.002829 0.001896 0.000964 3.12E-05 -0.0009

4/15/2011 0.005124 0.004515 0.003907 0.003298 0.002689

5/15/2011 0.007345 0.007048 0.006751 0.006453 0.006156

6/15/2011 0.00964 0.009663 0.009685 0.009708 0.009731

7/15/2011 0.01186 0.012191 0.012521 0.012852 0.013182

20 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%

3/15/2011 0.001876721 0.00142023 0.00096374 0.00050725 5.07559E-05

4/15/2011 0.004085815 0.003996263 0.00390671 0.00381716 0.003727609

5/15/2011 0.006222707 0.006486701 0.0067507 0.00701469 0.007278684

6/15/2011 0.008429884 0.009057599 0.00968531 0.01031303 0.010940746

7/15/2011 0.010564974 0.011543119 0.01252126 0.01349941 0.014477554

The CF
BTP
consistently underestimates the pricing factor across all coupon rates when the
delivery date is equal to or beyond the mid-point. When delivery is made soon after the last
coupon payment date, high coupon bond prices are overestimated. This is due to two factors: the
BTP discount component, b, and the accrued interest, f. The BTP approach to component b
overestimates the discount rate due to the addition of 1 in the time estimate, x, shown in equation
2. At the same time, component f results in a lower accrued interest estimate than the
conventional approach. When the coupon rate is high the conventional approach produces
accrued interest which is, on average, 1.9019% higher than the BTP approach, however, when
the coupon rate is low, the conventional accrued interest is 1.142% higher than the BTP
approach. Since accrued interest is subtracted from the annuity stream, this has the effect of
lowering the conventional conversion factor below that of the BTP when coupon rates are high.
However, the influence of accrued interest is not strong enough at low coupon rates to offset the
influence of the lower discount applied to the conventional annuity stream. The result is the
domination of the higher discount to the present value of cash flows from the NCD back to the
DD at most coupon rates and delivery dates. While the annuity component, labeled a,
underestimates the CF
c
approach, this is offset at higher coupon rates even after multiplying by
916

the discount factor. In general, while the underlying cash BTP bond is characterized by semi-
annual payments, Eurex makes adjustments to the conversion factor which converts the semi-
annual payments to annual in some parts of the calculation but not all parts. This inconsistency is
a major source of the pricing discrepancy associated with the deviation of the CF
BTP
from the
CF
c
.
The implication is that, with few exceptions, the long will pay less in the BTP futures market
than for a like bond in the cash market.

EUROBONDS
The results in Table 6, generated by equations 3 and 6, indicate that for coupon rates equal to 6%
and higher, the CF
approach overestimates the price of delivering a bond other than the one
specified in the contract. This holds across all delivery dates and all possible contracts tested.
The difference is largest for the highest coupon bonds which deliver close to the LCD of the
underlying bond. Conversely, when the CF
method is applied to low coupon bonds, the Eurex

factor consistently underestimates the actual price with larger deviations occurring when the
delivery date is close to the NCD.

The sources of the conversion differences introduced by Euro-bond CFs are components b and f.
The Euro-bond approach is almost identical the BTP-bond approach to the conversion factor

Table 6 Eurex Euro Conversion Factor Results

5 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.011591 0.004468 -0.00265 -0.00978 -0.0169
4/15/2011 0.012177 0.005151 -0.00188 -0.0089 -0.01593
5/15/2011 0.012716 0.005776 -0.00116 -0.0081 -0.01504
6/15/2011 0.013244 0.006387 -0.00047 -0.00733 -0.01419
7/15/2011 0.013728 0.006942 0.000156 -0.00663 -0.01342

10 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.008214 0.00278 -0.00265 -0.00809 -0.01352
4/15/2011 0.008783 0.003454 -0.00188 -0.0072 -0.01253
5/15/2011 0.009306 0.004071 -0.00116 -0.0064 -0.01163
6/15/2011 0.009818 0.004673 -0.00047 -0.00562 -0.01076
7/15/2011 0.010284 0.00522 0.000156 -0.00491 -0.00997

15 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.005268 0.001307 -0.00265 -0.00662 -0.01058
4/15/2011 0.005823 0.001974 -0.00188 -0.00572 -0.00957
5/15/2011 0.006332 0.002584 -0.00116 -0.00491 -0.00866
6/15/2011 0.006828 0.003179 -0.00047 -0.00412 -0.00777
7/15/2011 0.007281 0.003718 0.000156 -0.00341 -0.00697

20 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.00349 0.000417 -0.00265 -0.00573 -0.0088
4/15/2011 0.004035 0.00108 -0.00188 -0.00483 -0.00779
5/15/2011 0.004535 0.001686 -0.00116 -0.00401 -0.00686
6/15/2011 0.005023 0.002276 -0.00047 -0.00322 -0.00597
7/15/2011 0.005467 0.002812 0.000156 -0.0025 -0.00515

917

except the Euro-bond pays interest annually while the BTP-bond pays interest semi-annual.
When the coupon rate is below 6% the accrued interest is 3.4003% higher using the conventional
approach than the Euro approach but the discount is 5.4167% higher. The discount factor
dominates at low coupon rates resulting in a higher conventional conversion factor. When the
coupon rate is 6% or greater, the accrued interest subtracted from the CF is 5.6671% higher in
the conventional model than the Euro model. In this case, the accrued interest factor dominates
the discount component resulting in a lower conventional conversion factor than the Euro
approach.

CONF BONDS
Table 7 shows the difference between the conventional conversion factor and the CONF factor,
using equations 6 and 4, when applied to the same selected coupon rates, delivery dates and
maturity dates as previously shown.
Table 7 Eurex CHF Conversion Factor Results

5 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -8.464E-05 -3.7492E-05 9.6584E-06 5.68089E-05 0.00010396
4/15/2011 -6.551E-05 -2.98161E-05 5.8821E-06 4.15803E-05 7.7279E-05
5/15/2011 -8.198E-05 -3.82594E-05 5.4658E-06 4.91911E-05 9.2916E-05
6/15/2011 -6.193E-05 -2.96016E-05 2.7223E-06 3.50461E-05 6.737E-05
7/15/2011 -7.888E-05 -3.85604E-05 1.7638E-06 4.2088E-05 8.2412E-05

10 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -6.162E-05 -2.59819E-05 9.6584E-06 4.52988E-05 8.0939E-05
4/15/2011 -4.797E-05 -2.10415E-05 5.8821E-06 3.28056E-05 5.9729E-05
5/15/2011 -6.034E-05 -2.74385E-05 5.4658E-06 3.83701E-05 7.1274E-05
6/15/2011 -4.582E-05 -2.15467E-05 2.7223E-06 2.69912E-05 5.126E-05
7/15/2011 -5.865E-05 -2.84431E-05 1.7638E-06 3.19706E-05 6.2178E-05

15 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -4.15E-05 -1.59E-05 9.66E-06 3.53E-05 6.09E-05
4/15/2011 -3.27E-05 -1.34E-05 5.88E-06 2.52E-05 4.44E-05
5/15/2011 -4.15E-05 -1.80E-05 5.47E-06 2.89E-05 5.24E-05
6/15/2011 -3.18E-05 -1.45E-05 2.72E-06 2.00E-05 3.72E-05
7/15/2011 -4.10E-05 -1.96E-05 1.76E-06 2.31E-05 4.45E-05

20 Years
DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -2.941E-05 -9.87821E-06 9.6584E-06 2.91951E-05 4.8732E-05
4/15/2011 -2.341E-05 -8.76484E-06 5.8821E-06 2.0529E-05 3.5176E-05
5/15/2011 -3.006E-05 -1.22988E-05 5.4658E-06 2.32305E-05 4.0995E-05
6/15/2011 -2.328E-05 -1.0277E-05 2.7223E-06 1.57215E-05 2.8721E-05
7/15/2011 -3.034E-05 -1.4288E-05 1.7638E-06 1.78155E-05 3.3867E-05

There is virtually no difference in the final result for the CF
CHF
and the CF
c
across all coupons
and all deliveries, however, there are two sources of component differences, caused by the
month-count, which serve to offset each other. The month-count differences result in a higher
CF
CHF
discount rate and lower accrued interest.

Although the differences in conversion factors are negligible, it is interesting to note that there
exists a systematic pattern in the sign of the differences. Across all maturities and all delivery
dates tested the CF
CHF
, is slightly overestimated at coupon rates below 6%. However, at coupon
918

rates above 6% the CF
CHF
is slightly underestimated. Even so, given the small magnitude of the
difference no beneficial trading strategies result. The end result is that the short has virtually no
incentive to select one coupon or delivery date due to the CONF conversion factor methodology
employed.
JGB
Applying equation 5 to a Feb 15 JGB results in the conversion differences shown in Table 8 for
the selected delivery dates and coupon rates.
Table 8 NYSE-Euronext JGB Conversion Factor Results

5 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.0053038 0.0075594 0.0098566 0.0120708 0.0143265
4/15/2011 0.0005076 0.0052249 0.0099028 0.0146595 0.01937686
5/15/2011 0.000525 0.0052446 0.0099506 0.0146837 0.01940322
6/15/2011 0.0005038 0.0052457 0.0099991 0.0147294 0.01947132
7/15/2011 0.001753 0.0059004 0.0100477 0.0141951 0.01834242

10 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.0014378 0.0031693 0.0049422 0.0066323 0.00836378
4/15/2011 -0.00213 0.0014368 0.0049641 0.0085704 0.01213721
5/15/2011 -0.002134 0.0014335 0.0049876 0.0085687 0.01213627
6/15/2011 -0.002167 0.0014167 0.0050116 0.0085834 0.01216679
7/15/2011 -0.001248 0.0018938 0.0050356 0.0081773 0.01131901

15 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 0.005519783 0.002740968 3.60E-06 -0.00281666 -0.005595477
4/15/2011 0.003055341 0.001547963 1.12E-06 -0.00146679 -0.002974172
5/15/2011 0.003069572 0.001541584 3.95E-08 -0.00151439 -0.003042381
6/15/2011 0.003064076 0.001525981 -5.36E-07 -0.00155021 -0.003088304
7/15/2011 0.003733304 0.001865975 -1.35E-06 -0.00186868 -0.00373601

20 Years

DD C = 4% C = 5% C = 6% C = 7% C = 8%
3/15/2011 -0.001456 -0.0007468 3.597E-06 0.0006711 0.00137998
4/15/2011 -0.003304 -0.0016315 1.118E-06 0.0017126 0.00338468
5/15/2011 -0.003325 -0.0016558 3.95E-08 0.0016829 0.00335229
6/15/2011 -0.003361 -0.0016867 -5.356E-07 0.0016625 0.00333706
7/15/2011 -0.002889 -0.0014453 0.0050356 0.0014426 0.00288663

The CF
JGB
tends to underestimate the conventional factor for most of the observations. There are
isolated exceptions for the longer maturity bonds. Generally, the results suggest there are only
restrictive opportunities for the short to gain and the long tends to be the beneficiary of the
discrepancies.
The underlying JGB accrued interest is based on an actual/365 day-count, however, the CF
is
based on a month-count. Therefore, the application of equation 6 results in differences which
becomes larger as the delivery date approaches a new coupon payment date. The conversion
factor approach for the JGB results in negligible deviations from the conventional approach
when the coupon rate is 6% regardless of the delivery date.

919

V. Implications of CF Discrepancies
To further test the implications of using a conventional conversion factor versus an exchange
approximation, a random day in each month of 2011 was selected to determine the CTD selected
by each method presented. For each day studied, the population of deliverable bonds for three
futures contracts was used to determine the cheapest-to-deliver bonds using all five exchange
conversion factor methods. The bond futures contracts studied consisted of a 5-year contract to
be delivered March 2012, 10-year contract to be delivered Jun 2012 and 15-year contract to be
delivered March 2012.
The CTD was determined in the usual manner shown below:
NB
d
= (F
o
*CF
d
+ AI
d
) (B
d
+ AI
d
) = F
o
*CF
d
B
d
7

NB
d
= the net benefit to the short on the delivery day;
F
o
= the futures contract price on the contract date;
CF
d
= the conversion factor calculated on the delivery day based on the delivered bond
(normally the cheapest-to-delivery);
AI
d
= the accrued interest calculated on the delivery day; and,
B
d
= the price of the cash bond selected by the short to deliver.

Equation 7 was applied to all deliverable bonds using each exchange CF and the conventional
CF, the bond producing the highest net benefit to the short was identified as the CTD. This
procedure was repeated for each of the three futures contract. A total of thirty cash T-bonds were
identified as deliverable against the 5-year futures, twenty-two bonds were deliverable against
the 10-year contract and ten bonds against the 15-year contract. While the data is adequate to
determine whether the various conversion factors select different CTD cash bonds, it is not
appropriate for determining the exact value the benefits accruing to traders of the foreign bond
futures because data on the underlying cash bonds was not available. Instead, cash UST bonds
were used as a proxy for the BTP, CONF, Euro and JGB.
Table 9 shows the CTD selected by the conventional and CME approach when applied to the 5-
year, 10-year and 15-year futures. The highlighted cash bonds are those identified as the CTD by
the conventional method, the bolded bonds are the CTD identified by the exchange conversion
method and the highlighted/bold bonds were selected by both methods.
When both methods select the same bond, another consideration is added to the futures traders
decision of whether to reverse or deliver. For example, both methods select the 8.875 Feb 2019
bond to deliver against the June 2012 10-year futures. Since the difference is positive, this
suggests that if the long holds the futures and accepts delivery of the underlying bond, he will
pay a price for the bond that is $226.73 less than its precise value assuming cash yields are 6% at
the time. Cash bonds are price in the market according to the conventional method but futures
on those bonds are not. This suggests that when CF
c
> CF
UST
the long may be able to gain by
holding until delivery rather than reversing. This is only a potential gain; the actual gain will
depend on market yields at the time of delivery. If yields are less than 6% the long will gain
more and if yields are greater than 6% the long will gain less.
Likewise, when the difference between the conventional and exchange conversion method is
negative the short has a potential additional gain. For example, Table 9 shows that the
conventional and UST conversion methods both select the 6.375 Aug 2027 to deliver against the
15-year Mar 2012 futures. If the short elects to deliver rather than reverse he has the potential to
gain $23.54 depending on market yields at the time. Like the long, the short needs to consider the
920

impact of the application of an approximate conversion factor method as part of his overall
decision process.
Table 9 CME UST Conversion Factor Differences
5-Year
T-Bond
Futures
10-Year
T-Bond
Futures
15-Year
T-Bond
Futures

Mar (2012) Jun (2012) Mar (2012)
Deliverable Conversion Deliverable Conversion Deliverable Conversion
Maturity Coupon Difference Maturity Coupon Difference Maturity Coupon Difference
5/15/2016 5.125 0.000939717 12/31/2018 1.375 -0.015536963 8/15/2027 6.375 -0.000235381
5/15/2016 7.25 -0.001095734 2/15/2019 2.75 -0.002638137 11/15/2027 6.125 2.34001E-05
5/31/2016 1.75 0.005569261 2/15/2019 8.875 0.002267255 8/15/2028 5.5 0.000263467
5/31/2016 3.25 0.003679927 5/15/2019 3.125 0.00242525 11/15/2028 5.25 -0.000404052
6/30/2016 1.5 3.7982E-06 8/15/2019 3.625 0.002082279 2/15/2029 5.25 0.000287178
6/30/2016 3.25 1.76467E-05 8/15/2019 8.125 -0.001685999 8/15/2029 6.125 -9.36568E-05
7/31/2016 1.5 0.002977222 11/15/2019 3.375 0.00209398 5/15/1930 6.25 -9.8795E-06
7/31/2016 3.25 0.001829346 2/15/2020 3.625 -0.00185357 2/15/1931 5.375 0.000218607
8/15/2016 4.875 -0.001094075 2/15/2020 8.5 0.001862714 2/15/1936 4.5 0.000329884
8/31/2016 1 -0.006500066 5/15/2020 3.5 0.002104582 2/15/2037 4.75 0.000312901
8/31/2016 3 -0.003855544 5/15/2020 8.75 -0.00202117
9/30/2016 1 0.089952723 8/15/2020 2.625 0.002671373
9/30/2016 3 -8.36617E-06 8/15/2020 8.75 -0.002062916
10/31/2016 1 -0.00612929 11/15/2020 2.625 0.002544878
10/31/2016 3.125 -0.003597626 2/15/2021 3.625 -0.001679798
11/15/2016 4.625 -0.001298516 2/15/2021 7.875 0.001341656
11/15/2016 7.5 0.001382988 5/15/2021 3.125 0.002226728
11/30/2016 0.875 -0.003112908 5/15/2021 8.125 -0.00141593
11/30/2016 2.75 -0.002037835 8/15/2021 2.125 0.00288921
12/31/2016 0.875 0.00014411 8/15/2021 8.125 -0.001497749
12/31/2016 3.25 2.86847E-05 11/15/2021 2 0.002945896
1/31/2017 3.125 0.001839246 11/15/2021 8 -0.001255898
2/15/2017 4.625 0.00128471
2/28/2017 3 0.003606386
3/31/2017 3.25 4.66826E-05
4/30/2017 3.125 0.001835104
5/15/2017 4.5 0.001424064
5/15/2017 8.75 -0.002406279
5/31/2017 2.75 0.004077583
6/30/2017 2.5 0.006126942
In the case of the 5-year UST futures contract, the CME method selects a different bond as CTD
than the conventional method. This is common across all exchanges and allows the short another
avenue for potential gain. The short can track the CTD identified by both methods to determine
the best deliverable bond. Since F
o
*CF
UST
is locked-in as the shorts cash inflow when the
contract is written, the only variable remaining is the cost of the cash bond to fulfill the terms of
the contract. When CF
UST
> CF
c
, or the difference is negative, then when yields are lower than
6% both CTD bonds will cost more. If the duration of the CTD identified by the CF
c
is lower
than the alternate, the short will gain a minimum of $109.57/contract. However, if yields are
greater than 6% the CTD bonds will cost the short less than the numbers in Table 9 suggest. The
largest gains will accrue when the CTD bond identified by the conventional method has a higher
duration than does the CTD identified by the exchange method.
When the results of Table 9 are combined with those presented in Table 4, the implication for the
5-year futures is that the short can elect to deliver the 7.25 May 2016 instead of the 5.125 May
921

2016 and potentially gain $109.5 or more if yields increase above 6%. This gain will be
enhanced because the short is delivering in a month (March) that is after the coupon mid-point
(Feb). Table 4 indicates that this delivery date combined with the size of the coupon payment
will benefit the short thus increasing the possible gain.
When trading the 10-year futures, the short will deliver the bond identified by both methods, i.e.
the 8.825 Feb 2019. The results in Tables 4 and 9 indicate a benefit to the long in terms of both
the delivery month and the approximate conversion method employed. If yields remain low, the
long has the potential to gain more than $226.73 by holding until maturity rather than reversing
his position.
Alternatively, the 15-year CTD, 6.375 Aug 2027, selected by both the exchange and
conventional methods benefits the short by a small margin of $23.58. According to the results in
Table 4, the March delivery will have little to no impact on the results.
Table 10 Eurex BTP Conversion Factor Differences
5-Year
T-Bond
Futures
10-Year
T-Bond
Futures
15-Year
T-Bond
Futures

Mar
(2012) Jun (2012)
Mar
(2012)
Maturity Coupon CF Difference Maturity Coupon CF Difference Maturity Coupon CF Difference
5/15/2016 5.125 0.006876362 12/31/2018 1.375 0.002174205 8/15/2027 6.375 -0.037315016
5/15/2016 7.25 0.016777935 2/15/2019 2.75 0.011431498 11/15/2027 6.125 0.034166428
5/31/2016 1.75 -0.009988014 2/15/2019 8.875 0.06630687 8/15/2028 5.5 -0.01753064
5/31/2016 3.25 -0.004841906 5/15/2019 3.125 -0.001920486 11/15/2028 5.25 0.028672599
6/30/2016 1.5 -0.012404314 8/15/2019 3.625 0.019531169 2/15/2029 5.25 -0.016780194
6/30/2016 3.25 -0.01049036 8/15/2019 8.125 0.059354508 8/15/2029 6.125 -0.019351905
7/31/2016 1.5 -0.014015723 11/15/2019 3.375 -0.001195494 5/15/1930 6.25 0.034904336
7/31/2016 3.25 -0.016410454 2/15/2020 3.625 0.019780815 2/15/1931 5.375 -0.017073304
8/15/2016 4.875 0.006343913 2/15/2020 8.5 0.062410334 2/15/1936 4.5 -0.013992673
8/31/2016 1 -0.012148338 5/15/2020 3.5 -0.000674948 2/15/2037 4.75 -0.014801513
8/31/2016 3 -0.000318936 5/15/2020 8.75 0.008995482
9/30/2016 1 -0.014263718 8/15/2020 2.625 0.011376443
9/30/2016 3 -0.026222783 8/15/2020 8.75 0.064319308
10/31/2016 1 -0.011460143 11/15/2020 2.625 -0.002101982
10/31/2016 3.125 0.000644194 2/15/2021 3.625 0.020249638
11/15/2016 4.625 0.004732464 2/15/2021 7.875 0.056574888
11/15/2016 7.5 0.017740044 5/15/2021 3.125 -0.000931849
11/30/2016 0.875 -0.012298508 5/15/2021 8.125 0.007523573
11/30/2016 2.75 -0.006010969 8/15/2021 2.125 0.007787802
12/31/2016 0.875 -0.012440714 8/15/2021 8.125 0.058514884
12/31/2016 3.25 -0.010218542 11/15/2021 2 -0.002552799
1/31/2017 3.125 -0.015922015 11/15/2021 8 0.007171427
2/15/2017 4.625 -0.022595676
2/28/2017 3 -0.020514602
3/31/2017 3.25 0.007439474
4/30/2017 3.125 0.001049548
5/15/2017 4.5 0.004361525
5/15/2017 8.75 0.023038732
5/31/2017 2.75 -0.005728157
6/30/2017 2.5 -0.01048378
Table 10 outlines the results for the BTP futures contract traded on Eurex. For this sample, the 5-
year futures contract is the only one with a common CTD bond identified by both methods, the
7.25 May 2016. The long has the potential to gain 1,677.79 if the position is held until and
922

delivery accepted. This gain will be partially offset by the placement of the delivery date and
coupon relative to the mid-point.
The CTD for the 10-year BTP contract differs for the Eurex and conventional methods. The short
will likely select the 1.375% Dec 2018 bond indicated by the conventional CF because this bond
offers a lower benefit to the long than does the 8.875% Feb 2019 indicated by the exchange. The
results in Table 5 indicate that the positioning of the delivery month at the coupon mid-point and
the size of the payment will serve to slightly improve the benefit to the long.
Like the 10-year BTP futures, the conventional and Eurex CF methods select a different CTD
bond for the 15-year contract. In this case, the short will prefer to deliver 6.375 Aug 2027
selected by the conventional method, rather than the 6.125 Nov 2017 selected by the exchange
because the potential gain is 3,731.50 as oppose to losing 3,416.64; a difference of 7,148.14.
The impact of the delivery date, which is approximately one-month after the last coupon
payment, will have little to no impact on the result as shown in Table 5.
Table 11 Eurex Euro Conversion Factor Differences
5-Year
T-Bond
Futures
10-Year
T-Bond
Futures
15-Year
T-Bond
Futures

Mar
(2012) Jun (2012) Mar (2012)
Maturity Coupon CF Difference Maturity Coupon CF Difference Maturity Coupon CF Difference
5/15/2016 5.125 -0.05155349 12/31/2018 1.375 -0.037138586 8/15/2027 6.375 -0.006425532
5/15/2016 7.25 -0.055007413 2/15/2019 2.75 -0.045330282 11/15/2027 6.125 -0.067851104
5/31/2016 1.75 -0.023625545 2/15/2019 8.875 -0.060626519 8/15/2028 5.5 -0.04008592
5/31/2016 3.25 -0.005823061 5/15/2019 3.125 -0.067982329 11/15/2028 5.25 -0.073461513
6/30/2016 1.5 -0.030020726 8/15/2019 3.625 0.005754516 2/15/2029 5.25 -0.138432414
6/30/2016 3.25 -0.013497399 8/15/2019 8.125 0.059281088 8/15/2029 6.125 -0.015840555
7/31/2016 1.5 -0.033552525 11/15/2019 3.375 -0.021690372 5/15/1930 6.25 -0.063814156
7/31/2016 3.25 -0.021403407 2/15/2020 3.625 -0.046653679 2/15/1931 5.375 -0.136084223
8/15/2016 4.875 -0.015899617 2/15/2020 8.5 -0.060597151 2/15/1936 4.5 -0.155099686
8/31/2016 1 -0.039298973 5/15/2020 3.5 -0.070696129 2/15/2037 4.75 -0.149143262
8/31/2016 3 -0.030397717 5/15/2020 8.75 -0.123086447
9/30/2016 1 -0.041494469 8/15/2020 2.625 -0.004950619
9/30/2016 3 -0.037401116 8/15/2020 8.75 0.065745983
10/31/2016 1 -0.043756606 11/15/2020 2.625 -0.023880336
10/31/2016 3.125 -0.044669976 2/15/2021 3.625 -0.04584068
11/15/2016 4.625 -0.050740802 2/15/2021 7.875 -0.059451369
11/15/2016 7.5 -0.055413757 5/15/2021 3.125 -0.065983807
11/30/2016 0.875 -0.045587094 5/15/2021 8.125 -0.117566576
11/30/2016 2.75 -0.050872989 8/15/2021 2.125 -0.00943318
12/31/2016 0.875 -0.047528609 8/15/2021 8.125 0.057825398
12/31/2016 3.25 -0.060070116 11/15/2021 2 -0.025154353
1/31/2017 3.125 -0.066867377 11/15/2021 8 -0.002306175
2/15/2017 4.625 -0.083846656
2/28/2017 3 -0.07234074
3/31/2017 3.25 0.011015988
4/30/2017 3.125 0.001543548
5/15/2017 4.5 -0.049925978
5/15/2017 8.75 -0.058566839
5/31/2017 2.75 -0.010396052
6/30/2017 2.5 -0.019119933
Table 11 reveals that with very few exceptions the conversion method used by Eurex for the
Euro bond futures overestimates that of the conventional method. In addition, the size of the
923

potential advantage to the short is very large, reaching a maximum of 7,346.15 for the 15-year
contract. Even the 5-year contract indicates a large potential gain of 5,500.74.
In addition, the results reveal that both the conventional and Eurex CF methods select the 7.25
May 2016 deliverable bond against the 5-year futures contract resulting in a conversion gain to
the short. When combined with the results of Table 6, the timing of the delivery and the coupon
indicate that further gains to the short may be possible.
Likewise, both CF methods select the same, 8.875 Feb 2019 bond for the 10-year futures
resulting in a large potential benefit of 6,062.65 for the short. The coupon and the delivery
month are such that the short is also likely to gain from the timing of the payments as shown in
Table 6.
Table 12 Eurex CHF Conversion Factor Differences
5-Year
T-Bond
Futures
10-Year
T-Bond
Futures
15-Year
T-Bond
Futures

Mar
(2012) Jun (2012)
Mar
(2012)
Maturity Coupon CF Difference Maturity Coupon CF Difference Maturity Coupon
CF
Difference
5/15/2016 5.125 0.003175943 12/31/2018 1.375 0.000168008 8/15/2027 6.375 0.002211716
5/15/2016 7.25 -0.0041417 2/15/2019 2.75 -0.000752408 11/15/2027 6.125 -0.014207074
5/31/2016 1.75 8.93811E-05 2/15/2019 8.875 0.000700852 8/15/2028 5.5 -0.029696566
5/31/2016 3.25 5.89565E-05 5/15/2019 3.125 -0.00071161 11/15/2028 5.25 -0.053395
6/30/2016 1.5 4.69006E-05 8/15/2019 3.625 -0.000610302 2/15/2029 5.25 -0.051745207
6/30/2016 3.25 2.91233E-05 8/15/2019 8.125 0.000456455 8/15/2029 6.125 -0.00662204
7/31/2016 1.5 9.31429E-05 11/15/2019 3.375 -0.000618239 5/15/1930 6.25 -0.009641913
7/31/2016 3.25 5.75286E-05 2/15/2020 3.625 -0.000513322 2/15/1931 5.375 -0.048051288
8/15/2016 4.875 -0.000346379 2/15/2020 8.5 0.000578743 2/15/1936 4.5 -0.080135386
8/31/2016 1 0.000154021 5/15/2020 3.5 -0.000575234 2/15/2037 4.75 -0.070780159
8/31/2016 3 9.28654E-05 5/15/2020 8.75 0.000747554
9/30/2016 1 0.000102056 8/15/2020 2.625 -0.000800567
9/30/2016 3 6.1219E-05 8/15/2020 8.75 0.000566488
10/31/2016 1 0.000152136 11/15/2020 2.625 -0.00074674
10/31/2016 3.125 8.69449E-05 2/15/2021 3.625 -0.000483626
11/15/2016 4.625 -0.000374214 2/15/2021 7.875 0.00041529
11/15/2016 7.5 0.000449984 5/15/2021 3.125 -0.000630198
11/30/2016 0.875 0.000103366 5/15/2021 8.125 0.000560869
11/30/2016 2.75 6.49565E-05 8/15/2021 2.125 -0.000861479
12/31/2016 0.875 0.000154117 8/15/2021 8.125 0.000399199
12/31/2016 3.25 8.10196E-05 11/15/2021 2 -0.000832927
1/31/2017 3.125 0.000111769 11/15/2021 8 0.00039719
2/15/2017 4.625 -0.000330879
2/28/2017 3 -2.89006E-05
3/31/2017 3.25 5.71271E-05
4/30/2017 3.125 2.94046E-05
5/15/2017 4.5 0.005033507
5/15/2017 8.75 -0.00876684
5/31/2017 2.75 6.52693E-05
6/30/2017 2.5 3.46904E-05
For the 15-year futures, the conventional and exchanges methods select different deliverable
bonds which provides the short with the opportunity to select the one most beneficial to his
position. If the short delivers the bond selected by the conventional method, the 5.25 Nov 2028,
he has the potential to gain 7,346.15. However, given that the coupon rate is less than 6%, the
924

results in Table 6 indicate the shorts advantage may be slightly offset given the placement of the
delivery month in the payment cycle.
Table 12 displays the results of the conversion factor differences for the 5-year, 10-year and 15-
year contracts written on the Swiss government issued bonds denominated in Swiss francs, CHF.
The Eurex CONF method, selected the 7.25 May 2016 as the CTD against the 5-year contract as
did the conventional approach. The numbers indicate a relatively small gain to the short due to
the conversion differences. Table 7 results are so small that the contract delivery date is unlikely
to affect the findings in any meaningful way.
The CF methods select the same CTD for the 10-year CONF contract as well but the results
indicate a benefit to the long rather than the short. The potential for gain to the long is relatively
small at just over 70.00 and the delivery month adds little to this conclusion.
Table 12 results disclose two possibilities for the CTD: the 5.25 Nov 2028 indicated by the
conventional method and the 6.375 Aug 2027 indicated by the exchange method. Given these
two choices, the short will choose the 6.375 Aug 2027 because the potential gain is 5,339.50 as
opposed to losing 221.12.
Table 13 NYSE-Euronext JGB Conversion Factor Differences
5-Year
T-Bond
Futures
10-Year
T-Bond
Futures
15-Year
T-Bond
Futures

_Mar (2012) Jun (2012) Mar (2012)
Maturity Coupon CF Difference Maturity Coupon CF Difference Maturity Coupon 2 Difference
5/15/2016 5.125 -0.000284456 12/31/2018 1.375 -0.015570305 8/15/2027 6.375 7.67463E-05
5/15/2016 7.25 0.000370551 2/15/2019 2.75 -0.00081779 11/15/2027 6.125 -0.030320111
5/31/2016 1.75 8.9656E-05 2/15/2019 8.875 0.000697879 8/15/2028 5.5 -5.21099E-05
5/31/2016 3.25 5.82917E-05 5/15/2019 3.125 -0.000723285 11/15/2028 5.25 -0.026116827
6/30/2016 1.5 4.71106E-05 8/15/2019 3.625 -0.000583992 2/15/2029 5.25 -8.49452E-05
6/30/2016 3.25 2.87874E-05 8/15/2019 8.125 0.000496831 8/15/2029 6.125 3.66344E-05
7/31/2016 1.5 0.000156763 11/15/2019 3.375 -0.000639106 5/15/1930 6.25 -0.03092502
7/31/2016 3.25 9.52565E-05 2/15/2020 3.625 -0.018514513 2/15/1931 5.375 -5.8011E-05
8/15/2016 4.875 -0.002531112 2/15/2020 8.5 0.000569246 2/15/1936 4.5 -0.033483355
8/31/2016 1 -0.012808817 5/15/2020 3.5 -0.00046267 2/15/2037 4.75 -0.037981361
8/31/2016 3 -0.007685346 5/15/2020 8.75 0.000893913
9/30/2016 1 1.11022E-16 8/15/2020 2.625 -0.000777204
9/30/2016 3 -1.11022E-16 8/15/2020 8.75 0.000608673
10/31/2016 1 6.89886E-05 11/15/2020 2.625 -0.004620871
10/31/2016 3.125 4.01266E-05 2/15/2021 3.625 -0.004728684
11/15/2016 4.625 -0.000426132 2/15/2021 7.875 -0.005260206
11/15/2016 7.5 0.000434037 5/15/2021 3.125 -0.004484189
11/30/2016 0.875 3.51313E-05 5/15/2021 8.125 -0.004955686
11/30/2016 2.75 2.23764E-05 8/15/2021 2.125 -0.004460666
12/31/2016 0.875 0.000104751 8/15/2021 8.125 -0.005335575
12/31/2016 3.25 5.61988E-05 11/15/2021 2 -0.004296753
1/31/2017 3.125 9.67262E-05 11/15/2021 8 -0.004984574
2/15/2017 4.625 -0.000365606
2/28/2017 3 -3.99464E-05
3/31/2017 3.25 -0.016193606
4/30/2017 3.125 2.90719E-05
5/15/2017 4.5 -0.000450352
5/15/2017 8.75 0.000783862
5/31/2017 2.75 6.48209E-05
6/30/2017 2.5 0.006134311
925

Finally, Table 13 discloses the results for the JGB futures contracts. As with most of the other
CF methodologies, including the conventional approach, the NYSE-Euronext approach selects
the 7.25 May 2016 as the CTD against the 5-year contract. The potential advantage favors the
long but it is very small. Even so, given the gain to the long indicated in Table 8, the long has the
potential to benefit from the delivery process.
For the 10-year futures contract the conventional approach identifies the 1.375 Dec 2018 as the
CTD while the exchange approach selects the 8.875 Feb 2019. Given these two choices the short
will select the Dec 2018 because he has a potential advantage of 15,570.31 as opposed to losing
697.88.
16
A brief review of Table 8 indicates that the timing of the delivery and the coupon rate
are such that the short may enhance his gain.
For the 15-year JGB contract both the conventional and NYSE-Euronext approach select the
6.375 Aug 2027 as the CTD. This selection will not likely result in any benefit to the long or
short since the difference is so small. Reviewing Table 8 only confirms this conclusion since the
timing of the delivery and the coupon do not contribute anything to the final result.
There is no discernible pattern to the tendency of the conventional method selecting a different
deliverable bond than the exchanges method. However, longer maturity contracts show greater
divergence than the shorter maturity. That is, four of the five methods select the same CTD
against the 5-year contract while only three of the five do so for the 10-year contract. When the
contract is 15-years just two of the five select the same CTD. Even so, when faced with the
decision to reverse or deliver a bond futures contract, both the long and short can consider the
peculiarities of the conversion technique use by the exchange in order to exploit any benefits that
may result.
VI. Conclusion
Conversion factors are employed by exchanges around the globe in order to price the delivery of
bonds which do not meet the requirements of the hypothetical bond described in the contract
specifications. The practice of exchanges to allow cash bonds to be delivered which do not have
a coupon rate equal to the contract rate necessitates a method to equalize the price the delivered
bond relative to the contracted bond. The approaches used by all three exchanges employ
different versions of estimating the present value of a 100% par bond. However, the
implementation of the estimation procedures differs widely among the exchanges and result in
varying degrees of pricing precision relative to the conventional approach.
When the exchange-employed CF selects the same CTD as does the conventional approach,
traders can use the information as an input into determining the optimality of holding the
contract until delivery or reversing. When the CF
c
>CF
ex
, then the futures will likely be
underpriced relative to the deliverable bond which can benefit the long if he chooses to hold until
delivery. The extent of the difference depends on market yields at the time of delivery. In
addition, while CF
c
-CF
ex
can be used as a signal of the possibility of gain, the actual gain will be
a function of market yields since the cash market is not fixed at 6% as implied by the
calculations. Naturally, the reverse holds as well so that when the conventional approach is less
than the exchange approach, the short may benefit from holding until delivery rather than reverse
his position.
In addition to CF differences, this paper provides insight into the benefits and costs of selecting
futures which delivers at various points in the coupon payment cycle. This information can be
most helpful if used prior to establishing a position.

16
Yen amounts are based on a 1,000,000 par value.
926

The results presented in this paper will change when a different day is evaluated, when a
dissimilar deliverable bond set is evaluated or when yields change. Even so, the basic ideas
remain the same. That is, exchanges employ approximations to determine the CFs used in pricing
futures and these estimates provide information to traders when compared to the conventional
approach. Such information can be useful in determining: (i) the optimality of reversing or
delivering, and, (ii) the identity of the actual CTD.
The summary below highlights the findings presented in Tables 4-8; CR stands for coupon rate:

UST-bond BTP-bond Euro-bond CFH-bond JGB

Past Prior Past Prior Past Prior Past Prior Past Prior
5-Yr.
High CR - + + -/+ - - + + + +
Low CR + - + + + + - - + +

10-Yr.
High CR - + + -/+ - - + + + +
Low CR + - + + + + - - +/- +/-

15-Yr.
High CR - + + + - - + + - -
Low CR + - + + + + - - + +

The terms past and prior refer to the delivery date relative to the coupon midpoint. The positive
sign indicates the difference between the conventional and exchange conversion factors and
signifies a potential advantage to the long. The negative sign denotes a potential advantage to the
short. The summary confirms that no two exchange methods produce a pattern of benefits that is
the same. Instead, each CF method produces a unique set of factors to be considered.
The evidence provided for the cheapest-to-deliver bond is summarized below for each contract
and each exchange:
CTD Bonds
5-Yr 10-Yr 15-Yr
CME UST 7.25 May 16 8.875 Feb 19 6.375 Aug 27
CF Selection Model Conventional Both Both
Eurex BTP 7.25 May 16 1.375 Dec 18 6.375 Aug 27
CF Selection Model Both Conventional Conventional
Eurex Euro 7.25 May 16 8.875 Feb 19 5.25 Nov 28
CF Selection Model Both Both Conventional
Eurex CONF 7.25 May 16 8.875 Feb 19 5.25 Nov 28
CF Selection Model Both Both Conventional
NYSE-Euronext JGB 7.25 May 16 1.375 Dec 18 6.375 Aug 27
CF Selection Model Both Conventional Both
As the term of the contract lengthens, the advantage to the short becomes more frequently
identified by the conventional approach. In no case does the exchange approach alone identify
the bond of maximum benefit to the short. When both the conventional and exchange approaches
927

identify the same CTD bond, both the long and the short can benefit by ascertaining the
difference and using it to decide whether it is optimal to reverse or deliver.
Alternatively, when the conventional method selects a different CTD than does the exchange
method, the short can use this knowledge to determine whether a bond other than the one
identified by the exchange method is more to his advantage.

References
Arak, M., L. Goodman, and S. Ross. 1986. The Cheapest to Deliver Bond on the Treasury
Bond Futures Contract. Advances in Futures and Options Research, vol. 1:4974.
Benninga, S., & Wiener, Z. (1999). An investigation of cheapest to deliver on treasury bond
futures contracts. Journal of Computational Finance, 2, 3555.
Hegde, Shantaram P. 1988. An Empirical Analysis of Implicit Delivery Options in the T-Bond
Futures Contract. Journal of Banking and Finance, vol. 12, no.3 (September):469-492.
Hegde, Shantaram P. 1990. An Ex-Post Valuation of the Quality Option Implicit in the T-Bond
Futures Contract. Journal of Banking and Finance, vol. 14, no. 4 (October):741-760.
Helmer, Michael L. 1990. The Quality Delivery Option in T-Bond Futures Contract. Journal
of Finance, vol. XLV, no.5 (December):920-924.
Kane, Alex and Alan J. Marcus. 1984. Conversion Factor Risk and Hedging the T-Bond Futures
Market. Journal of Futures Markets, vol. 4, no. 1 (Spring):55-64.
Klemkosky R.C. and D.J. Lasser. 1985. An Efficient Analysis of the T-Bond Futures Market.
Journal of Futures Markets, 5:607-20.
Livingston, Miles. 1987. The Effective Coupon Level on Treasury Bond Futures Delivery.
Journal of Futures Markets, vol. 7, no. 3 (June):303-309.
Meisner, J.F. and J.W. Labuszewski. 1984. Treasury Bond Futures Delivery Bias. Journal of
Futures Markets, Vol. 7, no. 3 (Winter):569572.
Resnick, B.G., and E. Hennigan. 1983. The Relationship Between Futures and Cash Prices for
U.S. Treasury Bonds. Review of Research in Futures Markets, no. 2:282-99
Ritchken, P. and L. Sankarasubramanian. 1992. Pricing the Quality Option in T-Bond Futures.
Mathematical Finance. vol. 2:197-214.

928

Appendix
Day-Count Conventions
There are three different actual/actual day-count practices in use since there is no central
authority defining day-count conventions. The need for day-count conventions is a direct
consequence of interest-earning investments but even within a given sovereignty different
day-count conventions are used for short-term versus long-term securities and government
issues versus private issues. In the U.S., for example, the day-count convention for corporate
bond and municipal issues is 30/360 while for Treasury issues it is actual/365. Different
conventions were developed to address often conflicting requirements, including ease of
calculation, constancy of time period (day, month, or year) and the needs of the accounting
department. This development occurred long before the advent of computers.
Inspired by the approaching introduction of the euro currency, on July 16
th
, 1997 the
International Swaps and Derivatives Association, Inc. (ISDA), along with Cedel and
Euroclear, published a joint statement on market day-count conventions for the euro. The joint
statement was subsequently supported by both the European Commission and the European
Central Bank.
17

The joint statement identified and clarified three different interpretations of the actual/actual day-
count method as it applied to accrued interest. The terminology and definitions have
subsequently become widely used. The three different interpretations of actual/actual are:
18

1. AFB Method: the denominator is 365 in non-leap periods and 366 in leap-periods.
The joint statement suggested that the AFB Method be known as Actual/Actual
(Euro).
2. ISDA Method: the denominator varies depending on whether a portion of the
relevant calculation period falls within a leap year. In a leap year, the portion of the
accrued interest falling within the leap year is determined, using actual days in the
numerator and 366 in the denominator. This fraction is added to the fraction of
accrued interest falling within the non-leap period.
3. ISMA Method: the denominator is the actual number of days in the coupon period
times the number of coupon periods in the year (either one or two). The joint
statement suggested that the ISMA Method be known as the Actual/Actual (Bond)
The bonds studied in this paper use the following actual/actual conventions:

U.S. T-Bond BTP Bond Euro-Den. Bonds CONF Bonds JGB
Coupon Frequency: Semi-Annual Semi-Annual Annual Annual Semi-Annual
Day-Count Convention: AFB ISDA AFB ISMA actual/365

17
Then known as the European Monetary Institute.

18
http://www.isda.org/c_and_a/pdf/mktc1198.pdf

929

On Collecting Social Security Benefits after Age 66

Charles J. Higgins, Ph.D.
Dept. Finance/CIS
Loyola Marymount University
1 LMU Drive, Los Angeles, CA 90045
chiggins@lmu.edu

No liability inferred nor implied
930

For those with expectations of a normal longevity, electing to collect Social Security benefits at
or after age 66 is currently recognized as appropriate. However, the question of postponing
starting payments after the age of 66 is more problematic. This article will show that discounting
the various payment elections becomes moot after a 6 percent discount rate.
For those born after 1943, U.S. Social Security benefits may begin at age 62 through age 70.
Payments which start at age 62 are 75 percent of payments starting at age 66, decreasing 6.25
percent per year. Payments rise at 8 percent each year for payments starting after age 66 up to
age 70. Payments starting at age 70 are 132 percent of payments which start at age 66. Social
Security benefits may be taxable as ordinary income at rates ranging from zero to 85 percent (see
IRS publication 915). They may also affect IRA contributions for both deductible Traditional
and Roth IRAs in differing ways (see IRS publication 590).
One could consider discounting the future Social Security benefit payments to present values.
With inflation rates and current market returns low, discounting for inflation or interest rates
would be small. But one may increase discount rates to provide risk adjustment for
contemplated consumption utility preferences, for the ability to spend and enjoy, for possible
default scenarios, and/or for unexpected increases in interest rates or inflation.
Social Security payments for years starting at age 62 through 70 are respectively: 75, 81.25,
87.5, 93.75, 100, 108, 116, 124, and 132 percent of age 66 values. The cumulative total present
values were determined by P
t
/(1+k)
t
where P
t
is the payment at year t and k is the discount rate
and were figured for remaining years to the age of 100. Particular computations are presented in
Tables 3 and 4 for 1 percent and 6 percent discount rates for benefits starting at age 66.
For a zero percent discount rate, starting benefits at age 66 were desirable for those with an
expected remaining longevity of less than 13 years; starting benefits at age 70 were desirable for
those with an expected remaining longevity greater than 20 years; for longevities half way in
between, starting at age 68 was appropriate.
Increasing the discount rates raised the various longevities with the greatest present values per
the following table:
Table 1, Longevity with Greatest Present Value by Discount Rate
0 1 2 3 4 5 6 7
Starting Age
66 less than 79 80 81 82 84 87 91 100
68 82.5 84 85 87 89 93 100 -
70 more than 86 87 89 92 95 - - -

What is notable here is that any scenario that discounts Social Security benefits greater than 6
percent generally results in a choice to start collecting at the age of 66 regardless of cause. That
is to say, the decision to elect collecting Social Security benefits at the age of 66 is generally
correct and becomes more so as one increases discount rates regardless of reason. See U.S.
Social Security Benefits Planner and various other references listed at the end of this article. See
Table 7 for life expectancies.

Another way to view the results is to present the ratio of the present value of the accumulated
benefits for those electing a 70 year benefit payment divided by the present value of the
accumulated benefits for those electing a 66 year benefit payment for various longevities, or:

931

Table 2, Ratio Present Value Benefits Starting at 70 Divided by at 66
Longevity Age 75 80 85 90 95 100
Discount
0% 73% 94% 104% 110% 114% 116%
1 72 92 102 108 111 114
2 70 90 100 105 109 111
3 69 89 98 103 106 109
4 68 87 95 100 104 106
5 66 85 93 98 101 103
6 65 83 91 95 98 100

Not considered were reinvestment scenarios; see referenced articles at end of this article for such
computations. For those who contemplate collecting Social Security benefits prior to the age of
66, see Tables 5 and 6. Here are graphics for each payment election and various discount rates:
Not considered herein were cost of living adjustments (COLA). Note that COLA adjustments
the last two years were zero. For 2012, a 3.6 percent increase in benefits was announced.
Moreover, if one were to consider possible increases in future Social Security benefit payments,
one could make the adjustment by using the rate of (k-g)/(1+g) where k is the discount rate and g
is the growth rate in future benefit payments. Because 1+g is de minimis here, one can use k-g
as COLA adjusted discount rate. An example would be a discount rate of 5 percent and a COLA
of 2 percent; the effective net growth adjusted discount rate would be about 3 percent. Such
adjustments do not substantially change the analysis herein.
A succinct conclusion is that if one does live beyond ones expected longevity, the lost
opportunity of higher payouts by delaying the start of benefits is not large. As discount rates
rise, the differences between the age 66 to 70 election choices are not great amounting to some 6
percent above or below regardless of longevity. This present value analysis goes toward a bird
in the hand is worth two in bush and as such prefers a 66
th
year benefit as the preference for
present value considerations increases. Graphic 2 displays the present values for benefits starting
at age 62 in two year increments then at age 66 in one year increments. Notice that benefits
starting at age 66 generally are preferred when not dominated by either longevity or present
value considerations. Given most people face uncertain longevities, the likelihood of greater
present values bears a strong consideration.

