Does Shari’ah Screening Cause Abnormal

Returns? Empirical Evidence from Islamic

Equity Indices

Dawood Ashraf

Journal of Business Ethics

ISSN 0167-4544
Volume 134
Number 2

J Bus Ethics (2016) 134:209-228

DOI 10.1007/s10551-014-2422-2

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at [Link]”.

1 23
 Author's personal copy
J Bus Ethics (2016) 134:209–228
DOI 10.1007/s10551-014-2422-2

Does Shari’ah Screening Cause Abnormal Returns?

Empirical Evidence from Islamic Equity Indices
Dawood Ashraf

Received: 8 December 2013 / Accepted: 8 October 2014 / Published online: 17 October 2014
Ó Springer Science+Business Media Dordrecht 2014

Abstract Islamic equity funds are subject to the screen- Introduction

ing criteria for stock selection imposed by the principles of
Islamic jurisprudence (Shari’ah). Equities must pass three Investment based on ethical beliefs is the process of
basic screens: revenue source, business activity, and selecting and/or managing an investment portfolio with the
financial factors to be included in an Islamic fund. How- aim to maximize the return for investors by only investing
ever, screening criteria are not universal especially for the in securities which adhere to the religious and/or ethical
financial factors. One can use financial ratios based on beliefs of the investors. Investment based on religious
either the book-value of total assets or the market-value of beliefs is not a rarity in global financial markets. However,
equity for screening of stocks. This may not only result in a investment based on Islamic principles is a special case
different portfolio composition but also entail diverse re- since it enables a large population of faith-adhering Mus-
balancing and monitoring costs. The performance of 29 lims to invest in capital markets.
Islamic equity indices (IEIs) versus conventional indices Investment companies, both in Islamic as well as in non-
from four major international index providers using dif- Islamic countries, provide investors with the opportunity to
ferent Shari’ah screening criteria are analyzed in a single invest in Shari’ah compliant mutual funds and exchanged
as well as in a multi-equation framework. The use of a traded funds (ETF). Ernst and Young (2011) estimates that
multi-equation framework has the added advantage of there are currently more than 700 mutual funds worldwide,
utilizing the information content of different screening investing in Shari’ah compliant securities. Globally, assets
criteria adopted by different index providers. The empirical under management (AUM) of Islamic mutual funds rep-
findings suggest that the difference in screening criteria resent less than 1 % of the USD 25.92 trillion global
does not significantly affect the performance of IEIs. mutual funds market. The aggregate AUM of Islamic
Returns deviation, if any, stems from the relative riskiness mutual funds has grown from USD 3.3 billion in 2003 to
of the IEI as compared with the relevant benchmark. Work USD 58 billion in 2010 (Ernst and Young 2011).
needs to done to streamline the quantitative screening Performance measurement is one of the major chal-
criteria to avoid confusion among the investing public. lenges for investors in Islamic capital markets. Investors
require Shari’ah compliant equity benchmark indices to
Keywords Islamic investments  Equity indices  compare returns performance of Islamic equity fund man-
Shari’ah screening criteria  Performance measurement agers that matches closely with the objective of the funds.
To cater for this need, all the major index providers:
JEL Classification G11  G15  G24  O16 Financial Times Stock Exchange (FTSE), Standard and
Poor’s (S&P), Dow Jones (DJ) and Morgan Stanley Capital
International (MSCI), now build and provide Islamic
D. Ashraf (&) equity indices (IEIs) data based on independent Shari’ah
Islamic Research & Training Institute (Member of Islamic
screening criteria at global, regional and country level.
Development Bank Group), PO Box 9201, Jeddah 21413,
Kingdom of Saudi Arabia However, each of these index providers follows a slightly
e-mail: dashraf@[Link] different screening criterion for inclusion of equities in IEIs