932

Graphic 1, Benefits at 62, 64, 66, 68, 70 at Various Discount Rates

933

Graphic 2, Present Value of Benefits at Various Discount Rates

934

Table 3, Total Social Security present value of payments at 1 %

Pay Pay Pay Pay Pay Total Total Total Total Total Yr+ %70/66
100 0 0 0 0 99.01 0.00 0.00 0.00 0.00 1 0
100 108 0 0 0 197.04 105.87 0.00 0.00 0.00 2 0
100 108 116 0 0 294.10 210.70 112.59 0.00 0.00 3 0
100 108 116 124 0 390.20 314.48 224.06 119.16 0.00 4 0
100 108 116 124 132 485.34 417.24 334.43 237.14 125.59 5 25.88%
100 108 116 124 132 579.55 518.98 443.71 353.96 249.94 6 43.13%
100 108 116 124 132 672.82 619.71 551.90 469.61 373.06 7 55.45%
100 108 116 124 132 765.17 719.45 659.03 584.13 494.96 8 64.69%
100 108 116 124 132 856.60 818.20 765.09 697.50 615.65 9 71.87%
100 108 116 124 132 947.13 915.97 870.11 809.76 735.15 10 77.62%
100 108 116 124 132 1,036.76 1,012.77 974.08 920.90 853.47 11 82.32%
100 108 116 124 132 1,125.51 1,108.62 1,077.02 1,030.95 970.61 12 86.24%
100 108 116 124 132 1,213.37 1,203.51 1,178.95 1,139.90 1,086.59 13 89.55%
100 108 116 124 132 1,300.37 1,297.47 1,279.86 1,247.78 1,201.43 14 92.39%
100 108 116 124 132 1,386.51 1,390.49 1,379.78 1,354.58 1,315.13 15 94.85%
100 108 116 124 132 1,471.79 1,482.60 1,478.71 1,460.33 1,427.70 16 97.00%
100 108 116 124 132 1,556.23 1,573.79 1,576.66 1,565.04 1,539.16 17 98.90%
100 108 116 124 132 1,639.83 1,664.08 1,673.63 1,668.70 1,649.51 18 100.59%
100 108 116 124 132 1,722.60 1,753.48 1,769.65 1,771.34 1,758.77 19 102.10%
100 108 116 124 132 1,804.56 1,841.99 1,864.72 1,872.97 1,866.95 20 103.46%
100 108 116 124 132 1,885.70 1,929.62 1,958.84 1,973.58 1,974.06 21 104.69%
100 108 116 124 132 1,966.04 2,016.39 2,052.04 2,073.20 2,080.11 22 105.80%
100 108 116 124 132 2,045.58 2,102.30 2,144.31 2,171.84 2,185.11 23 106.82%
100 108 116 124 132 2,124.34 2,187.36 2,235.67 2,269.50 2,289.07 24 107.75%
100 108 116 124 132 2,202.32 2,271.57 2,326.12 2,366.19 2,392.00 25 108.61%
100 108 116 124 132 2,279.52 2,354.95 2,415.68 2,461.92 2,493.91 26 109.40%
100 108 116 124 132 2,355.96 2,437.51 2,504.35 2,556.71 2,594.81 27 110.14%
100 108 116 124 132 2,431.64 2,519.25 2,592.14 2,650.56 2,694.71 28 110.82%
100 108 116 124 132 2,506.58 2,600.17 2,679.07 2,743.48 2,793.62 29 111.45%
100 108 116 124 132 2,580.77 2,680.30 2,765.13 2,835.47 2,891.56 30 112.04%
100 108 116 124 132 2,654.23 2,759.64 2,850.34 2,926.56 2,988.52 31 112.59%
100 108 116 124 132 2,726.96 2,838.18 2,934.71 3,016.75 3,084.53 32 113.11%
100 108 116 124 132 2,798.97 2,915.96 3,018.24 3,106.04 3,179.58 33 113.60%
100 108 116 124 132 2,870.27 2,992.96 3,100.94 3,194.45 3,273.69 34 114.06%

935

Table 4, Total Social Security present value of payments at 6 %

100 0 0 0 0 94.34 0.00 0.00 0.00 0.00 1 0
100 108 0 0 0 183.34 96.12 0.00 0.00 0.00 2 0
100 108 116 0 0 267.30 186.80 97.40 0.00 0.00 3 0
100 108 116 124 0 346.51 272.34 189.28 98.22 0.00 4 0
100 108 116 124 132 421.24 353.05 275.96 190.88 98.64 5 23.42%
100 108 116 124 132 491.73 429.18 357.74 278.29 191.69 6 38.98%
100 108 116 124 132 558.24 501.01 434.88 360.76 279.48 7 50.06%
100 108 116 124 132 620.98 568.77 507.66 438.56 362.30 8 58.34%
100 108 116 124 132 680.17 632.70 576.32 511.96 440.43 9 64.75%
100 108 116 124 132 736.01 693.00 641.10 581.20 514.14 10 69.85%
100 108 116 124 132 788.69 749.90 702.20 646.52 583.67 11 74.01%
100 108 116 124 132 838.38 803.57 759.85 708.14 649.27 12 77.44%
100 108 116 124 132 885.27 854.20 814.24 766.28 711.16 13 80.33%
100 108 116 124 132 929.50 901.97 865.54 821.12 769.54 14 82.79%
100 108 116 124 132 971.22 947.04 913.95 872.87 824.62 15 84.91%
100 108 116 124 132 1,010.59 989.55 959.61 921.68 876.58 16 86.74%
100 108 116 124 132 1,047.73 1,029.66 1,002.69 967.73 925.60 17 88.34%
100 108 116 124 132 1,082.76 1,067.49 1,043.33 1,011.17 971.85 18 89.76%
100 108 116 124 132 1,115.81 1,103.19 1,081.67 1,052.15 1,015.48 19 91.01%
100 108 116 124 132 1,146.99 1,136.86 1,117.84 1,090.82 1,056.64 20 92.12%
100 108 116 124 132 1,176.41 1,168.63 1,151.96 1,127.29 1,095.46 21 93.12%
100 108 116 124 132 1,204.16 1,198.60 1,184.15 1,161.70 1,132.09 22 94.02%
100 108 116 124 132 1,230.34 1,226.88 1,214.52 1,194.17 1,166.65 23 94.82%
100 108 116 124 132 1,255.04 1,253.55 1,243.17 1,224.79 1,199.25 24 95.56%
100 108 116 124 132 1,278.34 1,278.72 1,270.20 1,253.68 1,230.01 25 96.22%
100 108 116 124 132 1,300.32 1,302.46 1,295.69 1,280.94 1,259.02 26 96.82%
100 108 116 124 132 1,321.05 1,324.85 1,319.75 1,306.65 1,286.40 27 97.38%
100 108 116 124 132 1,340.62 1,345.98 1,342.44 1,330.91 1,312.22 28 97.88%
100 108 116 124 132 1,359.07 1,365.91 1,363.85 1,353.80 1,336.58 29 98.35%
100 108 116 124 132 1,376.48 1,384.71 1,384.05 1,375.39 1,359.56 30 98.77%
100 108 116 124 132 1,392.91 1,402.45 1,403.10 1,395.75 1,381.25 31 99.16%
100 108 116 124 132 1,408.40 1,419.19 1,421.08 1,414.97 1,401.70 32 99.52%
100 108 116 124 132 1,423.02 1,434.98 1,438.03 1,433.09 1,421.00 33 99.86%
100 108 116 124 132 1,436.81 1,449.87 1,454.03 1,450.20 1,439.20 34 99.99%

936

Table 5, Total Social Security present values starting at age 62 at 1 %

75 0 0 0 0 $75.00 $0.00 $0.00 $0.00 $0.00 0 0
75 0 0 0 0 $149.26 $0.00 $0.00 $0.00 $0.00 1 0
75 87.5 0 0 0 $222.78 $85.78 $0.00 $0.00 $0.00 2 0
75 87.5 0 0 0 $295.57 $170.70 $0.00 $0.00 $0.00 3 0
75 87.5 100 0 0 $367.65 $254.79 $96.10 $0.00 $0.00 4 0.00%
75 87.5 100 0 0 $439.01 $338.04 $191.24 $0.00 $0.00 5 0.00%
75 87.5 100 116 0 $509.66 $420.47 $285.45 $109.28 $0.00 6 0.00%
75 87.5 100 116 0 $579.61 $502.08 $378.72 $217.47 $0.00 7 0.00%
75 87.5 100 116 132 $648.88 $582.89 $471.07 $324.60 $121.90 8 18.79%
75 87.5 100 116 132 $717.45 $662.89 $562.50 $430.66 $242.59 9 33.81%
75 87.5 100 116 132 $785.35 $742.11 $653.03 $535.67 $362.09 10 46.11%
75 87.5 100 116 132 $852.57 $820.53 $742.66 $639.65 $480.41 11 56.35%
75 87.5 100 116 132 $919.13 $898.19 $831.41 $742.59 $597.55 12 65.01%
75 87.5 100 116 132 $985.03 $975.07 $919.28 $844.52 $713.53 13 72.44%
75 87.5 100 116 132 $1,050.28 $1,051.19 $1,006.27 $945.43 $828.37 14 78.87%
75 87.5 100 116 132 $1,114.88 $1,126.56 $1,092.41 $1,045.35 $942.07 15 84.50%
75 87.5 100 116 132 $1,178.84 $1,201.18 $1,177.69 $1,144.28 $1,054.64 16 89.46%
75 87.5 100 116 132 $1,242.17 $1,275.06 $1,262.13 $1,242.22 $1,166.10 17 93.88%
75 87.5 100 116 132 $1,304.87 $1,348.21 $1,345.73 $1,339.20 $1,276.45 18 97.82%
75 87.5 100 116 132 $1,366.95 $1,420.64 $1,428.50 $1,435.22 $1,385.71 19 101.37%
75 87.5 100 116 132 $1,428.42 $1,492.35 $1,510.46 $1,530.29 $1,493.89 20 104.58%
75 87.5 100 116 132 $1,489.27 $1,563.35 $1,591.60 $1,624.41 $1,601.00 21 107.50%
75 87.5 100 116 132 $1,549.53 $1,633.65 $1,671.94 $1,717.61 $1,707.05 22 110.17%
75 87.5 100 116 132 $1,609.19 $1,703.25 $1,751.48 $1,809.88 $1,812.05 23 112.61%
75 87.5 100 116 132 $1,668.25 $1,772.16 $1,830.24 $1,901.23 $1,916.01 24 114.85%
75 87.5 100 116 132 $1,726.74 $1,840.39 $1,908.22 $1,991.69 $2,018.93 25 116.92%
75 87.5 100 116 132 $1,784.64 $1,907.95 $1,985.42 $2,081.25 $2,120.85 26 118.84%
75 87.5 100 116 132 $1,841.97 $1,974.83 $2,061.86 $2,169.92 $2,221.75 27 120.62%
75 87.5 100 116 132 $1,898.73 $2,041.06 $2,137.55 $2,257.71 $2,321.65 28 122.27%
75 87.5 100 116 132 $1,954.93 $2,106.62 $2,212.48 $2,344.63 $2,420.56 29 123.82%
75 87.5 100 116 132 $2,010.58 $2,171.54 $2,286.67 $2,430.70 $2,518.50 30 125.26%
75 87.5 100 116 132 $2,065.67 $2,235.82 $2,360.13 $2,515.91 $2,615.46 31 126.62%
75 87.5 100 116 132 $2,120.22 $2,299.46 $2,432.86 $2,600.27 $2,711.46 32 127.89%
75 87.5 100 116 132 $2,174.23 $2,362.46 $2,504.87 $2,683.81 $2,806.52 33 129.08%
75 87.5 100 116 132 $2,227.70 $2,424.85 $2,576.17 $2,766.51 $2,900.63 34 130.21%
75 87.5 100 116 132 $2,280.64 $2,486.62 $2,646.76 $2,848.40 $2,993.81 35 131.27%
75 87.5 100 116 132 $2,333.06 $2,547.77 $2,716.65 $2,929.47 $3,086.07 36 132.28%
75 87.5 100 116 132 $2,384.96 $2,608.32 $2,785.85 $3,009.75 $3,177.41 37 133.23%
75 87.5 100 116 132 $2,436.35 $2,668.27 $2,854.37 $3,089.22 $3,267.85 38 134.13%
62-70 66-70

937

Table 6, Total Social Security present values starting at age 62 at 6 %

75 0 0 0 0 $75.00 $0.00 $0.00 $0.00 $0.00 0 0
75 0 0 0 0 $145.75 $0.00 $0.00 $0.00 $0.00 1 0
75 87.5 0 0 0 $212.50 $77.87 $0.00 $0.00 $0.00 2 0
75 87.5 0 0 0 $275.48 $151.34 $0.00 $0.00 $0.00 3 0
75 87.5 100 0 0 $334.88 $220.65 $79.21 $0.00 $0.00 4 0.00%
75 87.5 100 0 0 $390.93 $286.03 $153.94 $0.00 $0.00 5 0.00%
75 87.5 100 116 0 $443.80 $347.72 $224.43 $81.78 $0.00 6 0.00%
75 87.5 100 116 0 $493.68 $405.91 $290.94 $158.92 $0.00 7 0.00%
75 87.5 100 116 132 $540.73 $460.81 $353.68 $231.70 $82.82 8 15.32%
75 87.5 100 116 132 $585.13 $512.60 $412.87 $300.36 $160.95 9 27.51%
75 87.5 100 116 132 $627.01 $561.46 $468.71 $365.14 $234.66 10 37.42%
75 87.5 100 116 132 $666.52 $607.55 $521.39 $426.24 $304.19 11 45.64%
75 87.5 100 116 132 $703.79 $651.04 $571.08 $483.89 $369.79 12 52.54%
75 87.5 100 116 132 $738.95 $692.06 $617.97 $538.28 $431.68 13 58.42%
75 87.5 100 116 132 $772.12 $730.76 $662.20 $589.58 $490.06 14 63.47%
75 87.5 100 116 132 $803.42 $767.27 $703.92 $637.99 $545.14 15 67.85%
75 87.5 100 116 132 $832.94 $801.72 $743.29 $683.65 $597.10 16 71.69%
75 87.5 100 116 132 $860.79 $834.21 $780.42 $726.73 $646.12 17 75.06%
75 87.5 100 116 132 $887.07 $864.87 $815.46 $767.37 $692.37 18 78.05%
75 87.5 100 116 132 $911.86 $893.79 $848.51 $805.71 $736.00 19 80.71%
75 87.5 100 116 132 $935.24 $921.07 $879.69 $841.88 $777.16 20 83.10%
75 87.5 100 116 132 $957.31 $946.81 $909.11 $876.00 $815.98 21 85.24%
75 87.5 100 116 132 $978.12 $971.09 $936.86 $908.19 $852.61 22 87.17%
75 87.5 100 116 132 $997.75 $994.00 $963.04 $938.56 $887.17 23 88.92%
75 87.5 100 116 132 $1,016.28 $1,015.61 $987.73 $967.21 $919.77 24 90.50%
75 87.5 100 116 132 $1,033.75 $1,036.00 $1,011.03 $994.24 $950.53 25 91.95%
75 87.5 100 116 132 $1,050.24 $1,055.23 $1,033.02 $1,019.73 $979.54 26 93.27%
75 87.5 100 116 132 $1,065.79 $1,073.37 $1,053.75 $1,043.79 $1,006.92 27 94.48%
75 87.5 100 116 132 $1,080.46 $1,090.49 $1,073.32 $1,066.48 $1,032.74 28 95.58%
75 87.5 100 116 132 $1,094.30 $1,106.64 $1,091.77 $1,087.89 $1,057.10 29 96.60%
75 87.5 100 116 132 $1,107.36 $1,121.88 $1,109.18 $1,108.09 $1,080.08 30 97.54%
75 87.5 100 116 132 $1,119.68 $1,136.25 $1,125.61 $1,127.14 $1,101.77 31 98.40%
75 87.5 100 116 132 $1,131.30 $1,149.81 $1,141.10 $1,145.11 $1,122.22 32 99.20%
75 87.5 100 116 132 $1,142.27 $1,162.60 $1,155.72 $1,162.07 $1,141.52 33 99.93%
75 87.5 100 116 132 $1,152.61 $1,174.67 $1,169.51 $1,178.07 $1,159.72 34 100.62%
75 87.5 100 116 132 $1,162.37 $1,186.05 $1,182.52 $1,193.16 $1,176.89 35 101.25%
75 87.5 100 116 132 $1,171.57 $1,196.79 $1,194.80 $1,207.40 $1,193.10 36 101.84%
75 87.5 100 116 132 $1,180.26 $1,206.92 $1,206.38 $1,220.83 $1,208.38 37 102.38%
75 87.5 100 116 132 $1,188.45 $1,216.48 $1,217.30 $1,233.50 $1,222.80 38 102.89%
62-70 66-70

938

Table 7, Life expectancies per Social Security Administration

Male Remain/ Female Remain/
Age Prob. 100,000 Expect Prob. 100,000 Expect

62 0.013289 83,217 19.40 0.008322 89,895 22.31
63 0.014326 82,111 18.66 0.009046 89,147 21.49
64 0.015453 80,935 17.92 0.009822 88,340 20.69
65 0.016723 79,684 17.19 0.010698 87,473 19.89
66 0.018154 78,351 16.48 0.011702 86,537 19.10
67 0.019732 76,929 15.77 0.012832 85,524 18.32
68 0.021468 75,411 15.08 0.014103 84,427 17.55
69 0.023387 73,792 14.40 0.015526 83,236 16.79
70 0.025579 72,066 13.73 0.017163 81,944 16.05
71 0.028032 70,223 13.08 0.018987 80,537 15.32
72 0.030665 68,254 12.44 0.020922 79,008 14.61
73 0.033467 66,161 11.82 0.022951 77,355 13.91
74 0.036519 63,947 11.21 0.025147 75,580 13.22
75 0.040010 61,612 10.62 0.027709 73,679 12.55
76 0.043987 59,147 10.04 0.030659 71,638 11.90
77 0.048359 56,545 9.48 0.033861 69,441 11.26
78 0.053140 53,811 8.94 0.037311 67,090 10.63
79 0.058434 50,951 8.41 0.041132 64,587 10.03
80 0.064457 47,974 7.90 0.045561 61,930 9.43
81 0.071259 44,882 7.41 0.050698 59,109 8.86
82 0.078741 41,683 6.94 0.056486 56,112 8.31
83 0.086923 38,401 6.49 0.062971 52,942 7.77
84 0.095935 35,063 6.06 0.070259 49,608 7.26
85 0.105937 31,699 5.65 0.078471 46,123 6.77
86 0.117063 28,341 5.26 0.087713 42,504 6.31
87 0.129407 25,024 4.89 0.098064 38,776 5.87
88 0.143015 21,785 4.55 0.109578 34,973 5.45
89 0.157889 18,670 4.22 0.122283 31,141 5.06
90 0.174013 15,722 3.92 0.136190 27,333 4.69
91 0.191354 12,986 3.64 0.151300 23,610 4.36
92 0.209867 10,501 3.38 0.167602 20,038 4.04
93 0.229502 8,297 3.15 0.185078 16,680 3.76
94 0.250198 6,393 2.93 0.203700 13,593 3.50
95 0.270750 4,794 2.75 0.222541 10,824 3.26
96 0.290814 3,496 2.58 0.241317 8,415 3.05
97 0.310029 2,479 2.44 0.259716 6,384 2.87
98 0.328021 1,711 2.30 0.277409 4,726 2.70
99 0.344422 1,149 2.19 0.294054 3,415 2.54
939

References
Bernard, Tara Siegel Collect Now, or Later? Timing Your Social Security Benefits
http://www.nytimes.com/2009/07/11/your-money/11retire.html
Hebelerel, Henry K. Start Social Security at 62, 66, or 70?
http://analyzenow.com/Articles/Social%20Security/Social%20Security%20Articles/Start%20Soc
ial%20Security%20at%2062,%2066%20or%2070%205-9-02.pdf
Investopedia, Maximize Your Social Security Benefits
http://www.investopedia.com/articles/retirement/08/cost-of-living.asp#axzz1qN1D2Fa3
Reichstein, William and William Meyer Social Security: When to Start Benefits and How to
Minimize Longevity Risk
http://www.fpanet.org/journal/CurrentIssue/TableofContents/SocialSecurityWhentoStartBenefits
Social Security Benefits Planner http://www.ssa.gov/planners/taxes.html
Social Security Administration Longevities http://www.ssa.gov/oact/STATS/table4c6.html
Spiegelman, Rande When Should You Take Social Security?
http://www.schwab.com/public/schwab/resource_center/expert_insight/retirement_strategies/pla
nning/when_should_you_take_social_security.html

940

Systematic Risk Estimation: OLS v. State-Space Methods

Joseph Callaghan
Professor of Accounting
Department of Accounting and Finance
School of Business Administration
Oakland University, Rochester, MI 48309, USA
callagha@oakland.edu

Liang Fu*
Assistant Professor of Accounting
Department of Accounting and Finance
School of Business Administration
Oakland University, Rochester, MI 48309, USA
liangfu@oakland.edu

Jing Liu
jliu@oakland.edu

*Corresponding author
941

Abstract
An important concept in risk management is systematic risk (). This paper focuses on the
dynamic feature of and introduces a state space model and Kalman filter framework for
estimation. This framework is used to model the structure of time-varying to follow any
stochastic times-series process. We investigate these properties of by comparing the forecast
errors of under conventional estimation with those under Kalman filtering estimation. Our
results suggest that Dow Jones Industry Average index companies, improvements in the
estimation of systematic risk are not obtained by a more realistically dynamic modeling of
systematic risk.
Key Words: Time-varying systematic risk , Kalman filter estimation, forecast errors of

I. Introduction and Literature Review
In finance, capital asset pricing model (CAPM hereafter) is used to determine the relationship
between an asset's risk and its required rate of return. The model identifies two types of risks that
are associated with an asset: one is systematic risk, which is also called market or non-
diversifiable risk; the other one is idiosyncratic risk, which is diversifiable (Sharpe 1964 and
Lintner 1965). The CAPM model is mathematically represented as
ER
ER

where ER
is the expected return on the capital asset; R
is the risk-free rate of interest often

surrogated by the interest yield on US government bonds; beta
is the sensitivity of the

expected excess asset returns to the expected excess market returns; and ER
is the expected
return of the market. The key parameter in the above CAPM model is coefficient. It measures
the part of the asset's statistical variance that cannot be removed by the diversification provided
by the portfolio of many risky assets, in other words, is a measure of an asset's non-
diversifiable risk, systematic risk, or market risk. Literature in finance and accounting has then
been using the CAPM model extensively to test the relationship between risk and return.
However, conventional adoption of CAPM model assumes the stationary of the risk factor .
This is a rather nave assumption. Ohlson and Rosenberg (1984) find two kinds of stochastic
variation in : there is an attendency for s to converge rather slowly towards the stationary
mean, and in that sense it is clear that the model of s requires a "memory"; that is modeled by a
stationary first-order, autoregressive process. At the same time, in each period, there is a purely
random (serially independent) perturbation in . Bos and Newbold (1984) use monthly data of
stocks trade on the New York Exchange from the period of 10 years covering from January 1970
to December 1979 and find that the assumption that is fixed parameter does not hold; however,
there is lack of strong evidence documenting that follows first-order autocorrelation. In
explaining the lack of autocorrelation in , Bos and Newbold (1984) indicate that a longer series
of observations would have led to more powerful tests and it is also likely that the autoregressive
behavior would be more manifest in data observed at shorter time intervals rather than monthly
stock data in their study. The post-1984 finance literature has been extensively focusing on the
conditional version of the CAPM, where the model allows the ratio of expected market risk
premium to market variance, the conditional expected excess returns, and the risks to change
over time. Strong evidence, for example, Bodurtha and Mark (1991) and Ng (1991), is
documented about the time-varying property of . More recently, by using a post-1963 sample
data, Ang and Chen (2003) document that using a conditional CAPM with time-varying beta
explains the portfolio returns equally well with incorporating additional risk factors, such as
book-to-market.
942

The accurate estimation of is not only important for the equity market to estimate the
systematic risk associated with the individual stock, it is also crucial in determining yield spread
of corporate bond in the bond market. Conventionally, the yield spread of corporate bond is
explained by expected default loss and tax premium; however, as documented in Elton, Gruber,
Agrawal, and Mann (2001), while expected default only accounts for a small fraction and state
taxes explain a substantial portion of the yield spread, the remaining portion of the spread is
related to the risk premium. The rationale is that since the return on corporate bonds is riskier
than the return on government bonds, investors should require a premium for the higher risk,
thus a large part of the risk on corporate bonds is systematic rather than diversifiable (Elton et
al. (2001), pp.247).
Systematic is also of special interest for the bond/debt market. One aspect is that the recent
study start using corporate bond yield spreads to gauge investors return expectations. Standard
asset pricing theory proposes that investors should demand an ex-ante premium for acquiring
risky securities. However, given the unobservability nature of the ex-ante risk premium,
empirical studies in finance use ex-post averaged stock returns as a proxy for expected stock
returns. The rationale is that the average return catches up and matches the expected return on
equity securities for sufficiently long horizon; therefore, the ex-post average excess equity
returns provide an estimate for the expected equity risk premium. However, this practice has
important limitation, that is, average realized return need not converge to the expected risk
premium in finance samples. Therefore, finance literature has been calling for a better measure
of expected return to test asset pricing theory (Elton (1999)). Campello, Chen, Zhang (2008)
develop an alternative measure of risk premium based on bond yield spreads data and find that
the market beta is significantly priced in the cross section of expected returns. The use of bond
yield data is logically built up on the argument that a firms systematic risk is incorporated into
both its stock and bond prices, and after controlling for default risk, firms with higher systematic
risk should have higher yield spreads. This linkage between systematic risk and the
determination of bond yield spreads emphasizes the importance of understanding beta.
The 2007 onward credit risk with respect to mortgage-back security problems have significantly
affected the US financial market as well the global financial market. Conventional credit models
typically focus on the default risk of the borrower, but largely overlook the risk associated with
the value of the collateral. Frye (2000) argues that the value of the collateral fluctuates with
economic conditions; this implies that if the economy takes a downturn, the creditor can
experience not only obligors default, but as well as the damage of the value of the collateral.
The importance of the paper is that it confirms the relevance of beta estimation to the credit risk
decision from a yet different but insightful perspective.
The banking industry is one of the main users of the credit risk models. However, even though
that systematic risk factors have been incorporated into the models of probability of default, the
integration of credit risk and other components of the risk exposure, such as interest risk and
systematic risk, is still largely overlooked. Allen and Saunders (2003) survey different credit risk
models to examine how macroeconomic and systematic risk effects are incorporated into
measures of credit risk exposure. They find that cyclical effects on loss given default and
exposure at default have largely ignored the effects from systematic risk factors. The practice in
banking industry helps expanding the usefulness of our study. Given that assessing risk is of
ultimate importance in bank industrys lending business, our study helps the industry realize that
systematic risk beta may not be stationary and it evolves over time; their role in measuring risk
exposure is quite dynamic.
943

This paper focuses on investigating the time-varying property of by applying a more general
form of CAPM mode, namely, the market model. The use of market model is without the loss of
generality with respect to the emphasis of this paper. The key difference between the CAPM and
market model is whether the return of the capital asset and return of the market is adjusted by
risk-free asset. In particular, market model is depicted in the following way
R

where R
is the single asset return, R
is the return of the market index in which the asset is

traded,
is an additive constant,
is the beta coefficient which indicates the sensitivity of R
to
the changes in R
and
is the error component (a random variable, with mean zero and

variance equal to
, which assumes independently and identically distributed values). The

model assumes that the risk of the portfolio can be divided into two components: the first
component can be reduced by the mean of diversification; the second component depends on
market risk. The parameter
can be interpreted as representing the relationship between the

riskiness of the single asset and the market index volatility; as it is given by the covariance
between the respective returns over the variance of the index returns.
In order to incorporate the time-varying property of the beta coefficient in the market model, the
literature has been utilizing different modeling techniques. Brooks, Faff, McKenzie (1998) and
Faff, Hillier, and Hillier (2000) examine the performance of modeling techniques that estimate
time varying . They focus on the three most commonly used approaches in the literature:
Bollerslevs (1990) multivariate generalized ARCH model, the time varying heteroskedastic
market model suggested by Schwert and Seguin (1990) and time dependent betas estimated with
the Kalman Filter algorithm. Choudhry and Wu (2008) investigates the forecasting ability of four
different GARCH models and the Kalman filter method for the time-varying . Moreover, the
time-varying property of is also investigated in portfolio optimization (Berardi, Corradin, and
Saommacampagna, 2002) and its relationship to macroeconomic variables (Fraser and
Groenewold,1997). Although a wide range of modeling techniques can be used to characterize
the time-varying essence of , Kalman filter method is the mostly commonly adopted.
A dynamic system such as the time-varying can be represented in a general form known as the
state space model. In the engineering literature of the 1960s, "state space" was developed by
control engineers to describe system that vary through time. The general form of a state space
model defines an observation (or measurement) equation and a transition (or state) equation,
which together express the structure and dynamics of a system. A state space model can be used
to incorporate unobserved variables into, and estimate them along with, the observable model to
impose a time-varying structure of the market model . Additionally, the structure of the time-
varying can be explicitly modeled within the Kalman filter framework to follow any stochastic
process. The Kalman filter is a recursive algorithm which allows one to upgrade model estimates
using new information (Harvey, 1994). In other words, Kalman filter recursively forecasts
conditional s from an initial set of priors, generating a series of conditional intercepts and
coefficients for the market model.
The Kalman filter method estimates the conditional , using market model
R

the above market model represents the observation equation of the state space model. However,
the form of the transition equation depends on the form of stochastic process that s are assumed
to follow. In other words, the transition equation can be flexible, such as using mean-reverting
first-order autoregressive process with a white noise process, v
, where:

944

or a random walk, where

or a random coefficient, where

to be consistent with Ohlson and Rosenbergs finding that is slowly converge towards a
stationary mean, this paper assumes the first characterization of where follows a mean-
reverting first-order autoregressive process.
The state-space representation of the dynamics of the market model is then given by the
following system of equations
R

In the literature, in the market model is used in that they are used for some future period, e.g.,
expected future returns. However, since future s are not known, it is common practice to use s
estimated over some recent period. Then future forecasted returns are projected from the
estimated s. By comparing the forecast errors derived from the difference between the observed
return and the forecasted return across different modeling techniques, researchers document
which model performs better. An important assumption in this practice is that the s estimated
over a recent period are appropriate as an input into calculations that require projections in to the
future. While this procedure would be acceptable if the s were known to be stable over time, it
is probably inappropriate given the widespread evidence that s are time-varying. This method is
particularly vulnerable when
s are estimated in order to measures market reactions to events in

that the very event may be causing
s to shift. The literature on estimating using alternative

modeling techniques document that, in general, there is considerable time variation in . The
empirical results for alternative modeling techniques are quite mixed. Using Australian and UK
returns data respectively, Brooks et al. (1998) and Faff et al. (2000) find that while all three
models (ARCH, time-varying heteroskedastic market model, and Kalman filter) perform
adequately in capturing movements in , Kalman filter algorithm is in general supported based
on in-sample forecast errors. Choudhry and Wu find that measures of forecast errors
overwhelmingly support the Kalman filter approach relative to four different GARCH models.
However, Frees (2004) find contrary results. He considers 90 US firms from the insurance
carriers, for each firm, he uses sixty months of data ranging from 1995 through 1999. The result
shows that Kalman filter approach shows little significant superiority.
Given the conflicting results from the literature and the questionable procedure to use a time-
varying to derive forecasted future returns as a foundation of comparison across different
models, this paper more directly investigates the time-varying properties of . The comparison
of the forecastability of across alternative models is derived with the forecast errors of itself,
rather than from the comparison of the forecast errors of the future returns. Since the prior
literature seems to suggest that Kalman filter algorithm performs better than the forms of
GARCH and time-varying heteroskedastic market model, this paper focuses on two alternative
modeling techniques in estimating and forecasting : one is simply the ordinary least square
(OLS) model by the market model, and the other one is the state-space model using Kalman
filter.
The challenge of direct comparison of the forecast error of s lies in the measurement of
coefficient. Since is not observed, the benchmark (or realized) has to be derived from certain
statistical process. Since of a security is defined as the covaribility of the security returns with
945

the market returns divided by the variance in market returns; therefore, we derive benchmark
by using looking-back return data with market model. More specifically, the benchmark at
any given point in time is calculated by regressing forward-looking individual stock returns
against market returns. The rationale of this approach is based on the argument that if investors
were to have perfect information about the future, the covaribility of the individual stock and the
market in any given point of time would incorporate all information from this point onward in
time. Therefore, by looking-back, the market model would capture the ideal information
environment and the realized would serve the purpose of benchmark. This rational
expectations approach to beta estimation establishes a suitable benchmark from which to
compare forecasted betas
under different estimation methods. This is the first paper to our

knowledge to derive benchmark in such a simple and straightforward way. A similar method of
deriving realized beta is employed by Hooper, Ng, and Reeves (2007). Given that is the
covariance of a security with the market divided by the variance of the market; therefore, the
benchmark would then to use the realized covariance of a security and the market divided by
the realized variance of the market. Hooper et al. use this approach to examine the forecastability
of monthly beta. They focus on Dow Jones Industrial Average stocks and find that the out-of-
sample forecasting shows a dramatic reduction of forecast error of beta on average by over 80%,
relative to the industry standard of the constant model (the industry standard for over thirty years,
following Fama and Macbeth, 1973).
II. Data and methodology
We use daily stock returns data for the period January 1, 1998 to March 31, 2007 with the data
sourced from the Center for Research in Security Prices (CRSP). We use twenty-five companies
in the Dow Jones Industrial Average (DJIA). The DJIA is a value-weighted average of 30
companies and since 1928 it has been an extensively-employed indicator of the stock market
performance. The companies that are included in the DJIA are large-cap companies with sizable
market value and liquidity. Due to the incompleteness of data, five companies were excluded
from the sample. Table 1 lists the sample of DJIA Companies used in this study.
As previously mentioned, forecast errors between the benchmark and
are the key

measures of the paper. Two alternative modeling techniques are adopted in this paper. For OLS
model, we run market model by regressing individual daily stock return against market return,
where the latter is represented by value-weighted market return. The literature often uses rolling
time period in estimation. To avoid data overlapping and the autocorrelation among forecast
errors, we use non-overlapping daily stock return for any given company in the sample to
generate quarterly average forecast of series. This
is labeled as
uses quarterly
return information; however, the criticism of this process is that it fails to take into consideration
the evolution of itself. As new information comes in every trading day, , as a risk measure, is
updating itself to adjust the risk implication. To address this criticism, we introduce Kalman
filter approach to estimate coefficient to incorporate the time-varying property. To
operationalize Kalman filter approach, an initial prior of coefficient is required. We use daily
stock return from 1998 to 1999 to derive the initial estimate of . With this prior, we use all
return data from 1998 to 2000 to derive the first forecast under Kalman filter algorithm. This
is labeled as
. Both
and
are derived from historical market information;

however, the key difference is that
reflects only the most recent three months market

information, while Kalman filter approach allows
to update continuously based on the

whole prior history.

946

Table 1: Sample of DJIA companies
Count Ticker Company
1 MMM 3M Co
2 AA Alcoa Inc
3 MO Altria Group Inc
4 AIG American International Group Inc
5 AXP American Express Co
6 BA Boeing Co
7 CAT Caterpillar Inc
8 C Citigroup Inc
9 KO Coca Cola Co
10 DIS Disney Walt Co
11 DD Du Pont E I de Nemours & Co
12 GE General Electric Co
13 GM General Motors Corp
14 HPQ Hewlett Packard Co
15 HD Home Depot Inc
16 IBM International Business Machines Corp
17 INTC Intel Corp
18 JNJ Johnson & Johnson
19 JPM Jpmorgan Chase & Co
20 MCD Mcdonalds Corp
21 MRK Merck & Co Inc
22 MSFT Microsoft Corp
23 PFE Pfizer Inc
24 UTX United Technologies cOrp
25 WMT Wal Mart Stores Inc

The
series from both models are then compared with the benchmark , which is labeled as
. As mentioned in the prior section,
is derived from the market model; however, rather than

using historical quarterly daily stock and market returns,
is generated based on forward-

looking quarterly daily stock and market returns. Specifically, the last quarter in our sample
period is the first quarter of 2007, this quarter is treated as the holdout period for the derivation
of the last benchmark in the
series.
The forecast error of , is the difference between the benchmark and
, the forecast errors

from the two models are

A group of measures derived from the forecast error are designed to evaluate forecasts. Mean
errors (ME)
ME

1
n

ME

1
n

are used to assess whether the models over or under-forecast series. Table 2 reports the mean
forecast errors by company.

947

Table 2: forecast errors by company
Company

3M Co 0.00538 0.21328
Alcoa Inc 0.01449 0.34762
Altria Group Inc 0.03077 0.12705
AIG Inc 0.01479 0.00717
American Express Co 0.00127 -0.12212
Boeing Co 0.00328 0.19392
Caterpillar Inc 0.03067 0.18916
Citigroup Inc -0.01420 -0.30913
Coca Cola Co 0.02417 -0.19633
Disney Walt Co -0.00792 0.14842
Du Pont 0.01502 -0.03359
General Electric Co -0.01298 -0.03794
General Motors Corp 0.00711 0.16299
Hewlett Packard Co -0.03722 0.16025
Home Depot Inc -0.03360 -0.00791
IBM Corp -0.00805 -0.09408
Intel Corp 0.00663 0.06536
Johnson & Johnson 0.02519 -0.24055
Jpmorgan Chase & Co -0.01290 -0.09815
Mcdonalds Corp 0.02786 0.13944
Merck & Co Inc 0.03569 -0.44874
Microsoft Corp -0.03420 -0.26123
Pfizer Inc -0.00542 -0.59922
United Technologies Corp 0.01513 0.18307
Wal Mart Stores Inc 0.01423 -0.20075
As Table 2 presents, the mean forecast errors, regardless of the modeling technique can be either
positive or negative. And the magnitude of ME
and ME
differ across companies. Closer

investigation of Table 2 shows that approximately half companies in the sample have lower
mean forecast errors when is forecasted under Kalman filter approach; which indicates that
is not superior than
in reducing the mean forecast errors of by company counts.

However, the accuracy of the forecast is not clear given the sign mixture in the mean forecast
error.
The accuracy of each forecast is measured by a particular loss function. Two popular loss
functions are Mean squared error (MSE) and Mean absolute error (MAE):
MSE
1
n

MAE
1
n
|

where j OLS, KAL. The lower the forecast error measure, the better the forecasting
performance. However, it does not necessarily mean that a lower MSE (MAE) indicates a
superior forecasting ability, since the difference between the MSEs (MAEs) across alternative
modeling techniques may not be significantly different from zero. Therefore, it is important to
check whether any reductions in MSEs (MAEs) are statistically significant.
Diebold and Mariano (1995) develop a test of equal forecast accuracy to test whether two sets of
forecasts error have equal mean value. Let L
, where j OLS, KAL be the loss function

(either MSE or MAE) for two competing models OLS and Kalman filter approach. To determine
if one model predicts better than another, we test the following null hypothesis
H
: EL
EL

948

against the alternative
H
: EL
EL

The Diebold-Mariano test is based on the loss differential
d

The null of equal predictive accuracy is then
H
: Ed
0
The Diebold-Mariano test statistic is
S
d
vard

where
d

1
n
d

is the mean of the difference between loss functions and where
vard
covd
, d

is the asymptotic variance of d
. According to Diebold and Mariano (1995), results of Monte

Carlo simulation experiments show that the performance of this statistic is good, even for small
samples and when forecast errors are non-normally distributed. Diebold-Mariano test is further
modified by Harvey, Leybourne, and Newbold (1997), where the latter is found to perform better
than the original Diebold and Mariano when the forecast errors are autocorrelated. Since we use
nonoverlapping data to derive
and benchmark , we are confident that Diebold and Marianos

test provides consistent and unbiased estimators.
Table 3: Benchmark , mean-squared errors and mean-absolute errors
Company

3M Co 0.83098 0.08902 0.10939 0.24167 0.24967
Alcoa Inc 1.31914 0.12089 0.19033 0.29285 0.37141
Altria Group Inc 0.50394 0.13253 0.14742 0.29942 0.28127
AIG Inc 0.99552 0.12751 0.10651 0.25363 0.26645
American Express Co 1.18144 0.05449 0.08692 0.18219 0.24349
Boeing Co 1.04666 0.16003 0.12727 0.29788 0.26744
Caterpillar Inc 1.22408 0.10096 0.11451 0.24663 0.29636
Citigroup Inc 1.11057 0.04148 0.17339 0.16223 0.35734
Coca Cola Co 0.53478 0.10182 0.16589 0.24839 0.32682
Disney Walt Co 1.13758 0.11789 0.17633 0.26309 0.35675
Du Pont 0.95824 0.05368 0.10674 0.17088 0.25705
General Electric Co 1.09473 0.10657 0.12408 0.27885 0.28932
General Motors Corp 1.12205 0.29967 0.16898 0.40193 0.31804
Hewlett Packard Co 1.32384 0.32918 0.30079 0.50628 0.44789
Home Depot Inc 1.16729 0.17289 0.09051 0.32394 0.23626
IBM Corp 0.97713 0.18291 0.13343 0.35376 0.29481
Intel Corp 1.66424 0.13214 0.22026 0.27244 0.41116
Johnson & Johnson 0.46275 0.12959 0.19723 0.29784 0.36801
Jpmorgan Chase & Co 1.31553 0.07386 0.14242 0.20485 0.33823
Mcdonalds Corp 0.73056 0.18351 0.12520 0.31325 0.30592
Merck & Co Inc 0.64771 0.13794 0.30074 0.30162 0.46146
Microsoft Corp 1.09683 0.08745 0.18746 0.23385 0.37401
Pfizer Inc 0.75078 0.15172 0.49559 0.31099 0.61013
United Tech Corp 1.04164 0.18892 0.15305 0.28233 0.22958
Wal Mart Stores Inc 0.82383 0.07587 0.08289 0.24424 0.22513

949

Prior to formally present the test results of Diebold-Mariano, Table 3 provides the benchmark
(
), mean-squared errors and mean absolute errors between the benchmark and
s derived
from both modeling techniques, respectively; i.e., MSE
and MAE
, where j OLS, KAL.

Closer investigation from Table 3 shows that regardless of the loss function, the forecast errors
between the benchmark and the OLS forecast tends to be smaller than that between the
benchmark and the Kalman filter approach forecast. Specifically, out of twenty-five sample
companies, seventeen companies has smaller forecast errors with respect to OLS forecasts under
mean-squared error loss function specification; while the number is sixteen for mean-absolute
error loss function specification.
However, the descriptive statistics from Table 3 do not readily testify that the forecastibility of
OLS forecast is superior, since the difference between the MSEs (MAEs) across alternative
modeling techniques may not be significantly different from zero. Table 4 presents the test
results of the Diebold-Mariano statistics. This table reports the t-statistics and the p-values of
two loss functions. A significant p-value with a positive t-statistics is an indicator that

performs better than
in reducing the forecast errors. A significant p-value with a negative t-

statistics is, on the other hand, an indicator that
has superior forecastibility than
in
producing smaller forecast errors.
Table 4: Diebold-Mariano statistics
Company
|MSE
|MAE
t-statistics p-value t-statistics p-value
3M Co -0.63 0.5328 -0.16 0.8735
Alcoa Inc -1.53 0.1389 -1.24 0.2269
Altria Group Inc -0.44 0.6616 0.36 0.7243
AIG Inc -0.12 0.9071 -0.32 0.7529
American Express Co -0.76 0.4563 -0.78 -0.4443
Boeing Co 1.97 0.0613 0.86 0.4015
Caterpillar Inc -0.40 0.6916 -0.89 0.3807
Citigroup Inc 0.86 0.4007 1.52 0.1426
Coca Cola Co -1.58 0.1271 -1.24 0.2288
Disney Walt Co -0.75 0.4596 -0.86 0.3996
Du Pont -2.97 0.0071 -3.17 0.0044
General Electric Co -0.65 0.5202 -0.30 0.7662
General Motors Corp 1.17 0.2537 1.33 0.1978
Hewlett Packard Co 0.35 0.7275 0.69 0.5001
Home Depot Inc 2.39 0.0254 2.42 0.0239
Ibm Corp 1.22 0.2352 15.71 <.0001
Intel Corp -1.60 0.1224 -2.02 0.0556
Johnson & Johnson -1.03 0.3157 -0.98 0.3366
Jpmorgan Chase & Co -1.23 0.2317 -1.54 0.1389
Mcdonalds Corp 1.25 0.2235 -0.29 0.7754
Merck & Co Inc -1.76 0.0915 -1.44 0.1653
Microsoft Corp -1.50 0.1483 -1.58 0.1291
Pfizer Inc -2.30 0.0316 -2.50 0.0205
United Tech Corp 1.38 0.1813 1.29 0.2113
Wal Mart Stores Inc -0.31 0.7582 0.43 0.6679
Consistent with Table 3, more than half of the companies have a negative t-statistics regardless
of the specification of the loss function. This means that OLS forecast performs better than its
Kalman filter approach counterpart. However, not all these companies show a significant
differential between two measurements. For the majority of them, although the t-statistics are
negative, p-values are not significant. This indicates the indifference between two modeling
950

approaches in forecasting . Two companies in the sample have significant results from Diebold-
Mariano statistics. These two companies are Du Pont and Pfizer. Take Pfizer as an example, the
t-statistics for mean-squared error and mean-absolute error loss functions are -2.30 and -2.50.
They are both significant at 5% level. The negative t-statistics with significant p-value indicate
that OLS forecast for Pfizer performs well in explaining the relationship between the individual
stock and the market return as reflected by smaller forecast errors. Incorporating all prior
information to allow to evolve over time does not add more power to the coefficient.
Therefore, conventionally simple market model is sufficient to estimate the risk measure for
this particular company.
We do find a couple of companies in the sample that shows the superiority of Kalman forecast .
One company is Home Depot Inc. In both loss function specifications, the t-statistics are 2.39
and 2.42, with a 5% significant p-value. The other company is IBM. Notice that IBM is quite
different from Home Depot, as well as the majority of other companies in the sample. For the
majority of the companies in the sample, the test results from two different loss functions are
quite consistent. More specifically, either both specifications show significance or they both fail
to show such differential significance. Moreover, the sign of the t-statistics tend to be consistent
across alternative loss functions. However, IBM has a t-statistics of 15.71, which is highly
significant under mean absolute error specification, on the other hand, if mean-squared error
function is applied; the t-statistics is only 1.22, which fails to differentiate two models. This
means that if we define the loss function to be mean absolute error, Diebold-Mariano test
provides the result that the forecast derived from Kalman filter algorithm outperforms its OLS
counterpart in reducing the forecast errors.
III. Conclusion
The employment of Kalman filter approach is based on the criticism that market model assumes
that is stationary over time, while in fact it is likely time-varying. Kalman filter algorithm
allows to evolve over time and takes into consideration of updating new information
temporally. While this paper do not find the significant difference between two modeling
techniques in terms of reducing the forecast errors of directly, it is consistent with Frees
(2004) that more sophisticated engineering approach may not contribute to forecast accuracy of
coefficient in the market model. There are a few other reasons that may contribute to the lack of
significant differential on two models. The derivation of benchmark in this paper is based on
the market model; the same process is also utilized for calculating OLS forecast. This implies
that the procedure would bias against finding favorable results for forecast from Kalman filter
algorithm. The other possible explanation would be that the sample consists of twenty-five DJIA
companies with stable long-term performance and liquidity. These companies tend to have large
analysts following and are actively traded. Therefore, information channels with respect to the
sample companies are transparent and open; leading to investors having consistent expectations
for the companies performance on the market. All these perspective imply that an average risk
measurement may well be the best estimate for forecast. OLS with market model helps
generating an estimated coefficient that reflects the averaging effect. This may explain why the
introduction of Kalman filter algorithm does not lead to significant forecast error differentials.
Moreover, the specification of the loss functions weighs positive and negative forecast errorr
equally. In prior literature, where forecast errors of the returns are considered, this may not
impose a problem; however, when forecast error of itself is considered, the sign matters. The
markets reaction for positive versus negative forecast errors would very likely be different. The
application of mean-squared error and mean-absolute error loss functions fail to take into
951

consideration the implication for the sign differences. Last but not least, Ohlson and Rosenberg
(1984) point out that not only there is an a tendency for s to converge towards the stationary
mean, but also at the same time, in each period, there is a purely random (serially independent)
perturbation in . Kalman filter algorithm is silent on how to address a purely random
perturbation term. Rubinstein (1973) documents the use of mean-variance security valuation
theorem to model the effect of capital structure on the value of a firm. Rubinstein further extends
to model the factors that determine . The systematic risk, , although unobservable, its ex post
values is shown to be a function of a firm's debt-to-equity ratio, corporate tax rate, and other
underlying fundamental operation aspects. Let
and
be the leveraged and unleveraged

systematic risk and
be the corporate tax rate, the following equation provides the

relationship between and a firms capital structure:
in general represents operational risks. Risks associated with operations include the
uncertainty in demand, the fluctuation in the inputs, and the cost structure between variable and
fixed costs. Therefore, without considering these other accounting information on determining ,
employing Kalman filter method to just count for the time-varying property of may not fully
represent the characteristic of .
There are a few potential criticisms with respect to the dynamic system of equations used in
Kalman Filter estimation. We argue in the paper that, using the market model rather than CAPM
results in no loss of generality. However, given that CAPM is more theory-based, the adoption of
CAPM estimation would be a next natural step. Such estimation would result in a zero expected
intercept, as R
f
would be subtracted from both market and individual returns in beta estimation.
In the paper, we do not model the intercept term because our interest is on the coefficient.
However, given our sample is from different industries, it is likely that explains industry-
specific effect or unusual firm performance.
The evolution equation in the Kalman Filter estimation is designed to follow mean-reverting
autoregressive stochastic process. A few issues may of particular concern. Firstly our results may
be sensitivity to different assumptions of s evolution; secondly, we have not performed tests on
the stationarity of the OLS-generated series (e.g., Dickey-Fuller or Phillips-Perron tests), and
begs the question of whether really follows an autoregressive stochastic process.Although it is
not reported, the first-order autoregressive coefficients on evolution is significant, implying
that there is a rationale for specifying as following first-order autoregressive process. That
being said, it remains unknown what is the mean value of OLS derived in the evolution
equation.
A few extensions may also shed light on our study. First, in much of the finance literature,
monthly returns are widely used for CAPM-type of study because this periodicity is less
susceptible to noise. A less frequent interval on returns might be adopted and provide different
results. Also, the sample period spans a 10-year range, within which, 1999-2000, was a
noticeable Internet bubble. Although the sample firms are typically not internet companies, the
effect of an overall bubble may cause a beta regime shift. Further, the sample is from DJIA index
, and the selection criteria for DJIA index inclusion may be restricted to certain firms that having
certain attributes that are not representative of stocks generally. It would be interesting to expand
the sample to S&P 500 firms or smaller ones to see if the time-series behavior of varies, and
whether Kalman Filter estimators perform better for stocks with certain characteristics. Finally,
given that both equity and debt can be thought of as contingent claims written on the same set of
952

productive assets (and hence should share common risk factors), it would be interesting to see
how a more comprehensive understanding on the evolution of would help determining the
systematic risk premium, which is a significant component of return differences.

References
Ang, A., and Cheng J. (2003). CAPM over the long-run: 1926-2001. Working paper.
Berardi, A., Corradin, S., and Sommacampagna, C. (2002). Estimating value at risk with the
Kalman Filter. Working paper.
Bollerslev, T. (1990). Modelling the coherence in short run nominal exchange rates: a
multivariate generalized ARCH model. Review of Economics and Statistics, 72(3), pp.
498-505.
Bos, T., and Newbold, P. (1984). An empirical investigation of the possibility of stochastic
systematic risk in the market model. Journal of Business, 57(1), pp. 35-41.
Bodurtha, J., and Mark, N. (1991). Testing the CAPM with time-varying risks and returns.
Journal of Finance, 46(4), pp. 1485-1505.
Brooks, R., Faff, R., and McKenzie, M. (1998). Time-varying beta risk of Australian industry
portfolios: A comparison of modeling techniques. Australian Journal of Management,
23(1), pp.1-22.
Choudhry, T., and Wu, H. (2008). Forecasting ability of GARCH vs Kalman Filter method:
evidence from daily UK time-varying beta. Journal of Forecasting, 27(8), pp. 670-689.
Diebold, F., and Mariano, R. (1995). Comparing predictive accuracy. Journal of Business &
Economic Statistics, 13(3), pp. 253-263.
Elton, E., Gruber, M., Agrawal, D., and Mann, C. (2001). Explaining the rate spread on
corporate bonds. Journal of Finance, 56(1), pp. 247-277.
Faff, R., Hillier, D., and Hillier, J. (2000). Time varying beta risk: an analysis of alternative
modeling techniques. Journal of Business Finance & Accounting, 27(5), pp. 523-554.
Fama, E., and MacBeth, J. (1973). Risk, return, and equilibrium: empirical tests. Journal of
Political Economy, 81(3), pp. 607-636.
Fraser, P., and Groenewold, N. (1997). Time-varying CAPM betas: Kalman filter estimates and
their relationship to macroeconomic variables. International Congress on Modelling and
Simulation, 97(3), pp. 1242-1247.
Frees, E. (2004). Longitudinal and panel data: analysis and applications in the social sciences.
Cambridge University Press, Cambridge.
Harvey, A. (1994). Forecasting, structural time series models and the Kalman filter. Cambridge
University Press, Cambridge.
Harvey, D., Leybourne, S., and Newbold, P. (1997). Testing the quality of prediction mean
squared errors. International Journal of Forecasting 13, pp. 281-291.
Hooper, V., Ng, K., and Reeves, J. (2007). Monthly forecasts of systematic risk: an evaluation.
Working paper.
Lintner, J. (1965). The valuation of risky assets and the selection of risky investments in the
portfolios and capital budgets. Review of Economics and Statistics 47(1), pp. 13-37.
Ng, L. (1991). Tests of the CAPM with time-varying covariances: a multivariate GARCH
approach. Journal of Finance 46(4), pp. 1507-1521.
Ohlson, J., and Rosenberg, B. (1982). Systematic risk of the CRSP equal-weighted common
stock index: A history estimated by stochastic-parameter regression. Journal of Business,
55(1), pp. 121-145.
953

Rubinstein, M. (1973). A mean-variance synthesis of corporate financial theory. Journal of
Finance, 28(1), pp. 167-181.
Schwert, G., and Seguin, P. (1990). Heteroscedasticity in stock returns. Journal of Finance,
45(4), pp. 1129-1155.
Sharpe, W. (1964). Capital asset prices: a theory of market equilibrium under conditions of risk.
Journal of Finance, 19(3), pp.425-442.

954

Operational Risk and Large Internal Frauds at Financial Institutions

Benton E. Gup, Ph.D.
Chair of Banking
The University of Alabama
Room 200 Alston Hall
Tuscaloosa, AL 35487
bgup@cba.ua.edu

955

Abstract
Operational risk depends, in part, on corporate culture. Corporate culture and risk management
policies at financial institutions are established by top management. But the execution of risk
management policies involves personnel at every level of the organization. History reveals that
insider abuse and criminal misconduct has occurred at all levels of management and operations.
This article focuses on nine case studies of domestic and international frauds committed by
insiders of financial institutions that each resulted in losses of billions of dollars. The study finds
that the best way to rob a financial institution is to own and control it, or to be a securities trader.
Large scale trading frauds were uncovered only after losses involving billions of dollars were
incurred. Most of them occurred at locations that were long distances from corporate
headquarters. All of the trading losses suggest a lack of internal controls and management
turning a blind eye as long as long as they made a profit.
Keywords: Financial institutions, operational risk, corporate governance, fraud, illegal behavior,
JEL Classifications: G20, G30, K4

Operational Risk and Corporate Culture
Operational risk is the risk of loss resulting from inadequate or failed internal processes, people
and systems, or from external events. It includes fraud, damage to physical assets, legal risks,
and other factors that result in unexpected losses. Management policies can have a major impact
on operational risk. Management policies start with the Board of Directors. The Basel
Committee on Banking Supervisions Principles for Sound Management of Operational Risk
states that the Board of Directors of a bank should establish a corporate culture that provides
appropriate standards and incentives for professional and responsible behavior throughout the
organization.
19
Corporate culture can be defined as what determines how decisions are made in
an organization.
20
Alternatively, it can be defined as the way we do things around here. It is
established by top management. However, corporate culture may not be consistently applied
throughout the entire organization. This is especially true for large, complex organizations.
Equally important, corporate culture varies from bank to bank. It has played a significant role in
the losses incurred by various banks and other types of financial institutions.
21

Corporate culture also depends on national cultures which affects attitudes of bankers and
borrowers toward corruption.
22
Transparency International publishes an annual Corruption
Perception Index.
23
Countries are ranked on a score of 10 (highly clean) to 0 (highly corrupt).
The 2011 Corruption Perception Index for 183 countries revealed that the vast majority of the

19
Principles for Sound Management of Operational Risk, June 2011, Basel Committee on Banking Supervision.
20
Brooks, Douglas W. 2010, Creating a Risk Aware Culture, Enterprise Risk Management: Todays Leading
Research and Best Practices for Tomorrows Executives, John Fraser and Betty J. Simkins, Editors, John Wiley &
Sons, New Jersey, 87.
21
For additional information, see Gup, Benton E. 2007. Basel II: Operational Risk and Corporate Culture, Benton
E. Gup, Editor, Corporate Governance in Banking: A Global Perspective, Edward Elgar, Northampton Mass., 134-
150.
22
For additional information on corruption and bank lending, see: Barth, J. R., Lin, C., Lin, P. & Song, F. M. 2009.
Corruption in bank lending to firms: Cross-Country micro evidence on the beneficial role of competition and
information sharing. Journal of Financial Economics, 91 (3). 361-388. Beck, T. Demirg-Kunt, A., & Levine, R.
2006. Zheng, X., El Ghoul, S., Guedhami, O., Kwok, C., Collectivism and Corruption in Bank Lending, Working
Paper presented at the Midwest Finance Association Annual Meeting, February 23, 2012.
23
For additional information about the Corruption Perception Index, see Transparency Internationals website:
http://www.transparency.org/policy_research/surveys_indices/cpi/2010/results
956

countries had a score of 5 or less. New Zealand (9.5), Denmark (9.4) and Finland (9.4) had the
highest scores, while North Korea (1.0) and Somalia (1.0) had the lowest scores. The United
States had a score of 7.1. Corruption can be defined as immoral or dishonest behavior.
Corruption is not necessarily illegal. However, fraud is illegal. Fraud consists of deceitful
practices used with the intent to deprive another person of his right, or in some manner to injure
that person.
The Basel Committees Principles say that the Board of Directors should approve and review
their banks risk appetite and tolerance statement for operational risk that articulates the nature,
types, and level of operational risk that the bank is willing to assume.
24
Operational risk is
defined as the risk of loss resulting from inadequate or failed internal processes, people and
systems or from external events.
25
Operational risk includes, but is not limited to fraud,
accounting issues, legal issues, damage to physical assets, business practices, and so on. Banks
with a strong culture of risk management and ethical business practices are less likely to
experience potentially damaging operational risk event and are better placed to deal with those
events that do occur.
26
Accordingly, banks must identify, assess, and manage the risks that they
face as part of their Enterprise Risk Management.
27
However, Enterprise Risk Management has
limitations. The limitations include controls that can be circumvented by management or by two
or more people who collude to override them.
28

Risk management starts at the top of the organization. But the execution of risk management
policies involves personnel at every level of the organization. History reveals that insider abuse
and criminal misconduct in general and fraud in particular has occurred at all levels of
management and operations. During the late 1980s and early 1990s, about 60 percent of the
frauds reported by financial institutions in the United States were committed by insiders.
29
Since
then, the share of internal frauds has declined and been replaced by external frauds such as
mortgage and payments systems frauds.
30

This article focuses on selected frauds resulting in losses in excess $1 billion committed by
insiders of financial institutions -- top management, securities traders, loan officers, and other
insiders that resulted in large losses in recent years.
31
It also reveals some interesting

24
Ibid. 5- 6.
25
Ibid. 5.
26
Ibid. 7.
27
For a discussion of Enterprise Risk Management, see Beasley, Mark S., Bruce C. Branson, and Bonnie V.
Hancock, 2010, Developing Key Risk Indicators to Strengthen Enterprise Risk Management, The Committee on
Sponsoring Organizations of the Treadway Commission (COSO), Durham N.C., 2004. Enterprise Risk
Management Integrated Framework, Executive Summary, December. The Committee on Sponsoring
Organizations of the Treadway Commission (COSO), Durham N.C. September, 2004.; also see Fraser, John and
Betty J. Simkins, Editors, 2010, Enterprise Risk Management: Todays Leading Research and Best Practices for
Tomorrows Executives, John Wiley & Sons, New Jersey.
28
Committee of Sponsoring Organizations of the Treadway Commission, 2004. Enterprise Risk Management-
Integrated Framework, Executive Summary, September, 5.
29
FBI, Financial Institution Fraud and Failure Report, Fiscal Years 2006 and 2007.
30
Perkins, Kevin L., Assistant Director Criminal Investigative Division, FBI, 2010. Statement Before the Senate
Judiciary Committee, Washington, D. C. September 22; Eaglesham, Jean, 2011. Financial Crimes Bedevil
Prosecutors, The Wall Street Journal, December 5, C1; FBI, 2010 Mortgage Fraud Report, 2011, Year in Review,
August, 10. See this report for details on additional information about current fraudulent schemes and techniques
17-22.
31
Only those frauds where there is sufficient information are discussed here. For information about pre-1995 bank
frauds, see Gup, Benton E., 1995. Targeting Fraud: Uncovering and Deterring Fraud in Financial Institutions,
Probus Publishing Company, Chicago, IL.
957

similarities to previous large scale frauds. Fraud is defined as the intentional perversion of the
truth or the purpose of inducing another person or entity in reliance upon it to part with
something of value or to surrender a legal right.
32
Fraud must be material in the sense that it
results in a significant financial loss, injury, or damage. Although a strong corporate culture
promoting ethical behavior is established by top management, the pervasiveness of internal
frauds suggests that it has not been applied equally at all levels of the organizations.

Top Management Frauds
There is a saying that the easiest way to rob a bank is to own it. That also applies to other types
of financial institutions. In this section, we examine recent frauds by top management of a
mortgage lender, banks, and securities firms.

Taylor, Bean & Whitaker (TBW) and Colonial Bank. Lee Bentley Farkas was the former
chairman and owner of Taylor, Bean & Whitaker (TBW) located in Ocala, Florida. TBW was
one of the largest privately held mortgage lending corporations in the United States. Farkas
masterminded a $2.9 billion fraud scheme that contributed to the failure of TBW, Colonial Bank,
and its parent company Colonial BancGroup Inc.
33
Colonial Bank, headquartered in
Montgomery Alabama had total asset of $25.8 billion. It was the 25
th
largest bank in the United
States out of more than 7,000 banks - and the largest bank failure in 2009.
TBW was licensed by the Florida Office of Financial Regulations Division of Finance.
34
It was
also subject to examination by mortgage regulators in Arizona, Georgia, Idaho, Illinois,
Massachusetts, Mississippi, New Jersey, North Carolina, Pennsylvania, Vermont and the District
of Columbia.
35

Farkas bragged that he could rob a bank with a pencil.
36
And that is what he did. The scheme
began in 2002, when Farkas and his co-conspirators transferred money between TBW accounts
at Colonial Bank, and ran overdrafts of more than $100 million in those accounts to cover
shortfalls of cash. This required the cooperation of Catherine Kissick, the Senior Vice President
of Colonial Banks Mortgage Lending Warehouse Division (MWLD) in Orlando, Florida, and
Teresa Kelly, MWLDs Operational Supervisor. MWLD was the largest warehouse lender in the
country with total assets of $4.9 billion in 2009. It earned Colonial substantial fee income for
keeping mortgages on their books for a short time.
37
TBW was MWLDs largest customer, and it
accounted for no less than 21 percent of Colonial BancGroups net income.
38

32
Barnett, Cynthia, 2002. The Measurement of White-Collar Crime Using Uniform Crime Reporting (UCR) Data,
U.S. Department of Justice, Federal Bureau of Investigation, Criminal Justice Information Services (CJIS) Division,
March.
33
Former Chairman of Taylor, Bean & Whitaker Convicted for $2.9 Billion Fraud Scheme that Contributed to the
Failure of Colonial Bank. 2011. Department of Justice, Press Release, April 19.
http://www.fdic.gov/bank/individual/failed/colonial-al_q_and_a.html
Colonial Bank was closed on August 14, 2009. Colonial BancGroup Inc. was a one-bank holding company.
http://www.fdic.gov/bank/individual/enforcement/2009-08-48.pdf
34
Florida Office of Financial Regulations Division of Finance, http://www.flofr.com/Finance/default.aspx
35
Carr, Susan Latham, 2011. Audit: It wasn't just Taylor Bean that led to Platinum bank's failure, Ocala.com,
September 19.
36
Former Chairman of Taylor, Bean & Whitaker Convicted for $2.9 Billion Fraud Scheme That Contributed to the
Failure of Colonial Bank. FBI Washington Field Office, op. cit.
37
OKeefe, Brian. 2009. The Man Behind 2009s Biggest Bank Bust, CNNMoney.com. October 12.
38
Securities and Exchange Commission V. Catherine I Kissick, United States District Court for the Eastern District
of Virginia, Alexandria Division, Civil Action File No. 1:11cv:215, March 2, 2011.
958

Farkas also had help from four TBW employees: Desiree Brown (Treasurer), Paul Allen (Chief
Executive Officer), Ray Bowman (President), and Sean Ragland, (Senior Financial Analyst).
TBWs most valuable asset, and primary source of income was the right to service mortgages
that it had originated and sold to the Federal National Mortgage Association (Fannie Mae) and
the Federal Home Loan Mortgage Corporation (Freddie Mac) and to other investors.
In 2002, TBW began running overdrafts in its master bank account at Colonial Bank because of
its inability to meet operating expenses which included payroll, mortgage servicing payments,
and other obligations. TBW covered up the overdrafts by sweeping funds between various
accounts, and then through the sale of fictitious mortgage loans to the bank.
39
The scheme
resulted in Colonial Bank buying more than $1.5 billion of fictitious and impaired mortgage
loans from Ocala Funding, a mortgage lending facility controlled by TBW. They included loans
that had been sold to other investors and faked pools of mortgage backed securities.
40
Colonial
Bank unknowingly reported the worthless loans on their books at their face value. Collectively,
TBW misappropriated $1.4 billion from Colonials MWLD in Orlando.
TBW also misappropriated $1.5 billion from Ocala Funding, a mortgage lending facility that it
controlled. Ocala Funding sold asset-backed commercial paper to Deutsche Bank, BNP Paribas,
and other financial institutions. Ocala Funding was required to maintain cash and mortgage
collateral equal to the value of the outstanding commercial paper. However, some of the funds
were diverted from Ocala Funding to TBW to cover their operating losses, and the sale of assets
to Fannie Mae and Freddie Mac. To cover up the diversion of funds, Ocala Funding sent false
information to Deutsche Bank, BNP Paribas, and other institutions.
Additionally, the co-conspirators at TBW misappropriated more than $1.5 billion in mortgage
loans at Ocala Funding LLC, a mortgage lending facility that was owned by TBW. This caused
Colonial Bank and the Federal Home Loan Mortgage Corporation (Freddie Mac) to believe the
each had ownership in the same mortgage loans worth hundreds of millions of dollars.
41

Colonial BancGroup Inc. needed additional capital due to heavy real estate loan losses. In the fall
of 2008, it applied for $570 million financial aid through the U.S. Treasury Departments
Troubled Asset Relief Program (TARP). As part of the application process, Colonial BancGroup
submitted financial data that Catherine Kissick knew was materially false as a result of TBWs
fraudulent scheme. The TARP application was approved for $553 million, contingent on
Colonial raising $300 million in private capital which did not happen. Colonials application for
the TARP funds caught the attention of the Special Inspector General of TARP and Federal
Bureau of Investigations (FBI), and that was beginning of the end of Colonial.
42
The fraudulent
loans and misrepresentations contributed to Colonials failure. But they were not Colonials only
problems.
Colonial Bank dates back to 1981. It has a history of aggressive growth through acquisitions. In
2009, Colonial Bank had 346 offices located in Alabama, Georgia, Florida, Texas, and Nevada.
Most of these are states that had large increases in population and increased demand for real
estate loans. About 85 percent Colonials loan portfolio was concentrated in real estate, a large
portion of which were acquisition, development, and construction (ADC) loans. Their ADC

39
This cover-up was known as Plan B. Former Treasurer and President of Taylor, Bean & Whitaker Each
Sentenced to Prison for Fraud Scheme,2011. FBI Washington Field Office, June 10.
40
Former Chairman of Taylor, Bean & Whitaker Convicted for $2.9 Billion Fraud Scheme That Contributed to the
Failure of Colonial Bank.2011. FBI Washington Field Office, Press Release, April 19.
41
Former Colonial Bank Senior Vice President Sentenced to Eight Years in Prison for Fraud Scheme, 2011. FBI,
Washington Field Office, Press Release, June 17.
42
Historic Bank Fraud Case, 2011. FBI Podcasts and Radio, July 15.
959

loans, high risk security investments, and MWL related loans were highly concentrated in high-
growth real estate markets in Florida, Georgia, and Nevada.
43
These states were very negatively
impacted by the economic downturn that began in 2007. Colonials MWL assets totaled $5.2
billion, and represented 20 percent of the banks total assets. Mortgage loans with TBW
accounted for 63 percent of the MWL assets.
44

Given the concerns about Colonials liquidity and declining asset quality, the Alabama State
Banking Department seized Colonial Bank in August, 2009.
45
Then Colonial BancGroup filed
for bankruptcy. The failure of Colonial Bank resulted in a loss of $3.8 billion to the Federal
Deposit Insurance Fund.
46
An FDIC report on Colonials failure said that Colonial failed
because the Board of Directors (Board) and management did not develop and implement
adequate risk management practices pertaining to its significant concentration in acquisition,
development, and construction (ADC) loans and investments in higher risk, mortgage-backed
securities, loan underwriting, credit administration, and risk analysis and recognition, and
mortgage warehouse lending operations. Colonials decision to aggressively pursue rapid growth
by acquiring 25 banks and high growth residential markets, such as Florida, Nevada, and
Georgia, along with concentrating the loan portfolio in ADC loans, eventually led to larges
losses and capital and liquidity deficiencies.
47

Farkas misappropriated more than $20 million from TBW and Colonial Bank to support his
extensive life style. He bought a $28 million private jet, vacation homes in Maine and Key West,
expensive antique cars, restaurants and bars. He was sentenced to 30 years in prison, and ordered
to forfeit $38.5 million. The other conspirators pled guilty.
48
In addition, a series of judgments
and orders were issued directing Farkas and his co-conspirators to pay, jointly and
severally, a total of approximately $3.5 billion in restitution.
49

Catherine Kissick was sentenced to eight years in prison, and the others received lesser
sentences. Interestingly, Kissick did not profit directly from the scheme. Although she earned a
substantial salary as Senior Vice President of Colonial Banks MWLD, she did not receive any
payments in connection with the bogus transactions.
50
At first, she tried to do the right thing by
alerting her superiors at Colonial bank and working with TBW to fix the problems. However, her
superiors at Colonial Bank did not take corrective actions to resolve the problems. One could say
that they turned a blind eye to the bogus transactions. Eventually the losses grew beyond their
ability to cover them up. Stronger internal controls at Colonial Bank would have mitigated their
losses.

43
Material Loss Review of Colonial Bank, Montgomery, Alabama, 2010. Office of the Inspector General, Federal
Deposit Insurance Corporation, Report No. MLR-10-021, April. 2.
44
Ibid. 10-12.
45
Alabama State Banking Department Takes Possession of Colonial Bank, 2009. Press Release, State of Alabama
State Banking Department, August 14.
46
Former TBW CEO Sentenced to 40 Months in Prison for Fraud Scheme, 2011. FBI, Washington Field Office,
Press Release, June 21.
47
Material Loss Review of Colonial Bank, Montgomery, Alabama, op. cit., 30-31.
48
Former Chairman of Taylor, Bean & Whitaker Sentenced to 30 Years in Prison and Ordered to Forfeit $38.5
million.2011. FBI Washington Field Office, Press Release, June 30. In the Mater of Colonial Bank Montgomery
Alabama, Notice Of Charges And Of Hearing, Federal Deposit Insurance Corporation, FDIC-09-402c & b.
49
United States Department of Justice, United States v. Lee Bentley Farkas
Court Docket Number: 1:10-cr-200. No date given, http://www.justice.gov/criminal/vns/caseup/farkasl.html
50
United States of America vs. Catherine Kissick. Case Number 1:11CR88, United States District Court for the
Eastern District of Virginia, Honorable Leonie M. Brinkema, Position of the United States With Respect to
Sentencing, June 17, 2011. FBI, 2008 Financial Crimes Report, 7.
960

Bernard L. Madoff Investment Securities LLC. (BLMIS). Bernard Madoff founded the
broker-dealer securities company Bernard L. Madoff Investment Securities LLC. (BLMIS) in
1960. He was a securities broker-dealer, and a prominent member of the securities industry.
Madoff served as Vice Chairman of the National Association of Securities Dealers (NASD) in
the early1990s.In addition to the broker-dealer securities company, he was a part owner of
Cohmad Securities Corporation, a registered broker-dealer firm. It was located in Madoff broker-
dealer firms offices on the 18
th
floor of the Lipstick Building in New York. Cohmad was paid
millions of dollars in fees for bringing new clients to Madoffs investment advisory business.
The investment advisory business was a gigantic Ponzi scheme. Madoff went to great lengths to
keep the investment advisory business separate from the broker-dealer business. He also went at
great lengths to avoid publicity for his investment business. New clients only head about it by
word of mouth from other clients. It was like a private club.
51
The investment advisory business
was located in a secret office space was located on the 17
th
floor of the Lipstick Building in New
York.
Madoff registered as an investment advisory firm with the Securities and Exchange Commission
in 2006. At that time, the investment advisory firm had more than $17 billion in assets under
management for 23 clients.
52
.Although he registered the investment advisory firm with the SEC,
he did not disclose that information to the Financial Institutions Regulatory Authority (FINRA),
that was responsible for examining the investment advisory firm.
53
FINRA is the largest
independent regulator of all securities firms doing business in the United States. Therefore
FINRA was not aware of it during their 2007 examination of the firm.
54
FINRA learned about
investment advisory firm after FBI arrested Madoff on December 11, 2008.
Madoff told his investment advisory clients that he invested in stocks and options using a split
strike conversion strategy that promised would yield above market rates of return. The strategy
consisted of buying a basket of common stocks from the Standard & Poors 100 Index before
an expected run-up in the price of the index, and then selling the stocks after then run-up.
However, instead of buying stocks, the money was deposited in bank accounts, and the Madoff
firm never executed a single securities trade for the investment advisory business. All of the
client account statements, trade confirmations and other documents relating to the investment
advisory business were wholly fabricated and completely fictitious.
55

According to the Securities and Exchange Commission (SEC), at the end of 2008, virtually all of
the billions of dollars in assets of the investment advisory business were missing.
56
The criminal
complaint charged that he was running a $64 billion Ponzi scheme, and that the business had
been insolvent for years. He provided his clients with monthly statements showing transactions
that never occurred. For decades, he used funds from new clients to pay redemptions of existing

51
Bandler, James and Nicholas Varchaver, 2009. How Bernie did it. CNNMoney, April 30.
52
Report of the 2009 Special Review Committee on FINRAs Examination Program in Light of the Stanford and
Madoff Schemes, 2009. FINRA Special Review Committee, September. 50.
53
The information should have been disclosed in FINRAs Investment Advisor Registration Depository (IARD).
54
Report of the 2009 Special Review Committee on FINRAs Examination Program in Light of the Stanford and
Madoff Schemes, ibid., 53-56.
55
Ibid.. 37.
56
SEC Charges Bernard L. Madoff for Multi-Billion Ponzi Scheme,2008. Press Release, Securities and Exchange
Commission, December 11, 2008-293.
961

clients. During the course of the fraud, there were approximately 90,000 disbursements of
fictitious profits to Madoff investors totaling $18.5 billion.
57

In March, 2009, Madoff was charged with eleven felony counts: securities fraud, investment
advisor fraud, mail fraud, wire fraud, three counts of money laundering, false statements,
perjury, filing false SEC statements, and theft from an employee benefit plan.
58
Madoff, age 72,
pled guilty to all counts, and was sentenced to 150 years in jail. This is the longest prison sentences in the
history of financial crimes.
59
He traded his 4,000 square foot penthouse in New York, for a 76 square
foot jail cell.
60
Was it worth it?
Madoff did not operate the scheme alone. The Ponzi scheme involved six or more co-
conspirators within the firm who falsified books and records, created fraudulent accounting
statements, committed bank fraud, and other crimes.

Stanford Financial Group. Robert Allen Stanford (a.k.a. R. Allen Stanford) was born in Mexia,
Texas. He has dual citizenship with the United States and the commonwealth of Antigua. He was
the first American to be knighted by the Governor General of Antigua and Barbuda. Sir Allen
Stanford was known as prominent financier, philanthropist, and professional sports sponsor. In
addition, he was chairman and sole owner of the Stanford Financial Group (SFG) headquartered
in Houston, Texas. The SFG and its affiliated companies: Stanford Capital Management, the
Stanford Group Company, the Stanford International Bank, the Stanford Trust Company, and the
Bank of Antigua, provided wealth management services to customers in some 140 countries.
It is alleged that Stanford and his co-defendants sold Certificates of Deposit (CDs) issued by
Stanford International Bank Ltd. (SIBL) located in Antigua.
61
The CDs were marketed by other
entities of the SFG. Furthermore, it is alleged that Stanford ran a giant Ponzi scheme that robbed
people of $7.2 billion, and jeopardized the economy of the island nation of Antigua and
Barbuda.
62
More than $1.6 billion of those funds were used for personal loans for Stanford.
63

Because the bank was located in Antigua, it was not subject to U.S. regulatory oversight. There
also was concern that the CDs issued by that bank and sold to investors in the United States were
not subject to U.S. federal securities laws. The CDs promised double digit returns, and investors
were told that the funds were invested in safe liquid investments. But, in fact, the funds were
invested offshore in risky illiquid assets such as real estate, stocks, and unsecured loan. In 2005,
the Securities and Exchange Commission expressed concern whether the CDs sold by the SFG
were considered to be securities subject to U.S. laws, but it also stated that Stanfords firm

57
United States Bankruptcy Court Southern District of New York, In re: SIPA Liquidation
Bernard L. Madoff Investment Securities LLC, No. 08-01789 (BRL), Debtor. (Substantively Consolidated):
IRVING H. PICARD, as Trustee for the Liquidation of Bernard L. Madoff Investment Securities LLC,
Adv. Pro. No. 09-1172 (BRL) Plaintiff, v. Estate Of Stanley Chais, et al.
58
United States Department of Justice, Southern District of New York, 2009. United States v. Bernard L. Madoff
(09 Cr. 213, DC) and Related Cases.
59
Perkins, Kevin L., ibid.
60
James Bandler and Nicholas Varchaver, How Bernie did it. op. cit.
61
The co-defendants named in the U.S. v. Stanford, case (Court Docket Number #H-09-342) are Laura Pendergest-
Holt, Gilberto Lopez, Mark Kuhrt, and Leroy King.
62
Chilton, Bart, 2009. Ponzimonium, Remarks by Commissioner Bart Chilton of the Commodities Futures
Trading Commission before the American Bar Associations Committee on Derivatives and Futures Law Students.
March 20; Stanford Financial Group Executives and Former Chairman of Antiguan Bank Regulator Indicted for
Fraud and Obstruction. 2009. United States Department of Justice, Press Release, June 19.
63
United States Department of Justice, Pending Criminal Division Cases, United States v. Robert Allen Stanford et.
al., Court Docket Number: H-09-342. (no date), http://www.justice.gov/criminal/vns/caseup/stanfordr.html
962

engaged in practices that are commonly associated with fraud schemes. Consequently, the Dallas
office of the Financial Industry Regulatory Authority (FINRA), initiated a cause exam.
64

Cause examinations are initiated because of an investor complaint, or surveillance-identified
issue, referrals from other regulators, or for other reasons.
R. Allen Stanford was charged in the United States with bribing the bank regulator in Antigua,
money laundering, and obstructing an investigation by the Securities and Exchange Commission.
He was scheduled to go on trial in federal court in Houston in January 2011. However, U.S.
District Judge David Hittner postponed the trial until January 2012 after finding Stanford
mentally unfit to assist with his defense due to drug problems. Following a six week long trial
that ended in March 2012, Stanford was found guilty on 13 counts including conspiracy to
commit wire and mail fraud, conspiracy to obstruct a U.S. Securities and Exchange Commission
(SEC) investigation, obstruction of an SEC investigation, and conspiracy to commit money
laundering. He was sentenced to 110 years in jail.
65

Bank of Credit and Commerce International (BCCI). BCCI is also known as the Bank of
Crooks and Criminals International. BCCI was founded by Agha Hassan Abedi, who grew up in
India but fled to Pakistan shortly after the two countries were partitioned in 1947.
66
He became a
prominent figure in Pakistan banking. Abedi organized United Bank, which became one of
Pakistans largest banks with overseas branches. In the early 1970s, Pakistani banks were
nationalized, which limited his plans for growth. Consequently, he organized a new bank BCCI
- in Abu Dhabi. Bank of America loaned him $2.5 million and owned 24 to 30 percent of the
shares at different times. Sheik Zayed, whose family owned and ruled Abu Dhabi, deposited at
least $50 million in the bank. By 1990, Sheik Zayed, his family, and the government of Abu
Dhabi were the principal shareholders of the bank.
The parent holding company, BCCI (Holdings), was located in Luxembourg. The bank (BCC
S.A.), was split in two parts: BCCI S.A. in Luxembourg, and BCCI S.A. (Overseas) in the Grand
Cayman Islands. Neither BCCI nor its subsidiaries conducted banking business in Luxembourg.
Instead the used the London office for most of the European and Middle East operations. The
Office in the Grand Cayman Islands was used for business in other countries, including the
United States. BCCI made extensive use of multiple layers of entities, nominees, banks-within-
banks, and other devices to hide the identity and operations form government regulators.
Because of its complex structure, BCCI had no primary bank regulators. It also had no central
bank to bail it out when it got in trouble.
BCCI had about $24 billion in assets. It operated in 73 countries. To put this in perspective,
there are about 270 countries in the world
67
. In the United States, BCCI had agencies licensed in
California, New York, and Florida. It also had interests in other banks, and a commodities
futures firm, Capcom Financial Services. BCCI owned 20 percent of Financial General
Bankshares (FGB), one of the 50 largest bank holding companies in the United States in the
1970s. BCCI also had ownership interests in other banks and financial institutions. It made

64
FINRA Special Review Committee, Report of the 2009 Special Review Committee on FINRAs Examination
Program in Light of the Stanford and Madoff Schemes, op. cit., 25-28.
65
U.S. Department Justice, 2012, Allen Stanford Convicted in Houston for Orchestrating $7 Billion Investment
Fraud Scheme, Justice News, March 6.
66
For additional details, see Gup, Benton E., 1995. Targeting Fraud: Uncovering and Deterring Fraud in Financial
Institutions, ibid. Revised Edition, Chapter 3, BCCI-The Bank of Crooks and Criminals International.
67
Central Intelligence Agency Publications: https://www.cia.gov/library/publications/index.html
963

extensive use of the multiple layers of entities, nominees, banks-within banks, and other means
to hide its identity from government regulators.
Without going into details about the crimes, BCCIs criminality included fraud by BCCI and
BCCI customers involving billions of dollars; money laundering in Europe, Africa, Asia, and the
Americas; BCCIs bribery of officials in most of those locations; support of terrorism, arms
trafficking, and the sale of nuclear technologies; management of prostitution; the commission
and facilitation of income tax evasion, smuggling, and illegal immigration; illicit purchases of
banks and real estate; and panoply of financial crimes limited only by the imagination of its
officers and customers.
68

In 1983, U.S. Customs service learned that a Jordanian arms dealer was one of the largest
customers in BCCIs Miami office. Columbian drug dealers and General Manuel Noreiga were
also BCCI customers making extensive use of their services for money laundering. BCCIs
Florida operations were closed in 1989, and the Federal Reserve issued a cease and desist order
to BCCI (Holdings). In 1991, the Bank of England closed BCCIs operations. A Price
Waterhouse U.K. report stated that BCCI may never have been profitable in its entire history. It
was a giant Ponzi scheme. Various people involved with BCCI were indicted, fined, and served
jail time. Abedi died in 1995.

Securities Traders Frauds
Jerome Kerviels $6.8 billion trading at Socit Gnrale SA: Jerome Kerviel was a 31 year
old trader at Socit Gnrale SA (SocGen), Frances second largest bank. He began working at
SocGen in the back office were trades are processed. He moved from there to an apprentice
trader position. As a trader, Kerviel used European futures to place bets outside his trading limit
for more than two years. During that period the bank profited from unauthorized trades.
However, when the market turned against him, Kerviels trades resulted in huge losses that led to
his downfall in 2008. At his trial, Kerviel said It wasnt me who invented these techniques
others did it, too, These practices were known and recognized by management.
69

The bank turned a blind eye to his unauthorized trades as long as they were making a profit.
SocGen admitted that it had weaknesses in its management and risk control systems. It failed to
follow through on at least 74 internal alerts about Kerviels trading activities. Daniel Bouton, the
banks former chief executive chairman, called Mr. Kerviel an evil genius whose actions had
been a catastrophe that nearly destroyed the 145-year-old bank.
70

At age 33, Kerviel was convicted on all counts of breach of trust, forgery and unauthorized use
of computer systems. The court sentenced him to five years, with two suspended, and barred him
from working in financial services.
71
He was also ordered to repay $6.6 billion (4.9).

Kweku Adobolis $2.3 Billion Trading Loss at UBS AG
Following the $6.8 billion trading loss in 2008 by Jerome Kerviel, a derivatives trader at Societe
Gererale SA, the Financial Services Authority (FSA) warned banks to monitor traders positions

68
The BCCI Affair:1992. A Report to the Committee on Foreign Relations, Senator John Kerry and Hank Brown,
United States Senate, December, 102d Congress, 2d Session Senate, Executive Summary.
69
Clark, Nicola and Katrin Rennhold, 2010.A Socit Gnrale Trader Remains a Mystery as His Criminal Trial
Ends, The New York Times, June 25.
70
Ibid.
71
Clark, Nicola,2010. Rouge Trader at Socit Gnrale Gets 3 Years, The New York Times, October 8.
964

for material risks. The FSA is the regulator of the financial services industry in the United
Kingdom (UK).
Most firms were satisfied that their basic controls and governance surrounding trading, risk
management and settlement activities were effective.
72
UBS Chief Executive Officer Oswald
Gurebel said that UBS had one of the best risk-management units in the industry.
73
He was
wrong! According to the Financial Institutions Regulatory Authority in Washington, the
supervisory system for locating and marking orders at UBS Securities LLC was significantly
flawed.
74
A smart employee, Kweku Adoboli who understood the risk-management system,
figured out a way to abuse it. UBS AG, is a Swiss financial services firm that provides wealth
management, investment banking, and other services to clients worldwide. It has total assets of
more than $1.4 billion, offices in more than 50 countries, and it employs more than 65,000
people.
75

Since 2009, and prior to the investigation of Kweku Adoboli that is discussed below, the FSA
has punished at least eight individuals or firms for improper trading activity, including making
unauthorized trades and fudging or fabricating numbers in an attempt to conceal investment
losses.
76
It fined UBS $12.8 million for unauthorized trades in their international wealth
management division.
77

Kweku Adoboli, was a 31 year old junior securities trader at UBSs Delta One desk in London.
Delta Ones trading activities included synthetic exchange traded funds (EFTs), and other
securities. Because UBS rules prohibit speculative trading, Adoboli used fictitious trades in ETFs
to hide his speculative gains and losses.
78
He established a mirror trading book using phantom
positions in ETFs and securities that mirrored the indexes to match the parallel trading books in
order to minimize glaring gains and losses.
Adobolis losses increased as the European debt crisis worsened, and the U.S. government debt
was downgraded by Standard and Poors. In early September 2011, risk control officers at UBS
began to question some of Adobolis trades. Subsequently, the losses were revealed. In 2011,
Adoboli was charged with fraud and false accounting in connection with unauthorized
speculative trading in various S&P 500, DAX, and Eurostoxx index futures during a three month
period that resulted in losses of $2.3 billion.

Nick Leesons $1.3 billion Loss for Barings
79

Francis Baring & Co. was established in London in 1792, making it the oldest merchant bank in
the United Kingdom. It built a reputation for financing corporate assets in emerging markets,
including the Louisiana Purchase in the in early 1800s. It almost failed in 1890 because of bad

72
Fornado, Lindsay and Ben Moshinsky, 2011. Regulators Begin Probe of UBS Trading Loss, Bloomberg,
September 16.
73
Broom, Giles, 2011. Ermotti Inherits UBS in Disarray, Bloomberg, September 26.
74
Gallu, Joshua, 2011. UBS Wil Pay $12M for Supervision Failures, Bloomberg, October 25.
75
See UBS, About Us: http://www.ubs.com/global/en/about_ubs/about_us.html
76
Enrich, David, Cassell Bryan-Low, and Sara Schaefer Muoz, 2011. U.K. Sets Its Sights On Rogue Traders,
The Wall Street Journal, September 22, C1, C2.
77
Ibid.
78
Mollenkamp, Carrick and Dana Cimilluca, 2011. UBS Loss Reveals Gaps, The Wall Street Journal, September
20, C1, C2.
79
For additional information, see Gup, Benton E.,1998. Bank Failures In The Major Trading Companies of the
World: Causes and Remedies, Quorum Books, Westport Ct., 50-51. Also see: Leeson, Nicholas William, Edward
Whitley, 1996. Rouge Trader: How I Brought Down Barings Bank and Shook the Financial Worlds, Little Brown,
London, UK.
965

loans in Argentina. Subsequently, as part of the restoration, the name was changed Barings
Brothers and Co. in the late1800s.
Barings Brothers and Co. had little experience when it in trading when it hired a 28 year old
trader, Nicholas W. Leeson, in 1992. By 1994, Nick was the director, and chief trader of Barings
Futures Singapore. In a twenty eight day period, he traded stock index futures in the Singapore
International Monetary Exchange (SIMEX) to bet more than $1 billion on the future direction of
the Nikkei Index of Japanese stocks. He lost $1.4 billion, and Barings had capital of $500
million. Subsequently, Barings failed, and Nick Leeson was sentenced in Singapore to a jail term
of six and one-half years.

Yasuo Hamanakas $2.6 Billion Loss for Sumitomo Corp
80

Hamanaka was the chief copper trader at Sumitomo Corp. headquartered in Japan. Over a ten
year period, he executed $2.6 billion worth of unauthorized off-the- book trades both on the
London Metal Exchange (LME) and outside the LME. By 1995, Sumitomo controlled 100% of
the copper in LME approved warehouses, which forced copper prices up, and generated large
profits. At one point in time, Hamanaka controlled 5 percent of the worlds copper, and he was
given the nickname "Mr. Five Percent."
His scheme unraveled after the LME imposed trading limits on copper in 1996, and copper
prices declined. When the scandal was uncovered, the cost to Sumitomo was $2.6 billion.
Sumitomo accused several senior UK metals brokers who worked for Credit Lyonnais Rouse
the banks London metals brokerage arm for turning a blind eye to the trades and dishonestly
assisting Hamanaka in manipulating copper prices.
81
In the United States, the Commodities
Futures Trading Commission (CFTC) charged Global Minerals and Metals, Merrill Lynch &
Co., and others with aiding, abetting, and assisted the worldwide manipulation and attempted
manipulation of copper prices.
82

In 1997, Hamanaka was sentenced to eight years in prison for fraud in manipulating copper
prices and forging his superiors signatures on letters to other dealers. He was released from
prison in 2005.

Toshihidi Iguchi Lost $1.1 Billion Trading Bonds for Daiwa Bank
Toshihidi Iguchi was a bond trader at Daiwa Banks New York office. Daiwa Bank Group is
headquartered in Osaka, Japan. Iguchi lost $1.1 billion in unauthorized trades of United States
bonds over an 11 year period (1984-1995). He covered his losses by selling bonds in customers
accounts and then forged trading slips and documents to make it appear is if the securities were
still there.
83
Iguchi sent a 30 page letter to the President of Daiwa Bank confessing his losses.
Iguchi, age 45, plead guilty to fraudulently doctoring bank records and embezzlement. In 1996,
hhe was fined $2 million and sentenced to four years in jail. He also was ordered to pay $2
million in fines and $570,000 in restitution.
84

80
What Happened: A Copper Trader Losses Big, 1997. The New York Times, January 2.
81
Bowers, Simon, 2004, UK Metal Brokers Accused of Rouge Trade Complicity, The Guardian, October 18.
http://www.guardian.co.uk/business/2004/oct/04/japan.internationalnews
82
Commodities Futures Trading Commission, Enforcement Proceeding, 1999. Release: #4265-99 (CFTC Docket #
99-11), May 20.
83
WuDunn, Sheryl, 1995. International Business: Daiwa Concedes It Told Trader To Keep Dealing, The New
York Times, Business Day, October 21.
84
Former Daiwa Trader Sentenced In Cover-Up of $1.1 Billion Loss,1996. New York Times, Business
Day, December 17.
966

Daiwa Bank was aware of Iguchis losses as early as 1993. In that year, Daiwa Bank hid some
of its records and temporarily removed Iguchi from its New York Branch in order to pass a
Federal Reserve Inspection.
85
Daiwas officials acknowledged that they allowed Iguchi to
continue trading weeks after he reported large losses.
86
The bank was required by federal and
state banking regulations to report any significant banking improprieties and other actions that
significantly affect the banks previous financial statements. However, the bank did not do so.
The Federal Reserve reported the matter to the Japanese Finance Ministry, which took no action
against the bank. In fact, the Japanese government was aware of the fraud six weeks before the
Federal Reserve reported it to them.
87
In February 1996, Daiwa Bank was fined $340 million
and barred from doing business in the United States.
88
However, in 2003, the Federal Reserve
changed its position, terminated in enforcement action, and allowed Daiwa to establish a
representative office in New York, New York.
89

Common Features of the Frauds
The common features of the frauds at Taylor, Bean & Whitaker (TBW), Bernard L. Madoff
Investment Securities LLC, Stanford Financial Group, and the BCCI fraudulent schemes are that
the owners of these closely held institutions were the primary fraudsters, and they were richly
rewarded for their efforts. Farkas, Madoff, Stanford, and Abedi successfully enlisted other
people to help them commit the frauds. The motives of some of the accomplices are not clear.
Nevertheless, the fact that the perpetrators were making large profits helped to insulate them
from investigation into the legitimacy of the sources of those profits. The perpetrators made
extensive use of both covert and complex organizations to shield their frauds from detection by
law enforcement agencies and regulators. If the regulators are not specifically looking for frauds,
the odds are that they will not find them.
The common features of the securities traders that committed frauds (Jerome Kerviel at Societe
Gererale, France; Kweku Adoboli at UBS AG, Switzerland; Nicholas W. Leeson at Barings
Brothers and Co. UK; Yasuo Hamanaka at Sumitomo Corp.; and Toshihidi Iguchi at Daiwa
Bank, Japan) are that most of them were located at distant locations from their corporate
headquarters. The affected companies turned a blind eye as to what the traders did as long as
they were making a profit, and/or could cover up their losses. They had weak operational risk
controls.