123
 Author's personal copy
210 D. Ashraf

based on the interpretation of the Shari’ah board’s Another major contribution of this study is the use of a
deliberation. multi-equation framework for the capital asset pricing
Each criterion comprises at least three basic screens: model (CAPM) rather than a single-equation framework
business activity, sources of revenue, and financial posi- usually applied in such analysis. The single-equation
tion. The first two screens are qualitative in nature and are framework is less efficient due to the existence of con-
generally the same for all indices. However, major differ- temporaneous cross equation correlation of error terms that
ences lie in the screening criteria based on financial factors. may arise due to the application of differences in Shari’ah
Under the financial screening criteria, investors are allowed screening practices, timing of rebalancing of IEIs and the
to invest in stocks of those companies that use limited debt, duration of investment horizons. The multi-equation
have fewer accounts receivable, and hold lower amounts of framework based on ‘‘Seemingly Unrelated Regression’’
cash and interest bearing securities as compared with the (SUR) model takes into account the contemporaneous
book value of total assets or 12–36 months trailing market correlations and allows for differences in the relative per-
value of equity. In the case of FTSE and MSCI IEIs, these formance of IEIs from different index providers. The
ratios are based on the book value of total assets while in coefficient estimates and quality of tests performed for
the case of DJ and S&P IEIs, the base value is the last 24 or empirical analysis is superior in the case of a multi-equation
36 months trailing market value of equity, respectively.1 framework as compared with a single-equation framework.
Varying Shari’ah screening standards not only leads to a The empirical findings suggest that IEIs provide similar
different portfolio composition (Derigs and Marzban 2008; risk adjusted returns as their conventional benchmark
Rahman et al. 2010) but also different return performance returns over the sample period from December 2000 to
(Derigs and Marzban 2009; Adamsson et al. 2014). May 2012. However, there is an evidence of performance
Outside the US, there is little evidence that varying deviation among IEIs based on different Shari’ah screening
Shari’ah standards have an impact on portfolio perfor- criteria adopted by four major index providers. IEIs using
mance. Even for the US, the published research does not the book-value of total assets as a divisor for calculating
cover a period beyond 2007 (Derigs and Marzban 2009). financial ratios performed, on a nominal basis, better than
There is a need for a study that investigates the impact of IEIs using market value of equity as a divisor. Two of the
different Shari’ah screening criteria on the return perfor- IEIs based on DJ Shari’ah screening criteria report the
mance of Islamic equity investments for wider geographic source of deviation in the form of a statistically significant
regions and for longer sample periods, especially the period positive abnormal return among IEIs holding small and
of the recent financial crisis.2 To address this question, mid-cap equities from the US. However, those IEIs
equity index level monthly performance data for 58 indices reporting positive abnormal returns also exhibit a steeper
(29 IEIs and 29 conventional indices) has been collected decline in returns during the peak of the financial crisis. On
and matched from four major equity index providers: the other hand, IEIs following any other Shari’ah screening
MSCI, S&P, FTSE and DJ from December 2000 to May criteria attributes return deviation to the relative riskiness
2012. Analyzing indices rather than investment funds has of IEIs with the benchmark index.
clear advantages; empirical results will not be biased by The findings of this paper are of interest to policy
transaction costs nor the specific characteristics of a given makers, fund managers and the general investing public.
investment fund compared to another (Ashraf and Mo- Findings of this study imply that the relative performance
hammad 2014; Schroder 2007). These broader criteria help of Islamic equity investments should be at least as good as
to reduce the selection bias in the interpretation and gen- conventional investments in the long term. There is no
eralization of results. evidence of a general over-performance associated with
The empirical evidence from this study is more robust specific Shari’ah screening criteria.
than previous studies due to the matching of IEIs with the Work needs to be done in streamlining Shari’ah
benchmark from where IEIs usually draw its equities. This screening criteria to avoid confusion among the investing
helps in reducing the tracking error as both IEIs and con- public. Shari’ah scholars setting standards for equity
ventional indices not only share the same universe of assets screening may come with unified standards. Both policy
but also use a similar index construction methodology. makers and fund managers can benefit by adhering to a
unified screening criterion without any significant loss to
1
There is no stipulated provision in Shari’ah that allows investment return performance. This will help to attract more faith-
in equities with financial leverage of 33 % or lower. It is mostly a adhering investors to invest in capital markets.
decision by the Shari’ah board which allows for a certain level of
The remainder of the paper is organized as follows.
financial leverage to be either based on total assets or market value of
equity. ‘‘IEIs—A Brief Review’’ section provides a review of IEIs,
2
In a recent unpublished study, Adamsson et al. (2014) compared the relevant literature is discussed in ‘‘Literature Review’’
performance of IEIs published by different index providers. section, empirical estimation methodology is presented in