Conclusion
Operational risk has played a major role in bank frauds. Federal Reserve Governor Susan Bies
said, History has shown that foreign affiliates are particularly vulnerable to internal control
problems. The failure of Barings Bank and the losses incurred by institutions such as Daiwa
Bank and, most recently, Allied Irish Bank were directly related to a failure in proper control of
foreign affiliates. In most cases these losses could have been avoided by ensuring that the banks
control procedures--for example, segregation of duties and the conduct of thorough and

85
Daiwa Says Regulators Saw Irregularities in 93, 1995. The New York Times, Business Day, October 7.
86
WuDunn, Sheryl, 1995. Key Dates in Scandal, NYDailyNews.com. November 3.
87
WuDunn, Sheryl, 1995. Japanese Delayed Letting U.S. Know of Big Bank Loss, The New York Times,
October 12.
88
Former Daiwa Traded Sentenced in Cover-Up of $1.1 Billion Loss, ibid.
89
Federal Reserve Joint Press Release, 2003. Regulators Announce Termination of Enforcement Orders
Against Daiwa Bank, May 5.

967

independent internal audits--were functioning properly. Even under a risk-focused approach,
supervisors must ensure that bank management is providing adequate controls for all aspects of
the business, even those that appear relatively minor in scope.
90

Federal Reserve Chairman, Alan Greenspan said ...the whole system of supervision proceeds
upon the basis of trust, whether in terms of the veracity of representations or reports filed by
management or transparency with regard to any material developments affecting the financial
condition of the institution. Supervisors need to trust the ability of bank management to carry out
their duties in a responsible and honest manner with adherence to systems and operational
controls to ensure the safe and sound conduct of business.
91
That didnt happen in these cases.
In the case of Diawa Bank, where trader Iguchi lost $1.1 billion, a court in Japan found that 11
senior executives of the bank were responsible for their failure to supervise their staff and detect
fraud.
92
The corporate culture supported the belief that someone making a lot of money must
know what he is doing, and we are not going to question it. They also conspired to cover up the
losses once they were discovered. They were ordered to pay $775 million in damages to their
own bank. The former president of the New York branch of Diawa was ordered to pay more than
$500 million. If penalizing top managers who failed in their duties to detect fraud was given
greater publicity, it might serve as an incentive to promote better corporate governance with
respect to deterring and detecting frauds by insiders.

References
Alabama State Banking Department Takes Possession of Colonial Bank, Press Release, State
of Alabama State Banking Department, August 14, 2009.
Bandler, James and Nicholas Varchaver, How Bernie did it. CNNMoney, April 30, 2009.
Barnett, Cynthia, The Measurement of White-Collar Crime Using Uniform Crime Reporting
(UCR) Data, U.S. Department of Justice, Federal Bureau of Investigation, Criminal Justice
Information Services (CJIS) Division, March 2002.
Barth, J. R., Lin, C., Lin, P. & Song, F. M. 2009. Corruption in bank lending to firms: Cross-
Country micro evidence on the beneficial role of competition and information sharing.
Journal of Financial Economics, 91 (3). 361-388.
Beasley, Mark S., Bruce C. Branson, and Bonnie V. Hancock, Developing Key Risk Indicators
to Strengthen Enterprise Risk Management, The Committee on Sponsoring Organizations
of the Treadway Commission (COSO), Durham N.C., December, 2010.
Beck, T. Demirg-Kunt, A., & Levine, R. 2006. Bank supervision and corruption in lending.
Journal of Financial Economics, 53 (8)
Bies, Susan S., Remarks by Governor Susan S. Bies At the World Bank Group's Finance Forum
2002, Chantilly, Virginia, Banking Supervision and Its Application in Developing
Countries, June 20, 2002.

90
Bies, Susan S., 2002.Remarks by Governor Susan S. Bies At the World Bank Group's Finance Forum 2002,
Chantilly, Virginia, Banking Supervision and Its Application in Developing Countries, June 20. The fraud at
Allied Irish Bank was by a rouge currency trader, John Rusnak, who lost who lost $691 million over a period of
five years. For additional details, see: AIB fraud 'going on for years',2002 BBC News, Business, February 20.
91
Greenspan, Alan, 1996. Statement to the Congress- Allen Greenspan on the Daiwa Bank Scandal, New Federal
Reserve Initiatives to Regulate U.S. Branches of Foreign Banks, Federal Reserve Bulletin, February.
92
Grammeraticas, Damian. 2000,Bank Bosses to Pay $775m Fraud Charges, BBC News, September 20.
968

Brooks, Douglas W. Creating a Risk Aware Culture, Enterprise Risk Management: Todays
Leading Research and Best Practices for Tomorrows Executives, John Fraser and Betty J.
Simkins, Editors, John Wiley & Sons, New Jersey, 2010, pp. 87.
Broom, Giles, Ermotti Inherits UBS in Disarray, Bloomberg, September 26, 2011.
Bowers, Simon, UK Metal Brokers Accused of Rouge Trade Complicity, theguardian, October
18, 2004.
Carr, Susan Latham, Audit: It wasn't just Taylor Bean that led to Platinum bank's failure,
Ocala.com, September 19, 2011.
Central Intelligence Agency Publications: https://www.cia.gov/library/publications/index.html
Chilton, Bart, Ponzimonium, Remarks by Commissioner Bart Chilton of the Commodities
Futures Trading Commission before the American Bar Associations Committee on
Derivatives and Futures Law Students. March 20, 2009.
Cimilluca, Dana, Deborah Ball, and Carrick Mollenkamp, UBS Raises Tally on Losses, The
Wall Street Journal, September 19, 2011, C1, C3.
Clark, Nicola, Rouge Trader at Socit Gnrale Gets 3 Years, The New York Times, October
8, 2010.
Clark, Nicola and Katrin Rennhold, A Socit Gnrale Trader Remains a Mystery as His
Criminal Trial Ends, The New York Times, June 25, 2010.
Committee of Sponsoring Organizations of the Treadway Commission, Enterprise Risk
Management- Integrated Framework, Executive Summary, September 2004, p. 5.
Commodities Futures Trading Commission, Enforcement Proceeding, Release: #4265-99 (CFTC
Docket # 99-11), May 20, 1999.
Daiwa Says Regulators Saw Irregularities in 93, The New York Times, Business Day,
October 7, 1995.
Eaglesham, Jean, Financial Crimes Bedevil Prosecutors, The Wall Street Journal, December 5,
2011.
Enrich, David, Cassell Bryan-Low, and Sara Schaefer Muoz, U.K. Sets Its Sights On Rogue
Traders, The Wall Street Journal, September 22, 2011,
FBI, 2008 Financial Crimes Report, 2009
FBI, Financial Institution Fraud and Failure Report, Fiscal Years (FY) 2006 and 2007. August
2011.
Federal Reserve Press Release, Approval of Proposal of Daiwa Bank Limited, February 28,
2003.
Federal Reserve Joint Press Release, Regulators Announce Termination of Enforcement Orders
Against Daiwa Bank, May 5, 2003.
Former Chairman of Taylor, Bean & Whitaker Convicted for $2.9 Billion Fraud Scheme that
Contributed to the Failure of Colonial Bank. Department of Justice, Press Release, April
19, 2011.
Former Chairman of Taylor, Bean & Whitaker Convicted for $2.9 Billion Fraud Scheme That
Contributed to the Failure of Colonial Bank. FBI Washington Field Office, Press Release,
April 19, 2011.
Former Chairman of Taylor, Bean & Whitaker Sentenced to 30 Years in Prison and Ordered to
Forfeit $38.5 million. FBI Washington Field Office, Press Release, June 30, 2011.
Former Colonial Bank Senior Vice President Sentenced to Eight Years in Prison for Fraud
Scheme,
1
FBI, Washington Field Office, Press Release, June 17, 2011.
969

Former Daiwa Trader Sentenced In Cover-Up of $1.1 Billion Loss, New York Times,
Business Day, December 17, 1996.
Former TBW CEO Sentenced to 40 Months in Prison for Fraud Scheme, FBI, Washington
Field Office, Press Release, June 21, 2011.
Former Treasurer and President of Taylor, Bean & Whitaker Each Sentenced to Prison for
Fraud Scheme, FBI Washington Field Office, June 10, 2011.
Fornado, Lindsay and Ben Moshinsky, Regulators Begin Probe of UBS Trading Loss,
Bloomberg, September 16, 2011.
Fraser, John and Betty J. Simkins, Editors, Enterprise Risk Management: Todays Leading
Research and Best Practices for Tomorrows Executives, John Wiley & Sons, New Jersey,
2010.
Gallu, Joshua, UBS Will Pay $12M for Supervision Failures, Bloomberg, October 25, 2011.
Greenspan, Alan, Statement to the Congress- Allen Greenspan on the Daiwa Bank Scandal,
New Federal Reserve Initiatives to Regulate U.S. Branches of Foreign Banks, Federal
Reserve Bulletin, February, 1996.
Gup, Benton E., Bank Failures In The Major Trading Companies of the World: Causes and
Remedies, Quorum Books, Westport Ct., 1998.
Gup, Benton E., Basel II: Operational Risk and Corporate Culture, Benton E. Gup, Editor,
Corporate Governance in Banking: A Global Perspective, Edward Elgar, Northampton
Mass., 2007, pp. 134-150.
Gup, Benton E., Targeting Fraud: Uncovering and Deterring Fraud in Financial Institutions,
Probus Publishing Company, Chicago, IL, 1995.
Historic Bank Fraud Case, FBI Podcasts and Radio, July 15, 2011.
In the Mater of Colonial Bank Montgomery Alabama, Notice Of Charges And Of Hearing,
Federal Deposit Insurance Corporation, FDIC-09-402c & b.
Irans Biggest Bank Fraud of $2.6 Billion unearthed, 19 arrested. Jafria News.Com. September
19, 2011.
Leeson, Nicholas William, Edward Whitley, Rouge Trader: How I Brought Down Barings Bank
and Shook the Financial Worlds, Little Brown, London, UK, 1996.
OKeefe, Brian. The Man Behind 2009s Biggest Bank Bust, CNNMoney.com. October 12,
2009.
Material Loss Review of Colonial Bank, Montgomery, Alabama, April 2010. Office of the
Inspector General, Federal Deposit Insurance Corporation, Report No. MLR-10-021, p.2.
Mollenkamp, Carrick and Dana Cimilluca, UBS Loss Reveals Gaps, The Wall Street Journal,
September 20, 2011, C1, C2.
Perkins, Kevin L., Assistant Director Criminal Investigative Division, FBI, Statement Before the
Senate Judiciary Committee, Washington, D. C. September 22, 2010.
Principles for Sound Management of Operational Risk, Basel Committee on Banking
Supervision, June 2011.
Report of the 2009 Special Review Committee on FINRAs Examination Program in Light of
the Stanford and Madoff Schemes, FINRA Special Review Committee, September 2009.
SEC Charges Bernard L. Madoff for Multi-Billion Ponzi Scheme, Press Release, Securities
and Exchange Commission, December 11, 2008
Securities and Exchange Commission V. Catherine I Kissick, United States District Court for the
Eastern District of Virginia, Alexandria Division, Civil Action File No. 1:11cv:215, March
2, 2011.
970

Stanford Financial Group Executives and Former Chairman of Antiguan Bank Regulator
Indicted for Fraud and Obstruction. United States Department of Justice, Press Release,
June 19, 2009.
The BCCI Affair: A Report to the Committee on Foreign Relations, Senator John Kerry and
Hank Brown, United States Senate, December 1992, 102d Congress, 2d Session Senate,
Executive Summary.
UBS, About Us: http://www.ubs.com/global/en/about_ubs/about_us.html
United States Bankruptcy Court Southern District of New York, In re: SIPA Liquidation
Bernard L. Madoff Investment Securities LLC, No. 08-01789 (BRL), Debtor. (Substantively
Consolidated): IRVING H. PICARD, as Trustee for the Liquidation of Bernard L. Madoff
Investment Securities LLC,
Adv. Pro. No. 09-1172 (BRL) Plaintiff, v. Estate Of Stanley Chais, et al.,
United States Department of Justice, United States v. Lee Bentley Farkas
Court Docket Number: 1:10-cr-200. No date given
United States Department of Justice, Pending Criminal Division Cases, United States v. Robert
Allen Stanford et. al., Court Docket Number: H-09-342. No date given.
United States Department of Justice, Southern District of New York, United States v. Bernard L.
Madoff (09 Cr. 213, DC) and Related Cases. 2009.
United States of America vs. Catherine Kissick, Case Number 1:11CR88 , United States District
Court for the Eastern District of Virginia, Honorable Leonie M. Brinkema, Position of
the United States With Respect to Sentencing, June 17, 2011.
Weardan, Graeme. The Biggest Rouge Traders in History, thegardian, January 24, 2008.
What Happened: A Copper Trader Losses Big, The New York Times, January 2, 1997.
WuDunn, Sheryl, Japanese Delayed Letting U.S. Know of Big Bank Loss, The New York
Times, October 12, 1995.
WuDunn, Sheryl, International Business: Daiwa Concedes It Told Trader To Keep Dealing,
The New York Times, Business Day, October 21, 1995.
WuDunn, Sheryl, Key Dates in Scandal, NYDailyNews.com. November 3, 1995.
Zheng, X., El Ghoul, S., Guedhami, O., Kwok, C., Collectivism and Corruption in Bank
Lending, Working Paper presented at the Midwest Finance Association Annual Meeting,
February 23, 2012.

971

Macroeconomic Factors, Market Volatility and the Performance of Large
Cap Funds During the 2008-2011 Period

Dr. Abhay Kaushik
Assistant Professor of Finance

Department of Accounting, Finance, and Business Law
Radford University,
akaushik@radford.edu

972

Abstract
This study evaluates 744 large cap funds over the period 2008-2011. The time period of this
study is very critical as this was the worst global recession since the great depression. We
specifically analyze the performance of large cap funds because they tend to be more stable and
they are supposed to be better prepared to position themselves against the market volatility that is
mainly driven by fear factor and other macroeconomic factors. Results of this study show that
large cap funds cannot be used as a hedging tool during the severe market downturn. Results also
suggest that most of the negative performance is a function of extreme market volatility and not
due to fund specific attributes.
Key Words: Mutual Funds; Performance
JEL Classification: C15

I. Introduction
According to the National Bureau of Economic Research (NBER), the U.S. economy entered
into recession in December of 2007 and this recession ended in June of 2009. Even though
recession officially ended in middle of 2009, still the financial markets kept experiencing
slowdown and a very high level of volatility. On an average, over the period 1/2008 12/2011,
S&P 500 Index, a well-known market benchmark, experienced up or downturn 43 times out of
48 months whereas the Dow Jones Industrial Average, the portfolio of 30 blue chip firms and a
worldwide indicator of financial markets, experienced up or downturn 40 times during this same
period
93
. According to the financial press, for market participants the period from 2008 to 2011
is one of the most eventful periods of their lifetime. Middle of 2008 saw the meltdown in the
financial system starting with the collapse of Lehman Brothers and Bear Sterns followed by the
possible bankruptcy of AIG. Other banks, like the Citi Corporation, Wachovia, and other
financial situations were threatening a wide spread panic in the banking sector and across the
entire financial system. In order to stabilize the system and to avoid panic, the U.S. government
passed the Bailout Bill in 2008 that allowed the U.S. Treasury to create a $700 billion Troubled
Asset Relief Program (TARP) to help those failing banks. For the first time, rating of the U.S.
dropped from AAA to AA+ and the Dow Jones Industrial Average dropped by 600 points the
next trading day.
European nations that are a part of European Union were also facing daunting challenges during
the same time period. The market started casting doubts on the European sovereign debt
beginning of September 2009. The first trigger was pulled on December 8, 2009 when Greece
became the first European nation whose sovereign debt was downgraded by Fitch rating agency.
In order to maintain stability in the Euro zone and to avoid default by Greece, European nations
and the IMF announced 110 billion bail-out plan for Greece. While the world financial markets
and the European financial markets in particular were hoping to edge out of crisis, sovereign debt
crisis moved to Portugal. In May 2010, European Union nations pledged to add nearly $1 trillion
rescue package to support the struggling European Union nations. The contagious sovereign debt
crisis kept moving and by June 2010 Spain became the third European Union nation facing
default. Ireland requested $100 billion bailout in November 2010 followed by Portugal in April
2011. Come July 2011, another European Union member nation, Italy, started showing troubles
with its sovereign debt.

93
Edwards and Caglayan (2001) use 1% monthly change (12% annual equivalent) as a proxy for upturn and
downturn (Bull and Bear) signs.
973

In the midst of all these debt crises and domestic political problems brewing in the U.S.,
financial markets witnessed an upheaval that has not been witnessed since the great depression.
Equity markets suffered one of the largest single day drops and volatility became a common
synonym with the equity markets around the globe.
The main objective of this research is to analyze the performance of large cap equity funds
during this tumultuous time period. One of the main reasons people invest in mutual funds is to
get diversification. The reason this research is focused on large cap funds because large cap firms
are supposed to be stable firms with a strong financial track record. We believe that large cap
funds are better prepared to position themselves against the market volatility that is mainly
driven by fear factor and other macroeconomic factors. Our belief is supported by the end of year
2011 newsletter/prospectus release by BlackRock, a large investment fund with roughly $3.5
trillion assets under management that says Investing in large-capitalization stocks is the most
efficient way to participate in earnings from large US companies. These stocks have the potential
for more stable earnings than that of small-or mid-capitalization stocks, and their prices tend to
be less volatile. Since we are analyzing funds as opposed to individual stocks, it is interesting to
evaluate whether active fund management has selectivity skills to rebalance the portfolios and
can investors rely on the active fund management to hedge their risks during a volatile period?
Since large cap funds may differ significantly from each other therefore we conduct cross-
sectional analysis to evaluate the impact of funds attributes on the abnormal performance of
these funds.
The remainder of this paper is as follows. Section 1 reviews the existing literature, Section 2
describes the data and descriptive statistics, methodology is explained in Section 3, results are
presented in Section 4, and Section 5 concludes this paper.

II. Literature Review
Results of performance of actively managed mutual funds are mixed at best and inconclusive at
worst. Staring with Jensen (1968), many researchers (for example Henriksson (1984), Ippolito
(1989), Busse et al. (2010) to name a few) have documented that actively managed funds do not
outperform the passive indices after deducting fees and expenses. However, there are quite a few
studies that support the value added effects of active fund management. Grinblatt and Titman
(1992) analyzed 279 funds and conclude persistence of performance in their sample funds. The
major implication of their study was the ability of funds to consistently outperform the market
indices. Their results suggest superior selectivity skills of fund managers. Similarly, Coggin,
Fabozzi, and Rahman (1993) also found evidence of strong selectivity skills in active
management. Their findings show that, on average, active funds generate superior alpha
regardless of the benchmarks. Gruber (1996) documented the smart money effect. His results
show that performance is persisted for certain funds and rational investors are able to see it. In
other words, superior performance can only continue if fund managers have strong selectivity
skills and if so then smart investors will know this ability and they will pour their funds into
those funds. Gruber (1996) findings further strengthened by the findings of Zheng (1999) who
examined 1,826 diversified mutual funds and documented that funds that receive higher flow of
fund in the previous period tend to perform superior in the subsequent periods compared to funds
that experienced outflow of fund in the previous period. Findings of Carhart (1997) suggest that
fund managers have no skills and any abnormal performance is only because of momentum
effect. His findings suggest that if momentum is the only reason to earn superior performance
than why to give money to fund managers and incur expenses when any individual can beat the
974

passive index by following cheap momentum strategies. Huij and Post (2011) analyze emerging
market funds and find superior performance by those funds. Analyzing roughly 137 emerging
market funds over the 1993 to 2006 period, their research suggests that emerging market funds
not only outperform the passive benchmarks, but also beat their domestic U.S. counterparts and
this performance is only slightly contributed by momentum effects. They also document
persistence of performance in these funds. Another strand of research analyzed the performance
of funds that have narrow focus and find superior performance by those funds. For example,
Kacperczyk al. (2005) show that funds that have narrow industry focus tend to earn superior
alpha compared to a broad diversified portfolio. Several studies have documented that large
funds perform better because they have abilities to enjoy economies of scale and these funds also
have more equity cushion in case of adverse market conditions or a sudden redemptions by the
investors (for example please refer to Amenc, Curtis, and Martellini (2004), De Souza and
Gokcan (2003) among others).
II.1 Data
Majority of the data is taken from Morningstar Direct database. First all the funds are screened
on domicile basis. Any fund that is U.S. domiciled and classified as U.S. equity fund under the
Morningstar Category is selected as the sample fund. Since this research is focused only on the
performance of large cap funds therefore funds that belong to large cap core or large cap growth
or large cap value categories are selected as the sample funds. It is also important to separate
equity funds from balanced equity funds therefore if a fund has less than 90 percent investment
in stocks is screened out from the sample selection process. Further, any fund that is classified as
specialty fund or index fund or institutional fund or fund of funds is also removed from the
selection process. Following the literature on mutual funds, we selected the oldest share class as
the sample fund in case a fund has multiple share classes in the dataset
94
. Finally, consistent with
the literature on mutual funds, we removed all those funds that do not have at least 36 monthly
observations. Altogether 744 funds are selected in the final sample. Selection of benchmark is
very important in order to analyze the true abnormal performance of funds and existing literature
is divided on what should be the most appropriate benchmark. Morningstar analysts follow
closely each funds prospectus and holdings and recommend the choice of the most appropriate
benchmark. This study uses the benchmark suggested by Morningstar analysts. Finally, monthly
factors mimicking the passive portfolios such as Fama-French factors --SMB (difference in
returns between small and large capitalization stocks), HML (difference in returns between high
and low book-to-market stocks), and Carhart momentum factor, MOM (difference in returns
between stocks with high and low past returns), are taken from Kenneth Frenchs website.
II.2 Descriptive Statistics
Large cap core category has the highest number of funds (238) whereas the minimum sample
funds are from large cap value category (191). Altogether, 744 funds existed in the sample over
the period of this study. Interestingly, average asset under management is also highest in large
cap core category. Expense ratio is pretty consistent across all three categories at around 1.15
percent per year. Turnover ratio varies across all three categories. It seems growth funds
experienced the highest rebalancing of portfolios. On average, a manager stays roughly 6.50
years with a fund and this number is pretty consistent across all three categories.

94
Mutual funds are allowed to issue multiple classes of fund, however, these different share classes have claims on
the same assets and they differ only in terms of their load and fee structure.
975

Table 1: Descriptive Statistics
The table shows descriptive statistics across all three categories over the period 2008-2011. TNA
is the average monthly net asset under management, Expense Ratio is the average annual
expenses charged by the fund including 12b1 fee, Turnover ratio is the minimum of aggregated
sales or aggregated purchases of securities divided by the average 12-month total net assets of
the fund, Managerial Tenure is the average number of years managers stay with the fund, and
TTOP is the percentage of investment by a fund in its top 10 percent holdings.
Large Cap Value Large Cap Growth Large Cap Core
Number of Funds 191 315 238
TNA (million $) $1,211.02 $1,357.42 $1,517.44
Expense Ratio 1.16% 1.21% 1.12%
Turnover Ratio 81.05% 91.05% 62.39%
Managerial Tenure (years) 6.60 6.52 6.56
TTOP 31.33% 33.57% 31.32%
Most of the funds within their category may have homogenous fund objective, but they can differ
significantly from each other in terms of their fund specific attributes. Cross-sectional analysis
does show the impact of these fund specific variables on abnormal performance of funds, but it is
equally important to know how these variables react to each other. In order to have a better
understanding of cross-sectional attributes of the funds, we estimate the correlation among these
key fund specific variables across all three categories. Table 2 panel A shows that most of these
variables under large core category are either negatively and less correlated to each other. For
example, total net asset is negatively correlated with expense ratio, turnover ratio, and funds
investment in its top 10 percent holdings whereas it is positively correlated with manager tenure.
Similar correlations are observed for large cap growth and large cap value categories. Majority
of these funds have negative and/or very low correlations. In fact, this lack of correlation is
interesting as it suggests that a funds performance is very much dependent on the interplay
among its key attributes.
Table 2: Correlation Matrix
The table shows correlation among key fund specific variables. Description of those variables is
explained in Table 1 above. A positive coefficient suggests positive correlation between those
variables whereas a negative coefficient shows the negative correlation between those variables.
A coefficient of +1 is the perfect positive correlation and a -1 coefficient shows the perfect
negative correlation between those variables.

Panel A: Large Cap Core Funds
TNA Expense Ratio Turnover Ratio Manager Tenure TTOP
TNA 1 -0.216594512 -0.0796 0.1613 -0.0678
Expense Ratio -0.2166 1 0.0977 0.0313 0.0051
Turnover Ratio -0.0796 0.0977 1 -0.1854 -0.1418
Manager Tenure 0.1613 0.03130 -0.1854 1 0.1041
TTOP -0.06779 0.0051 -0.1418 0.1041 1

976

Panel B: Large Cap Growth Funds
TNA 1 -0.2105 -0.0741 0.1683 -0.0711
Expense Ratio -0.2105 1 0.1122 0.0409 0.1270
Turnover Ratio -0.0741 0.1122 1 -0.1108 -0.1280
Manager Tenure 0.1683 0.0409 -0.1108 1 0.0977
TTOP -0.0711 0.1270 -0.1280 0.0977 1

Panel C: Large Cap Value Funds
TNA 1 -0.2996 -0.1210 0.1290 -0.0585
Expense Ratio -0.2996 1 0.0805 -0.0087 0.0849
Turnover Ratio -0.1210 0.0805 1 -0.2597 0.0009
Manager Tenure 0.1290 -0.0087 -0.2597 1 0.2301
TTOP -0.0585 0.0849 0.0009 0.2301 1

III. Methodology
We use the single factor market model and the four factor model to estimate abnormal
performance of these funds. The excess fund return is regressed on the excess market return to
capture the effect of market risk on the alpha (abnormal performance) of a fund. However, a vast
amount of mutual fund literature exists that shows that market risk alone does not explain the
true performance of active management and it is important to control other known biases such as
small stocks minus big stocks premium, high market to book minus low market to book premium
and the momentum effects. In other words, if these biases are not controlled then a fund manager
can be considered to be a superior fund manager by simply following cheap equity premiums
and momentum strategies.

Jensens (1968) alpha is perhaps the best known primary model.
r
it
- r
ft
=
it
+
i
*RMRF
t
+
i, t
(1)
Where:
r
it
- r
ft
is the excess return on fund i over the Treasury bill rate,
i
is the measure of the portfolio's performance (Jensen's Alpha),
RMRF
t
= RM
t
- RF
t
is the excess return on the market, and
i
= is the unconditional measure of risk.

The four-factor model of Carhart (1997), which adjusts fund excess return for the Fama-French
(FF) factors SMB (the difference in returns between small and large capitalization stocks), HML
(the difference in returns between high and low book-to-market stocks ) and Carharts
momentum effect.

r
it
- r
ft
=
i
+
1i
* RMRF
t
+
2i
* SMB
t
+
3i
*HML
t
+
4i
*MOM
t
+ 5i*INDRF
t
+
i, t
(2)
Where :
RMRF
t
is the excess monthly return (market return net of 3 month T-Bill return) on
comparative benchmarks explained in Table 2,
SMB
t
is the difference in returns between small and large capitalization stocks and HML
t

is the difference in returns between high and low book-to-market stocks.
MOM
t
is the difference in returns between stocks with high and low past returns.
977

a. Cross Sectional Analysis
It is well documented in the mutual fund literature that fund specific variables are very critical to
funds abnormal performance (Chordia (1996), Carhart (1997), Wermers (2000 and 2003),
Nanda, Narayanan, and Warther (2000) among others). Moreover, different funds in the same
category may follow same investment objective, but they differ significantly from each other in
terms of their management style and cost structure therefore it is more important to do cross-
sectional analysis to estimate abnormal performance of a fund across its fund specific variables.
We use Fama and MacBeth (1973) methodology to conduct cross-sectional analysis. Consistent
with Carhart (1997), Brown, Harlow, and Starks (1996), among others, we use 1 year of prior
monthly returns to estimate the beta loadings of the four-factor model, and use this rolling
window to estimate alpha for each month
95
. This method considers that beta changes over time,
and time-varying alphas incorporate this effect.
We use rolling window of past 12 months to estimate beta loadings by using equation (2)
above and then use the following model to estimate alpha for each fund each month with 12
th

month as first period with rolling alpha.
it
= r
it
- r
ft
-
1i
* RMRF
t
-
2i
* SMB
t
-
3i
*HML
t
-
4i
*MOM
t

Finally, we use Fama and MacBeth (1973) methodology and regress
it
on fund specific
variables such as expense ratio, turnover ratio, size, funds investment in its top 10 percent
holdings, and average tenure of manager with the fund.
Model
it
=
0
+
1
Expense Ratio
it
+
2
Turnover Ratio
it
+
3
Size
it
+
4
TTOP
it
+
5
Manager_Tenure
i

+
it

Where:
it
is the monthly abnormal performance of fund of fund i
Expense Ratio
it
is the management, administrative, and 12b-1 fees as percent of
total net assets of fund i in month t
Turnover Ratio
it
is the minimum (of aggregated sales or aggregated purchases of
securities), divided by the average 12-month total net assets of the
fund, also used as a proxy for transaction costs associated with
rebalancing the portfolio of fund i in month t
Size
it
is the monthly log of Total Net Assets of fund i in month t
TTOP
it
is the monthly investment of fund in its top 10% holdings of fund i
in month t
Manager
Tenure
i
is the average number of years a manager stays with fund i
Expense ratio and turnover ratio are reported only on annual basis therefore in accordance with
the existing literature, annual expense ratio and turnover ratio are divided by 12 to estimate
monthly expense ratio and turnover ratio respectively. Manager tenure is one figure per fund
therefore it is constant per fund per month.
IV. Empirical Results
Empirical results show that market conditions played a very big role on the performance of large
cap funds. Moreover, the relationship between market conditions and performance was

95
Carhart (1997) and others use 3 years (36 months prior monthly returns) as the pre-estimation period. This study
is evaluating the performance over 2008-2011 period, a short four year window therefore we use only 1 year as pre-
estimation period.
978

consistent across all three categories of large cap funds. Results also show that even the large cap
funds that mainly invest in stable firms are highly susceptible to market conditions and cannot be
considered as a hedge against market downturn.
Results in table 3 panel A show that alpha is negative across large cap core and growth
categories and positive but statistically insignificant for large cap value stocks. Table 3 panel B
documents similar findings. On average, large cap core funds, generate negative alpha of 0.134
percent per month (-1.61 percent per year) after controlling for risk and style adjustments. The
abnormal performance is positively related to the market conditions. On average, a 1 percent
increase (decrease) in market returns increase (decrease) funds returns by 0.99 percent. This
association is statistically highly significant. Large cap growth funds generated negative alpha of
0.237 percent per month or -2.84 percent per year whereas large cap value funds generated
negative alpha of 0.051 percent per month or -0.612 percent per year. In all three cases, market
excess return is highly positive and highly statistically significant. Moreover, both models
explain a great deal of relationship between funds excess return and explanatory variables. The
adjusted R
2
is more than 90 percent across all categories for both models. It is also interesting to
find that large cap value stocks have the lowest negative alpha among all three categories and in
general, it is not very high compared to other studies in the mutual fund literature. This finding is
particularly important because large cap value funds tend to invest not only in large firms, but
also in those firms that are financially and fundamentally quite stable. The findings support the
notion that value stocks can hold their value even under extreme market volatility and can be
used as a hedge against extreme market swings.
Table 3: Performance of Funds
Following table shows the abnormal performance of funds across three different categories. The
abnormal performance () is based on the single factor and four-factor models.
r
it
- r
ft
is the dependent variable and it the excess monthly return of fund i over one month U.S. T-
Bill return. INDRF
t
is the excess monthly return of comparative benchmark over the one month
U.S. T-Bill return. SMB, HML, and MOM are monthly returns of size (the difference in returns
between small and large cap stocks), book to market (the difference in returns between high and
low book-to-market stocks), and momentum (the difference in returns between stocks with high
and low past returns) portfolios respectively. Alpha is expressed in percent per month. Results
are based on Newey-West heteroscedasticity and autocorrelation adjusted standard errors.

Panel A: Model: r
it
- r
ft
=
i
+
1i
* INDRF
t
+
i, t

alpha MKTRF N Adj. R
2

Large Cap Core -0.09351*** 1.00338*** 11,351 0.9349
Large Cap Growth -0.194*** 1.03515*** 15,067 0.9266
Large Cap Value 0.00639 0.96053*** 9,096 0.9324

*, **, *** indicate statistical significance at the 10%, 5% and 1% levels, respectively.

979

Panel B: Model: r
it
- r
ft
=
i
+
1i
* INDRF
t
+
2i
* SMB
t
+
3i
*HML
t
+
4i
*MOM
t
+
i, t
alpha MKTRF SMB HML MOM N Adj. R
2

Large Cap Core -0.00134*** 0.99629*** 0.04500*** -0.04186*** -0.01858*** 11,351 0.9356
Large Cap Growth -0.00237*** 1.01837*** 0.08742*** -0.00874* -0.00283 15,067 0.9275
Large Cap Value -0.00051*** 0.99819*** 0.05759*** -0.18091*** -0.00812*** 9,096 0.9377


Cross-Sectional Results
Results in table 4 show that, on average, fund specific variables do not affect abnormal
performance of funds across all three categories. The interesting finding lies with the expense
ratio. Generally, expenses reduce funds abnormal performance, but in our sample funds,
expenses have no significant impact on funds abnormal performance. It is commonly known
fact that mutual fund investors seek diversification and expertise of fund managers and therefore
willing to pay loads and fees to fund management. However, it is also documented in the
literature that these fees and loads are generally much higher and overweigh the benefits
achieved from active management. The non-negative impact of expenses, in part, suggests that
fund management of large cap funds evaluated in this study fulfilled its basic fiduciary duties. In
general, findings show that large cap funds abnormal performance over the period of this study is
affected by macroeconomic factors and extreme sell off by market participants in the wake of
extreme volatility rather than fund specific attributes.
Table 4: Cross-sectional Analysis: Impact of Fund Specific Variables on Abnormal
Performance
This table reports the multivariate Fama and MacBeth (1973) cross-sectional regressions for each
month over the period 01/2008-12/2011. The dependent variable is the monthly intercept
calculated using 4 factor beta loadings on the prior 12 monthly returns. The independent
variables are: Expense Ratio is management, administrative, and 12b-1 fees as percent of total
net assets, Turnover is the minimum of aggregated sales or aggregated purchases of securities
divided by the average 12-month total net assets of the fund, Size is the log (Total Net Assets),
TTOP is the monthly investment of a fund in its top 10% holdings, and Manager Tenure is the
average number of years a manager stays with a fund. The reported estimates are time-series
averages of monthly cross-sectional regression slope estimates as in Fama and MacBeth (1973).

980

Model:
it
=
0
+
1
Expense Ratio
it
+
2
Turnover Ratio
it
+
3
Size
it
+
4
TTOP
it
+
5

Manager_Tenure
i
+
it

Intercept
Expense
Ratio
Turnover
Ratio Size TTOP
Manager
Tenure N ADJ. R
2

Large Cap Core -0.00279** 0.6661 0.00208 0.00017* -0.00051 0.0000 4,966 0.048
Large Cap Growth -0.00201 -0.8858 0.01000 0.00015 0.00038 0.00006 6,824 0.062
Large Cap Value -0.001 1.782 -0.014 -0.000 -0.001 -0.000 4,119 0.041


V. Conclusion
This study analyzes the abnormal performance of large cap funds over the period 2008-20011
across all three categories, large cap core, large cap growth, and large cap value. The period from
2008 and 2011 is one of the eventful periods in the modern history. The world economy suffered
the worst recession since the great depression. This study specifically evaluates the performance
of large cap funds because they invest in large and stable firms and these funds are supposed to
weather the market volatility under the extreme market conditions. The findings of this research
suggest that large cap funds cannot be used as a hedging tool during the economic meltdown.
However, results also show a big variation in terms of abnormal performance across all three
categories. Large cap value stocks are least affected by the market volatility and in conjunction
with other hedging instruments, they can be used as a good hedging tool in the severe market
downturn. Results also show that abnormal performance is highly affected by market
movements. Most of the negative performance is not contributed by fund specific attributes, but
they are contributed by the macroeconomic factors and increased market volatility.

References
Amenc, Nol, Susan Curtis, and Lionel Martellini, 2004, The Alpha and the Omega of Hedge
Funds Performance, Working Paper, EDHEC (February).
Brown, Keith C., W. V. Harlow, and Laura T. Starks, 1996, Of tournaments and temptations: An
analysis of managerial incentives in the mutual fund industry, The Journal of Finance,
51(1), 85110.
Busse, Jeffery. A., Amit Goyal, and Sunil Wahal, 2010, Performance and persistence in
institutional investment management, The Journal of Finance 65(2), 765790.
Carhart, Mark M., 1997, On persistence in mutual fund performance, The Journal of Finance,
52(1), 5782.
Chordia, Tarun, 1996, The structure of mutual fund charges, Journal of Financial Economics 41,
3-39.
Coggin, T. Daniel, Frank J. Fabozzi, and Shafiqur Rahman, 1993, The investment performance
of U.S. equity pension fund managers: An empirical investigation. The Journal of Finance,
48(3), 10391055
981

De Souza Clifford, and Gokcan Suleyman, 2004, Allocation Methodologies and Customizing
Hedge Fund Multi-Manager Multi-Strategy Products, Journal of Alternative Investments,
Spring 2004, 7-21.
Edwards, Franklin. R. and Mustafa. O. Caglayan, 2001, Hedge Fund and Commodity
Investments in Bull and Bear Markets, Journal of Portfolio Management, Summer, 97-108.
Fama, Eugene, F., and Kenneth R. French, 1993, Common Risk Factors in the Returns on
Stocks and Bonds, Journal of Financial Economics, 33, 3-56.
Fama, Eugene, F., James MacBeth, 1973, Risk, Return, and Equilibrium: Empirical Tests,
Journal of Political Economy, Vol. 81, Issue 3, 607-636.
Grinblatt, Mark, and Sheridan Titman, 1992, The persistence of mutual fund performance, The
Journal of Finance, 47(5), 19771984.
Gruber, Martin J., 1996, Another Puzzle: The Growth in Actively Managed Mutual Funds. The
Journal of Finance 51:783-810.
Henriksson, Roy D., 1984, Market timing and mutual fund performance: An empirical
investigation, The Journal of Business, 57(1), 7396.
Huij, Joop, and G. T. Post, 2011, On the Performance of Emerging Markets Equity Mutual
Funds, Emerging Markets Review, 12, 238-249.
Ippolito, Richard A., 1989, Efficiency With Costly Information: A Study of Mutual Fund
Performance, 1965-1984, The Quarterly Journal of Economics 104: 1-23.
Jensen, Michael C., 1968, The performance of mutual funds in the period 1945-1964, Journal of
Finance, 23(2), 389416.
Kacperczyk, Marcin.T., Clemens Sialm, and Lu Zheng, 2005, On Industry Concentration of
Actively Managed Equity Mutual Funds, Journal of Finance 60, 1983-2011.
Nanda, Vikram, M. P. Narayanan, and Vincent A. Warther, 2000, Liquidity, investment ability,
and mutual fund structure, Journal of Financial Economics 57, 417-443.
Wermers, Russ, 2000, Mutual Fund Performance: An Empirical Decomposition into Stock-
Picking Talent, Style, Transactions Costs, and Expenses, Journal of Finance 55: 1655-
1695.
Wermers, Russ, 2003, Is Money Really Smart? New Evidence on the Relation Between Mutual
Fund Flows, Manager Behavior, and Performance Persistence, Working paper.
Zheng, Lu, 1999, Is Money Smart? A Study of Mutual Fund Investors Fund Selection Ability,
Journal of Finance 54: 901933.

982

Cash Payout and Long Run Performance of Mergers

Wei Zhang
Department of Finance and International Business
School of Business, Ithaca College
Ithaca, NY 14850
wzhang@ithaca.edu

Alka Bramhandkar
Department of Finance and International Business
School of Business, Ithaca College
Ithaca, NY 14850
abramhandkar@ithaca.edu

This research started when Zhang was a doctoral student at Washington State University.
983

Abstract
For a large sample of mergers and acquisitions announced between 1978 and 2007, we explore
how the magnitude of cash paid in mergers and acquisitions affects acquirer long-run abnormal
returns over the three-years after deal completion. We find that the magnitude of cash payout is
positively associated with the long run abnormal returns. This relation holds even after
controlling for various deal and firm characteristics, such as acquirer size, relative size, deal
attitude, and stock payment. The evidence indicates that the market at the announcement period
under-reacts to the information contained in the magnitude of cash payout in mergers and
acquisitions.
I. Introduction
Mergers and acquisitions (M&A) play an important role in corporate strategies for companies
from around the world. Because mergers involve dramatic organizational changes when two
independent entities are combined together, it is imaginable that the ex post integration process is
complex, compared to any incremental investment decisions. Such a complex process makes it
very difficult, if not impossible, to form a reliable expectation from the very beginning when the
decision is made. As a result, knowing the likelihood of success of an M&A decision is an
important question for corporate executives, board of directors, and investors in general. In the
past, researchers have focused on the market reactionthe change of stock prices around the
announcement periodto gauge the expected value accrued to acquiring firms (Jensen and
Ruback, 1983; Andrade, Mitchell, and Stafford, 2001). Relatively more recently, however,
empirical evidence shows that market reaction at the announcement period is probably not the
whole story, as researchers start to document that abnormal stock returns exist after the merger is
completed (see Agrawal and Jaffe, 2000 for a review of the literature during the late 80s and
early 90s). The academic debate goes on regarding the extent to which these long-run abnormal
returns after mergers, together with those after many other decisions, indicate the rejection of the
efficient market efficiency hypothesis (Fama, 1970; Fama, 1998). From managements
perspective, however, the existence of long run abnormal returns suggests that it is imperative to
learn the long term impact of merger decisions in general.
In this paper, we extend our early work Zhang and Bramhandkar (2011) to the long term
implication. Zhang and Bramhandkar (2011) study the relation between cash payout in mergers
and acquisitions, and document a significant positive relation between the magnitude of cash
payout and the announcement returns to the acquirers. The relation continues to hold
significantly, even after controlling for various firm and deal characteristics such as firm size,
relative size, deal attitude, merger terms, and in particular, method of payment, indicating that
the magnitude of cash payout contains incremental information beyond the method of payment.
Following the literature on long term returns, we hope to examine the long run implications of
cash payout.
While one would think cash payout (defined as the ratio of cash payment in the transaction value
to the market capitalization of the acquirer firm) is easy to understand, therefore the market
should have incorporated the whole information at the announcement, leading to no long-term
abnormal returns, the literature on long run returns does suggest that method of payment, a
closely related M&A term to cash payout, affects long run returns (Rau and Vermaelen, 1998;
Lougran and Vijh, 1997). The main reason that method of payment, specifically the use of stock
to pay for the acquisition, is that it signals the overvaluation of the acquiring firm at the
announcement. However, as Shleifer and Vishny (2003) argue, the market is inefficient, giving
executives, who attempt to maximize long-term shareholder value, the opportunity to take
984

advantage of the wedge between the current stock price and its long run intrinsic value. In this
setting, firms with overvalued equity would buy hard assets with real value for the benefit of
long term shareholders. Over time when the market learns the intrinsic value of the acquirers,
their stock prices drop, leading to long-run underperformance.
We follow the argument put forth in Zhang and Bramhandkar (2011) regarding how cash payout
affects merger performance. Specifically, value can be created by signaling and incentive
realigning by the amount of cash payout. The magnitude of cash payout conveys further positive
information about acquirer firm value beyond what method of payment indicates. The method of
payment determines the choice between stock and cash. Stock payment signals overvaluation of
equity while cash payment signals otherwise. While the signal is clear cut in method of payment,
it does not take into consideration the continuous information in the spectrum of payment
structure. That is, a higher amount of cash paid out in acquisitions signals that the acquirer
management is more confident about the value of the target and its capability of maintaining the
firms liquidity and avoiding bankruptcy. Cash payout, therefore, serves as a channel through
which the market learns incremental information about the firm value. To summarize, since cash
payout may be costly to the acquirer firm in terms of decreased liquidity and increased
probability of bankruptcy, the ability of paying out cash signals the value of the acquisition or
the value of the acquiring firm itself. Second, paying out a sizable amount of cash depletes the
free cash flows available at the discretion of acquirer management, leading to better alignment
of the managements interest with shareholders. If the cash paid out in acquisitions is raised
through issuing bonds or borrowing directly from banks, the acquirers leverage will increase,
which also better align interests of managers and shareholders. The importance of cash in
altering managerial behavior is studied extensively in the literature. Too little cash or too much
debt will cause under-investment, or the debt overhang problem (Myers, 1977). On the other
hand, too much cash will induce managers to deviate from maximizing shareholder value by
over-investing (Jensen, 1986). According to the agency theory, managers maximize their own
interest instead of shareholder value. Managers prefer larger offices, luxury corporate jets, and
other private benefits. These actions are more likely to take place when there are sufficient
resources whose use is under managers control. Cash, as the most liquid asset, is used at
managerial discretion. If investors expect that a portion of the cash left in managers hands will
be squandered, taking the cash away from them by paying for acquisitions increases shareholder
value. Paying out cash in acquisitions does not only decrease the resources under managers
control, it also increases the likelihood that the managers raise external financing later, a process
in which the firm and its managers are monitored by the market. In this sense, paying out cash
commits managers to better post-acquisition performance.
For a sample of 1,504 completed mergers and acquisitions between publicly traded acquirers and
targets during the years 1978 to 2007, we find that acquirer long run buy-and-hold abnormal
returns over the three years after deal completion are significantly positively correlated with the
magnitude of cash paid out in the total consideration. This relation exists even after controlling
for method of payment, relative size, and many other deal characteristics. The findings confirm
our conjecture that the market at the announcement period does not take into consideration the
full information contained in the magnitude of cash payout, leading to further drift in stock
prices.
Our paper begins with the literature review and research methodology. The subsequent sections
describe the data, and present the results of regression analyses before discussing the
contributions and limitations of our study.
985

The literature on long-run abnormal returns is motivated by the desire to check the viability of
the efficient market hypothesis (Fama, 1998). Long-run abnormal returns emerge when investors
either under-react or over-react to the information contained in a corporate event, be it mergers,
seasoned equity offering, or dividend initiation. It is well documented that merger decisions
involve investor under-reaction, leading to stock price drift (See Agrawal and Jaffe (2000) for
the review of the empirical work.). As a result, it is interesting to address the long term value
implications of merger-related decision making.
An important issue in investigating long-run returns is the methodological debate. The event-
study literature reveals that when the valuation impact of an event is measured over a relatively
short time period, the methodology in estimating equilibrium returns is not an issue (Kothari and
Warner, 2007). When abnormal returns are measured over a long time period, for example three
years, the bad model problem becomes severe. The debate over the best methodologies remains
an important academic issue. For our purpose, we follow the method that is relatively more
popularly accepted and used in the literature when measuring long run returns (Barber and Lyon,
1997; Ma, Whidbee, and Zhang, 2011).
On the cash payout side, we follow our earlier work Zhang and Bramhandkar (2011).
III. Sample and Data
III.1 Sample Selection
The sample is originally drawn from Securities Data Corporation (SDC) Platinum online
Mergers and Acquisitions Files from year 1978 to 2007, where both acquirer and target are
public firms. For later analysis, it is required that both the acquirer and target firms have
sufficient stock price data in the Center for Research in Security Prices (CRSP) at the University
of Chicago in order to calculate shareholder wealth change during a specified event window.
We start with 7,419 deals from SDC that satisfy the following criteria: deal value is greater than
$1 million; the transaction form is Acquisition of assets, Merger, or Acquisition; both
target and acquirer are US firms and both are publicly traded; the announcement year is between
1978 and 2007. After deleting deals where the acquirer and target are the same firm, we are left
with a total of 7,344 deals.
We further require the acquirer and target stocks to have share code of 10 or 11 (common
stocks); both have book value of equity available from the most recent quarterly financial
statements (from Compustat) immediately before the announcement; both firms have market
capitalization information at the month end before the announcement. This reduces the sample
size to 4,751. We then exclude incomplete deals and deals in which the acquirer has owned more
than fifty percent of the target before the announcement. Further, we require the 3-year long-run
buy-and-hold period abnormal returns exist, and the deal size relative to the acquirer market
capitalization is between 1% and 200%. Our final sample size is 1,504.
Distribution of number of deals year by year is presented in Table 1. From the first column, it is
clear that there are significantly more deals in mid-to-late 90s, when the historic merger wave
occurred. There are a total of 131 No cash deals out of a total of 265 deals during the 80s,
indicating that in 51% deals (1 131/265) at least some cash is paid. Comparatively, this ratio is
32% (1 519/763) for the 90s. This pattern is consistent with the notion that mergers during the
90s are more likely to use equity, possibly driven by overvalued equity (Shleifer and Vishny,
2003). Between the low cash and high cash series, a relatively higher proportion of deals in late
80s is classified as high cash, compared to the 90s and the 2000s. Significantly fewer deals take
place in 2007, probably due to the recent recession.
986

Table 1: Distribution of sample deals year-by-year
The table lists, for every year from 1978 to 2007, the total number of deals, the number of deals
paying no cash, the number of deals with cash payout lower than the sample median, and the
number of deals with cash payout higher than the sample median.

Year Num. Obs. No cash Low cash High cash
1978 4 0 0 4
1979 2 0 0 2
1980 3 1 1 1
1981 17 12 0 5
1982 10 10 0 0
1983 22 21 1 0
1984 31 24 2 5
1985 35 9 11 15
1986 49 14 11 24
1987 42 18 10 14
1988 32 9 10 13
1989 24 13 5 6
1990 20 12 4 4
1991 32 22 5 5
1992 33 21 5 7
1993 28 16 10 2
1994 65 47 10 8
1995 99 67 16 16
1996 104 69 13 22
1997 128 94 15 19
1998 137 107 19 11
1999 117 64 28 25
2000 98 49 17 32
2001 87 50 21 16
2002 41 18 17 6
2003 68 27 26 15
2004 64 21 27 16
2005 55 15 21 19
2006 49 16 20 13
2007 8 0 4 4
Total 1504 846 329 329

III.2 Data Description
For the sample deals, we construct the following variables. The main data sources are CRSP,
Standard & Poors Compustat, and SDC.
BHAR
We define BHAR following the methodology in Ma, Whidbee, and Zhang (2011). BHARs are
calculated using the reference portfolio approach. Each month, we sort all NYSE common stocks
into size deciles based on their month end market capitalization. These deciles are further sorted
into quintiles using book-to-market ratios. We then place AMEX and NASDAQ firms into one
of these 50 portfolios based on their month-end size and book-to-market ratios. We calculate the
monthly return for each of the 50 reference portfolios by averaging the monthly returns across all
stocks in each portfolio. These portfolio returns are then used as benchmarks to calculate buy-
and-hold abnormal returns (BHARs) for our sample firms. The following formula states the
definition more formally.
987

BHAR 1 R
,
1 R

Where R
,
is acquirer js return over the t-th month after deal completion; R
,
is the
corresponding control portfolio returns for acquirer j over the t-th month after deal completion.
We measure the abnormal returns over the 36 months after deal completion. That is, if a deal is
completed on July 4, 1998, the long-run period is the 36 months from August 1998 to July 2001.
Cash payout
We define cash payout, or the magnitude of cash payout in mergers and acquisitions, as the cash
amount paid in acquisitions divided by the acquirers market capitalization before the
announcement (see equation 1). Method of payment, on the other hand, focuses on the
percentage of payment in stock or cash. We use cash percentage as the measure of method of
payment, defined in equation (2).
Cash payout = cash paid / acquirer market cap (1)
Cash percentage = cash paid / deal value (2)
The following example illustrates the differences between cash payout and cash percentage.
Compare two deals, A and B, with the same deal size of $500 million, of which 80% is paid in
cash. In deal A, the market capitalization of the acquirer is $1 billion, while that of deal B is $10
billion. In acquiring the same firm, the acquirer in deal A pays out cash which is as large as 40%
of its market capitalization, while the acquirer in deal B pays out cash only 4% of its market
capitalization. The information on cash payout and cash percentage is shown in the exhibit
below.
Deal Acquirer Market cap ($mil) Deal size ($mil) Cash (%) Cash Payout
A 1,000 500 80% 40%
B 10,000 500 80% 4%
The two deals have the same method of payment as reflected in cash percentage (80%), yet the
magnitude of cash payout is dramatically different. It is not hard to understand that paying out
cash that is 40% of market capitalization reveals far more favorable information about the firms
future performance than paying out cash that is only 4% of market cap. Thus magnitude of cash
payout conveys more information than method of payment does.

Other variables
SDC provides information on the stock percentage, cash percentage in the method of payment.
We also define dummy variables for any stock, and pure cash. Any stock is equal to one if the
stock percentage is positive in the method of payment; similarly, pure cash is equal to one if all
payment is in cash. Transaction value is defined as the total value of consideration paid by the
acquirer, excluding fees and expenses. Moeller, Schlingemann, and Stulz (2004) report that
acquirer size is an important determinant of acquirer abnormal returns in acquisitions. We,
therefore, include a dummy variable acq is small, which is equal to one if the acquirers market
capitalization is below the twenty-fifth percentile of all NYSE firms covered in CRSP in the
announcement year.
There is an ongoing debate in the literature over whether firm diversification creates or destroys
value. Matsusaka (1993; 2001) shows that diversification is a value-maximizing strategy.
However, other researchers (for example, Lang and Stulz, 1994; Berger and Ofek, 1995) argue
that diversification destroys firm value. While in-depth discussion of diversification is beyond
the scope of this paper, to control any effect of diversifying acquisitions on buyer returns, we
988

define a dummy variable related, which takes the value one if the acquirer and target share two-
digit Standard Industry Code (SIC).
Relative size has been found to explain the acquirer returns (Asquith, Brunner and Mullins,
1983). We define relative size as the ratio of transaction value to acquirer firm market
capitalization. To account for the skewed distribution of this variable, we use ln(1+relative size)
in regressions. Financial industry deal dummy takes value of one when both acquirer and target
are in the financial industry (primary SIC code is between 6000 and 6999). An acquirer is
defined as having a toehold if it owns over 5% of target shares before the acquisition
announcement. Target termination fee used is a dummy variable, which is equal to one if in the
deal target termination fee is used. Lockup is a dummy variable, which takes the value one if
lockup option is used in the deal; hostile is a dummy variable if the deal attitude is hostile or the
deal is unsolicited; tender offer is a dummy variable if the deal is recorded as a tender offer in
SDC. The Appendix contains a complete list of variables used in this study.

III.3 Summary Statistics
Table 2 presents the sample means for the whole sample(All), the subsample with no cash
payment (subsample No Cash), the subsample with cash payout below the sample median
(Low Cash) and the subsample with cash payout above the sample median (High Cash). The
last four columns present the differences between the Low Cash and No Cash subsamples,
and between the High Cash and Low Cash subsamples, along with the p-values of the t-tests
testing equal means of the two subsamples.
Panel A shows that on average the BHAR is -8.017%, which is largely consistent with the recent
literature (e.g., Agrawal and Jaffe, 2000; Ma, Whidbee, and Zhang, 2011). The acquirer BHARs,
however, vary significantly across the three subsamples. In the No cash subsample, the average
acquirer BHAR is -14.869% while that of the High cash subsample is 8.091% and the Low
cash subsample has a mean of -6.505%. The difference between the Low Cash and High
Cash subsamples is significant at the 1% confidence level.
Presented in Panel B, the amount of cash paid is on average 13.5% of acquirer market
capitalization, indicating the economic significance of the amount of cash paid in mergers and
acquisitions. The cash payout in the half subsample of low cash payout is 11.6%, while that of
the half subsample of high payout is almost 50.3%. The differences between the Low and No
and between the High and Low subsamples are all significant at the 1% level.
Panel C reports these statistics for variables of method of payment, deal size, and other deal
terms. On average the percentage of cash payment in deals is 28%, and that of stock is about
59%. About 18% deals pay all cash; in about 68% of the deals at least some stock is paid for the
deal; on average the relative size is 47.8%.The distributions of these variables across the three
subsamples indicate that these variables are correlated with cash payout. Tender offers constitute
17.4% of deals. High cash subsample has the highest proportion of tender offers, while no cash
subsample has the lowest proportion. The average deal value is about $2.17 billion; acquirer and
target are related in about 69% of the deals; in 8.5% of the deals both acquirer and target are
financial firms; the acquirer owns a toehold (over 5% ownership in target) in 3.1% of deals.
Target termination fee is used in 55% of deals; lockup is used in 22% of the deals; only 3.3% of
the deals are hostile; 13.8% acquirers are classified as small firms. The distributions of these
variables across three subsamples indicate that these variables are correlated with cash payout.
Table 2: Summary statistics of deal characteristics
The sample contains 1,504 transactions between publicly traded targets and acquirers that are announced
between 1978 and 2007. All variables are defined in the Appendix. For each variable listed in the first
989

column, the table presents its mean values for the whole sample, the subsample with no cash payout, with
low cash payout and with high cash payout, in the next four columns, respectively. The next two columns
present the difference in the mean values between the low cash and no cash subgroups and the p-value
testing equal means; the last two columns report the difference in the mean values between the high cash
and low cash subgroups and the p-value testing equal means.

All
(1,504)
No cash
(846)
Low cash
(N=329)
High cash
(N=329)
Low
No p-value
High
Low p-value
Panel A: Acquirer buy-and-hold abnormal returns (BHARs)
BHAR -8.017 -14.869 -6.505 8.091 8.363 0.013 14.596 0.005
Panel B: Magnitude of cash paid in the transactions
Cash payout 0.135 0.000 0.116 0.503 0.116 0.000 0.387 0.000
Panel C: Deal characteristics
Cash percentage (%) 27.587 0.000 57.745 68.367 57.745 0.000 10.622 0.000
Stock percentage (%) 58.560 87.969 28.663 12.834 -59.305 0.000 -15.829 0.000
Pure cash 0.180 0.000 0.347 0.474 0.347 0.000 0.128 0.001
Any stock 0.683 0.920 0.489 0.267 -0.430 0.000 -0.222 0.000
Tender offer 0.174 0.041 0.331 0.356 0.290 0.000 0.024 0.512
Relative size 0.478 0.485 0.250 0.685 -0.235 0.000 0.435 0.000
Deal value ($bil) 2.173 2.666 1.549 1.531 -1.118 0.005 -0.018 0.959
Related 0.691 0.713 0.653 0.675 -0.059 0.053 0.021 0.564
Financial industry 0.085 0.091 0.103 0.052 0.012 0.528 -0.052 0.013
Toehold 0.031 0.022 0.040 0.043 0.017 0.153 0.003 0.844
Tgt termination fee used 0.549 0.521 0.635 0.532 0.114 0.000 -0.103 0.007
Lockup 0.221 0.285 0.134 0.143 -0.151 0.000 0.009 0.735
Hostile 0.033 0.018 0.036 0.070 0.019 0.098 0.033 0.056
Acq is small 0.138 0.118 0.070 0.255 -0.048 0.007 0.185 0.000

The overall evidence suggests that cash payout in mergers and acquisitions is correlated with
acquirer abnormal returns. At the same time, various deal and firm characteristics that the
literature suggests affect long run acquirer returns are also correlated with cash payout. Whether
the correlation between cash payout and BHAR is driven by other variables should be
determined by running multiple regressions. Before conducting regression analysis, however, we
should be cautious about the potential multicollinearity problem, as relative size, cash
percentage, acquirer size, and cash payout variables could be positively correlated.
Table 3: The correlation coefficient matrix
This table contains Pearson correlation coefficients matrix for the variables used in the main
analysis. The sample contains 1,504 transactions between publicly traded targets and acquirers
that are announced between 1978 and 2007. The pair-wise Pearson correlation coefficients are
presented in the table with the p-value testing its statistical significance in the brackets.
Cash percentage Ln(1+relative size) Acq is small
Cash payout 0.554 0.394 0.149

[<.0001] [<.0001] [<.0001]
Cash percentage

-0.118 0.066

[<.0001] [0.098]
Ln(1+relative size)

0.166

[<.0001]
Acq is small

990

To address this concern, we examine the Pearson correlation coefficient matrix of cash payout,
cash percentage, ln(1+relative size), and acquirer is small. The results are presented in Table 3.
Cash payout is positively correlated with cash percentage, relative size, and the dummy variable
acquirer is small. The strongest correlation is between cash payout and percent of cash paid,
with a coefficient of 0.554, significant at the 0.01% level. According to Berry and Feldman
(1990), multicollinearity problem arises when correlation coefficients exceed 0.80. Therefore, for
the sample and data we are working with, multicollinearity is not a major concern.

IV. Regressions Analysis
To examine the ultimate relation between cash payout and BHAR, we present several multiple
regression models in Table 4. The dependent variable is BHAR. The independent variables are
commonly used in the literature of mergers and acquisitions. The main independent variable of
interest is cash payout.
Table 4: Multiple regressions of acquirer abnormal returns
Three ordinary least squares (OLS) regressions are reported. The sample contains 1,504
transactions between publicly traded targets and acquirers that are announced between 1978 and
2007. The dependent variable is BHAR. All variables are defined and contained in the Appendix.
T-statistics based on heteroskedasticity-robust standard errors are reported in parentheses.
***
,
**
,
*
denote statistical significance at the 1%, 5% and 10% confidence levels, respectively.

Independent variables Model (1) Model (2) Model (3)
Cash payout

32.651
***
32.203
**

(2.58) (2.57)
Cash percentage 0.132
**
0.016 0.041

(1.97) (0.20) (0.50)
Any stock -8.292 1.717 -4.261

(-1.38) (0.35) (-0.76)
Ln(1+relative size) 16.527
*
-0.434 0.154

(1.96) (-0.05) (0.01)
Tender offer -13.308
***

-13.106
***

(-2.68)

(-2.62)
Acq is small 26.926
***
26.372
***
26.080
***

(3.90) (3.92) (3.86)
Related 4.861

4.488

(1.36)

(1.26)
Financial industry 17.890
***

17.997
***

(3.35)

(3.36)
Toehold -0.838

0.751

(-0.11)

(0.10)
Tgt termination fee used -1.231

-0.547

(-0.36)

(-0.16)
Lockup 2.609

2.801

(0.71)

(0.77)
Hostile 9.941

5.856

(1.30)

(0.74)
Constant -18.477
***
-17.535
***
-17.155
***

(-3.16) (-3.38) (-2.92)
Adj. R
2
0.042 0.041 0.049

In Model (1), we regress BHAR on the control variables only. We find that cash percentage has a
positive coefficient, significant at the 5% level. This is consistent with the literature which
991

suggests that using cash sends out positive signals about the value of the deal. In similar vein,
using stock signals that the stock is overvalued, leading to marginally lower abnormal returns.
Relative size has a positive coefficient, significant at the 10% level. Mixed results are
documented in the literature regarding the relation between relative size and acquirer abnormal
returns. For example, Jarrell and Poulsen (1989), Asquith, Bruner, and Mullins (1983) report a
positive coefficient of relative size on acquirer returns in a sample of successful tender offers;
Travlos (1987) reports a negative coefficient; Boone and Mulherin (2008) report positive
coefficients in their OLS regressions but positive and negative coefficients in treatment
regressions, depending on estimation windows. Our study focuses on long run returns. In our
study, relative size has a positive and significant coefficient. Consistent with Moeller,
Schlingemann, and Stulz (2004) that smaller acquirers earn higher abnormal announcement
returns, we find smaller acquirers also earn higher long run returns. Acquirer returns are lower in
tender offers, inconsistent with Lougran and Vijh (1998). The coefficients on target termination
fees and lockup options are not significant. Deals in the financial industry are associated with
higher abnormal returns. By and large, results in Model (1) are consistent with the prior
literature.
In Model (2), we examine the effect of cash payout after controlling for cash percentage, any
stock payment, relative size, and the indicator variable for smaller acquirers. As expected, cash
payout has a positive coefficient with a t-statistics of 2.58, significant at the 1% level. The
coefficient on cash percentage becomes statistically insignificant, although it remains positive.
The dummy variable acq is small has a significant positive coefficient as well. Results in
Model (2) indicate that cash payout contains information beyond what method of payment (cash
percentage and stock payment) represents. After controlling for the magnitude of cash payout,
the other variables are no longer significant. Further, consistent with results in Table 3, there
does not appear to have a multicollinearity problem. In Model (3), we regress BHAR on cash
payout and all the control variables. The coefficient of cash payout remains positive and
significant at the 1% level. The coefficients of other variables are largely similar to Model (1).

V. Conclusion
In a large sample of mergers and acquisitions announced between 1978 and 2007, we study the
effect of cash payout on acquirer long run abnormal returns. We find that the magnitude of cash
paid out in acquisitions is significantly positively correlated with long run abnormal returns. The
result indicates that the magnitude of cash payout in acquisitions conveys distinctive information
beyond what is contained in method of payment or the choice between cash and equity. Further,
the market at the announcement under-reacts to the information, which also affects the long run
returns.
The implications of our results to various parties involved in merger negotiations are clear.
Acquirer managers and their advisers (such as the investment banks they hire) need to recognize
the signaling effect of cash payment. They also need to keep in mind that the magnitude of cash
payout is an important determinant of wealth changes experienced by their stockholders when
they make the decisions on choosing means of investment financing. In particular, it affects stock
returns not just at the announcement, but also over the long term.

992

References
Agrawal, Anup, and Jeffrey F. Jaffe, 2000. The post-merger performance puzzle. In: Cooper, G.,
Gregory, A. (Eds.), Advances in Mergers and Acquisitions, Vol. 1. Amsterdam: Elsevier,
pp. 7-41.
Andrade, Gregor, mark L. Mitchell, and Erik Stafford, 2001. New Evidence and Perspectives on
Mergers. Journal of Economic Perspectives, 15, 103-120.
Asquith, Paul, Robert F. Brunner, and David W. Mullins Jr., 1983. The Gains to Bidding Firms
from Mergers. Journal of Financial Economics, 11, 121-139.
Barber, Brad M., and John D. Lyon, 1997. Detecting Long-Run Abnormal Stock Returns: The
Empirical Power and Specification of Test Statistics. Journal of Financial Economics, 43,
341-371.
Berger, Philip G., and Eli Ofek, 1995. Diversifications Effect on Firm Value. Journal of
Financial Economics, 37, 39-65.
Berry, William D., and Stanley Feldman, 1990. Multiple Regression in Practice. Newbury Park,
CA: Sage Publications.
Boone, Audra L., and J. Harold Mulherin, 2008. Do Auctions Induce A Winner's Curse? New
Evidence from the Corporate Takeover Market. Journal of Financial Economics, 89, 1-19.
Fama, Eugene F., 1970. Efficient Capital Markets: A Review of Theory and Empirical Work,
Journal of Finance, 25, pp. 383-417.
Fama, Eugene F., 1998. Market Efficiency, Long-Term Returns, and Behavioral Finance,
Journal of Financial Economics, 49, pp. 283-306.
Jarrell, Gregg A., and Annette B. Poulsen, 1989. The Returns to Acquiring Firms in Tender
Offers: Evidence from Three Decades. Financial Management,18, 12-19.
Jensen, Michael, 1986. Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.
American Economic Review, 76, 323-329.
Jensen, Michael C., and Ruback, Richard S., 1983. The market for corporate control: The
scientific evidence. Journal of Financial Economics 11, 5-50.
Kothari, S.P., and Jerold B. Warner, 2007. Econometrics of Event Studies. In Espen Eckbo
(Eds.) Handbook of Corporate Finance, empirical corporate finance, Volume 1. North-
Holland pp. 3-36.
Lang, Larry, and Rene M. Stulz, 1994. Tobins Q, Corporate Diversification, and Firm
Performance. Journal of Political Economy, 102, 1248-1280.
Loughran, Tim, and Anand M. Vijh, 1997. Do Long-term Shareholders Benefit From Corporate
Acquisitions? Journal of Finance 52, 1765-1790
Ma, Qingzhong, David Whidbee, and Wei Zhang, 2011. Value, Valuation, and the Long Run
Performance of Merged Firms. Journal of Corporate Finance 17, 1-17.
Matsusaka, John G., 1993. Takeover Motives during the Conglomerate Merger Wave. RAND
Journal of Economics, 24, 357-379.
Matsusaka, John G., 2001. Corporate Diversification, Value Maximization, and Organizational
Capabilities. Journal of Business, 74, 409-431.
Moeller, Sara B., Frederik P. Schlingemann, and Rene M. Stulz, 2004. Firm Size and the Gains
from Acquisitions. Journal of Financial Economics, 73, 201-228.
Myers, Stewart C., 1977. Determinants of Corporate Borrowing. Journal of Financial
Economics, 5, 147-175.
Rau, P. Raghavendra, and Theo Vermaelen, 1998. Glamour, value and the post-acquisition
performance of acquiring firms. Journal of Financial Economics 49, 223-253.
993

Shleifer, Andrei, and Robert.W. Vishny, 2003. Stock Market Driven Acquisitions. Journal of
Financial Economics, 70, 295-311.
Travlos, Nickolaos G., 1987. Corporate Takeover Bids, Method of Payment, and Bidding Firms
Stock Returns. Journal of Finance, 4, 943-967.
Zhang, Wei, and Alka Bramhandkar, 2011. Cash Payout in mergers and Acquisitions. Journal of
Current Research in Global Business, Winter, 14, 33-43.

Appendix: Variable definitions and data source in brackets.
Acq is small is a dummy variable that takes the value one if the acquirers market capitalization
is below the 25
th
percentile of the NYSE firms of the same year the deal is announced; [CRSP]

Any stock is a dummy variable equal to one if stock payment is positive in the transaction;
[SDC]

BHARs are calculated using the reference portfolio approach. Each month, we sort all NYSE
common stocks into size deciles based on their month end market capitalization. These deciles
are further sorted into quintiles using book-to-market ratios. We then place AMEX and
NASDAQ firms into one of these 50 portfolios based on their month-end size and book-to-
market ratios. We calculate the monthly return for each of the 50 reference portfolios by
averaging the monthly returns across all stocks in each portfolio. These portfolio returns are then
used as benchmarks to calculate buy-and-hold abnormal returns (BHARs) for our sample firms.
[CRSP, Compustat]

Cash payout is defined as the amount of cash paid out in the deal divided by the market
capitalization of the acquirer before the announcement; [SDC, CRSP]

Cash percentage is the percentage of deal value made in cash; [SDC]

Complete is equal to one if the deal finally completes; [SDC]

Deal value ($bil) is the total transaction value as recorded in SDC; [SDC]

Financial industry is a dummy equal to one if both acquirer and target are from financial industry
(SIC between 6000 and 6999); [SDC]

Hostility is equal to one if the deal attitude is hostile; [SDC]

Lockup is a dummy equal to one if lockup option is used; [SDC]

Pure cash is a dummy variable, which is equal to one if the payment is all in cash; [SDC]

Relative size is the transaction size divided by acquirer market capitalization before the
announcement; [SDC, CRSP]

Related is a dummy that takes the value one if the acquirer and target share 2-digit primary SIC
codes; [SDC]
994

Tender offer is a dummy variable equal to one if the deal is a tender offer; [SDC]

Tgt termination fee used is a dummy variable equal to one if target termination fee is used in the
deal; [SDC]

Toehold is a dummy variable equal to one if the percentage of target firm shares owned by the
acquirer at the announcement date is above 5%. [SDC]

995

Too Critical to Fail: What Saved the Public Company Accounting Oversight
Board?

Stanley A. Leasure
Associate Professor of Business Law
Missouri State University
J.D. University of Tulsa College of Law

Jana Ault Phillips*
Per Course Instructor of Business Law
Missouri State University
J.D. Washington University in St. Louis School of Law

*Corresponding Author
996

Abstract
Sarbanes-Oxley established the Public Company Accounting Oversight Board vested with broad
powers to regulate accounting firms auditing publicly traded companies. The Supreme Court
found that portion of the Act requiring a showing of cause prior to removal of Board members by
the SEC to be unconstitutional. However, rather than vitiating Sarbanes-Oxley in its entirety, the
Court severed the defective portions, leaving the remainder intact. The result was preservation of
the PCAOB. This Article addresses the policy issues surrounding the decision to preserve the
Board and its impact.

I. Introduction
In the wake of several major accounting scandals Congress passed the Sarbanes-Oxley Act in
2002 (Act or Sarbanes-Oxley). Among other things, it established the Public Company
Accounting Oversight Board (Board or PCAOB) vested with broad powers to regulate
accounting firms auditing publicly traded companies.
96
The Securities and Exchange
Commission ("SEC") was empowered to appoint Board members and to remove them, but only
for good cause.
97
The President was empowered to appoint SEC Commissioners and to remove
them for "inefficiency, neglect of duty, or malfeasance in office."
98
Neither Board members nor
SEC Commissioners are subject to removal at will.
The constitutionality of these removal provisions was challenged in a lawsuit brought against the
Board after it issued a report critical of the auditing procedures of Beckstead and Watts, LLP, a
registered accounting firm.
99
The firm and the Free Enterprise Fund (collectively FEF) of
which the firm was a member, filed suit alleging that the provision of Sarbanes-Oxley mandating
that the SEC can only remove Board members for cause violates Article II of the Constitution.
100

The question was whether the two separate layers of protection from at will removal violated
Article II which vests executive power in the President.
101

In a case called the most important separation-of-powers case regarding the President's
appointment and removal powers . . . in the last 20 years,"
102
the United States Supreme Court,
in a majority opinion written by Chief Justice Roberts, and joined by Justices Scalia, Kennedy,
Thomas and Alito declared that the provision requiring a finding of cause antecedent to removal
of Board members by the SEC violates the United States Constitution.
103
This holding also
raised the very significant question of whether the Court should spare the Board by severing the
defective portions of the Act and leaving the rest intact.
Three amici, National Association of State Boards of Accountancy (NASBA), Institutional
Investors, and former chairmen of the SEC were the strongest advocates for severance and
preservation of the Board. NASBA argued that the negative secondary effects of abolishing the
Board justified severance and preservation of the remainder of Sarbanes-Oxley, along with the

96
15 U.S.C. 7211.
97
15 U.S.C. 7211(e)(6); 15 U.S.C. 7217(d)(3).
98
Humphrey's Executor v. United States, 295 U.S. 602, 620, 55 S.Ct. 869, 79 L.Ed. 1611 (1935).
99
Id. at 3148.
100
Id. at 3147.
101
Id.
102
Free Enterprise Fund v. Public Company Accounting Oversight Board, 537 F.3d 667, 685 (3
rd
Cir 2008)
(Kavanaugh, J., dissenting).
103
Free Enterprise Fund v. Public Company Accounting Oversight Board, 130 S.Ct. 3138, 3147 (2010).
997

Board it created.
104
NASBA described the havoc which would be wreaked if the Board was
swept away as a result of this litigation.
105
The Court agreed, fashioning a remedy which: struck
that portion of Sarbanes-Oxley requiring a showing of cause as a condition precedent to the
removal of Board members; declared that Board members serve at the will of the SEC; and
preserved the remainder of Sarbanes-Oxley and the Board.
106

This Article initially provides an overview of the constitutional, statutory, and common law
foundation upon which Free Enterprise Fund was decided.
107
Next, the majority opinion of the
Court is considered in detail.
108
Following that, the myriad policy issues surrounding the decision
to sever only the unconstitutional portions of the Sarbanes-Oxley Act, thereby preserving the
Board are considered.
109

II. Review of The Jurisprudential Foundation
According to the Court, the "for cause" predicate to the removal of members of the Board in
juxtaposition to those same type requirements for the removal of SEC Commissioners violates
the Appointments Clause and the constitutional doctrine of separation of powers.
110
The
interface between the Sarbanes-Oxley; Supreme Court precedent; the Appointments Clause and
the doctrine of separation of powers is essential to an analysis of the issues in this case; the
Court's decision; and the potential negative secondary effects of the decision predicted by the
dissent.

A. Constitutional Provisions
The majority evaluated the issues, extant case law and relevant statutory provisions through the
prism of the separation of powers doctrine and the Appointments Clause. The separation of
powers doctrine is embodied in Article II, sections 1 and 3:
The executive Power shall be vested in a President of the United States . . .
111

***
. . . [H]e shall take Care that the Laws be faithfully executed, and shall
Commission all the Officers of the United States.
112

104
Brief of National Association of State Board of Accountancy as Amicus Curie in Support of Respondents, Free
Enterprise Fund v. Public Accounting Oversight Board, 130 S.Ct. 3138 (2010) (No. 08-861), 2009 W.L. 3404247 at
30. The National Association of State Boards of Accountancy is made of all fifty of the state boards of accountancy
plus the District of Columbia and the territories of Puerto Rico, Guam, the Northern Mariana Islands, and the Virgin
Islands. Id. at 1.The Council of Institutional Investors also filed an amicus brief urging severance. It was joined on
the brief by the American Federation of Labor and Congress of Industrial Organizations; the California Public
Employees Retirement System; the California State Teachers' Retirement System; CFA Institute; the Consumer
Federation of America; the Los Angeles County Employees Retirement Association; the Nathan Cummings
Foundation; Thomas P. DiNapoli, Comptroller of the State of New York, as Trustee of the New York State
Common Retirement Fund and as Administrative Head of the New York State and Local Retirement Systems; the
Public Employees Retirement Association of Colorado; the Sacramento County Employees' Retirement System; and
the Teachers Insurance and Annuity Association-College Retirement Equities Fund (hereinafter collectively referred
to as Institutional Investors).
105
Id. at 31.
106
Id.
107
See infra Part I.
108
See infra Part II.
109
See infra Part III.
110
Id. at 3151-3161.
111
U.S. Const. art. II 1, cl. 1.
112
998

The Appointments Clause is found in Article II 2 of the Constitution:
He shall . . . nominate, and by and with the Advice and Consent of the
Senate, shall appoint . . . Officers of the United States, whose Appointments are
not herein otherwise provided for, and which shall be established by Law: but the
Congress may by Law vest the Appointment of such inferior Officers, as they
think proper, in the President alone, in the Courts of Law, or in the Heads of
Departments.
113

B. Statutory Provisions
Central to the decision were the statutory provisions establishing the Board and the scope of its
power, governing the appointment and removal of its members, providing for the removal of
Commissioners of the SEC, and establishing its relationship with the SEC were central to the
decision.
Sarbanes-Oxley established the Board and determined its status, duties and, membership.
114
The
Act gives the Board broad jurisdiction including the regulation of all firms auditing publicly
traded companies.
115
The power of the Board to regulate the accounting industry vis--vis
publicly traded companies is extensiveincluding, among other things, accounting standards,
documentation requirements, audit reports, professional ethics and independence.
116
The Board
is vested with enforcement authority to conduct inspections and investigations of registered
accounting firms and the power to sanction violators.
117
The Boards adjudicatory powers are
addressed in Title 5.
118

The Act defines the relationship between the SEC and the Board. The SEC must approve the
Board's budget and has financial control over the Board.
119
The SEC has prior approval and
amendment authority over the rules of the Board. In addition, it may review and, in certain
circumstances, even remand sanctions imposed by the Board.
120
Other important provisions
relate to the operation of the Securities and Exchange Commission including: relief from final
orders;
121
rulemaking procedures;
122
sanctions;
123
and judicial review.
124

C. Supreme Court Precedent
The Constitution provides: "Congress may by law vest the appointment of such inferior officers
as they think proper in the president alone, in the courts of law, or in the heads of
departments.
125
In United States v. Perkins
126
the United States Supreme Court considered the
distinction between an "officer" and an "at will employee." It broadly construed the definition of
"inferior officers", for purposes of separation of powers analysis.
127

113
114
15 U.S.C. 7211(a)-(c)(6); 15 U.S.C. 7211(e)(1); 15 U.S.C. 7211(e)(6)).
115
15 U.S.C. 7212(a) and (f).
116
15 U.S.C. 7213(2).
117
15 U.S.C. 7215.
118
5 U.S.C. 554; 5 U.S.C. 556; 5 U.S.C. 557.
119
15 U.S.C. 7211(f)(6) and 7219(b).
120
15 U.S.C. 7217.
121
15 U.S.C. 78y.
122
15 U.S.C. 78s.
123
15 U.S.C. 78ff.
124
15 U.S.C. 78y.
125
116 U.S. 483, 21 Ct.Cl. 499, 6 S.Ct. 449, 29 L.Ed. 700 (1886).
126
Id.
127
United States v. Perkins, 116 U.S. 483, 6 S.Ct. 449, 29 L.Ed. 700 (1886).
999

In Myers v. United States
128
the Court considered whether the President has the exclusive power
to remove executive officers appointed with the advice and consent of the Senate.
129
The
linchpin of the majority's decision was the Presidents mandate to execute the laws and his power
to remove executive officers in light of his power to appoint them. The Supreme Court viewed
the President's executive power to include the authority to use subordinates, acting at his
discretion, in the execution of the laws.
130
Just as his power to appoint administrative officials is
essential to his execution of the laws, so is his ability to remove those for whom he cannot
continue to be responsible.
131
The reasoning in Myers is pragmatic; those who select executive
subordinates to help them carry out their responsibilities need the self-same authority to remove
them.
132

Section 1 of the Federal Trade Commission Act (FTC Act) provided: Any commissioner may
be removed by the president for inefficiency, neglect of duty or malfeasance in office.
133
In
Humphreys Executor v. United States, the Supreme Court addressed the potential limitations on
the power of the president to remove Federal Trade Commissioners.
134
The Court found the FTC
Act to be definite and unambiguous, making removal limited to circumstances in which the
stated grounds were present.
135
The Court also held legislative intent to allow the term to
proceed uncurtailed, absent one or more of the stated causes is evidenced by the definite term
established in the statute.
136
The intent to insulate the Commission from political influence in the
exercise of its quasi-judicial and quasi-legislative powers was also considered significant.
137
The
language of the statute and its legislative history evidenced congressional intent to create a
commission of experts independent of executive authority and benefitted by experience
extended through length of service and in carrying out their duties.
138
As such, section 1 of the
Federal Trade Commission Act was held to preclude their removal except in the face of
inefficiency, neglect of duty, or malfeasance in office.
139

The Court further determined that Congressional limitations on the power of the President to
remove commissioners were further determined to be constitutional rendering the President
unable to remove commissioners absent a statutory basis.
140
The Court distinguished between
officers of the type at issue hereFederal Trade Commissionersand executive officers.
141
It
characterized Myers as limited to the question of the removal of executive officers.
142

In Morrison v. Olson
143
the United States Supreme Court reviewed a constitutional challenge to
the Ethics and Government Act of 1978 (the Ethics Act) based on the contention that it
violated the Appointments Clause by vesting appointment of independent counsel in the Special

128
272 U.S. 52, 47 S.Ct. 21, 71 L.Ed. 160 (1926).
129
Myers v. United States, 272 U.S. 52, 47 S.Ct. 21, 71 L.Ed. 160 (1926).
130
Id.
131
Id.
132
Id.
133
Id. at 620.
134
Humphreys Executor v. United States, 295 U.S. 602, 55 S.Ct. 869, 79 L.Ed. 1611 (1935).
135
Id. at 623.
136
Id.
137
Id. at 624.
138
Id. at 625-626.
139
Id. at 626.
140
Id. at 631-632.
141
Id. at 631-632.
142
Id. at 627, 628 (citing Myers v. United States, 272 U.S. 52, 47 S.Ct. 21 (1926)).
143
487 U.S. 654, 108 S.Ct. 2597, 101 L.Ed. 2d 569 (1988).
1000

Division, a court created by the Ethics Act.
144
The Supreme Court first determined that the
independent counsel is an inferior officer for purposes of Appointments Clause analysis since he
is subject to removal by the Attorney General, a superior officer.
145
The Court reasoned that
Congress was authorized to vest appointment power in the Special Division pursuant to the
Appointments Clause: . . . Congress may by Law vest the Appointment of such Inferior
Officers, as they think proper, in the President alone, in the Courts of Law, or in the Heads of
Departments (emphasis added).
146

The Court found the Special Division's authority to be derived from the Appointments Clausea
source of judicial authority separate and apart from Article III. Inasmuch as the Appointments
Clause allows Congress to delegate the power to appoint independent counsel it does not violate
Article III.
147

The Court next addressed the contention that the provisions creating the Special Division violate
Article III because they threaten the impartial and independent federal adjudication of claims
within the judicial power of the United States. The Court examined the validity of the Ethics
Act under a separation of powers rubric, dividing its analysis into two parts: whether the
limitation on the Attorney General's authority to remove independent counsel only on a showing
of good cause impermissibly interfered with the President's constitutional duties and whether
the Ethics Act interferes with the ability of the President to supervise the prosecutorial powers of
the independent counsel in violation of the separation of powers doctrine.
148

With respect to the removal issue, the Court focused on the fact that the independent counsel is a
less essential "inferior officer" for Appointments Clause analysis obviating the need to make the
decision terminable at will by the President.
149
The Court also found no impermissible burden on
the ability of the President to control the actions of the independent counsel since the President
still retains, through the Attorney General, the ability to remove the independent counsel upon a
showing of good cause.
150

Finally, the Court rejected the contention that the Ethics Act, in violation of the constitutional
doctrine of separation of powers, taken as a whole, unduly interferes with the function of the
Executive Branch since Congress, by passing the Ethics Act, did not: seek to increase its power
and decrease that of the executive branch; aggrandize power in the judiciary at the expense of the
executive branch; and did not impermissibly undermine the powers of the executive branch so as
to disrupt the proper balance between the branches of government.
151

III. Free Enterprise Fund v. Public Company Accounting Oversight Board
A. Background
This case originated in the United States District Court for the District of Columbia where the
Free Enterprise Fund (FEF) and a Nevada accounting firm sued the Public Company Accounting
Oversight Board (the Board). They contended that the Sarbanes-Oxley Act (the Act) by which
the Board was created violated the Constitutions separation of powers principles, the

144
Morrison v. Olson, 487 U.S. 654, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988).
145
Id. at 671.
146
Id. at 672, 676, 20 (citing U.S. Const. art. II, 2, cl. 2).
147
Id. at 676, 678-679.
148
Id. at 685.
149
Id. at 691.
150
Id. at 692-693.
151
Id.
1001

Appointments Clause, and the non-delegation doctrine.
152
The District Court, on joint motions
for summary judgment, found that the FEF's facial challenge to the Board was nothing but a
hypothetical scenario and that the FEF failed to establish an absence of any circumstances under
which Sarbanes-Oxley would be valid.
153
FEFs complaint was dismissed and it appealed to the
District of Columbia Circuit of the United States Court of Appeals.
The Circuit Court ruled that the creation of the Board violated neither the Appointments Clause
nor the Separation of Powers doctrine.
154
Its decision was based on the previously established
constitutionality of independent agencies and what it characterized as comprehensive control
over the Board by the SEC, all of which precluded a showing that the statutory scheme so
restricts the Presidents control over the Board as to violate separation of powers.
155
The
Circuit Court decidedover the dissent of Justice Cavanaughthat Sarbanes-Oxley did not
violate the Appointments Clause or the doctrine of Separation of Powers. It affirmed, in toto, the
District Court's summary judgment in favor of the Board and the United States (which had
intervened in support of the Act).
156
The United States Supreme Court granted certiorari to
determine whether Sarbanes-Oxley violated Article II by providing Board members with two
separate layers of protection from removal.
157

B. The Majority Opinion
Chief Justice Roberts wrote the majority opinion in this case of first impression. He addressed
the constitutionality of restricting the President's ability to remove a principal official when that
officials ability to remove an inferior officervested with the authority to determine policy and
enforce lawsis similarly restricted.
158
The Court held that this multi-level insulation from
removal violates Article II as an impermissible interference with the ability of the President to
take Care that the Laws be faithfully executed."
159
The Court addressed the following issues:
separation of powers;
160
severability;
161
and, the Appointments Clause.
162

The Court found the multilevel limitations on removal of Board members in contravention of the
Constitution's separation of powers doctrine:
We hold that such multi-level protection from removal is contrary to Article IIs
vesting of the executive power in the President. The President cannot take Care
that the Laws be faithfully executed if he cannot oversee the faithfulness of the
officers who executed them. Here, the President cannot remove an officer who
enjoys more than one level good-cause protection, even if the President
determines that the officer is neglecting his duties or discharging them
improperly. That judgment is instead committed to another officer, who may or
may not agree with the Presidents determination, and whom the President cannot
remove simply because that officer disagrees with him. This contravenes the

152
Free Enterprise Fund v. Public Company Accounting Oversight Board, No. 06-0217, 2007 WL 891675 at *1
(D.D.C. Mar. 21, 2007).
153
Id. at *6 (citing United States v. Salerno, 481 U.S. 739, 745 (1987)).
154
Free Enterprise Fund v. Public Accounting Oversight Board, 532 F.3d 667, 685 (D.C. Cir. 2008).
155
Id. at 685.
156
Id.
157
Free Enterprise Fund v. Public Co. Accounting Oversight Bd., 130 S.Ct. 3138, 3149 (2010).
158
Id. at 3147.
159
Id.
160
Id. at 3151-3161.
161
Id. at 3161-3162.
162
Id. at 3162-3164.
1002

Presidents constitutional obligation to insure the faithful execution of the
laws.
163

The Court divided its analysis of the dual for-cause limitations on the removal of Board members
into several inquiries: under the removal limitations, could the President take Care that the
Laws be faithfully executed as required under Article II?;
164
did the Act violate Article II which
makes the President responsible for executive branch actions, by requiring him to delegate
ultimate responsibility to others?;
165
and must the removal requirements imposed on the
executive branch be shown to enhance the power of the legislative branch to support a finding of
unconstitutionality?
166

The majority found Sarbanes-Oxley's limitations on removal of members of the Public Company
Accounting Oversight Board violative of the separation of powers doctrine.
167
The Court
carefully examined the landmark case of Myers v. United States.
168
Myers reaffirmed that in
furtherance of his duty to see that the laws are faithfully executed; the President is vested with
the general administrative control of those executing the laws.
169
According to Myers, this
responsibility compels the conclusion that the President must have the ability to remove "those
for whom he cannot continue to be responsible to carry out his duties.
170

With Myers as the cornerstone, the Court then explicated Humphreys Executor v. United
States.
171
It held that Congress couldwithout violating the principles laid down in Myers
limit the tenure of principal officers of some independent agencies to dismissal for good-
cause.
172
The fundamental difference between Myers and Humphreys Executor was that in
Myers the official worked in a unit of the executive department subject to the unfettered control
of the President while in Humphreys Executor, the officer employed in the Federal Trade
Commission was not a purely executive official but rather employed in an agency having
quasi-legislative and quasi-judicial duties on which Congress could impose removal
standards.
173
The Court also addressed the distinction between heads of department and inferior
officers, the subject of United States v. Perkins.
174
In Perkins the Court held that Congress could
impose removal restrictions on inferior officers whose appointment was vested in department
heads.
175

Based on its review of the statutory and common law, the Court held that the dual for-cause
removal limitations in Sarbanes-Oxley unconstitutionally exceeded the permissible restrictions
on the power of the President to remove executive officials.
176
The basis for the holding was that
the dual for-cause limitations not only insulate Board members from removal absent a showing
of good cause, but also foreclose the President's ability to make decisions as to whether good

163
Id. at 3147.
164
Id. at 3151-3154.
165
Id. at 3153-3155.
166
Id. at 3155-3157.
167
Id. at 3151. See 15 U.S.C. 7211(e)(6)(6); 7217(d)(3).
168
Id. at 3152 (citing Myers v. United States, 272 U.S. 52, 60, 47 S.Ct. 21, 71 L.Ed. 160 (1926)).
169
Id. at 3152.
170
Id. at 3152,
171
Humphreys Executor v. United States, 295 U.S. 602, 620, 55 S.Ct. 869, 79 L.Ed. 1611 (1935).
172
Id. at 3152 (citing Humphreys Executor v. United States, 295 U.S. 602, 620, 55 S.Ct. 869, 79 L.Ed. 1611
(1935)).
173
Id.
174
Id.
175
Id.
176
Id. at 3153.
1003

cause even exists.
177
Under the Sarbanes-Oxley rubric, the determination of good cause is vested
in commissioners of the Securities and Exchange Commission who are also shielded from the
direct control of the President. As such, according to the Court, the members of the Public
Company Accounting Oversight Board are not accountable to the President, resulting in a Board
for which the President is not responsible.
178
It is the insulation between the Commission and
the Board that creates the constitutional dilemma. Otherwise, the Commission would be able to
remove a Board member, thereby making the Commission responsible for actions of the Board
and the President could call the Commission to account for its supervision of the Board.
179

The double for-cause protection prevents the removal of a Board member by the Commission at
will and precludes the President from holding the Commission fully accountable for the actions
of the members of the Board.
180
This, the Court said, leaves the President impotent, except in
circumstances rising to inefficiency, neglect of duty, or malfeasance in office.
181
This
statutory scheme is repugnant to the President's duty to faithfully execute the laws and the
proscription against the delegation of his supervisory obligation.
182
The majority answered
several of the public policy questions raised by the dissent in support of the multi-level
protection that the dissent endorses which the Court considered the path to growth in the power
of the legislative branch.
183
Board members cannot be removed absent actions which amount to
willful violation or willful abuse or unreasonable failure.
184
The majority rejected this
unusually high" removal standard.
185
The dissent argued that the result sought by the majority
was accomplishedand the constitutional infirmity avoidedsince the Act vested the
Commission with the authority to establish Board functions. This, according to the dissent, vests
control over the removal of Board members in the Commission. The Court viewed this to be no
substitute since such a blunt instrument would require the Commission to destroy the Board in
order to fix it.
186
The Court also rejected the dissent's argument that adequate substitutionary
control could be exercised by the promulgation of SEC Rules or amendment to the rules of the
Board.
187