123
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 211

‘‘Islamic Stock Index Evaluation Methodology’’ section, principles or may be more relaxed, where investment is
followed by data sources and descriptive statistics in ‘‘Data allowed in companies under a specific threshold of
Sources and Descriptive Statistics’’ section and empirical impermissible activities. These different screening criteria
estimations are discussed in ‘‘Empirical Results’’ section. may lead to confusion among the investing community.
‘‘Summary and Conclusion’’ section concludes the paper This also raises an important question, does the existence
with a brief summary of the key findings of this study. of different screening criteria cause different return per-
formance and thus may induce fund managers to select a
benchmark index that follows slightly relaxed Shari’ah
IEIs—A Brief Review screening criteria? The general screening criteria for the
inclusion of any security in an IEI depend on two levels of
An investor in the Islamic equity market invests in equities screenings: qualitative and quantitative screenings. Quali-
of Shari’ah compliant companies3 or in a publicly offered tative screenings exclude shares of all such companies
portfolio consisting of these equities offered through unit engaged in activities strictly prohibited (haram) in Islam.
trusts, mutual funds, or ETFs. One of the major challenges These include companies whose major line of business is
for investors in the Islamic equity market, especially in unit dealing with financial transactions involving interest
trusts and mutual funds, is the performance measurement (reba), gambling activities (meiser), intoxicants (khumar)
of managers. A benchmark consisting of all equities may such as alcohol or similar drugs that can obscure one’s
not be appropriate to assess the performance of an Islamic judgment, pork and/or excessive risk taking (gharar) such
mutual fund manager. To address this shortcoming, the first as insurance and speculative investments. Quantitative
Islamic equity index was launched in 1999 by DJ followed screens are used to further screen companies with core
by FTSE in 2000. S&P and MSCI followed suit and activities as permissible (halal) under Shari’ah but have a
developed and launched indices based on Shari’ah princi- portion of revenue from non-permissible activities such as
ples in 2006 and 2007, respectively (Shanmugam and Za- borrowing or lending money on interest (reba) and/or has a
hari 2009). major proportion of assets in liquid form. These screens are
The determination of Shari’ah compliance rests with the based on arbitrary ratios and are quite controversial within
judgment of Shari’ah boards consisting of Islamic scholars. the Muslim community.7
Each board sets different rules to screen equities to ensure Exhibit 1 provides the screening criteria used by the four
the Shari’ah compliance. One of the main reasons for the major index providers MSCI, S&P, FTSE, and DJ. It is
existence of different Shari’ah screening standards is the evident from Exhibit 1 that each index provider uses a
absence of such guidelines from the main sources of Isla- slightly different criterion. Qualitative screening criteria
mic jurisprudence i.e., the Quran,4 the Hadith5 and the exhibit few differences, for example, S&P Shari’ah indices
Ijmaa.6 Shariah boards derive Shari’ah screening princi- specifically excludes businesses engaged in cloning, and
ples by using the principles of analogy (Qeyas) where the trading of precious commodities i.e., gold and silver as
Shari’ah scholars use the past rulings of a similar nature to cash on a deferred basis, while the other three are silent on
derive new rulings. Since Islamic investment operates these issues. There are some methodological differences as
within the existing capital market system and are required well such as the cleansing ratio for purification of revenues.
to adjust to the conventional norms, this may provide Three of the index providers: S&P and FTSE provide the
another explanation for the existence of varying Shari’ah percentage of income to be purified to the investor while in
standards. the case of MSCI; it applies a dividend adjustment factor in
There may be very conservative screening criteria, their return calculation.
where equity has to comply with more stringent Shari’ah On the other hand, in terms of quantitative screening
criteria, there are at least two major differences. One is the
3
choice of divisor for financial ratio calculations while the
Equities of Shari’ah compliant companies include companies
other is the tolerance level for non-permissible activities.
whose major source of revenue comes from permissible (halal)
activities. All companies are excluded who are predominantly DJ and S&P use the trailing market value of equity as a
engaged in any of the following non-permissible (haram) activities: divisor while MSCI and FTSE use the book-value of total
trading of alcohol, pork, pornography, gambling or from profit assets as a divisor to calculate the financial ratios. The
associated with the charging of interest on loans.
4
appeal for using the market capitalization for computation
The Muslim Holy book regarded by Muslims as the immutable and
of financial ratios is its capability for continuous
final revelation of Allah (Shanmugam and Zahari 2009).
5
Authentic saying and reported actions of Prophet Mohamed (peace
be upon him). 7
For a detailed discussion on the quantitative and qualitative screens
6
A general consensus of opinion by the majority of the Muslim for stock screening please see Obaidullah (2005), Derigs and Marzban
scholars. (2008), and Rahman et al. (2010).