Since a portion of the Act was deemed unconstitutional, the Court was next required to consider
whether the offending portion could be severed, sparing the rest of the Act. The Court held that
the unconstitutional dual for-cause limitations on removal of Board members were severable
from the remainder of Sarbanes-Oxley.
188
The Court reasoned that the unconstitutionality of the
removal limitations did not necessarily defeat or affect the validity of its remaining
provisions.
189
So, the Court decided to limit the solution to the problem by striking out only
the offending provisions and leaving the rest.
190

177
Id.
178
Id.
179
Id. At 3153-3154.
180
Id. at 3153.
181
Id. at 3154.
182
Id..
183
Id. at 3156.
184
Id.
185
Id. at 3158.
186
Id. at 3158-3159.
187
Id. at 3159.
188
Id. at 3161.
189
Id. 3161.
190
Id.
1004

The Court also ruled that after the removal of the unconstitutional provisions the members of the
Board would then be removable at the will of the Commission and segregated from the President
by only one level of good-cause tenure. This, according to the majority, makes the Commission
responsible for actions of the Board and subject to Presidential oversight.
191
The remainder of
Sarbanes-Oxley will be fully operative as law and, as such, sustainable.
192

The FEF raised additional challenges under the Appointments Clause: (1) Board members are
principal officers and, as such, must be appointed by the President and confirmed by the Senate;
(2) the Securities and Exchange Commission is not a department for purposes of analysis under
the Appointments Laws; and, (3) the full Commission cannot appoint Board members because
only the Chairman is the head.
193
The Court rejected them all.
194

FEF's argument that members of a Board must be appointed by the President and confirmed by
the Senate was held to be without merit. Inferior officers are defined as those who are supervised
by someone above them who has been appointed by the President and confirmed by the
Senate.
195
Under this definition, members of the Public Company Accounting Oversight Board,
when viewed consistently with the remainder of the majority opinion are removable at will by
the Commission. So, they fall within the definition of inferior officers subject to appointment by
a department head so authorized by Congress.
196

Next, FEF urged that the SEC is not a department for purposes of Appointments Clause
analysis.
197
The Court previously reserved the question of whether agencies such as the SEC are
departments for this purpose.
198
The Court took this opportunity to go ahead and make the
determination that the SEC is a department for Appointments Clause analysis. The Court held
this conclusion to be in line with the contemporary definition of a department; the early practice
of Congress; and prior case law.
199
The Court also found it important that the Securities and
Exchange Commission is a separate part of the executive branch and not contained within any
other unit of that branch.
200

Lastly, the Court disposed of the argument that the full Commission cannot, consistent with the
Constitution; appoint members since the moniker Head applies only to the Chairman of the
Commission.
201
In rejecting this argument, the Court relied on several factors including:
Commissioners do not report to the Chairman; the Chairman is appointed from among the
Commissioners by the President; and a multi-member body may be considered head of a
department.
202
In addition, the Court reasoned that the Appointments Clause contemplates
collective appointments by the courts and Congress.
203

C. The Dissenting Opinion
Justice Breyer authored a dissenting opinionjoined by Justices Stevens, Ginsburg and
Sotomayorin which he vigorously argued that Sarbanes-Oxley violates neither the separation

191
Id. at 3161.
192
Id. at 3161-3162.
193
Id. at 3162-3164.
194
Id.
195
Id. at 3162
196
Id.
197
Id. at 3161.
198
Id.
199
Id. at 3162-3163.
200
Id. at 3163.
201
Id.
202
Id.
203
Id. at 3163-3164 .
1005

of powers doctrine nor otherwise significantly interferes with the executive power of the
President.
204
He identified the Court's task as determining whether the circumstances
surrounding Sarbanes-Oxley justify Congress's limitation on the President's authority to remove
Board members.
205
This, he said, requires examination of: the Necessary and Proper Clause;
206

the structural separation of powers doctrine;
207
and the President's duty to take Care that the
Laws be faithfully executed.
208
Justice Breyer observed that the statutory language, legislative
history, and Supreme Court precedenttraditional starting points in cases like thisprovide no
clear answer.
209
Faced with a non-textual question, he focused on how the removal
requirements of Sarbanes-Oxley, when considered in context, will function in the real world.
210

In the final section of his dissent, Justice Breyer focused on a harm created by the majority he
called far more serious than any imaginable harm this for cause provision might bring
about.
211
This, according to Justice Breyer arises as a result of the Court's failed attempt to
create a bright line rule forbidding multi-layered for cause removal requirements.
212
The Court
failed to create a definitive rule, thereby injecting much uncertainty as to the scope of the
holding.
213
Justice Breyer positsvia a series of rhetorical questionshis list of horribles to
result from the majority opinion.
214
He summarized his concerns about these questions as
follows:
Thus, notwithstanding the majoritys assertions to the contrary, the potential
consequences of todays holding are worrying. The upshot, I believe, is a legal
dilemma. To interpret the Courts decision as applicable only in a few
circumstances will make the rule less harmful but arbitrary. To interpret the rule
more broadly will make the rule more rational, but destructive.
215

III. Sparing The Board
The Supreme Court was faced with the prospect of abolishing the Board if it held Sarbanes-
Oxley unconstitutional. To understand why the Court went to so much trouble to keep the Board
in placewhy the Board was too critical to failit is important to focus on why Congress
created the Board. This is the successful approach several amici took to convince the Court to
spare the Board.
The three most influential amicus briefs on this issue were filed by the NASBA, the Institutional
Investors and the Former Chairmen of the SEC (SEC Chairmen).
216
To set the stage for

204
Id. at 3164.
205
Id. at 3165-3166 .
206
U.S. Const. art. I 8, cl. 18.
207
Id. at 3165.
208
Id.
209
Id. at 3167. See also Carol J. Miller & Stanley A. Leasure, Post-Kelo Determination of Public Use and Eminent
Domain in Economic Development under Arkansas Law, 59 Ark. L. Rev. 43, 54-56 (2006); Stanley A. Leasure,
Arbitration after Hall Street v. Mattel: What Happens Next?, 31 U. Ark. Little Rock L. Rev. 273 (2009).
210
Id.
211
Id. at 3178.
212
Id. at 3177.
213
Id.
214
Id. at 3178.
215
Id. at 3182.
216
Brief of Institutional Investors as Amicus Curie in Support of Respondents, Free Enterprise Fund v. Public
Accounting Oversight Board, 130 S.Ct. 3138 (2010) (No. 08-861), 2009 W.L. 3404247. Id. at 2009 W.L. 3370511
Brief for Amici Curiae Former Chairmen of the Securities and Exchange Commission in Support of Respondents,
Free Enterprise Fund v. Public Accounting Oversight Board, 130 S.Ct. 3138 (2010) (No. 08-861), 2009 W.L.
3404248 (hereinafter referred to as "SEC Chairmen").
1006

demonstrating the importance of the Board, these amici emphasized real world examples of
deficiencies in accounting oversight before the passage of Sarbanes-Oxley and the negative
effect these deficiencies had on the investor public. In addition to that, they outlined the
fundamental reliance state boards of public accounting placed on the Board in performing their
regulatory functions. NASBA, Institutional Investors, and SEC Chairmen touted the Board as a
significant factor in curing the failures of self-regulation.
A. Pre Sarbanes-Oxley Oversight
Historically, the accounting industry was subject to the oversight of the Public Oversight Board
(POB).
217
The POB was created and funded by a private association of accountants, the
American Institute of Certified Public Accountants (AICPA).
218
The accounting industry
dominated the organizations that were supposed to be overseeing them.
219
The POB disciplinary
processes were slow and ineffective and suffered from a number of limitations.
220

The POB had no power to sanction auditors for deficiencies or incompetence found during
quality control reviews.
221
In twenty-five years of existence, the POB never sanctioned a major
accounting firm, even when peer reviews uncovered serious shortcomings in a firms audit
procedures.
222
In addition, the POB was funded by voluntary dues paid by AICPA member
firms that audited public companies.
223
SEC Chairmen noted that, at one point, these firms went
so far as to cut off funding for POB review of major accounting firms.
224

SEC Chairmen buttressed Institutional Investors' point that auditor objectivity was declining
concomitantly with the decline in POB effectiveness. As the Big 8 became the Big 5, these
firms increasingly emphasized highly profitable consulting services to public company clients, in
addition to their audit relationship.
225
It was argued that increased revenue potential from non-
audit services impaired auditor objectivity and compromised insistence on disclosure of negative
financial facts.
226

SEC Chairmen and Institutional Investors also argued that the integrity of United States capital
markets was at stake at a time when an estimated 55% of American families own stock in public
companies.
227
The self-regulation of the POB era did not work and amici considered the Board
to be the answer. It therefore follows, according to these advocates that preserving the Board
would fulfill the goal of enhancing investor protection which lay at the core of the passage of this
legislation.
228

B. Post Sarbanes-Oxley Regulation
The Boards clearly delineated authority to establish auditing, attestation, quality control and
independence standards, subject to SEC approval, provides the necessary structure around which
a reformed American investment market is being built.
229
Under Sarbanes-Oxley, large firms
those with more than 100 public audit clientsare inspected at least annually; smaller firms at

217
Brief of NASBA at 30.
218
Id.
219
Brief of Institutional Investors at 23.
220
221
222
Id. at 12.
223
Id. at 10.
224
Brief of SEC at 11.
225
Brief of SEC at 10. Brief of Institutional Investors at 10.
226
227
Brief of SEC at 2. Brief of Institutional Investors at 14.
228
229
Id. at 22.
1007

least once every three years.
230
Violations found as a result of Board inspections have been the
basis for state accountancy board discipline in at least a half dozen cases.
231
The Board has also
undertaken several enforcement actions, including, in December 2007, fining Deloitte & Touche
$1 million for knowingly permitting an incompetent auditor to handle the audit of a
pharmaceutical company.
232

The structure of the Board and its funding also contribute to its independence from the
accounting industry it regulates. Only two of the five members of the Board are allowed to be
accountants.
233
Subject to SEC approval, the Board is empowered to establish its own annual
budget and to fund that budget by collecting accounting support fees from public companies,
brokers and dealers.
234

NASBA focused on the criticality of the Boards continued existence vis--vis the regulation of
public accountancy at state and federal levels.
235
NASBA identified the regulatory void which
would be left if the Court eliminated the Board by striking down Sarbanes-Oxley in its
entirety.
236
Calling the Board vital to the protection of U.S. financial markets it argued that its
abolishment would place the burden of regulating the accounting industry on the shoulders of
state boards of public accountancy since the SEC has insufficient resources to fill the void.
237

NASBA noted that although state boards of public accountancy have the power to sanction
certified public accountants, they rely on the Board to perform a central role in both the
investigatory and disciplinary processes.
238

Sarbanes-Oxley requires the Board to perform inspections of public accounting firms engaged in
auditing publicly traded companies and report findings to appropriate state boards.
239
The
regulation of the practice of public accounting and the resulting protection of the public would
suffer, according to NASBA, if the Board fell.
240

From a more technical, though critically important, standpoint the nationwide organization of
accounting regulatory bodies also pointed out that decisions and standards of the Board have
been incorporated into public accounting statutes and rules in states across the country.
241
Others
have adopted disciplinary orders of the Board as prima facie evidence of violation of state

230
Id. at 28.
231
232
Brief of Institutional Investors at 30-31.
233
Id. at 15.
234
Id.
235
Id. In addition to making its public policy arguments, NASBA incorporated by reference the legal arguments
concerning the severability issue, made by Institutional Investors and discussed supra. Id.
236
Id.
237
Id.
238
Id.
239
Id. (citing 15 U.S.C. 7214 (c) (2); 15 U.S.C. 7214 (g) (1)). As the NASBA noted in its brief, the Board also is
authorized to refer its ongoing investigation to the state boards of public accountancy and violations uncovered by
the Board have supported disciplinary actions by several state boards of public accountancy. Id.
240
Id.
241
Id. at 29. The states which have adopted Board standards include Alabama, Arkansas, Connecticut, Kansas,
Louisiana, Mississippi, Missouri, Montana, New Hampshire, New Jersey, Pennsylvania, South Carolina, South
Dakota, Utah and Vermont. Id.
1008

accountancy laws.
242
In the same vein, Board standards are incorporated into the Uniform
Accountancy Act adopted by the AICPA and the NASBA.
243

The National Association of State Boards of Accountancy described the havoc which would be
wreaked if this litigation led to the demise of the Public Company Accounting Oversight Board:
[T]he Board . . . is an integral component of accountancy regulation in the United
States. It plays a vital role in regulating accounting firms and accountants who
audit publicly created companies. Any decision by this Court to judicially abolish
the Board would leave this segment of the accounting profession largely
unregulated. . . .
244

Institutional Investors addressed the severability issue from a different perspective.
245
The
answer is to excis[e] the invalid portions and, in accordance with congressional intent, preserve
the Board as a vital protector of the investing public, said Institutional Investors.
246
Its argument
centered on three points: the judicial preference for a limited solution to a limited problem;
247
the
presumption that a defective statute is presumed severable even without a severability clause;
248

and severability being a question of legislative intent.
249
Institutional Investors argued that the
Sarbanes-Oxley limitations on the removal of SEC Commissioners is remediable by rewriting
the statute to either give the SEC generic for-cause removal authority or leaving the SEC with
at will removal authority.
250
As described in Ayotte v. Planned Parenthood, this would result in
only striking that portion of the legislative effort necessary to remedy the constitutional defect
and thereby: upholding the intent of the elected legislative body; fixing the problem without
rewriting Sarbanes-Oxley in its entirety; and affirming congressional preference that the
constitutionally sound portions of Sarbanes-Oxley survive.
251

On the severability question, Institutional Investors argued that the Acts legislative history
supports the conclusion that Congress would have created the PCAOB even without restrictions
on the ability of the SEC to remove members of the Board.
252
The Act's purpose was to create a
regulatory framework separate and apart from the accounting profession to provide additional
protection to investors not previously available.
253
Enhanced SEC oversight of the Board
evidences the intent to make the Board independent of the accounting professionnot
independent of the SEC.
254
Furthermore, the legislative history is devoid of evidence of intent to

242
Id. States adopting Board auditing standards include: Alabama, Alaska, Arkansas, Colorado, Florida, Hawaii,
Idaho, Iowa, Kansas, Louisiana, Michigan, Minnesota, Montana, New Jersey, North Carolina, Oregon,
Pennsylvania, South Dakota, Tennessee, Texas, Vermont, Washington, and Wyoming. Id.
243
Id. at 30.
244
Id. at 31.
245
Brief of Institutional Investors as Amici Curiae Supporting Respondents at 32, Free Enterprise Fund v. Public
Company Accounting Oversight Board, 130 S.Ct. 3138 (2010) at 7-31 (No. 08-861), 2009 W.L. 3370511.
246
Id. at 32, citing Ayotte v. Planned Parenthood, 546 U.S. 320, 328-329, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006).
247
Brief of Institutional Investors at 32 (citing Ayotte v. Planned Parenthood, 546 U.S. 320, 328-329, 126 S.Ct. 961,
163 L.Ed.2d 812 (2006)).
248
Id. at 33.
249
Id. (citing Regan v. Time, Inc., 468 U.S. 641, 653, 104 S.Ct. 3262, 82 L.Ed.2d 487 (1984)).
250
Id. at 32 (citing Ayotte v. Planned Parenthood, 546 U.S. 320, 328-329, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006)).
251
Id. at 32-33. Congress passed Sarbanes-Oxley without a severability clause. However, the Court has indulged a
presumption in favor of severability. Id. at 33.
252
Id. at 32-33 (citing Regan v. Time, Inc. 468 U.S. 641, 653, 104 S.Ct. 3262, 82 L.Ed.2d 487 (1984)).
253
Id.
254
Id. at 34.
1009

inhibit executive branch authority over the Board. In fact, the removal power was never debated
and during hearings, mentioned only briefly.
Not missing an opportunity to mention a recent auditing scandal, Institutional Investors even
discussed proposed legislation that could give the Board additional authority over registered
firms, to better protect investors from Ponzi schemes like the Madoff fraud involving non-
registered auditors.
255
These demonstratedand perceivedadvantages over the old POB
system helped the amicus curiae convince the Court that the Board was too critical to fail.
IV. Conclusion
After determining that Sarbanes-Oxley violated Article II of the Constitution; what next the
Courts recognition that PCAOB was too critical to fail and the need to find a way to save it? Or
maybe the legal analysis of severability and the conclusion that the offending portion could be
severed, sparing the rest and, unintentionally, the Public Company Accounting Oversight Board?
Leaving nothing to chance, amicus briefs filed by the National Association of State Boards of
Accounting, Institutional Investors and SEC Chairmen presented the Court with compelling legal
and public policy arguments to save the PCAOB.
In the end, Free Enterprise Fund v. Public Company Accounting Oversight Board provides a
good history lesson on separation of powers, severability and the regulation of the accounting
industry. That notwithstanding, it did not result in any significant changes to the Sarbanes-Oxley
Act or the regulation of the accounting industry. Given the profound impact of the Public
Company Accounting Oversight Board on public company financial reporting generally, and the
public accounting profession specifically, the significance of the decision is obvious; its wisdom
remains in question.

255
Id. at 28.
1010

Exchange Rates and Stock Market Returns In Mexico: A Markov Regime-
Switching Approach

Dr. Ricardo Tovar-Silos
Department of Information Systems and Analysis
Lamar University
Beaumont, TX 77710
ricardo.tovar-silos@lamar.edu

1011

Abstract
In this paper, we apply a Markov regime-switching model to the time series of the Mexican
Peso-U.S. Dollar exchange rate and to the Mexican stock market returns. The existence of two
states is statistically significant in both cases. The exchange rate is characterized by extended
periods of low variance and slight appreciations followed by short-lived high variance,
depreciationary states. The stock market is characterized by long periods of low volatility and
positive returns followed by short periods of increased variance and negative returns. We found
that the states of these time series are in sync 71% of the time due to the persistence of the low
volatility state and evidence of positive correlation between the regimes.
Keywords: Exchange rates, stock prices, volatility, regime-switching, concordance, Mexico
JEL: C52, F31

I. Introduction
The empirical relationship between exchange rates and stock markets in emerging economies
such as Mexico is of great importance for both academia and investors. A large body of literature
seeks to explain the relationship between these two financial markets. We can classify these
theories in two main branches. The first line of thought argues that depreciation of the local
currency is beneficial to exporters which in turn lead to better prospects and higher valuations of
companies resulting in higher stock prices. In short, the empirical prediction is that exchange
rates and stock returns are positively correlated.
The second theory is based on the idea that during economic downturns stock prices on average
decrease. In emerging economies where capital mobility is not restricted such as Mexico,
adjustments in investment portfolios (inflows/outflows of foreign capital) occur. In particular,
foreign capital flies out of the country and a depreciation of the local currency is observed.
Domestic monetary policy may reinforce in the short run this depreciationary process if interest
rates decrease in an attempt to foster economic growth making less attractive to invest in local
(Peso-denominated) assets. The findings of Sosa (2009) seem to support this theory in the case of
Mexico. He found that the business cycles of Mexico and the U.S. exhibit a great degree of
synchronicity since the NAFTA was signed and that U.S. shocks explain a large share of
Mexicos macroeconomic fluctuations. Then, the expected and observed chain of events is as
follows: if the U.S. enters into an economic downturn or a recession Mexico will also be affected
and the country will experience a fall in the stock market along with a depreciation of the Peso
caused by the outflows of foreign capital. Other external factors, such as crises in other emerging
economies, can also increase the risk-aversion of investors who will adjust their portfolios in
consequence. The empirical prediction of this approach is that stock returns and exchange rates
are negatively correlated.
Most literature focusing on the relationship between exchange rates and stock markets, tries to
explain the direction of causality between these variables. The results are inconclusive and at
times contradictory just as the empirical predictions of the theories discussed above. Most of
these studies have relied on Granger causality tests. For example, Kutty (2010) found that, in the
case of Mexico, stock prices lead exchange rates in the short run and there is no long run
relationship between these two financial variables. In this paper, we propose an alternative
approach to the analysis: we model the non-linearities (cycles) of the time series with a Markov
regime-switching model as proposed by Hamilton (1989). We identify and date two distinct
states in the two time series. We then describe the characteristics of these two phases and
measure their degree of synchronicity. We found that in the case of Mexico negative stock
1012

returns periods are associated with currency depreciating periods 71% of the time and that the
states are significantly positively correlated. In other words, in the case of Mexico there is a
negative relationship between exchange rates and stock market returns.
It is well-documented that the time series of the exchange rate of the Mexican Peso is
characterized by the existence of two clearly defined states. Bazdresch and Werner (2005) found
evidence for two clearly identified regimes: one with an appreciating trend and low volatility,
and another with large depreciations and high volatility. Cheung and Miu (2008) have challenged
these findings arguing that one can falsely report statistically significant distinct regimes that are
in fact the ARCH effect.
The application of regime-switching models to the behavior of the stock market is also
widespread. Hamilton and Lin (1994) investigated the joint time-series behavior of monthly
stock returns and growth in industrial production in the United States. They studied a bivariate
regime-switching model and found that stock returns are well characterized by year-long
episodes of high volatility separated by longer quite periods whereas real output is subject to
abrupt changes in the mean associated with economic recessions. They also concluded that
economic recessions are the primary factor that drives fluctuations in the volatility of stock
returns. Schaller and Van Norden (1997) used an extension of Hamiltons (1989) Markov
switching techniques to describe and analyze stock market returns. They found very strong
evidence of switching behavior and identified two regimes: a low-return state and a high-return
state.
To the extent of our knowledge no paper has ever used Markov regime-switching models to
address the relationship between exchange rates and stock markets in Mexico. The primary
objective of our investigation is to analyze and describe the relationship between the Mexican
Peso exchange rate and the Mexican stock market. To attain our objective, we use Hamiltons
(1989) Markov regime-switching model. In particular, we update and confirm the results of
Bazdresch and Werner (2005) for the period 1994-2008. Then, we apply a similar model to the
Mexican stock market and we find evidence of the existence of two well-defined states: a
negative return-high volatility state and a high positive return-low volatility state. Finally, we
address the relationship between exchange rates and stock market returns in Mexico by
measuring the degree of synchronicity of the states in the two series and testing to determine if
they are correlated.
Our results seem to support the theory that economic downturns leading to drops in the stock
market coincide with periods of exchange rate depreciation that may be the result of foreign
capital being taken out of the country by risk-averse investors. The paper is structured in the
following way. In Section II the data is presented. Section III presents the model and results,
whereas Section IV analyzes the concordance and correlation between the two time series.
Section V concludes.
III. Data
Data on the daily exchange rate of the Mexican peso was obtained from Banco de Mexico for the
sample period January 1996 to December 2008. Also, data for the stock market was obtained
from Dow Jones News/Retrieval. It consists of daily closing of the index IPC, Mexicos market
capitalization weighted index of the leading 35-40 stocks. After eliminating some of the
incompatible data, 3,256 data points were generated.

1013

IV. Model
In this section, we start the analysis by testing to determine if there is evidence of distinct
regimes in exchange rates and stock market returns. In a formal sense, regime-switching
econometric models refer to a situation in which the realizations of a time series are drawn from
two different distributions, with some well defined stochastic process determining the likelihood
that each realization of the variable is drawn from a given distribution.
To determine whether there is regime-switching in the time series of exchange rates and stock
returns, we consider two specifications. The first specification is an AR(1) model which can
serve as a benchmark for the examination of regime-switching effects. For the purpose of this
study, we define x to be the exchange rate and r to be stock market returns. Specifically, we
consider the following model for the time series:
x

,
(1)
r

,

where
x
logx
logx
logr
logr

,
~ iid N0,
and
,
~ iid N0,

This is the specification used for the null hypothesis of no switching. The alternative hypothesis
implies that exchange rates and stock returns are drawn from distributions with different means
and variances:

x

,
(2)
r

,

where
,
~ iid N0,
,
and
~ iid N0,
,
, S
0,1
Under the alternative hypothesis, the distribution from which the variable is drawn is determined
by the state variable S
t
which in any given point in time is either 0 or 1.
Two important elements of this regime-switching model are the proportion of time the process
stays in each state and the probabilities of transition between different states. In order to use
these models for forecasting it is necessary to estimate these probabilities as well as the
coefficients for each regime. The probabilities of transition from one state to another are
expressed in a matrix P:

p

(3)
where p
ij
represents the probability of going to state j conditioned on being in state i. The
proportion of time spent on each state (ergodic probabilities) is easily computed from this matrix.
With this specification, we expect the variables exchange rate and stock market to go from one
state to another according to certain probabilities, and to remain in any of them for several
periods at a time.
Table I presents the likelihood for the null and the alternative hypotheses. As the table shows,
when we test (2) against the null (1) the likelihood ratio statistic is 1385.65 for exchange rates
and 898.89 for stock returns. Garcia (1992) shows that the 5% critical value for the likelihood
ratio statistics is 13.52 and the 1% critical value is 17.67. Our results therefore imply very strong
rejection of the null hypothesis of no switching for both time series.

1014

Table I. Test for Regime Switching in Exchange Rates and Stock Market Returns

Exchange Rate Stock Market Returns
No Switching Switching No Switching Switching
Log Likelihood 15165.569 15858.396 11800.612 12250.569
Likelihood Ratio 1385.65 898.89
This table reports tests of the null hypothesis of no switching against an alternative specification of switching in both
means and variances. The 5% and 1% critical values for this non-standard distribution are 13.52 and 17.67 as
tabulated in Garcia (1992).

The empirical results of the regime-switching specification are shown in Table II where we
report the means (
,

,
,
), the autoregressive coefficients (

x
,
r
) and the variances
(
,
,
,
) of the processes driving movements of exchange rates and stock market returns
under each regime.

Table II. Maximum Likelihood Estimates of the Regime-switching Models

Exchange Rate Stock Market
, -0.0082 0.1257

(0.0078) (0.025)
,
0.2485 -0.0721

(0.0787) (0.0838)
0.0358 0.0928

(0.0183) (0.018)
0.1501 0.9842

(0.0052) (0.0416)
1.8192 5.8577

(0.1797) (0.3148)
LF 15858.396 12250.569

i x, r, standard errors in parenthesis. LF is the log likelihood of the regime-switching models.
In the case of exchange rates, state 1 has a variance that is more than 12 times that of state 0. In
contrast, in the case of stock market returns, the variance in state 1 is only approximately six
times as large as the variance of state 0. For this reason, we call state 0 the low variance state.
Based on the ratio of the variances in the two states, it is clear that exchange rates exhibit greater
volatility than stock market returns. From the estimates of the means, we found that in the case
of exchange rates, the low variance regime mean
x,0
is

negative and small, implying that this
regime goes along with small daily appreciations in the exchange rate (0.008%). The mean in the
high variance state
x,1
is positive and much larger implying large depreciations (0.25%).
In the case of stock market returns, the mean return in state 0 is positive (0.13%) suggesting that
if state 0 were to persist for one year it would increase stock prices by about 40.3%. On the other
1015

hand, the mean return in state 1 is negative (0.07%), so that if state 1 were to persist for one year
then stock prices would be reduced by 17.7%. The results imply that the Mexican stock market
is characterized by a state in which risk (measured by the variance) is relatively low and
investors earn a substantial return and a state in which risk is substantially higher and investors
lose money.
Table III presents the probabilistic results. We consider the exchange rates results first. Note that
the low variance state is very stable; according to these estimates the Markov chain stays in this
regime for 76.92 days, on average. The high variance state is considerably less stable; on
average, the Markov chain stays in this regime 9.52 days at a time. In the case of stock market
returns, the low variance-positive return state is, as in the exchange rates case, very stable with
an average duration of 55.56 days. The high variance-negative return state on the other hand
shows and average duration of 27.03 days.
256

Table III. Regime-switching Model Probabilistic Results

Exchange Rate

Transition probability matrix P
Ergodic Probabilities Expected duration of system
in each regime
State 0 1
0 0.987 0.105 0.893 76.92
1 0.013 0.895 0.107 9.52

Stock Market

Transition probability matrix P
Ergodic Probabilities Expected duration of system
in each regime
State 0 1
0 0.982 0.037 0.671 55.56
1 0.018 0.963 0.329 27.03

From the estimates of the ergodic probabilities, it is clear that both time series are usually in the
low variance regime, in particular exchange rates stay there 89.3% of the time whereas stock
returns stay there 67.1% of the time. Given the high degree of persistence of the low variance
state, we expect the two time series to be in the same regime most of the time. In other words,
two time series may show a high degree of concordance if one state is very persistent but this
does not necessarily mean that they are correlated. We will formally compute the concordance
index and test for a significant positive correlation between the two series in the following
section.
Every day, each of the time series behaves according to either one of two regimes. Figure 1
shows the daily closing value of the exchange rate alongside the estimated probability of having
been a high variance regime day. These probabilities are the so-called smoothed probabilities,
computed by taking into account the entire sample of observations, and not just observations
prior to that point in time. In total, we found 30 periods of high volatility in exchange rates in the
sampled period 1996-2008. Some descriptive statistics of these 30 periods are shown in Table
IV. The distribution of the length of these periods is skewed right (not shown). The median
provides a better idea of the typical length of these periods and it is equal to 6 days which is less
than the mean of 9.52 days we showed before. There are two outliers in the distribution that

256
The mean length of time i is equal to 1/ (1- p
ii
).
1016

correspond to the periods of August 10, 1998-October 9, 1998, and September 10, 2008-
December 31, 2008 which correspond to the Russian financial crisis and the Global financial
crisis respectively.
Table IV. Descriptive Statistics of high-variance regime periods

Number of
periods
Mean Duration
(days)
Median
(days)
Outliers
Exchange
Rates
30 9.52 6 August 10,1998-October 9,1998
September 10,2008-December 31,2008
Stock
Returns
37 27.03 18 July 23,1998-December 15,1998
March 1,2000 to July 31,2000
August 24,2000 to January 11,2001

We perform a similar analysis in the case of stock market returns. Figure 2 shows the IPC index
alongside the smoothed probabilities of having been a high variance and negative return regime
day. In total, we found 37 periods of high volatility in the stock market in the sampled period
1996-2008. Some descriptive statistics of these 37 periods are shown in Table IV. As in the case
of the exchange rate, we find that the distribution is skewed right. The median length of these
periods is 18 days whereas, as shown before, the mean length is 27.03 days. Also, there are three
outliers that correspond to the periods of July 23, 1998 - December 15, 1998, March 1, 2000 -
July 31, 2000 and August 24, 2000 - January 11, 2001 which correspond to the Russian financial
crisis and to the dot-com bubble burst.

0
2
4
6
8
10
12
14
16
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
-
4
-
9
6
8
-
4
-
9
6
3
-
4
-
9
7
1
0
-
4
-
9
7
5
-
4
-
9
8
1
2
-
4
-
9
8
7
-
4
-
9
9
2
-
4
-
0
0
9
-
4
-
0
0
4
-
4
-
0
1
1
1
-
4
-
0
1
6
-
4
-
0
2
1
-
4
-
0
3
8
-
4
-
0
3
3
-
4
-
0
4
1
0
-
4
-
0
4
5
-
4
-
0
5
1
2
-
4
-
0
5
7
-
4
-
0
6
2
-
4
-
0
7
9
-
4
-
0
7
4
-
4
-
0
8
1
1
-
4
-
0
8
P
e
s
o
s

p
e
r

U
.
S
.

D
o
l
l
a
r
P
r
o
b
a
b
i
l
i
t
y
Figure 1. Exchange rate and probability of high volatility regime
1017

V. Concordance
In this section, we turn our attention to the degree of synchronicity of the regimes of exchange
rates and stock returns in Mexico. Harding and Pagan (2003) measure the degree to which two
business cycles are in sync by the percentage of time the two economies were in the same regime
their degree of concordance. We will apply the same methodology to analyze the extent to
which the fluctuations of exchange rates and stock markets are in sync. Specifically, the degree
of concordance between the fluctuations of the exchange rate and the stock market
C
,
T
S
,
S
,
1 S
,
1 S
,
(4)
where t denotes the period and T is the total number of periods. S
x,t
is the regime of the exchange
rate time series at time t and S
r,t
is the regime of the stock market return time series at time t. The
exchange rate-stock market concordance measure is reported in Table V, it indicates that these
time series are in the same state 71% of the time. Note that this concordance measure should be
interpreted relative to an expected value for C
x,r
under the null hypothesis that the regimes of
exchange rates and stock market returns are uncorrelated, which need not be zero if one regime
is more persistent than the other.
257
Table V also reports this expected concordance measure. The

257
For example, as it was noted before, for each time series the low variance regime is significantly more persistent
than the other, so even if they were uncorrelated, we could still expect a strong degree of concordance between the
two time series. Harding and Pagan prove that the degree of concordance C
x,r
can be rewritten as
,
1
2
, where
S
is the correlation coefficient between
,
and
,
,
,
and
,
, then under the null of no correlation (
S
=0) the expected value of C
x,r
,
,
1 2
.
0
5000
10000
15000
20000
25000
30000
35000
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
-
4
-
9
6
8
-
4
-
9
6
3
-
4
-
9
7
1
0
-
4
-
9
7
5
-
4
-
9
8
1
2
-
4
-
9
8
7
-
4
-
9
9
2
-
4
-
0
0
9
-
4
-
0
0
4
-
4
-
0
1
1
1
-
4
-
0
1
6
-
4
-
0
2
1
-
4
-
0
3
8
-
4
-
0
3
3
-
4
-
0
4
1
0
-
4
-
0
4
5
-
4
-
0
5
1
2
-
4
-
0
5
7
-
4
-
0
6
2
-
4
-
0
7
9
-
4
-
0
7
4
-
4
-
0
8
1
1
-
4
-
0
8
I
P
C
P
r
o
b
a
b
i
l
i
t
y
Figure 2. Mexican stock market (IPC) and probability of high volatility regime
1018

degree of concordance ( 0.712 ) is greater than the expected value ( E
C
,
= 0.6393 ) under the
null of no correlation but it still remains to be determined if this difference is statistically
significant. To see if the two series are correlated we test the following set of hypotheses:
H
0
:
S
=0 vs. H
a
:
S
0 (5)
where
S
is the correlation coefficient between S
,
and S
,
.
S
can be found from the regression:

S
,
a
S
,
u
(6)
The estimated value of
S
was found to be 0.262064 in this regression. The t-test was equal to
15.49 and provided a p-value equal to 0. However, one should be cautious about these results
because as pointed out by Harding and Pagan, when the null
S
=0 holds the error term inherits
the serial correlation properties of S
,
, which is strongly positively correlated and as known
positive serial correlation highly increases the chances of rejecting the null, unless inferences are
made robust to serial correlation as well as to any heteroskedasticity in the errors. Consequently,
we also report the robust t-test calculated with Newey-West standard errors. The corrected t-test
takes the value of 5.74 and it is significant at .01, so we reject the null that the regimes are
uncorrelated. We conclude that the cycles of the exchange rate and the stock market are 71% of
the time in the same state because the states are highly persistent and because they are positively
correlated. We leave for future research the analysis of the sources of correlation between the
cycles. Factors affecting the global economy may be a possible explanation for the sample period
of this study. Another potential explanation may be the link of the business cycles of Mexico and
the United States, its primary commercial partner.
Table V. Concordance and Correlation Results
Concordance Null Expected
Concordance
Correlation Standard
t-statistic
Robust
t-statistic
0.712 0.639 0.262 15.49* 5.74*
*Significant at .05

VI. Conclusions
In this paper, we estimated regime-switching models for the time series of the Mexican Peso
exchange rate and the Mexican stock market returns. Both time series are characterized by the
existence of two regimes: an extended low volatility regime with small appreciations in the case
of exchange rates and with positive returns in the case of the stock market, and another short-
lived regime characterized by high volatility and depreciations of the currency and average
negative returns. There is a high degree of concordance between the regimes of these two time
series caused mainly by the dominance of the low variance regime in both time series but also
caused by a positive correlation between the two time series. For future research we propose to
analyze the causes underlying the positive correlation found between the regimes of the two time
series.

References
Bazdresch, Santiago, and Alejandro Werner, 2005, Regime switching models for the Mexican
peso, Journal of International Economics 65(1), 185-201.
Cheung, C.S., and P. Miu, 2009, Currency Instability: Regime Switching versus Volatility
Clustering, Quarterly Journal of Finance and Accounting 48(1), 67-81.
Garcia, Rene, 1992, Asymptotic Null Distribution of the Likelihood Ratio Test in Markov
Switching Models, University of Montreal.
1019

Kutty, G., 2010, The relationship between exchange rates and stock prices: the case of Mexico,
North American Journal of Finance and Banking Research 4, 112.
Hamilton, James B., 1989, A New Approach to the Economic Analysis of Nonstationary Time
Series and the Business Cycle, Econometrica 57(2), 357-384.
Hamilton, James D. and G. Lin, 1996, Stock Market Volatility and the Business Cycle, Journal
of Applied Econometrics 11, 573-593.
Harding, D. and A. Pagan, 2002, Synchronization of Cycles, mimeo.
Huntley Schaller & Simon Van Norden, 1997, Regime switching in stock market returns,
Applied Financial Economics 7(2), 177-191.
Sosa, Sebastian, 2008, External Shocks and Business Cycle Fluctuations in Mexico: How
Important are U.S. Factors? IMF Working Papers 08/100, 1-31.

1020

Determinants of Foreign Direct Investment in Emerging Markets: Evidence
from Nigeria

Dr. Vigdis W. Boasson
Department of Finance and Law
Central Michigan University
Mount Pleasant, MI 48859, USA
boass1v@cmich.edu

Dr. Emil Boasson
Department of Business Information Systems
Central Michigan University
Mount Pleasant, MI 48859, USA
boass1e@cmich.edu

Mr. Hameed H. Salihu
Davenport University
Grand Rapids, MI 49512, USA

1021

Abstract
In this paper, we examine the determinants of the flow of Foreign Direct Investment (FDI) to
Nigeria. Specifically, we investigate the effects of exchange rate volatility on FDI flows while
controlling for macroeconomic factors. We first conducted Dickey-Fuller unit roots tests for our
time-series data to determine any trend elements in the sample data. We then conducted
Johansen cointegration test for the long run relationships among the determinants of the FDI in
emerging markets. Our findings show that a significant long-run relationship exists between FDI
and exchange rate volatility. Using Vector Error Correction regression model (VECM), we find
that exchange rate volatility has statistically significant negative impacts on the flow of FDI to
Nigeria. We also find that market size, exchange rate, infrastructural development, and world
commodity prices have positive and statistically significant impacts on the FDI flow to Africa.

KEYWORDS: Foreign Direct Investment, Exchange Rate Volatility, Macroeconomic factors,
Nigeria Economic Policy
JEL Classification: F21; F31; G11

I. Introduction
Foreign Direct Investment (FDI) has been an important engine of economic growth and
development for both developed and developing economies (Wafure and Nurudeen (2010),
Onyeiwu and Shretha (2004), Asiedu (2002)). The important role of FDI as a major source of
financing and engine of economic growth for developing countries have led to a great deal of
research as to what factors that matter for the decisions made by foreign investors to transfer
capital to developing countries. Identifying key FDI determinants could provide policy
implications for developing nations to attract FDI inflows. Despite the rapid growth of FDI
flows to developing countries, the pattern of growth is hugely uneven and many African nations
are left out of the FDI growth. In Nigeria, despite several reform efforts taken by the Nigerian
government to boost foreign direct investments, the flow of FDI to the country is still relatively
insignificant as compared to the FDI flows to other developing nations in East Asia and Latin
America. The poor and declining flow of FDI to African region motivates our quest for answers
to the following questions:
1. What are the key factors are that attract FDI flows to Nigeria?
2. What are the impacts of exchange rate volatility and other key macroeconomic
fundamentals on FDI flows to Nigeria?
3. What roles do FDI play in the Nigerian economic growth and development process and
what policies, programs and reform efforts are needed to improve these roles of FDI if
any?
Despite a relatively large body of literature on foreign direct investment, there is still a scarcity
in the research that focuses on the FDI in Africa and Nigeria in particular. Thus, this paper aims
to fill this void by focusing on identifying specific internal and external factors that determine
the FDI flows to Africa and to Nigerian in particular. The rationale for selecting Nigeria as our
study sample for the FDI flows to Africa is that Nigeria is by far the largest African country in
terms of population.
This paper tests the proposition that FDI flows to Nigeria is significantly influenced by exchange
rate volatility and macroeconomic fundamentals, notably internal and external macroeconomic
factors. In order to examine this proposition, this paper empirically tests the impacts of
exchange rate volatility, internal factors such as domestic market size, inflation, financial
1022

development, infrastructures, as well as external factors such world commodity price index,
degree of openness, and crude oil price, on the flows of FDI to Nigeria. Our sample period
ranges from 1970 to 2010. Employing unit root tests and Johansen cointegration tests, we find
that the flow of FDI to Nigeria is largely determined by exchange rate volatility as well as other
major macroeconomic variables and the results are statistically significant. Specifically, we find
that domestic market size, exchange rate volatility, domestic interest rate, infrastructures, and
world commodity price index have important influence on the flow of FDI to Nigeria. The
primary conclusion that follows from the empirical investigation is that exchange rate volatility
and major macroeconomic fundamentals have strong influence on the ability of Nigeria to attract
FDI to the country. This conclusion is supported by the fact that with more stable
macroeconomic conditions that follow the transition to a democratic government in Nigeria
in1999, more and more foreign firms have been transferring investment capital to the country.
Our contributions to the literature are can be summarized in four aspects. Firstly, our research
framework incorporates the empirical analysis of both internal and external factors that might
influence the flow of FDI to Nigeria. We examine the relationships among the FDI flows to
Nigeria and exchange rate volatility while controlling for major macroeconomic fundamentals.
The major macroeconomic fundamentals tested in this study include domestic market size,
inflation rate, domestic interest rate, infrastructures, per capita GDP, country risk, exchange rate,
world commodity price, degree of openness and crude oil price, in addition to exchange rate
volatility. Theoretically, domestic market size, infrastructures, per capita GDP, degree of
openness, commodity price and crude oil price are all expected to have positive relations with
FDI inflows to Nigeria. Exchange rate, domestic interest rate, inflation rate and exchange rate
volatility, on the other hand, are expected to relate negatively to FDI inflows to Nigeria.
Secondly, our study bridges the gap inherent in prior empirical literature on the determinants of
FDI flows to African region. Specifically, prior literature tends to focus more on cross-country
analysis of FDI determinants in Africa. Since the actual economic, political, social and cultural
conditions in African countries are significantly different from one country to another, cross-
country analysis may not provide reliable estimates useful for investment decisions making and
for policy formulations by an individual African country. Thirdly, most studies on FDI
determinants in Africa are very narrow in scope and analysis. Majority of these studies do not
consider the core macroeconomic fundamentals specific to African situation on an individual-
country basis. Recognizing these limitations, our study attempts to bridge these gaps by focusing
on country-specific macroeconomic fundamentals that influence the transfer of foreign capital
and investment decisions to African region. Finally, despite the fact that this study uses Nigeria
as the study sample, the approach employed for this research could easily be applied to the
analysis of other African countries.
The paper is structured as follows. Section 2 reviews the relevant literature. Section 3 presents
the data, working hypothesis, and measurements of variables in our statistical tests. Section 4
describes our research methods and regression models. Empirical results are presented and
discussed in section 5. Section 6 concludes the paper.

II. Research Context
Nigeria is the most populous country in the sub-Sahara Africa with a population of well over 150
million. The economic development of Nigeria has been closely tied to the increased revenue
derived from crude oil production and exploration following the discovery of the first oil well in
the Niger Delta region of the country by the British in the 1950s. The revenue from oil
1023

production and exploration accounts for over 95% of the countrys annual exports earnings, over
80% of federal government annual revenue, about 50% of the countrys GDP, over 90% of
foreign exchange earnings. Consequently, there has been a high influx of foreign investments
directed to the petroleum sector of the country due to the high potential profitability from crude
oil production and exploration in the global markets. By far, the petroleum industry and the oil
and gas (alongside other primary resources) sector in general account for more than 50% the
foreign direct investments into the country. Although other sectors such as the banking and
finance, manufacturing, agricultural, energy and education also attract foreign direct investments
to the country, the amount is significantly low and decreasing as a percentage of GDP when
compared to the oil and gas sector.
Despite the high volume of FDI concentration in the Oil and Gas sector of the country, the
amount of this FDI inflow compared to the aggregate flow of FDI to other developing nations in
Asia and Latin America is relatively insignificant. Since the 1990s FDI flows to developing
nations have increased more than 40% with countries in Asia and Latin America sharing
significant portions of the flows. However, the FDI flows to African region during the same
period have been decreasing. Since the 1990s, East Asia and the pacific have attracted more than
$50 billion in FDI inflows, Latin America and the Caribbean more than $18billion, Eastern
Europe and Central Asia more than $14Billion. In contrast, Sub-Saharan African region has
attracted less than $5billion for the same period.
The poor and declining nature of FDI to African region as compared to global flows of FDI to
developing nations despite the abundant availability of human and natural resources in the region
is a great concern. This is an indication that foreign investors do not find investing their capital
in the region attractive and worthwhile venture (Udeaja et al. 2008). African leaders have
proposed and adopted different measures aiming to attract more inflows of FDI to the region.
These measures though in part produced positive responses, the results are still very insignificant
compared to FDI flows to other developing regions of the world.
There is no doubt that FDI is vital for economic growth and development in any nation. The
increased flows of FDI from advanced economies to developing and emerging economies help
provide these FDI recipients countries with needed capital for investments, drives more
opportunities for domestic jobs creation and provides advanced managerial skills to local
personnel (Wafure and Nurudeen (2010)).
Given the significant role played by FDI in a countrys economic growth and development, it
becomes very important to critically assess what factors actually attract FDI flows to a particular
region or country. In the case of Nigeria with large reliance on oil and gas sector, examining
what factors that attract FDI to the oil and Gas sector as well as to all the other sectors of the
economy is important. This helps stabilize and attract more foreign capital into the country and
increase the pace of economic growth and development process in the country.
Several studies have attempted to examine what factors that attract FDI to certain regions.
Among those factors that are deemed to be important for attracting FDI flows to certain regions
include host-country market size, cost of domestic labor, government liberal policies for the
employment of expatriate staff, country risk, availability of good and adequate infrastructures,
degree of economic openness of a country, stable macroeconomic environment, trade and
exchange rate policies, and exchange rate volatility (Wheeler and Mody (1992), Basu and
Srinivasan (2002), and Omisakin et.al (2009)). Other studies suggest that institutional
framework, financial sector effectiveness, host countrys manpower and skills as well as overall
1024

macroeconomic stability and technological conditions are important factors that attract FDI
flows to a country (Omisakin et al. (2009)).
The factors that matter for FDI flows to developing countries as documented by prior literature
include structural reforms and host country domestic market size. Human capital development is
also documented to be a significant factor attracting FDI flows to developing countries (Zheng
and Markusen (1999); Dunning (2008), Chakrabarti (2001)). In addition, infrastructural
developments and degree of openness are found to be important for attracting FDI to developing
countries (Onyeiwu and Shrestha (2004)). Rogoff and Reinhart (2003) and Onyeiwu and
Shrestha (2004) document that price instability significantly impacts investments decisions as it
makes business planning and forecasts difficult to project. World commodity prices are found to
have a direct influence on the flows of FDI to African region as Africa experiences large inflow
of FDI to the region during rising prices of commodities in the global market and vice versa
(Rogoff and Reinhart (2003)). Nzotta and Okereke (2009) find that stable, efficient and well-
functioning financial systems significantly influence the flow of FDI to a region.
Market size hypothesis claims that large domestic market (as symbolized by large population and
fast growing domestic economic activities) serves as an attraction for the transfer of foreign
capital into a country. This is so as foreign investors believe that a large and growing domestic
market in the host country offers potential high revenue and profitability for their investments
(Obadan (1982); Bajo-Rubio and Sosvilla-Rivero (1994); Omisakin et al. (2009); Chakrabarti
(2001); and Wafure and Nurudeen (2010)). These studies support the argument that a large and
domestic market size plays a positive and significant role in attracting FDI to Nigeria.
Empirical study by Campos and Kinoshita (2008) found that structural reforms strongly
influence the flows of FDI to Latin American and Eastern Europe countries. The study shows
that financial sector reforms and privatization policies in these regions during the observed
period significantly the flow of FDI to these regions. What this study implies is that efficiency of
the banking and financial sector, trade liberalization, privatization of domestic economic
activities and good institutional settings are significant factors that attract large amount of FDI
flows to a country.
Human capital development is a necessarily relevant factor for any economy to achieve greater
and long-run sustainable productivity and economic growth. The amount and quality of human
capital influences both domestic and foreign investments of capital in a country. Studies by
Zhang and Markusen (1999), and Dunning (2008) argue that human capital development is
significant in the attraction of FDI to developing countries. The lack of relevant human capital
in terms of skills, education and good health needed for productive activities could potentially
reduce both the attraction and volumes of FDI flows to a country. In the empirically studies by
Noorbakhsh, Paloni and Youssef (2001), human capital is found to be a significant determinant
of FDI flows to developing nations.
Another argument for the impact of human capital on FDI flows to a country could be made in
terms of improved productive capacity of a country due to the availability of advanced
technologies. Continued development of the domestic human capital provides the needed
knowledge for the ease and efficient use of advanced technologies for productions. This situation
encourages more investments in the domestic economy and also serves as an attraction for
foreign investors to transfer their capital investments to the domestic economy.
A country that opens its economy to the rest of the world does so to create a smooth trading
relationship with the rest of the world. Such a country adopts less stringent trade policies and
lower tariffs on imported goods into the country. This atmosphere conveys positive message to
1025

foreign investors that the country is business-friendly and safe investments havens. In this vein,
the degree of trade openness of a country is, to a large extent, a determining factor of foreign
transfer of investments capital to the country from abroad. Particularly, trade openness is argued
to play very crucial role in attracting FDI flow to less developed countries, though the degree and
direction of this relationship is controversially debated in the literature (Omisakin et al. (2009)).
Omisakin et al. (2009) find that there is a positive and significant relation between FDI and trade
openness. This study suggests that even though the relationship might be controversial in the
short-run, there are long-run positive relations among foreign direct investment, trade openness
and output growth in Nigeria. Therefore, higher degree of openness encourages more inflow of
FDI to the country, which in turn leads to higher growth rate of output in the host country.
Some studies have also shown that developing nations that possess strong consumer potentials,
abundant natural resources, and labor cost advantages have been able to attract more FDI flows
into their countries relative to other countries without these factors. The fact that global flows of
FDI have been significantly increasing while the share of these aggregate flows to some
developing nations have been decreasing could be attributed to macroeconomic fluctuations and
more frequent policy regime changes prevailing in these countries which are not present in
industrial nations (Aguiar and Gopinath, 2007). Countries with relatively stable macroeconomic
conditions and political stability are potentially attractive to foreign investors doing business in
these countries. In general, macroeconomic factors such as economic and political instability,
economic size, external debt, inflation, corruption, poor state of infrastructural development, per
capita real income, inflation, world interest rate, credit rating, indigenization policy, natural
resources endowment, openness, and unstable macroeconomic policies are particularly observed
by various studies to determine the magnitude and attraction of FDI flows to Africa (Iyoha
(2001); Anyanwu (1998); Dinda (2009); Ekpo (1997); Obadan (1982); and Onyeiwu and
Shrestha (2004)).
A more recent study of the determinants of FDI in Nigeria by Wafure and Nurudeen (2010) finds
that domestic market size, deregulation, political instability, and exchange rate depreciation are
significant determinants of FDI flows to Nigeria. Other studies document that political and
macroeconomic stability, low growth, weak infrastructure, poor governance, inhospitable
regulatory environments, and ill-conceived investment promotion strategies are key factors that
negatively affect the flow of FDI to African region. Asiedu (2002, 2006) shows that natural
resources and large markets attract more FDI flows to African countries. The study also finds
that low inflation, good infrastructure, highly educated population, openness to trade, less
corruption, political stability and a reliable legal system play significant roles in FDI attraction to
African region.
Prior literature has mixed results for the role of exchange rate volatility in determining FDI flows
to the developing regions. Exchange rate volatility measures the macroeconomic stability of the
host countries. Theoretically, high volatility of the rate of exchange signals unstable
macroeconomic conditions to the potential foreign investors. This increases the uncertainty about
investments decisions and profitability forecast for the potential investors. Masayuki and
Ivohasina (2005) and Wafure and Nurudeen (2010) claim that a depreciation of the host country
exchange rate attracts more inflow of FDI to the country as it becomes relatively cheaper to
transfer capital and other productive equipments to the host country , merge with host countrys
firms or in certain cases acquire host country firms. Froot and Stein (1991) argue that
depreciation of the host countrys currency versus the investor home country currency positively
influences the decision of the investor to transfer capital to the host country. This is due to the
1026

potential increase in wealth and cheap acquisition of assets in the host country that follows the
currency depreciation. A depreciation of the host countrys currency relative to the investor
home country currency could potentially reduce expected values of both the price and nominal
returns on the assets invested. This situation could, therefore, discourage the decision by
investors to transfer investments capital abroad (Tokunbo and Lloyd (2009)). In contrast, some
studies argue that a higher and rising level of the host country currency increases the expected
future profitability from transferring capital to the host country. Hence, an appreciation of the
host country currency is a positive and significant attraction for FDI to the domestic economy.
As far as the relationship is concerned in the Nigerian case, Tokunbo and Lloyd (2009)
empirically investigate the effect of exchange rate volatility on FDI flows to Nigeria and find
that a positive and significant relationship exists between real inward FDI and exchange rate
volatility in Nigeria. Although the authors note the situation of sub-optimal investment ratio that
still exists in Nigeria could be attributed, among other factors, largely to exchange rate instability
following the adoption of flexible exchange rate policy in 1986.
Generally speaking, as far as the ability of African countries to attract FDI flows to the region is
concerned, both theoretical and empirical literature attempting to investigate what factors are
significant attractions for FDI inflows to the region conclude that degree of economic and trade
openness (though long hindered by policy failures by most African governments in the areas of
trade restrictions that negatively impacted participation in international production networks and
more efficient service activities by African businesses), problem of policy incredibility which
manifested in the nature of poor implementation of liberalization efforts and misapplication trade
policies, unfavorable and unstable tax regimes, the slow pace of public sector reform, inadequacy
of intellectual property protection, high levels of corruption both in public and private settings ,
over regulation and political risk , irresponsible fiscal and monetary policies leading to
unsustainable budget deficits, huge debts and inflationary pressures, rising domestic production
costs, exchange rate instability and volatility, sluggish financial liberalization, poor state
infrastructural development and the perceived location risk for long-term investments in more
dynamic non-traditional export sectors by foreign investors are responsible for Africas poor
macro performance in attracting FDI to the region (Asiedu (2002, 2006); Rogoff and Reinhart
(2003)).

III. The Data, Working Hypotheses and Variable Measurements
In specifying the empirical model and the choice of variables employed for the present analysis,
our research is grounded on the latest theoretical and empirical literature on FDI determinants.
In particular, our approach and estimation methods are largely grounded on the works of
Onyeiwu and Shrestha (2004), Asiedu (2002, 2006), Campos and kinoshita (2008), Cushman
(1985), Rogoff and Reinhart (2003), Basu and Srinivasan (2002), and Kyereboah-Coleman and
Agyire-Tettey (2008). The data on the variables of macroeconomic fundamentals and exchange
rate volatility were collected from the Central Bank of Nigeria (CBN), World Bank data bank
(WDI), IMF, Penn-World tables, and Nigerian Bureau of Statistics.
A. Dependent Variable: FDI
The FDI is measured by the annual net FDI inflows in Nigeria in current US dollar. Time series
data on FDI measure is obtained from WDI published by the World Bank.

1027

B. Variables for Macroeconomic Fundamentals
1. Market Size (GDPg)
There have been a number of studies on the importance of a countrys market size in its ability to
attract foreign capital. A large and growing domestic market is an attraction for a foreign
investor and increases the flow of FDI to the country from abroad (Basu and Srinivasan (2002);
and Onyeiwu and Shrestha (2004)). Grounded on the literature, we thushypothesize that there is
a positive relationship between FDI flows to Africa and growing domestic market size. Growth
rate of economic (activities as measured by GDP growth) is used to as a proxy for Market size in
this study. This is conventional in all empirical studies. Time series data on Nigeria GDP growth
rate is obtained from CBN stastical bulletin and WDI published by World Bank.
2. Infrastructures
The availability of good quality and reliable infrastructures is a core requirement for promoting
investments in a country and an attraction of FDI flows into a country. Good and reliable
infrastructures are a symbol of high level of economic development and they play an important
role in attracting foreign investment (Basu and Srinivasan (2002); (Asiedu (2002)). Grounded on
this stream of literature, we hypothesiz that there is a positive relation between the infrastructural
development and FDI flows to Africa. Electricity production (kWh) in Nigeria is used as a
proxy for infrastructural development. Intuitively, a good quality, stable and adequate supply of
electricity potentially reduces the cost of operating plant, machinery and other equipments used
for productive activities by firms. Consequently, this leads to a reduction in production costs as
well as the costs of doing business. Time series data on electricity production in Nigeria is
obtained from CBN statistical bulletin and WDI published by the World Bank.
3. Index of Openness
The Index of Opennes measures the degree of economic openness of Nigeria in the global
community. Higher degree of openness by a counttry in terms of economic actvities in the
global community encourages the transfer of capital investments from aboroad to the domestic
market. Higher degree of openness by a country in the forms of tarrifs reduction, absence of
capital control, less restrictive trade policies among others gives foreign investors strong
incentives to investment in the country(Onyeiwu and Shrestha (2004)). We hypothesize that the
degree of openness of the domestic economy to the rest of the world has a positve impact on the
FDI flows to Africa. The Index of Openness is computed as the ratio of import and export to
GDP. Times series data on Nigeria import, export and GDP is obtained from CBN statistical
bulletin and WDI published by the World Bank.
4. Inflation
Price instability signals very high risks of doing business and discourages both domestic and
foreign investors from committing their investmensts in a country. Price instability in the form
of high and unpredictable inflation is counter-productive for investment. It not only makes
business planning and forcasting a difficult task, but also leads to higher costs of doing business
and consequently lowers expected returns from investments (Rogoff and Reinhart (2003);
Onyeiwu and Shrestha (2004) ). Therefore, countries with a high and unpredictable level of
inflation are likely to have difficulties to attract FDI inflows. Thus, we hypothesize that there is
a negative relationship between FDI and the level of inflation. Time series data on inflation rates
for Nigeria is obtained from CBN and WDI.
5. Crude Oil Price (COP)
The abundant oil and other natural resources available in some African countries account for a
significant portion of the total FDI flows to Africa. Oil exporting countries in African account
1028

for the biggest attraction of FDI flows to the African region. Because the rate of return is the
highest in the petroleum industry, foreign investors tend to transfer more investment capital to
countries with abundant crude oil resources. Countries like Nigeria and Angola have been able
to attract very large inflows of FDI because of the very high rate of return in the petroleum sector
(Basu and Srinivasan (2002)). Supported by this literature, we hypothesize that there is a
positive relationship between FDI flows to Africa and the world crude oil price. The rising
world price of crude oil leads to more revenue and profitability in the petroleum sector, and
hence more potentials for the attraction of FDI into the country.
The Spot OPEC Reference Basket (ORB) Prices ($/barrel) is used in this paper to represent the
world price of crude oil. Time series data on Spot OPEC Reference Basket (ORB) Prices
($/barrel) is obtained from CBN statistical Bulletin and OPEC statistical bulletin.
6. World Commodities Price
World commodities prices have been found to have a sytematic impact on the flows of FDI to
African region. Africa typically experiences large inflows of FDI during rising prices of
commodities prices and very low inflows when commodities prices are decreasing (Rogoff and
Reinhart (2003)). Suported by this literature, we hypothesize that there is a positive relationship
between increasing FDI flows to Africa and rising world commodities prices. In this study, the
Index of Average World Prices (C.I.F.) of Nigeria's Major Agricultural Commodities in Naira
per ton (1985=100) is used as a proxy for World Commodities Prices. Time series data on
Indices of Average World Prices (C.I.F.) of Nigeria's Major Agricultural Commodities in Naira
per ton (1985=100) is obtained from CBN statistical bulletin.
7. Exchange Rate
Exchange rate is a vital macroeconomic variable. It plays a far reaching and significant role in
the effective and efficient operation of an economy. It is a key determinant in economic policy
formulations necessary for achieving and maintaining a strong and sustainable economic growth
and development. Based on this evidence, we hypothesize that there is a positive relationship
between the flows of FDI to Africa and an increase in exchange rate of the host countrys
currency. When the exchange rate increases, the host countrys currency depreciates while the
source countrys currency appreciates simultaneously (Giorgioni (2002)). A depreciation of the
host countrys currency following an increase in the exchange rate is an important attraction for
FDI as it becomes cheaper for foreign investors to transfer capital to the host country as well as
lowering costs of production in the host country.
Bilateral exchange rate between Nigeria Naira and the US dollar is adopted for this study and
time series data on Naira/Dollar exchange rate is obtained from CBN statistical bulletin and WDI
published by the World bank.
8. Financial Development
The financial system is largely responsible for the mobilization and allocation of excess funds
from the surplus side to the needed areas of the economy. These funds are then channeled to the
appropriate productive sectors of the economy and thereby ensuring availability of liquidity in
the system (Nzotta and Okereke (2009)). A strong, efficient and well-functioning financial
system is a very attractive factor for foreign investors.
In this study, we measure financial development by the level of financial deepening. A high
level of financial deepening is an impetus for economic growth and efficient financial system,
which in turn encourages the inflows of FDI to the domestic economy. Thus, we hypothesize
that the level of financial deepening has a positive impact on the FDI flows to Africa. Following
Asiedu (2002), we measure the level of financial deepening by the ratio of money and quasi
1029

money (M2) to GDP. Time series data is obtained from the CBN bulletin and WDI published by
the World Bank.
9. Exchange Rate Volatility
Exchange rate volatility is defined and measured by a number of ways in the empirical literature.
In most cases, different measures of volatility give different estimates and effects in the
empirical literature. This is one of the root causes of the controversy surrounding the effects of
exchange rate volatility in terms of direction and magnitude in international transactions and
foreign direct investment. There is so much controversy surrounding the degree and direction of
the impacts of exchange rate volatility on the flow of FDI to Africa, as is the case in the rest of
the world. It is still very unclear whether increased volatility of the rate of exchange increases or
decreases trade flows and all other international transactions including FDI movements from
abroad to African region. We hypothesize that an increase in exchange rate volatility has a
negative impact on FDI flows to Africa. The intuition behind our hypothesis is that the increased
level of exchange rate volatility leads to a high level of uncertainty which in turn negatively
affects investment decisions, and most likely international investment decisions involving the
transfer of capital to foreign countries. Therefore, we posit that there a negative theoretical
relationship between FDI flows to Africa and exchange rate volatility.
In this study, we measure the exchange rate volatility by estimating the quarterly standard
deviations from the monthly nominal official exchange rates between Nigeria and United States.
Time series data on the Nigerian naira vis--vis United States dollar is obtained from Central
Bank of Nigeria (CBN) statistical bulletin and WDI published by the World Bank.
C. Variables for Country Risks
1. Political Economy
Political economy is used as a proxy to capture the possible effects that a significant economic
and political event in Nigeria may have on the aggregate economic activities and the ability of
Nigeria as a country to attract FDI flows into the country. A value of 1(one) is assigned to any
year if a significant economic and political event(s) occurs and a value of 0(zero) if no major
economic and political events occur that could impact macro-level economic activities.
2. Regime Shift
An exchange rate regime change and a major trade policy shift could lead to a large impact on
FDI flows to a country. Thus we employ a dummy variable to capture the effects of a structural
change. A value of zero is assigned to each year during the period of fixed exchange rate regime
in Nigeria and pre-trade liberalization era; while a value of one (1) is assigned to each year
following the adoption of Structural Adjustment Program (SAP) in 1986 when floating exchange
rate system and trade liberalization policies were introduced in Nigeria.

IV. Research Method and Design
We employ time series econometric techniques to examine the impacts of exchange rate
volatility as well as major macroeconomic fundamentals on the flow of FDI to Nigeria. In time
series analyses, most macroeconomic data employed tend to contain trend elements and thus, it is
very crucial to remove these trend elements before proceeding to estimation stage. Time series
data containing trend elements tend to produce spurious regression results and unreliable
estimates. The implication of trend element in a variable is that the variable contains some
cyclical components that must be removed in order to obtain meaningful regression results.
Otherwise, hypotheses tests conducted on the basis of these spurious regression results are highly
unreliable. Most empirical studies using time series data employ unit root tests to determine
1030

whether a variable contains a trend element. We employ the Augmented Dickey-Fuller unit root
test, which is widely used in time-series analysis.
If the hypothesis for the presence of unit root is rejected, then the time-series is considered trend-
stationary, namely there are no trend elements in the series. This series is integrated in the order
of zero or I(0) and the OLS estimation technique can be used for the analysis. However, if the
hypothesis for the presence of unit root cannot be rejected, the series is said to be difference-
stationary, meaning that the series contains trend elements or considered nonstationary.
Nonstationary time-series data contain unit roots which suggest that common OLS estimation
technique should not be employed for analysis. The time-series data with unit roots are usually
integrated at higher orders than zero, such as I(1) or higher, until the unit roots are removed from
the series. When this occurs, further tests are required to test the likelihood of long-run
relationships among the variables employed in the analysis. This can be done using
cointegration test, which basically tests for long-run correlation of the variables employed in the
analysis.
The common cointegration technique is the Johansen cointegration technique which is employed
in our study. Vector Error Correction Model (VECM) is used to estimate the regression
equations resulting from Johansen cointegration analysis.
The empirical model in our study specifies FDI as a function of exchange rate volatility and key
macroeconomic fundamentals. The econometric model is specified as follows:
FDI
t
=f (Exchange Rate Volatility, Macroeconomic Fundamentals, Political-economic events,
Regime Shift)
FDI
NairaVol
LGDP
LIOP
LRP
LCOP
LElectric
LPCGDP
LINF
LWCPI
LCredit/GDP
LXRT
Pdum
Regdum
where,
FDI
t
represents foreign direct investment net inflows in in current US$ at time t;
NairaVol is the volatility of Naira (Nigerian currency) in relation to the foreign currency
proxied by U.S. dollar;
LGDP is the log of Gross Domestic Product;
LIOP measures the log of the Index of Openness;
LRP represents domestic interest rate;
LCOP (Crude Oil price) represents the log of the Spot OPEC Reference Basket (ORB) Prices
($/b) of crude oil;
LElectric represents the log of electricity production in Nigeria (kWh);
LPCGDP represents per capita GDP;
LINF represents the log of inflation (consumer prices in annual %);
LWCPI captures the log of the Indices of Average World Prices (C.I.F.) of Nigeria's major
agricultural commodities in Naira per ton (1985=100);
LCredit/GDP represents the Financial Deepening which is measured by the log of money and
quasi money (M2) as % of GDP;
LXRT is the log of the official exchange rate (Naira per US$, period average);
Pdum is a dummy variable that captures significant political and economic events in Nigeria;
Regdum represents Trade and Exchange Rate Regime dummy (0 = Fixed Regime; 1 =
liberalization and Flexible Regime).
1031

V. Empirical Results
In this section, the estimated FDI regression results are presented and discussed. Table I presents
the summary of the unit root test results based on Augmented Dickey-Fuller test. All the
included variables in the FDI equation were found to contain unit roots at their levels but became
stationary after the first difference at 1% level of significance. This implies that the trend
element or non-stationary nature of the included variables in the FDI equation has disappeared
after taking first differences of all the variables in their natural log forms. The results are
presented in the Table I.
Table I. Unit Root Analysis of FDI Determinants
With intercept Without intercept and Trend
t-statistic Decision Rule t-statistic Decision rule
LFDI -6.625*** I(1) -6.535*** I(1)
LGDP -5.069*** I(1) -4.996*** I(1)
LIOP -5.920*** I(1) -6.028*** I(1)
LRP -6.660*** I(1) -6.840*** I(1)
LXRT -5.482*** I(1) -5.606*** I(1)
LCOP -6.786*** I(1) -6.910*** I(1)
LWCPI 5.723*** I(1) 5.661*** I(1)
LCREDIT/GDP -5.625*** I(1) -5.607** I(1)
LNairaVol -4.826*** I(1) -5.683*** I(1)
LINF -6.107*** I(1) -6.011*** I(1)
LElectric -6.493*** I(1) -7.744*** I(1)
LPCGDP -5.055*** I(1) -4.987*** I(1)

Table I shows that the null hypothesis for the presence of unit root in all the variables for the FDI
equation is rejected at 1% significance level. As shown in Table I, the augmented DickeyFuller
(ADF) statistic is highly negative and the more negative it is, the stronger the hypothesis that
there is a unit root is rejected. The unit root test results strongly suggest that all the included
variables are integrated in the order of I(1). Since all variables are in the same order of
integration, the next step is to test for cointegration or the long-run relationship between net FDI
inflow to Nigeria and all its fundamentals. Cointegration test allows for the determination of
long-run relationships among the included variables in the FDI equation. Johansen cointegration
technique is employed in this this test and the results are presented in Table II.

1032

Table II. Unrestricted Johansen Cointegration Test Results
Maximum Rank Eigen Value Trace Statistics Critical Value Probability
0* 0.97949 749.09 95.75 0.0000
1 0.96955 605.27 69.82 0.0000
2 0.94329 476.08 47.86 0.0000
3 0.87094 369.90 29.80 0.0113
4 0.83584 294.14 15.495 0.0245
5 0.81290 227.28 3.842 0.0169

The results of the unrestricted Johansen cointegration test are reported in Table II. The standard
statistics used in the interpretation of the test are the eigenvalues and the trace statistics at given
levels of significance. The application of Johansen (1998) cointegration technique reveals a
number of cointegration equations among the variables. At 5% level of significance, the results
show that there is only one cointegrating equation based on the basis of trace statistic and suggest
that there exists a significant long-run equilibrium relationship between net FDI inflow to
Nigeria and macroeconomic fundamental variables in our model.
In addition to the long-run relationships among the variables estimated by the Johansen
technique, Vector Error Correction Model (VECM) technique is employed to test whether there
is a significant relationship between net FDI flows to Nigeria and all its fundamental
determinants. The estimated results of (VEC) model are presented in Table III.
The results from VECM estimation as reported in Table III show a strong and positive
relationship between net FDI inflows and macroeconomic fundamentals such as domestic market
size (LGDP), macroeconomic stability (LXRT), financial development (lCredit/GDP),
infrastructural development (LElectric), as well as external factors such as world crude oil price
(LCOP) and world price of primary agricultural commodities (LWCPI). The variables such as
exchange rate volatility (LNairaVol), domestic inflation rate (LINF), trade openness (LIOP),
GDP per capita (LPCGDP), and domestic interest rate (LRP) show a negative but statistically
significant relationship with net FDI inflows to Nigeria. Political and economic events seem to
affect the flow of FDI to Nigeria negatively and this result is statistically significant when trade
and exchange rate regime shifted in 1986 from fixed and controlled regime to trade liberalization
and flexible exchange rate regime. Trade liberalization and flexible exchange regime have
generated positive impacts on FDI flow to Nigeria and this result is statistically significant.
VECM technique employed in this study is estimated without exogenous factors.
The size of domestic market is a positive and significant attraction of FDI flows to Nigeria. The
coefficient on LGDP, the measure of domestic market size, is both positive and significant. This
shows that growth of Nigerian economy and rapid expansion of domestic economic activities are
significant attractions for foreign investors to transfer their capital to the country.

1033

Table III. VECM Results for FDI to Nigeria
Lag order 1, maximum likelihood estimates, observations 1972-2010 (T = 39).
Cointegration rank = 1
Case 2: Restricted constant beta (cointegrating vectors, standard errors in parentheses).

Cointegration Equation: CointEq1
LFDI (-1)

1.0000

LGDP (-1)

278.84
(37.241)
LIOP(-1) -3.5590
(0.55668)
LRP (-1) -1.1771
(0.16634)
LXRT (-1)

2.1936
(0.16934)
LPCGDP (-1) -282.52
(37.245)
LCOP (-1)

5.3517
(0.25157)
LWCPI (-1)

2.9526
(0.20922)
LCREDIT/GDP (-1)

0.23215
(0.21193)
LNairaVol(-1)

-0.098861
(0.00910)
C -8.9775
(1.0027)

The coefficient on LIOP, the measure of trade openness in Nigeria, shows an unexpected
negative sign which implies a negative relationship between FDI flows and the degree of trade
and economic openness in Nigeria. This might be due to the lack of confidence by foreign
investors in the adopted trade and other government policies relating to investments in the
country, hence a negative relationship result in this test.
Domestic interest rate (LRP), as measured by the difference between commercial prime lending
and the Treasury bill rate is negatively related to FDI flow to Nigeria, which is theoretically
expected. The coefficient on LRP is both negative and significant. The higher the spread
between prime lending rate and Treasury rate, the higher the cost of borrowing from banks in the
1034

country and the more cost of investing and doing business in Nigeria. A higher spread or interest
rate premium indicates higher cost of investments in Nigeria, and thus discourages both domestic
and foreign investments in the domestic economy. A lower spread or interest rate premium
encourages more domestic and foreign investments in the country. Hence, there is a justification
for the negative relationship between domestic interest rate and FDI flows to the country.
Exchange rate (LXRT) is both positively and significantly related to FDI flows to Nigeria as
theoretically expected. This is a sign of macroeconomic stability in Nigeria. An appreciation of
the Nigerian naira encourages more inflow of FDI to the country. This positively influences the
decision of foreign investors to choose Nigeria as the final destination for their capital
investments and hence, increased flow of FDI to the country.
GDP per capita (LPCGDP) is an indicator for the purchasing power of local consumers in the
host country (Furceri and Borelli (2008). A higher level of consumers purchasing power due to
improved economic conditions leads to increased demand for both domestic and imported goods
and services. Therefore, high and improved GDP per capita encourages both domestic and
foreign investments in the economy. This is an attraction for FDI flow to the host country as it
represents higher potential profitability and returns on investment. Unexpectedly, the coefficient
of LPCGDP in our FDI equation shows a negative but significant relationship between per capita
DDP and the flow of FDI to the country.
Significant portion of FDI flows to Nigeria is largely influenced by the abundant oil and other
natural resources endowment in the country. As expected, the coefficient on LCOP (world crude
oil price), a measure of natural resources revenue and rate of returns on investments in the
petroleum industry is both positive and statistically significant. This is consistent to the argument
that foreign investors tend to transfer more capital investments to Nigeria with rising crude oil
prices in the world market, as this leads to higher revenue and profitability on the invested
capital. This thus leads to more inflows of FDI to Nigeria.
World commodities prices (LWCPI), as expected, influence the flow of FDI to Nigeria
positively. The coefficient on LWCPI is both positive and statistically significant. It has been
observed that there are greater inflows of FDI to African region during periods of rising
commodities prices in the global markets and vice versa (Roggoff and Reinhart (2003)). As
expected, the positive influence of world commodities prices on FDI flows to African region is
empirically confirmed by this result. World commodities prices relate directly and significantly
to the flows of FDI to Nigeria.
Financial Deepening which is a proxy for financial development (LCredit/GDP), as measured by
the ratio of credit of the private sector to GDP, relates positively to FDI flows to Nigeria. The
coefficient on LCredit/GDP is positive though not statistically significant. Although the
relationship is not found to be statistically significant, it indicates that a dependable and
favorable credit environment is beneficial to both domestic and foreign investors who are willing
to invest in Nigeria. This is an attraction for FDI flows to Nigeria.
Infrastructural development proxied by electric power production in Kwh in Nigeria relates
positively and significantly with FDI inflow to Nigeria. The coefficient on LElectric is both
positive and statistically significant, which implies that electricity production is an important
determining factor of FDI flows to Nigeria. The improvement in electricity production,
generation and distribution in Nigeria has been a key factor boosting both domestic and foreign
investments in Nigeria. Good quality, stable and adequate production and distribution of electric
power in Nigeria helps reduce the cost of operating plant, machinery and other equipment used
for productive activities by businesses in Nigeria.
1035

Although there is so much controversy on the degree and direction of the impact of exchange
rate volatility on FDI flows to the host country, we expect a negative influence of exchange rate
volatility on FDI flows to Nigeria. As expected, the coefficient on LNairaVol is both negative
and statistically significant. This confirms with the theoretical arguments that increased
volatility leads to higher uncertainty which in turn negatively impacts investment decisions and
hence, reduces the flows of FDI to Nigeria.
As expected, there exists a negative empirical relationship between domestic inflation and the
flow of FDI to Nigeria. The coefficient on LINF, the measure of domestic inflation rate shows a
negative empirical sign, and is statistically significant. A high rate of inflation is an indication of
price instability and weak monetary and fiscal policies at stabilizing the economy. This
discourages foreign investment and hence, reduces the amount and frequency of FDI flow to the
country. Both negative and significant coefficient of domestic inflation points to the fact that
high rate of inflation in Nigeria is a key reason for the slow pace of FDI flow to the country
relative to other developing countries of Asia and Latin America.
The coefficient on Poldum, the dummy variable for significant political and economic events in
Nigeria negatively influences the flow of FDI to Nigeria. Frequent and major political-economic
crises discourage flow of investors to a country as this is a sign of instability in the country and
hence, undesirable for investments. This is confirmed by this study as the coefficient on Poldum
is both negative and statistically significant, which implies political and economic crises in
Nigeria significantly influence FDI flows to Nigeria but in a negative way.
The shift from trade control and fixed exchange rate regime to trade liberalization and flexible
exchange rate regime influences FDI flow to Nigeria positively and significantly. The
coefficient on Regdum, the dummy variable for trade and exchange rate regime change in
Nigeria is both positive and statistically significant.

VI. Summary and Conclusions
The objective of this study is to add to the existing literature on the determinants of FDI flows to
developing countries. Most existing studies on this topic tend to focus more on general factors
that affect all developing countries and fail to address country-specific factors that attract FDI to
a country. Realizing that there are country-specific factors that determine the ability of
individual country in attracting foreign investments, our study attempts to bridge this gap by
focusing on the factors specific to African and to Nigeria in particular that impact the flows of
foreign direct investments into the country. In doing so, both internal and external factors that
are specific to the country are considered in our analysis.
In the general analysis of the empirical investigations of the factors that determine the flow of
FDI to Nigeria, our empirical results show that domestic market size measured by the GDP,
inflation rate, exchange rate, domestic interest rate, world price of primary agricultural
commodities, exchange rate volatility, infrastructural development, political-economic events
and exchange rate and trade regime change are important determinants of FDI flows to Nigeria.
Degree of openness, financial sector development and per capital GDP all play an important role
in attracting FDI flows to Nigeria. By and large, our study shows that exchange rate volatility
and major macroeconomic fundamentals in Nigeria play an important role in attracting FDI
flows to the country.
The Nigerian domestic market size measured by GDP is a very important determinant of FDI to
the country. This argument is supported by the empirical results obtained for the coefficient of
GDP in this study.
1036

The results on the estimates of infrastructures, exchange rate, domestic interest rate and inflation
rate show that internal macroeconomic factors are crucial determinants of FDI flows to Nigeria.
The estimates of the coefficients on political and economic events and regime shift show that
various political events in the country over the past three decades are among the factors that
foreign investors examine critically before deciding to transfer their capital to the country.
Overall, our empirical results indicate that it is imperative for the government and policy-makers
in Nigeria and Africa to focus more on formulating and implementing policies that aim to
improve the political and economic environment for attracting FDI inflows to the region. When
adequate stability of policies and efficient infrastructures are put in place, both domestic and
foreign investors are likely to be encouraged to invest their capital in the country. Hence, the
findings of this study provide important policy implications for African governments and policy-
makers for undertaking strategies to attract FDI inflows to the country.
Our study contributes to literature in four areas. Firstly, our research incorporates the empirical
analysis of both internal and external factors that might influence the flow of FDI to Nigeria.
Secondly, our study bridges the gap inherent in prior empirical literature on the determinants of
FDI flows to African region. Specifically, prior literature tends to focus more on cross-country
analysis of FDI determinants in Africa. However, there is a huge variation in the actual
economic, political, social and cultural conditions among African countries. Thus, a cross-
country analysis may not provide reliable estimates specific to investment decisions making and
to policy formulations by an individual African country. Thirdly, most of prior studies do not
consider the core macroeconomic fundamentals specific to the African situation on an individual
country basis. Our study has bridged these gaps by focusing on country-specific macroeconomic
fundamentals that influence the flows of FDI to African region. Finally, the approach employed
in the analysis of FDI flows to Nigeria could easily be applied to the analysis of other African
countries.

Refrerences
Aguiar, Mark and Gita Gopinath, 2007, Emerging Market Business Cycles: The Cycle Is the
Trend, Journal of Political Economy, University of Chicago Press, vol. 115, pages 69-102.
Anyanwu, J.C. (1998). An econometric investigation of determinants of foreign direct
investment in Nigeria. In investment in the growth process: proceedings of the Nigerian
economic society conference 198, pp. 219 -240. Ibadan, Nigeria.
Asiedu, E., 2002, On the Determinants of Foreign Direct Investment to Developing Countries: Is
Africa Different? World Development, 30 (1): 107-19.
Asiedu, E., 2006, Foreign direct investment in Africa: The role of natural resources, market size,
government policy, institutions and political instability, World Economy, 29 (1), 63-77.
Bajo-Rubia,O. and S. Sosvilla-Rivero, 1994, An Econometric Analysis of Determinants of
Foreign Direct Investment in Spain, Southern Economic Journal, 61(1), pp. 104-117.
Basu, Anupam and Krishna Srinivasan, 2002, Foreign Direct Investment in Africa: Some Case
Studies. International Monetary Fund Working Paper.
Campos, Nauro F. and Yuko Kinoshita, 2008, Foreign Direct Investment and Structural
Reforms: Evidence from Eastern Europe and Latin America, IMF Working Papers,
International Monetary Fund.
Chakrabarti, A., 2001, The determinants of foreign direct investment sensitivity analyses of
cross-country regressions, Kyklos vol. 54, pp. 89-112.
1037

Cushman, D. O., 1985, Real Exchange Rate, Expectations, and the Level of Direct Investment.
The Review of Economics and Statistics , 297-308.
Dinda, S., 2009, Factors Attracting FDI to Nigeria: An Empirical Investigation. Madras School
of Economics, Chennai, India.
Dunning, J.H. and S.M. Lundan, 2008, Theories of foreign direct investment, in John H.
Economy, Cheltenham: Edward Elgar Publishing Limited, pp 79-115.
Ekpo, A.H., 1997, Determinants of foreign direct investment in Nigeria :evidence from time
series data, CBN economic and financial review, Vol.35, No.1, pp 59-78.
Furceri, Davide and Sara Borelli, 2008, Foreign Direct Investments and Exchange Rate
Volatility in the EMU Neighborhood Countries, Journal of International and Global
Economic Studies, 1(1), 42-59.
Froot, Kenneth A. and Jeremy Stein, 1991, Exchange rates and foreign direct investment: an
imperfect capital markets approach, The Quarterly Journal of Economics, 106 (6),1191-
217.
Giorgioni, G., 2002, Foreign Direct Investment and Exchnage Rate Volatility. CIBEF and
Liverpool John Moores University.
Iyoha MA, 2001, Policy simulation with a macro-econometric model of the Nigerian economy.
The Nigerian Economic Financial Review, 6(1): 29-56.
Kyereboah-Coleman, A., and Kwame F. Agyire-Tettey, 2008, Impact of macroeconomic
indicators on stock market performance: The case of the Ghana Stock Exchange, Journal of
Risk Finance, Vol. 9 Iss: 4, pp.365 - 378
Masayuki, H and F. Ivohasina, 2005, The Determinants of Foreign Direct Investment into Japan,
Kobe University Economic Review 51.
Noorbakhsh, Farhad, Alberto Paloni and Ali Youssef. (2001). Human Capital and FDI Inflows to
Developing Countries: New Empirical Evidence. Journal of World Development , 1593-
1610.
Nzotta, Samuel M. and Emeka J. Okereke, 2009, Financial Deepening and Economic
Development of Nigeria: An Empirical Investigation, African Journal of Accounting,
Economics, Finance and Banking Research, Vol. 5, No. 5.
Obadan, M.I.,1982, Direct foreign investment in Nigeria: An empirical analysis, African Studies
Review vol.1.
Omisakin, Olusegun, Oluwatosin Adeniyi, and Ayoola Omojolabi, 2009, Foreign Direct
Investment, Trade Openness and Growth in Nigeria. Journal of Economic Theory , 13-18.
Onyeiwu, Steve and Hemanta Shrestha, 2004, Determinants of Foreign Direct Investment in
Africa. Journal of Developing Societies , 89-106.
Rogoff, Kenneth and Carmen Reinhart, 2003, FDI to Africa: The Role of Price Stability and
Currency Instability. IMF Working paper.
Tokunbo, Osinubi S. and Amaghionyeodiwe A. Lloyd, 2009, Foreign Direct Investment and
Exchange Rate Volatility in Nigeria. International Journal of Apllied Economeretics and
Quantitative Studies , 83-116.
Udeaja A.E, Udoh .E. and F.S Ebong, 2008, Do Determinants of Foreign Direct Investment in
Nigeria Differ Across Sectors? An Empirical Assessment. Economic and Financial Review
of the Central Bank of Nigeria, Vol.46 (2).
Wafure, Obida Gobna, and Abu Nurudeen, 2010, Determinants of Foreign Direct Investment in
Nigeria: An Empirical Analysis. Global journal of Human Social Science , 26-34.
1038

Wheeler, D., and A. Mody, 1992, International investment location Decision. The case of US
firms, Journal of International Economics, 33:57-76.
Zhang, K.H. and J.R. Markusen, 1999, Vertical multinational and host country characteristics,
Journal of Development Economics, 59(2), 233262.

1039

Fairness Opinions, Expert Independence and Reputation in Mergers and
Acquisitions

Nina T. Dorata*
St. Johns University
Queens, NY 11439
doratan@stjohns.edu

Takeshi Nishikawa
University of Colorado-Denver
Denver, CO 80217
takeshi.nishikawa@ucdenver.edu

*corresponding author
1040

Abstract
Recent regulation prohibits an auditing firm from providing appraisal or valuation services or
fairness opinions to its audit clients but does not restrict valuation specialists employed by non-
audit firms to provide multiple advisory services. Using expert reports in connection with
mergers and acquisitions, this study provides an analysis of returns of experts providing multiple
services and their respective reputations. The results show that fairness opinions are increasing
in returns, but are decreasing in announcement period returns when public accounting auditors
were permitted to provide fairness opinions for their audit clients. Financial advisor reputation
matters if that advisor also delivers the fairness opinion.
Keywords: experts opinions, mergers, independence, reputation
I. Introduction
Fairness opinions became a matter of routine in mergers and acquisitions since 1985 when the
directors of Trans Union Corporation were found grossly negligent by failing to use sound
business judgment in connection with the sale of the company to Marmon Group Inc., which was
owned by the Pritzker family. Although the Delaware Supreme Court acknowledges that
fairness opinions are not essential to support a business judgment, the Court implies that with
one, better decisions result.
258
When directors make informed decisions, they are entitled to the
protections of the business judgment rule. The fairness opinion presents an evaluation by the
valuation specialist that the transactions terms and prices and ultimately expresses an opinion on
the fairness of the offer. The opinions provide comfort to combining firms that the terms of the
merging agreement are fair.
Although fairness opinions are commonplace in mergers and acquisitions, the practice reeks with
flaws. Davidoff (2006) contends that the court opinion in the Van Gorkom case was flawed
because it assumes that fairness opinion practices are inherently accurate and comparable. The
reliability of fairness opinions may be questionable because although the opinion identifies
documents, discussions, and observations used as a basis for the opinion, it is devoid of any
details that support the opinion due to the absence of independent verification of the information
supplied used to support the opinion. Also, information provided to the merging constituents is
only valid as of the opinion date with no requirement to update in the event of extended merger
completion dates (Sweeney and Sechler 1999).
In addition to the problems associated with the contents of fairness opinions, the structure of the
business providing such opinions could pose more problems with the reliability of fairness
opinion. First, majority of the fees involved with providing fairness opinion is paid upon the
completion of the merger which could give incentives to the provider of fairness opinion to
provide favorable view of the merger deal even to the detriment of shareholders. For example,
on February 13, 2007, the Delaware Chancery Court delayed a shareholder meeting to vote on
the merger between Caremark and CVS Corp. citing the failure to disclose in proxy materials
that delivery of a fairness opinion was a precondition to the advisors receipt of fees that are
contingent upon completion of the merger. Second and more importantly, most merging firms
may receive their fairness opinions from the valuation specialists (in case of pre-2001, from their
auditors) who already provide other advisory services. Therefore, conflicts of interests exist
when fairness opinions are provided to merging firms by their financial advisor due to
preexisiting business relationships that may impair judgments necessary for fairness opinions.
When a conflict of interest is combined with the contingent fee structure inherent in fairness

258
See Smith v. Van Gorkom 488A. 2d 858 (Del. 1985)
1041

opinions, the reliability of fairness opinions in merger and acquisition transactions is
questionable.
As a remedy for the potential conflicts of interest in fairness opinion and its dealings, the SEC in
2001 revised its rule about auditor independence requirements related to providing fairness
opinions. The SEC rule is further strengthened with passage of SOX. As a result, now the
auditors are not allowed to provide fairness opinions for their existing audit clients. However,
unlike auditors from public accounting firms, valuation specialists from non-audit firms are not
subject to independence standards making it common for the latter to provide both fairness
opinions and advisory services for not only the business combination transactions, but also in the
regular course of business. In this study, duality is the circumstance when the valuation
specialist providing the fairness opinion is also the financial advisor. This means that while
duality is not banned in fairness of opinion, only auditors from public accounting firms are
prohibited from providing such opinions for existing audit clients. These regulatory changes
surrounding fairness opinions provide a natural setting to empirically test the relative value of
auditors in providing these opinions compared to investment banks because after the regulatory
changes only investment bankers can function as dual provider of opinions and other advisory
services. By so doing, this paper extends the literature related to fairness opinions in connection
with mergers and acquisitions.
This study attempts to answer: 1). the relative value differences between fairness opinions
provided by auditors and those provided by investment bankers; and 2). how reputation of the
fairness opinion providers might affect the value of such opinions.
In a sample of 576 completed, nonfinancial industry mergers and acquisitions, the results of our
study show that overall fairness opinions do provide value to shareholders. However, the
analysis shows that a significant majority of cases when the fairness opinion is rendered, the
provider of the opinion also acts as the firms financial advisor for the merger. In the case prior
to the change in the and there is duality, returns are diminishing, suggesting that fairness
opinions provided by auditors were less welcomed by the market. Financial advisor reputation
matters, especially for target shareholders who only view positively those compromised financial
advisors with the best reputation.
This study is organized as follows. Part II provides the background and hypotheses, Part III, the
methodology. Part IV presents the results of the empirical tests and the summary and
conclusions are in Part V.

II. Background and Hypotheses
Auditors and M&A
SOX section 201 prohibits an auditing firm from providing appraisal or valuation services,
fairness opinions, or contribution-in-kind reports for existing audit clients effective May 6, 2003
(hereinafter the SOX date). The Public Company Accounting Oversight Board (PCAOB) and
the Securities and Exchange Commission (SEC) may exempt audit firms from the statutory
provision on a case-by-case basis, provided the service will be in the best interest of the public.
Section 201 permits non-audit clients to receive this service from a public accounting firm.
259

Before SOX, auditing firms were viewed as preferred providers of valuation work because of the
knowledge spillover to the audit process and vice versa. Auditors provided a plethora of services
due to their unique expertise in a firms business.
260
Although SOXs impact on mergers and

259
For a review of auditor conflicts, see Kaplan (2004).
260
See Gorman (1988) for a thorough discussion of business combination services CPAs provide.
1042

acquisition due diligence process could be mostly attributed to officer certification requirements
of Sections 302 and 906 and the internal control compliance reporting of Section 404, SOX
creates significant consulting opportunities for valuation specialists for services that were once
performed by auditors for their clients.
However, SOX was not the first to limit non-audit services auditors could provide. The SEC
with Rule 57-13-00 on February 5, 2001 (hereinafter the SEC date) revised auditor
independence requirements and provides that auditors are not independent if the auditor provides
fairness opinions on the adequacy of consideration exchanged in a business combination
transaction. Due to SOX and SEC regulation, investment bankers now face less competition
from auditors for merger and acquisition valuation and advisory services while imposing
significant compliance reporting that strains the due diligence process. In this study, the impact
of regulation is considered a control factor to capture not only its compliance effects, but also to
recognize easing competitive restrictions for investment bankers.
Empirical studies of independence impairment in non-audit services when the firms auditor
provides non-audit services are not plentiful, even before the passage of SOX.
261
Using UK data,
Antle et al. (2006) find knowledge spillover effects between non-audit services and audit
services expressed in fee data. Bugeja (2005) studied target firms in Australia, where law
requires fairness opinions for target firms. At the time of that study, Australian law permitted
auditors to provide fairness opinions for their audit clients. Bugeja (2005) study finds that
directors will still accept the findings of the fairness opinion irrespective of whether the fairness
opinion is from the auditor. Additionally, market reaction to auditor expert opinion was not
favorable compared with those opinions prepared by experts other than the auditors. From these
studies the empirical evidence shows that when the auditor provides multiple services including
fairness opinions, terms and pricing effects are apparent in the merger transaction.
Investment Bankers and M&A
The incentives for investment bankers to provide multiple services other than the fairness
opinion are immense. Knowledge spillover increases efficiency in providing multiple services.
Fees associated with multiple services only deepen the conflict of interest (DeMott 2004). As an
example, in the Bank of Americas acquisition of Fleet Boston Financial Corp., Goldman Sachs
acted in a dual role as a financial advisor and as provider of the fairness opinion for Bank of
America. Morgan Stanley provided identical services to Fleet Boston. Goldman Sachs received
a $25 million fee of which $3 was payable upon execution of the engagement letter, $5 million
for the fairness opinion, and the remaining $17 million payable upon completion of the merger.
Morgan Stanley received an identical fee from FleetBoston, the principal portion payable upon
completion of the merger.
262
Both investment bankers provided other services to the respective
firms in the ordinary course of business. With such fee structures as customary, there is reason
for concern that existing business relationships provide an incentive to find the deal fair (Leddy
and Walters 2005).
There is limited empirical evidence of the impact that investment banker fairness opinions have
on the merger and acquisition transaction. Kisgen et al. (2009) finds for both completed and
incomplete transactions, the fairness opinion is associated with lower announcement period

261
In this study, only eight acquirers and nine target firms used their auditor for advisory services in the business
combination. Auditors before the SEC date may provide purchase price allocation valuation services needed for
the purchase method of accounting. The rendering of these services is not typically disclosed.
262
See Form S-4, Registration of Securities, Business Combinations filed by Bank of America on December 4, 2003
with the Securities and Exchange Commission.
1043

returns for the acquirer, but having one does increase the likelihood that the deal closes. The
authors also find for their sample, a positive, but insignificant association between announcement
period returns and when the acquirer-obtained fairness opinion is through an independent
financial advisor. Conversely, Makhija and Narayanan (2007) find significantly higher
announcement period returns for both acquiring and target firms when independent financial
advisors deliver the fairness opinion compared with the sample of non-independent financial
advisors.
There is also limited evidence that fairness opinions influence bid premiums. In a study of
Australian mergers and acquisitions, Eddey (2001) finds no significant differences in bid
premiums when a fairness opinion is issued compared with when it is not issued. Conversely,
Makhija and Narayanan (2007) find lower bid premiums when the financial advisor provides the
fairness opinion and shareholders react negatively to the conflict of interest but positively react if
the opinion is provided by an independent reputable advisor. Kisgen et al. (2009) find
conflicting results for acquiring and target firms: no effect on bid premium for target firms that
obtain a fairness opinion, but lower bid premiums for acquirers who do the same, especially
when acquirers use multiple advisors.
In order to assess the relative value difference between fairness opinions provided by auditors for
their audit clients and those provided by valuation specialists, we use the effective date of SEC
Rule 57-13-00, February 5, 2001, to capture the relative difference. Given that between the time
periods when auditors were permitted and not permitted to provide fairness opinions for their
clients, regulators deemed auditor lack of independence issues severe enough to require
regulation. We expect that the market should behave in a similar fashion to that of regulators.
The impact of the conflict of interest the financial advisor who also provides the fairness opinion
is stated in the following hypothesis:
Hypothesis 1a. There is a negative association between announcement period
returns and when the acquirer or target firm financial advisor delivers the fairness opinion.
Hypothesis 1b. The negative association between fairness opinions and
announcement period returns is more pronounced for mergers and acquisitions that closed before
the SEC date, when auditors were not prohibited from providing fairness opinions for their audit
clients, than for those mergers and acquisitions that closed after the SEC date.
Investment Banker Reputation
Kisgen et al. (2009) posit that the better investment banker reputation the better quality of the
fairness opinion, which should lead to better pricing for acquiring firms in pending deals.
However, their empirical results do not support their hypothesis. Lower quality advisors should
not improve the deal pricing structure due to less useful information. An analysis of completed
deals could provide a different result while considering the perspectives of the merging firms. In
their sample that includes mergers in the financial services industry
263
, Makhija and Narayanan
(2007) find target announcement period returns increase as the reputations of financial advisors
providing fairness opinions improves. The authors found no reputational effects for acquirer-
obtained fairness opinions or when independence of the financial advisor is compromised for
either target or acquiring firms.
The reputation of the financial advisor may be vital in the situation when the financial advisor
also delivers the fairness opinion. The following hypotheses are set forth:
Hypothesis 2. There is a positive association between announcement period returns and
the acquiring and target firms financial advisors who have a high reputation.

263
See Note 8.
1044

Hypothesis 3. The association between announcement period returns and the reputation
of the financial advisor is modified when the financial advisor also delivers the fairness opinion.

III. Research Methodology
Univariate tests are used to test the significance in the mean differences between the samples
through two sample t-tests and the nonparametric Wilcoxon rank-sum test. The two sample t-
tests is vulnerable to the problems of nonnormality and outliers while the Wilcoxon test is less
restrictive and less susceptible to the influence of outliers. Multiple regression analysis is used to
test the association between the dependent variable and the test variables. Tests for
heteroscedasticity and multicollinearity are performed. Sensitivity analysis is done for the
alternate calculations of dependent and independent variables.
Sample
The sample of mergers and acquisitions is obtained from the Mergerstat database with the
requirement that all transactions are between U.S. publicly held firms with announcement and
completion dates between January 1, 1995 and December 31, 2006 and the acquiring firm
acquired control of the target firm through a stock acquisition. This procedure yielded 5,579
observations. Bid premium, deal value, financial advisors, auditors, and their respective roles are
obtained from Mergerstat. The observations from Mergerstat are complemented with financial
data from Compustat and stock data from CRSP. The ranking of advisor reputation follows that
used by Kisgen et al. (2009). See Appendix A for description of advisor rankings. The following
observations are deleted from the final sample (i) if there is any missing data from Mergerstat,
Compustat or CRSP; (ii) if an observation is accompanied with a negative bid premium; (iii) if
the firm is in the financial industry (SIC 6000 6999); or (iv) if counter-party is deleted due to
the aforementioned reasons. The last requirement is to make sure that the final sample is a clean
matched sample with both acquirer and target information. Similar to Kisgen et al. (2009), it is
noted that the Mergerstat database is not only incomplete, but also inaccurate. To ensure
accuracy, additional data is manually obtained and verified against the data provided by
Mergerstat through review of proxy and registration statements (Forms DEFM 14A, S-4)
disclosures in merger agreements found in the SECs EDGAR filings. With this screening, our
final sample consists of 576 acquirer-target matched observations
264
. A distribution of the
sample observations appears in Table 1 and Table 2 provides definitions of variables.
Panel A of Table 1 summarizes the sample value by year. Looking at the value of the deals,
there is a tendency that both mean and median deal value have increased over time, suggesting
some potential influence of inflation over time. Also, the distinct difference between mean and
median values clearly indicates that there are a few deals with extreme values . The deals are
clustered around 1998 and 1999, while some other years, except for 1995 and 2006, show
somewhat closer number of deals. This clustering of deals around 1998 and 1999 could be
attributed to the high performance of stock market around this time fueled by the dot-com stocks.

264
Unlike Makhija and Narayanan (2007) where the financial services industry represents nearly 29% of their
sample, this study excludes the financial services industry because they typically may use their own experts for
valuation services.
1045

Table 1: Descriptive Statistics
Panel A: Deal value by year ($millions)
Year # of
observations
Mean Median Minimum Maximum
1995 1 1,078 . . .
1996 49 54,498 2,184 9.11 811,405
1997 71 71,469 13,534 66.00 1,685,503
1998 106 287,518 8,755 8.82 8,142,982
1999 116 140,588 9,985 18.14 5,325,314
2000 60 492,996 17,025 49.73 13,275,549
2001 55 818,241 85,717 13.98 7,254,933
2002 26 720,023 12,086 16.16 6,705,013
2003 42 236,428 30,888 26.01 1,825,364
2004 34 1,308,293 347,292 152.90 12,368,035
2005 15 7,494,952 241,459 21.79 57,920,336
2006 1 447,385 . . .
Total 576
Panel B: Distribution by Industry
Industry SIC Acquirers Targets
Agricultural 100-900 1 3
Mining 1000-1400 18 23
Construction 1500-1731 6 9
Manufacturing 2000-3990 256 234
Transportation 4011-4899 49 42
Utilities 4900-4991 36 32
Wholesale/Retail 5000-5990 67 59
Other services 7000-8750 143 174
Total 576 576
Panel C: Deal characteristics
Cash only transaction 33.85%
Before SEC date 68.23%
A_Dual 28.82%
T_Dual 63.54%
Acquirer with fairness opinion 29.17%
Target with fairness opinion 68.92%
Panel D: Sample firm characteristics
Mean Median Std. Dev.
Acquirer Total Assets 11,760.4 1,776.4 42,027.7
Acquirer M/B 2.47 2.88 36.37
Target Total Assets 1,488.6 222.6 4,703.2
Target M/B 3.88 2.14 23.01
Prem 1 17.19 8 21.10
Prem 5 19.94 13 22.33
Prem30 21.21 14 20.58
Panel E: Advisor reputation
Advisor Reputation Acquirers Targets
Tier 1 230 179
Tier 2 87 150
Tier 3 259 247
Total 576 576
See Table 2 for variable definitions.
Panel B of Table 1 presents the distribution of the final sample by acquirers and targets
industries, respectively. More than half of the deals are involved either manufacturing (SIC 2000
to 3990) or other services industries (SIC 7000 8750). It is interesting to note that based on the
1046

distribution of samples by industry, mergers tend to happen within an industry most of the time.
This is indicated by the fact that given any industry in Panel B, there are comparable numbers
between acquirers and targets.

Table 2: Definitions of Variables
A_CAR2 = Acquirers cumulative abnormal returns over (-1,0) period.
T_CAR2 = Targets cumulative abnormal returns over (-1,0) period.
A_CAR3 = Acquirers cumulative abnormal returns over (-2,0) period.
T_CAR3 = Targets cumulative abnormal returns over (-2,0) period.
AFO = 1 if acquirer firm a fairness opinion, 0 otherwise.
TFO = 1 if target firm obtains a fairness opinion, 0 otherwise.
A_Dual =1 if the acquirer firm financial advisor also provides the fairness opinion,
0 otherwise.
T_Dual =1 if the target firm financial advisor also provides the fairness opinion, 0
otherwise.
SEC_Date = 1 if business combination closed before the SEC date of February 5, 2001, 0
otherwise.
Payment = 1 if the form of consideration used in the business combination was only cash, 0
otherwise.
A_MB = the acquiring firm market-to-book ratio for the business combination year.
T_MB = the target firm market-to-book ratio for the business combination year.
A_TA = the log value of the acquiring firm total assets for the business
combination year.
T_TA = the log value of the target firm total assets for the business combination year.
Premium_1 = bid premium calculated as t-1, where t is the announcement date.
ATier1 = 1 if acquirer uses a tier one advisor, 0 otherwise.
ATier2 = 1 if acquirer uses a tier two advisor, 0 otherwise.
ATier3 = 1 if acquirer uses a tier three advisor, 0 otherwise.
TTier1 = 1 if target uses a tier one advisor, 0 otherwise.
TTier2 = 1 if target uses a tier two advisor, 0 otherwise.
TTier3 = 1 if target uses a tier three advisor, 0 otherwise.

The deal characteristics are shown in Panel C of Table 1. About one-third of the deals use cash
as the sole form of consideration. Also, about 68 percent of our final sample took place before
the SEC date of February 5, 2001. Effective on this date, auditors are not allowed to perform
advisory functions for a firm if it acts as the auditor for the firm. Focusing on the fairness of
opinion, it is shown that close to 30 percent of acquirers obtained fairness opinions. In contrast,
target firms use fairness of opinion more frequently (about 69 percent). These occurrences of
fairness opinions are similar to Kisgen, Qian and Songs study (2009). Both acquirers and
targets show very similar percentages for duality and fairness opinions. The finding is due to the
possibility that the fairness opinions are provided by the financial advisor for the firm. The
correlations, shown in Tables 5 and 6, provide the evidence for this argument.
Panel D of Table 1 summarizes the descriptive statistics of sample firms. The average (median)
size of the acquirer is $11.8 billion ($1.8 billion), respectively. In contrast, the average (median)
1047

size of the target is $1.5 billion ($0.2 billion), respectively, which is considerably smaller than
the acquirers. Both acquirers and targets show the market-book ratio of more than 2, which
suggests that they both have some potential growth. Premium of the bid is around 20 percent
over the stock price on average, while the median premium is around 13 percent. Finally, Panel
E of Table 1 summarizes the advisor reputation. Surprisingly, for both acquirers and targets, the
most deals utilize advisors with lowest level of reputation. However, advisors with the best
reputation are more likely to be associated with acquirers, while this is not true for targets.

IV. Results
Table 3 summarizes the results of stock market reactions around the merger announcements. For
each announcement of the 576 mergers, daily abnormal returns (ARs) are calculated using a
standard event-study methodology for targets and acquirers separately. Market model estimates
are obtained by using 250 trading days of returns data, beginning 271 days before and ending 21
days before the announcement. Our benchmark market return is the CRSP equally-weighted
market return (using CRSP value-weighted market returns yield similar results). The ARs for
days -2, -1, and 0 (where day 0 is the announcement day) are shown in the top section of Table 3
and are accumulated to obtain a cumulative three-day abnormal return (i.e., CAR(-1,0))
surrounding the announcement date. The cumulative abnormal returns of 2-day (-1, 0) and 3-day
(-2, 0) are shown in the lower part of the Table 3. This study uses ARs of -2, -1, and 0 as
fairness opinions are typically made available on the day prior to the actual merger
announcement.
265

From Table 3, it is clear that targets enjoy a very significant positive mean and median
cumulative abnormal returns of about 22% and 16%, respectively. This is consistent with the
earlier studies that examined the effect of merger announcements on targets stock reactions (i.e.,
Dennis and McConnell (1986) and Huang and Walkling (1987)). Turning attention to acquirers
reactions, acquirer stockholders tend to react negatively on the merger announcements. The
results of CARs in Table 3 indicate that the mean and median returns are around -2% and -1.5%,
respectively. While prior studies
266
have shown inconclusive evidence on the acquiring firms
stock reactions around merger announcements, our results are consistent with the findings by
Moeller, Schlingemann, and Stulz (2005) in which they contribute more significant negative
returns on the frenzied merger activity period of 1998 to 2001 which our sample period overlaps.
Table 3: Event study results
Acquirer Target
Mean Median Mean Median
AR (-2) -0.24%
***
-0.19%
**
1.37%
***
0.47%
***
AR (-1) -0.17%
*
-0.19%

1.53%
***
0.71%
***
AR (0) -1.75%
***
-1.24%
***
20.69%
***
15.13%
***

CAR (-1, 0) -1.92%
***
-1.57%
***
22.22%
***
16.56%
***
CAR (-2, 0) -2.16%
***
-1.79%
***
23.59%
***
18.31%
***

Table 4 presents the univariate tests for the sample of mergers and acquisitions. First, the sample
is segmented by payment method. According to Panel A, approximately 68% of the sample

265
A sample of the observations used in this study finds that the fairness opinion date is dated one day prior to
announcement date. As a robustness check, including AR of one day after the announcement does not change our
results significantly.
266
For example, Jensen and Ruback (1983) find no price reactions by acquirers stocks, while Jarrell, Brickley and
Netter (1988) show a declining trend in the acquirers stock reactions from 1960s to 1980s.
1048

observations represent merger transactions occurring before the SEC date. During that period,
CARS for Target firms were significantly higher than their CARS after the SEC date. In Panel D
of Table 1, it is noted that about one-third of deals are cash only transactions. From the
perspective of acquirers shareholders, new stock issuance for acquisition results in dilution
effects, while only cash transactions do not result in dilution. Therefore, it might be beneficial
for acquirers shareholders if the deal only involves cash. Similarly, targets shareholders would
have more flexibility if they were paid by cash rather than stock of the combined firms. Results
of differences in CARs between two subsamples, shown in Panel B of Table 4, confirm our
arguments. Acquirers negative CARs around the merger announcements are dominated by the
firms who use stock as some or all of the payment. Target shareholders react more positively to
the announcements when the deals are cash only transactions. Looking at the premium bid
differences, it is noted that these differences in CARs are not driven by the premium bid as there
is no difference between the two subsamples. The size of the firm seems to matter. In other
words, the larger the acquirer, the more likely the all cash payment is employed. On the other
hand, for target firms, it is more likely to be cash transactions when the targets are smaller firms.
Next, univariate tests are separately performed for acquirers and targets. Specifically, deals are
segmented by duality, which is the circumstance when the financial advisor also renders the
fairness opinion. For acquirers, the results are summarized in Panels B and C of Table 4, while
Panels E and F are for targets.
Table 4: Univariate comparisons
P-values are reported for the tests of differences.
Panel A: SEC date
1 (n=393) 0 (n=183) Test of Differences
Mean Median Mean Median Mean Median
A_CAR2 -1.79% -1.68% -2.19% -1.09% 0.532

0.879
A_CAR3 -2.18% -1.82% -2.12% -1.77% 0.917

0.532
T_CAR2 20.73% 16.24% 25.42% 17.34% 0.038
**
0.676
T_CAR3 22.25% 18.49% 26.45% 17.08% 0.068
*
0.907
Prem_1 16.75 8 18.13 9 0.461

0.635
Prem_5 19.68 13 20.52 15 0.674 0.691
Prem_30 21.46 15 20.65 13 0.661 0.869
A_MB 2.30 3.05 2.84 2.40 0.868 0.002
***
T_MB 4.33 2.23 2.95 1.98 0.504 0.017
**
A_TA 8,440.52 1,440.61 18,889.81 2,842.65 0.005
***
0.0004
***
T_TA 1,288.85 229.23 1,917.49 218.96 0.135

0.586

Panel B: Payment
A_CAR2 0.81% 0.15% -3.31% -3.27% < 0.001
***
< 0.001
***

A_CAR3 0.33% -0.44% -3.44% -2.80% < 0.001
***
< 0.001
***

T_CAR2 29.56% 22.37% 18.46% 14.16% < 0.001
***
< 0.001
***

T_CAR3 30.84% 23.58% 19.87% 15.62% < 0.001
***
< 0.001
***

Prem_1 19.68 13 15.92 7 0.043
**
0.261
Prem_5 21.83 14 18.98 12 0.147 0.523
Prem_30 22.56 15 20.67 13 0.381 0.333
A_MB 2.93 2.67 2.24 3.02 0.829 0.117

T_MB 2.55 1.91 4.57 2.24 0.318 0.009
***
A_TA 17,926.8 2,411.0 8,604.3 1,569.5 0.012
**
0.002
***

T_TA 730.3 169.4 1,876.7 298.6 0.006
***
0.006
***

1049

Panel C: A_Dual
A_CAR2 -3.54% -3.37% -1.26% -0.82% 0.001
***
0.001
***
A_CAR3 -3.88% -3.02% -1.47% -1.41% < 0.001
***
0.001
***
Prem_1 17.65 13.5 17.01 7.5 0.743 0.122
Prem_5 21.33 17 19.38 12 0.345 0. 062
*
Prem_30 21.71 18.5 21.00 13 0.710 0.306
A_MB 3.76 2.58 1.95 2.98 0.585 0.021
**
A_TA 6,593.6 915.8 13,852.3 2,304.4 0.060
*
< 0.001
***

Panel D: A_Opinion
A_CAR2 -3.27% -3.04% -1.36% -0.96% 0.003
***
0.002
***
A_CAR3 -3.66% -2.78% -1.55% 1.44% 0.002
***
0.002
***
Prem_1 17.40 14 17.11 8 0.879 0.210
Prem_5 20.48 16.5 19.72 12 0.713 0.165
Prem_30 21.24 16 21.20 14 0.982 0.474
A_MB 3.50 2.59 2.05 2.99 0.665 0.013
**
A_TA 6,995.1 880.6 13,722.5 2,304.4 0.081
*
< 0.001
***

Panel E: T_Dual
T_CAR2 21.53% 16.11% 23.43% 18.25% 0.385 0.362

T_CAR3 22.71% 16.88% 25.12% 19.73% 0. 280 0.384

Prem_1 17.66 10 16.38 7.5 0.482 0.317
Prem_5 20.28 14.5 19.35 10 0.629 0.306
Prem_30 21.96 16.5 19.90 12.5 0.250 0.178
T_MB 4.44 2.15 2.91 2.13 0.440 0.879
T_TA 1,457.9 224.0 1,542.0 222.3 0.837

0.749
Panel E: T_Opinion
T_CAR2 21.88% 16.37% 22.98% 17.48% 0.385 0.653
T_CAR3 23.15% 17.08% 24.56% 19.19% 0.280 0.719
Prem_1 17.69 9 16.10 6 0.482 0.139
Prem_5 20.66 16 18.35 6 0.629 0.068
*
Prem_30 21.98 18 19.50 9 0.250 0.133
T_MB 4.41 2.17 2.74 2.10 0.440 0.295
T_TA 1,400.3 236.4 1,684.4 212.8 0.837

0.913

Noteworthy in Panels C and D is that the results in both Panels are very similar. As noted in the
Table 1, the firms advisor provides the fairness opinion (thus duality). For that reason, on the
remaining tests focus on the results from fairness of opinion. For acquirers, the use of fairness of
opinion seems to be associated with lower CARs. Also, the smaller acquirers use fairness of
opinion more frequently than larger acquirers do. Similar patterns are noticeable for targets. In
other words, the results for targets based on duality and on fairness of opinion are also very
similar. However, unlike acquirers, there are no differences in CARs and size of target firms.
Tables 5 and 6 present the Pearson Correlation Coefficients for target and acquiring firms,
respectively. The two-day CARs and three-day CARs have almost a perfectly positive
correlation for both acquirers and targets. This indicates that adding the extra day to the CAR
calculation does not affect the results. Due to the high correlation, our regression analyses will
be based on two-day CARs. The correlation between duality and fairness of opinion is highly
1050

significant and positive for both acquirers and targets, confirming our argument that the fairness
of opinions are mostly provided by the firms advisors. The correlation tables for targets and
acquirers firms, the target firm correlation indicates significant positive correlations between
returns and cash consideration and with bid premium. Significant negative correlations exist
between returns and SEC date, T_TA, and with T_MB. For the acquirer, returns and cash
consideration is also positive and significant. To avoid any collinearity problems, we only
employ one measure of premium, Prem_5, in the regression analyses in the following section
267
.
Table 5: Pearson Correlation Coefficients Target Firms

T_CAR2 T_CAR3 T_Dual T_Opinion
Sec
Date
Cash T_TA T_MB Prem_1 Prem_5 Prem_30
Tier_1 Tier_2 Tier_3
T_CAR 2 1 0.983
(< 0.001)
-0.036
(0.385)
-0.020
(0.630)
-0.086
(0.038)
0.208
(< 0.001)
-0.122
(0.003)
-0.029
(0.482)
0.122
(0.003)
0.101
(0.016)
-0.074
(0.075)
-0.042
(0.313)
0.000
(0.994)
0.039
(0.349)

T_CAR 3 1

-0.045
(0.280)
-0.025
(0.544)
-0.076
(0.068)
0.202
(< 0.001)
-0.115
(0.006)
-0.020
(0.631)
0.109
(0.009)
0.114
(0.006)
-0.070
(0.095)
-0.042
(0.317)
0.011
(0.791)
0.029
(0.483)

T_Dual 1 0.886
(< 0.001)
-0.168
(<
0.001)
-0.015
(0.728)
-0.009
(0.837)
0.032
(0.440)
0.029
(0.482)
0.020
(0.629)
0.048
(0.250)
-0.053
(0.208)
0.014
(0.740)
0.037
(0.378)

T_Opinion 1 -0.128
(0.002)
-0.019
(0.649)
-0.028
(0.503)
0.034
(0.422)
0.035
(0.404)
0.048
(0.250)
0.056
(0.181)
0.029
(0.481)
-0.037
(0.369)
0.006
(0.891)

SEC Date 1 -0.087
(0.037)
-0.063
(0.135)
0.028
(0.504)
-0.031
(0.461)
-0.018
(0.674)
0.018
(0.661)
0.120
(0.004)
0.031
(0.457)
-0.140
(< 0.001)

Cash 1 -0.115
(0.006)
-0.042
(0.318)
0.084
(0.043)
0.061
(0.147)
0.037
(0.381)
-0.060
(0.149)
0.010
(0.807)
0.047
(0.257)

T_TA 1 -0.013
(0.763)
0.019
(0.655)
0.013
(0.765)
-0.005
(0.900)
0.183
(<
0.001)
-0.060
(0.149)
-0.118
(0.005)

T_MB 1 0.003
(0.944)
0.025
(0.548)
-0.013
(0.754)
-0.012
(0.769)
-0.024
(0.565)
0.033
(0.432)

Prem_1 1 0.737
(<
0.001)
0.541
(<
0.001)
0.025
(0.543)
-0.061
(0.143)
0.030
(0.467)

Prem_5 1 0.595
(<
0.001)
0.069
(0.098)
-0.067
(0.106)
-0.005
(0.909)

Prem_30 1 0.038
(0.364)
-0.035
(0.398)
-0.004
(0.920)

Tier_1 1 -0.398
(< 0.001)
-0.582
(< 0.001)

Tier_2 1 -0.514
(< 0.001)

Tier_3 1
See Table 2 for definitions of variables.

267
Using Prem_1 in the regression models show similar results, while with Prem_30 findings loses some
significance.
1051

Table 6: Pearson Correlation Coefficients Acquiring Firms

A_CAR2 A_CAR3 A_Dual A_Opinion
Sec
Date
Cash A_TA A_MB Prem_1 Prem_5 Prem_30
Tier_1 Tier_2 Tier_3
A_CAR 2 1 0.914
(< 0.001)
-0.145
(< 0.001)
-0.123
(0.003)
0.026
(0.532)
0.275
(< 0.001)
0.022
(0.603)
0.012
(0.783)
-0.014
(0.775)
-0.018
(0.665)
0.011
(0.797)
0.066
(0.115)
-0.008
(0.850)
-0.059
(0.157)

A_CAR 3 1

-0.149
(< 0.001)
-0.130
(0.002)
-0.004
(0.917)
0.241
(< 0.001)
0.027
(0.512)
0.059
(0.156)
0.007
(0.876)
0.033
(0.433)
0.007
(0.873)
0.027
(0.515)
0.016
(0.709)
-0.038
(0.363)

A_Dual 1 0.916
(< 0.001)
-0.043
(0.299)
-0.277
(< 0.001)
-0.078
(0.060)
0.023
(0.585)
0.014
(0.743)
0.039
(0.345)
0.016
(0.710)
-0.034
(0.422)
0.031
(0.453)
0.011
(0.802)

A_Opinion 1 -0.079
(0.058)
-0.257
(< 0.001)
-0.073
(0.081)
0.018
(0.665)
0.006
(0.879)
0.015
(0.713)
0.001
(0.982)
-0.024
(0.565)
0.007
(0.873)
0.019
(0.651)

SEC Date 1 -0.087
(0.037)
-0.116
(0.005)
-0.007
(0.868)
-0.031
(0.461)
-0.018
(0.674)
0.018
(0.661)
0.092
(0.027)
-0.098
(0.019)
-0.020
(0.626)

Cash 1 0.105
(0.012)
0.009
(0.829)
0.084
(0.043)
0.061
(0.147)
0.037
(0.381)
0.001
(0.981)
0.057
(0.173)
-0.042
(0.315)

A_TA 1 0.013
(0.751)
0.072
(0.083)
-0.001
(0.975)
0.054
(0.197)
0.017
(0.691)
0.154
(< 0.001)
-0.127
(0.002)

A_MB 1 0.005
(0.911)
-0.075
(0.074)
-0.091
(0.028)
0.043
(0.300)
0.035
(0.403)
-0.068
(0.105)

Prem_1 1 0.737
(<
0.001)
0.541
(< 0.001)
-0.022
(0.602)
-0.010
(0.811)
0.029
(0.493)

Prem_5 1 0.595
(< 0.001)
0.012
(0.768)
-0.013
(0.763)
-0.003
(0.942)

Prem_30 1 -0.023
(0.589)
0.002
(0.959)
0.021
(0.620)

Tier_1 1 -0.344
(< 0.001)
-0.737
(< 0.001)

Tier_2 1 -0.381
(< 0.001)

Tier_3 1
See Table 2 for definitions of variables.

Target Firms Regression Analysis
Table 7 presents the regression analyses to test hypotheses 1 for the target firms. The variable
TFO is negative and significant. Significant results are found for the interaction between SEC
date and target fairness opinion and in the separate regressions before and after the SEC date,
suggesting that when the deal closed before the SEC date, the value of the fairness opinion also
diminishes returns. These findings support hypotheses 1. The interaction between bid premium
and fairness opinion is positive and significant, which suggests that for those target firms that
obtained a fairness opinion, the bid premium impact on returns increases.

1052

TABLE 7: Target Regressions
Dependent Variable T_CAR2
Before SEC
Date
After SEC
Date
Total Sample

Intercept 0.258 0.285 0.238 0.207 0.384
(9.021)*** (9.683)*** (8.178)*** (7.741)*** (7.556)***
TFO -0.032 -0.076 0.017 -0.050 0.013
-(2.394)* -(3.994)*** (0.971) -(3.667)*** (0.523)
SEC_Date -0.024 -0.024 -0.003
-(1.801) -(1.834) -(0.179)
Payment 0.090 0.093 0.093 0.093 0.097
(6.804)*** (7.275)*** (7.200)*** (6.890)*** (3.909)***
T_TA -0.036 -0.036 -0.035 -0.020 -0.100
-(4.851)*** -(4.964)*** -(4.761)*** -(2.467)* -(6.934)***
Premium_5 0.002 0.000 0.002 0.002 0.002
(5.346)*** (0.068) (5.409)*** (4.921)*** (3.807)***
Premium*TFO 0.002
(3.462)***
SEC_Date*TFO -0.060
-(3.738)***

Adjusted R
2
0.118 0.135 0.134 0.147 0.210
F-Value 16.400 15.990 15.820 17.900 13.090
N 576 576 576 393 183
*,**,*** Statistical significance at the 0.05, 0.01,0.001, respectively (2-tailed). T-values are in
parenthesis. See Table 2 for definition of variables.

Acquirer Regressions
Table 8 presents results for the hypothesis 1 for acquiring firms. There is a significant negative
association between AFO and returns. The interaction between bid premium and fairness
opinion is negative and insignificant. The separate regressions for before the SEC date show
consistent findings as above for AFO.
Table 8: Acquirer Regressions
Dependent Variable A_CAR2
Total Sample
Before
SEC Date
After SEC
Date

Intercept -0.012 -0.014 -0.011 -0.011 -0.029
-(1.375) -(1.638) -(1.335) -(1.129) -(2.752)**
AFO -0.011 -0.006 -0.022 -0.012 -0.004
-(2.845)** -(1.210) -(3.357)** -(2.447)* -(0.788)
SEC_Date 0.003 0.002 0.001

(0.776) (0.473) (0.266)

Payment 0.034 0.034 0.034 0.036 0.033
(10.456)*** (10.427)*** (10.193)*** (8.776)*** (8.805)***
A_TA -0.005 -0.004 -0.005 -0.005 0.001
1053

-(2.379)* -(2.192)* -(2.339)* -(1.889) (0.198)
Premium_5 0.000 0.000 0.000 0.000 0.000

-(1.624) -(0.364) -(0.998) -(0.593) -(2.474)**
Premium_5*AFO
0.000

-(1.593)

SEC_Date*AFO
0.012

(1.704)

Adjusted R
2

0.131 0.133 0.132 0.122 0.203
F-Value
18.370 15.710 15.520 14.570 12.580
N 576 576 576 393 183


Reputation Regressions
Table 9 presents the regression analyses for target advisor reputation effects and tests hypotheses
2 and 3, which follows the ranking procedure of Kisgen et al. (2009 (see Appendix A). Financial
advisors classified as tier one and two have a positive insignificant association with returns and
then a negative association with a tier three advisor, but still remains insignificant. When the
advisor reputation interacts with fairness opinion, the interaction variable is negative and
significant for tier 1 advisors and turns to positive and significant for tier three advisors. From
these results as the reputation of the financial advisor lessens, the more weight the fairness
opinion has on target firm returns. The market does view the use of a tier one advisor more
negatively than the use of a tier two or three advisor, when the target firm obtains a fairness
opinion. In separate regressions of returns on TFO by advisor tier, TFO is negative and
significant for tier one and tier two advisors, but positive and insignificant when the target firm
uses a tier three advisor (results not tabulated). The findings described above hold when the
merger and acquisition occurred before the SEC date (results not tabulated).
Table 9: Target Advisors Reputation Regressions
Dependent Variable T_CAR2
Intercept 0.275 0.274 0.034 0.257 0.264 0.319
(9.742)*** (9.650)*** (8.935)*** (9.003)*** (9.326)*** (10.266)***
TFO -0.036 -0.036 0.015 -0.020 -0.023 -0.081
-(2.756)** -(2.809)** -(2.604)** -(1.397) -(1.590) -(4.940)***
SEC_Date -0.030 -0.027 0.015 -0.031 -0.028 -0.036
-(2.276)* -(2.069)* -(1.606) -(2.332)* -(2.130)* -(2.776)*
Payment 0.091 0.087 0.015 0.095 0.085 0.095
(7.033)*** (6.811)*** (7.032)*** (7.344)*** (6.710)*** (7.568)***
T_TA -0.043 -0.040 0.009 -0.040 -0.040 -0.042
-(5.614)*** -(5.588)*** -(5.416)*** -(5.285)*** -(5.645)*** -(5.804)***
Premium_5 0.002 0.002 0.000 0.002 0.002 0.002
(5.516)*** (5.618)*** (5.862)*** (5.486)*** (5.779)*** (5.969)***
TTier1 0.018

0.055

(1.520)

(3.338)**

TTier2

0.002

0.055

(0.125)

(2.111)*

1054

TTier3

0.014

-0.076

-(0.976)

-(4.553)***
TTier1*TFO
-0.054

-(2.790)**

TTier2*TFO
-0.066

-(2.241)*

TTier3*TFO
0.088

(4.652)***

Adjusted R
2

0.130 0.127 0.125 0.153 0.133 0.161
F-Value
15.260 14.900 14.730 12.760 13.540 16.720
N 576 576 576 576 576 576

parenthesis. See Table 2 for definition of variable
1055

The reputation effect of acquirer financial advisors appears on Table 10 to test hypotheses 2 and
3. Tier one advisors is positive and significant and changes to negative and insignificant with
tier two and significant with tier three advisors. The interactions between advisor reputation and
fairness opinion are negative and significant for tier one and turns to positive and significant
when a tier two and tier three advisors are used. Similar to the target firm analyses, the findings
above hold in separate regressions for tier one and two advisors and for mergers and acquisitions
before the SEC date (results not tabulated). Collectively, the regression results for advisor
reputation of target and acquiring firms do support hypotheses 2 and 3.
Results for control variables
Returns are significantly lower for target firms, yet significantly higher for acquiring firms.
Returns are significantly greater when cash is the sole form of consideration for both acquiring
and target firms. Returns are significantly lower with large target firms. Target firms also
produce higher returns with higher bid premiums.
Table 10: Acquirer Advisors Reputation Regressions
Dependent Variable A_CAR2
Intercept -0.013 -0.016 -0.001 -0.015 -0.018 -0.003
-(1.569) -(1.972) -(0.143) -(1.944) -(2.180)* -(0.384)
AFO -0.012 -0.010 -0.012 -0.003 -0.012 -0.018
-(3.099)** -(2.735)** -(3.370)** -(0.605) -(2.883)** -(3.924)***
SEC_Date 0.000 0.002 0.001 0.000 0.003 0.002
(0.101) (0.749) (0.285) -(0.013) (0.884) (0.757)
Payment 0.036 0.037 0.035 0.036 0.037 0.036
(11.348)*** (11.712)*** (11.269)*** (11.607)*** (11.722)*** (11.668)***
A_TA -0.005 -0.004 -0.006 -0.005 -0.003 -0.005
-(2.665)** -(1.953)* -(2.864)** -(2.656)** -(1.638) -(2.694)**
Premium_5 0.000 0.000 0.000 0.000 0.000 0.000
-(1.896) -(1.534) -(2.095)* -(1.968)* -(1.519) -(1.833)
ATier1 0.010

0.015

(3.046)**

(4.123)***

ATier2

-0.002

-0.005

-(0.597)

-(1.314)

ATier3

-0.010

-0.014

-(3.055)**

-(3.664)***
ATier1*AFO
-0.021

-(3.109)**

ATier2*AFO
0.014

(1.991)*

ATier3*AFO
0.014

(2.064)*

Adjusted R
2

0.155 0.150 0.158 0.164 0.150 0.160
F-Value
18.550 17.870 18.920 17.160 15.500 16.680
N 576 576 576 576 576 576


1056

Sensitivity Results
The sensitivity analyses for alternative calculations of cumulative abnormal returns calculated
for a three day window and bid premium at one and thirty days prior to announcement date and
controlling for transactions before the effective date of the SOX regulation provide results that
are qualitatively similar to those previously reported.

V. Summary and Conclusions
Fairness opinions provide value to shareholders, but when conflicts of interest by financial
advisors exist, criticisms of fairness opinions appear valid. Recent regulation acknowledges the
criticism through enhanced disclosures but falls short of prohibiting conflicting services in spite
of extensive compliance statutes that strain the due diligence process. Furthermore, since the ban
on auditors providing valuation services for their audit clients, investment bankers and other
valuation specialists benefit greatly from less competition.
Our examination yields results that support the criticisms. When auditors were permitted to
provide fairness opinions to their audit clients, returns are diminishing. Advisor reputation has
an impact. As the reputation of the financial advisor lessens, the more weight the fairness
opinion has on target firm returns and when the target financial advisor also provides the fairness
opinion, that financial advisor must have the best reputation, otherwise shareholders do not place
significant weight on the value of the fairness opinion. Acquirer shareholders also benefit from
fairness opinions and from using financial advisors with the highest reputation, but the market
may penalize the firm when fairness opinions are provided by the financial advisor even though
that financial advisor may have the best reputation.

References
Anderson, D., Stokes, D., and I. Zimmer. 1993. Corporate takeovers and auditor switching.
Auditing, Vol. 12(1): 65.
Antle, R., Gordon, E., Narayanamoorthy, G., and L. Zhow. 2006. The joint determination of
audit fees, non-audit fees, and abnormal accruals. Review of Quantitative Finance and
Accounting, Vol. 27 (3): 235-
Beatty, R. P. 1989. Auditor reputation and the pricing of initial public offerings. The
Accounting Review, Vol. 64 (4): 693-709.
Bugeja, M. 2005. The independence of expert opinions in corporate takeovers: Agreeing with
directors recommendations. Journal of Business Finance and Accounting, Vol. 32 (9 and
10): 1861-1885
Davidoff, Steven, 2006. Fairness Opinions, American University Law Review 55 (6), 1557-
1625
Davis, A. 2004. NASD Scrutinizes conflicts in bankers fairness opinions. The Wall Street
Journal. June 11, 2004: A.1
DeMott, D. A. 2004. Bank conflicts raise threat of lawsuits over M&A. International
Financial Law Review, August: 27-29.
Dennis, D. and J.J. McConnell, 1986. Corporate mergers and security returns. Journal of
Financial Economics, Vol. 16, 143-188.
Eddey Ph. H. 2001. Independent experts reports in takeover bids. Accounting and Finance,
May: 1-18.
Gorman, J., 1988. M&As: The CPA as acquisition adviser. The Journal of Accountancy,
August: 36-42.
1057

Huang, Y. and R. Walkling, 1987. Acquisition announcements and abnormal returns. Journal
of Financial Economics, Vol. 19, 329-350.
Jarell, G.A., J.A. Brickley and J.M. Netter, 1988. The market for corporate control: The
empirical evidence since 1980, Journal of Economic Perspective, Vol 2, 49-68.
Jensen, M.C. and R.S. Ruback, 1983. The market for corporate control. Journal of Financial
Economics, Vol. 11, 5-50.
Kaplan, R. L. 2004. The mother of all conflicts: Auditors and their clients. Journal of
Corporation Law, Winter 2004 Vol. 29(2) 3: 363-383.
Kisgen, D.J., Qian J., and W. Dong (2009). Are fairness opinions fair? The case of mergers and
acquisitions. Journal of Financial Economics Vol. 91 (2) 179-207.
Leddy, P. J. and R. M. Walters. 2005. The growing storm over fairness opinions. Mergers
and Acquisitions, March Vol. 40 (3) 35-40.
Louis, H., 2005. Acquirers abnormal returns and the non-Big 4 auditor clientele effect.
Journal of Accounting and Economics, 40: 75-99.
Makhija, A. K. and R.P. Narayanan. 2007. Fairness opinions in mergers and acquisitions.
Fisher College of Business Working Paper Series, Fisher College of Business, The Ohio
State University.
Moeller, S. B., Schlingemann, F. P. and R. M. Stulz. 2005. Wealth destruction on a massive
scale? A study of acquiring-firm returns in the recent merger wave. The Journal of
Finance, Vol. 60 (2): 757-782.
Placenti, F. M. 2007. Fairness opinions and success fees can make for volatile mix: Delaware
court finds that success fees are contingent compensation for certain investment bank
fairness opinions. The Investment Dealers Digest IDD New York May 7.
Sarbanes-Oxley Act of 2002. Pub.L. No. 107-2004, 201 (a), 116 Stat. 745, 771-72.
Securities and Exchange Commission. 2001 Final Rule: Revision of the Commissions Auditor
Independence Requirements. File No. S7-13-00 found at
http://www.sec.gov/rules/final/33-7919.htm.
Schneider, A., Church, B. K., and K. M. Ely. 2006. Non-audit services and auditor
independence: A review of the literature. Journal of Accounting Literature, 25:169-211.
Sharma, D. S., and J. Sidhu. 2001. Professionalism vs. commercialism: The association
between non-audit services (NAS) and audit independence. Journal of Business Finance
& Accounting, Vol. 28 (5) 7 (6): 595-629.
Simunic, D., A., 1984. Auditing, consulting, auditor independence. Journal of Accounting
Research, Vol. 22 (2): 679-702.
Sweeney P. and C. S. Sechler. 1999. Who says its a fair deal? Journal of Accountancy,
August Vol. 188(2): 44-51.

1058

APPENDIX A: Advisor Reputation

Acquirer Top Tier Target Top Tier
Merrill Lynch & Co Goldman Sachs & Co
Salomon Smith Barney Merrill Lynch & Co
Morgan Stanley & Co Donaldson Lufkin & Jenrette
Donaldson, Lufkin & Jenrette Morgan Stanley & Co
Goldman Sachs & Co Salomon Smith Barney

Acquirer Second Tier Target Second Tier
Bear Stearns Credit Suisse FirstBoston
Credit Suisse FirstBoston Lehman Brothers
Lehman Brothers Bear Stearns & Co
Alex Brown &Sons Inc. Alex Brown & Sons
Lazard Freres Sandler ONeill
Keefe Bruyette & Woods Dillon, Read & Co
Piper Jaffray Keefe Bruyette & Woods
Sandler ONeill Piper Jaffray
JPMorgan & Co Paine Webber
Hambrecht & Quist McDonald Investments
Stifel Nicolaus & Co Broadview
Dillon, Read &Co JPMorgan & Co
Paine Webber Hambrecht & Quist
UBS Lazard Freres
Wasserstein Perella Group Robinson-Humphrey

Acquirer Third Tier Target Third Tier
Banc of America Securities Dain Rauscher
Chase Manhattan Bank Wasserstein Perella Group
Robinson-Humphrey CIBC
CIBC Deutsche Bank AG
Jefferies & Co Montgomery Securities
Ryan Beck & Co Robert W Baird & Co
Montgomery Securities Charles Webb & Co
Needham & Co Robertson Stephens & Co
Nations Banc Corp SG Cowen Securities
Robert W Baird & Co Prudential Securities

1059

Editorial Board
Editor-In-Chief
Prof. Y Rama Krishna, MBA, PhD

Associate Editor
Prof. Vinay Datar, Seattle University, USA

Senior Sub-Editor
Dr. Moustafa Abu El Fadl, Rhode Island College, USA

Members
Dr. Yu Hsing, Southeastern Louisiana University, USA
Prof. Richard J. Cebula, Jacksonville University, USA
Prof. Nur Adiana Hiau Abdullah,Universiti Utara Malaysia
Dr. Mohammad Talha, King Fahd University of Petroleum & Minerals, Saudi Arabia
Dr. Sorin A Tuluca, Fairleigh Dickinson University, USA
Dr. Nozar Hashemzadeh, Radford University, USA
Dr. Dar-Hsin Chen, National Taipei University, Taiwan
Dr. Hayette Gatfaoui, Rouen Business School, France
Dr. Emmanuel Fragnire, University of Bath, Switzerland
Dr. Shwu - Jane Shieh, National Cheng-Chi University, Taiwan
Dr. Raymond Li, The Hong Kong Polytechnic University, Hong Kong
Prof. Yang-Taek Lim, Hanyang University, Korea
Dr. Anson Wong, The Hong Kong Polytechnic University, Hong Kong
Dr. Shaio Yan Huang, National Chung Cheng University, Taiwan
Mr. Giuseppe Catenazzo, HEC-University of Geneva, Switzerland
Prof. J P Singh, Indian Institute of Technology (R), India
Dr. Jorge A. Chan-Lau, International Monetary Fund, Washington DC
Prof. Carlos Machado-Santos, UTAD University, Portugal
Dr. Osama D. Sweidan,University of Sharjah, UAE
Dr. David Wang, Chung Yuan Christian University, Taiwan
Prof. David McMillan, University of St Andrews, Scotland UK
Dr. Abdullah Sallehhuddin, Multimedia University, Malaysia
Dr. Burak Darici, Balikesir University, Turkey
Dr. Mohamed Saidane, University of 7 November at Carthage, Tunisia
Dr. Anthony DiPrimio, Holy Family University, USA

Advisors
Dr. Michail Pazarskis, Technological Educational Institute of Serres, Greece
Dr. Huseyin Aktas, Celal Bayar University, Turkey
Dr. Panayiotis Tahinakis, University of Macedonia, Greece
Dr. Ahmet Faruk Aysan, Bogazici University, Turkey
Dr. Leire San Jose, Universidad del Pas Vasco, Spain
Dr. Arqam Al-Rabbaie, The Hashemite University, Jordan
Dr. Suleyman Degirmen, Mersin University, Turkey
Prof. Dr. Ali Argun Karacabey, Ankara University,
1060

Dr. Moawia Alghalith,UWI, St Augustine, West Indies
Dr. Michael P. Hanias, School of Technological Applications, Greece
Dr. Mohd Tahir Ismail, Universiti Sains Malayasia, Malayasia
Dr. Hai-Chin YU, Chung Yuan University, Taiwan
Prof. Jos Rigoberto Parada Daza, Universidad de Concepcin, Chile
Dr. Ma Carmen Lozano, Universidad Politcnica de Cartagena, Spain
Dr. Federico Fuentes, Universidad Politcnica de Cartagena, Spain
Dr. Aristeidis Samitas, University of Aegean, Greece
Dr. Mostafa Kamal Hassan, University of Sharjah, College of Business, UAE
Prof. Muhammad Abdus Salam, State Bank of Pakistan, Pakistan

FDI Nigeria (1960-2010)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FDI Nigeria (1960-2010)

Uploaded by

Copyright:

Available Formats

ISSN 2229-6891

International Research Journal of

database, using the Standard Industrial

= the yen-denominated conversion factor;

method is applied to low coupon bonds, the Eurex

is the expected return on the capital asset; R

is the risk-free rate of interest often

is the sensitivity of the

is the single asset return, R

is the return of the market index in which the asset is

is the beta coefficient which indicates the sensitivity of R

is the error component (a random variable, with mean zero and

, which assumes independently and identically distributed values). The

can be interpreted as representing the relationship between the

s are estimated in order to measures market reactions to events in

s to shift. The literature on estimating using alternative

under different estimation methods. This is the first paper to our

are the key

are derived from historical market information;

reflects only the most recent three months market

to update continuously based on the

. As mentioned in the prior section,

is derived from the market model; however, rather than

is generated based on forward-

, the forecast errors

differ across companies. Closer

is not superior than

in reducing the mean forecast errors of by company counts.

, where j OLS, KAL be the loss function

. According to Diebold and Mariano (1995), results of Monte

and benchmark , we are confident that Diebold and Marianos

, where j OLS, KAL.

in reducing the forecast errors. A significant p-value with a negative t-

has superior forecastibility than

be the leveraged and unleveraged

be the corporate tax rate, the following equation provides the

), the autoregressive coefficients (

You might also like