123
 Author's personal copy
212 D. Ashraf

Exhibit 1 Shari’ah based screening criteria adopted by major index providers: MSCI, S&P, FTSE and DJ
Screen MSCI Islamic index series S&P Shari’ah indices FTSE Islamic index DJ Islamic indices

Revenue Directly or indirectly securing Non-permissible income Total interest and non- Income from impure
source income from interest bearing or other than interest income compliant activities income sources should not
prohibited activities \5 % of of the total revenue \5 % should not exceed 5 % of exceed 5 %
total income total revenue
Business Alcohol and tobacco including Same as MSCI Same as MSCI Same as MSCI
activity production and sales thereof
Pork related products: both Same as MSCI Same as MSCI Same as MSCI
manufacturer and retailers of
pork products
Financial service providers Same as MSCI Same as MSCI Same as MSCI
excluding Islamic banks and
Islamic insurance companies and
investment companies
developing financial products
based on Islamic principles
Entertainment: gambling, casinos. Advertising and media, Gaming and arms Entertainment, hotels,
Music, hotels, cinemas and adult gambling, pornography, manufacturing casinos/gambling,
entertainment cloning and trading of cinema, pornography,
gold and silver as cash on music etc
a deferred basis
Financial Total debt/total assets \33.33 % Debt/market value of equity Debt/total assets \33.333 % Debt/market value of
screening (36 months average) equity (trailing
\33 % 24 months average)
\33 %
Sum of a company’s cash and Sum of a company’s cash Cash and interest bearing Company’s cash plus
interest-bearing securities/total and interest-bearing items/total assets \33.333 % interest-bearing
assets \33.33 % securities/market value of securities/market value
equity (36 months of equity (trailing
average) \33 % 24 months average)
\33 %
Sum of a company’s accounts Sum of a company’s Accounts receivable and cash/ Sum of a company’s
receivables and cash/total assets accounts receivables/ total assets \50 % accounts receivables/
\33.33 % market value of equity market value of equity
(36 months average) (trailing 24 months
\49 % average) \33 %
Dividend MSCI apply a ‘‘dividend Dividend 9 (Non- FTSE does not purify the No such information is
purification adjustment factor’’ to all permissible revenue/total income rather it provides the available
ratio reinvested dividends. The revenue) - supplied to 5 % purification of dividends
‘‘dividend adjustment factor’’ is investors for charity to investors for purification
defined as: (total earnings - purposes at their end
(income from prohibited
activities ? interest income))/
total earnings
Source Author’s own compilation from documentation available on the websites of each index provider

monitoring of Shari’ah compliance status and indifference corporations are going-concerns and also it represents the
in the use of accounting principles (Derigs and Marzban replacement cost of assets.
2008). Obaidullah (2005) suggests that Islamic equity Figure 1 shows the impact of the financial leverage ratio
portfolios based on market capitalization criteria are on the likelihood of six randomly selected stocks, to be
deemed to be lagging behind conventional portfolios due to Shari’ah compliant using screening criteria as proposed by
the continuous trading of stocks—buying when the market all four major index providers from January 2002 to March
is rising and selling when the market is falling to meet the 2013. Each of these stocks complies with all other
Shari’ah compliance requirements. He further suggests that screening criteria as of March 2013. It is apparent from
the book-value of total assets is a better alternative since Fig. 1 that a stock may not comply with all the financial

123
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 213

.8 Canadian Natural Resources Intertek Group Plc Quanta Services Inc

.8

.8
.6

.6

.6
.4

.4

.4
.2

.2

.2
0

0
m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1
0 2 00 4 00 6 00 8 01 0 01 2 01 4 0 2 00 4 00 6 00 8 01 0 01 2 01 4 0 2 00 4 00 6 00 8 01 0 01 2 01 4
20 2 2 2 2 2 2 20 2 2 2 2 2 2 20 2 2 2 2 2 2

Emaar Properties Royal Dutch Shell Saudi Basic Industries

.8

.8

.8
.6

.6

.6
.4

.4

.4
.2

.2

.2
0

0
m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1 m1
0 2 00 4 00 6 00 8 01 0 01 2 01 4 0 2 00 4 00 6 00 8 01 0 01 2 01 4 0 2 00 4 00 6 00 8 01 0 01 2 01 4
20 2 2 2 2 2 2 20 2 2 2 2 2 2 20 2 2 2 2 2 2

Threshhold MSCI & FTSE S&P Dow Jones

Fig. 1 Financial leverage ratio comparison based on varying Sha- allow investments in stocks with the debt to total asset ratio of 33 %
ri’ah screening criteria. This figure presents the financial leverage or less. DJ and S&P allow investments in stocks of companies with
ratio of selected Shari’ah compliant stocks based on the definition of the ratio of debt to 24 and 36 months of trailing market value of
leverage provided by four major index providers. MSCI and FTSE equity, respectively of 33 % or less

leverage criteria of each of the index providers. Three These apparent differences in the Shari’ah screening
stocks: Canadian National Resources, Quanta Services criteria and their impact on portfolio composition and
Inc., and Saudi Basic Industries comply with the criteria of ultimately on their return performance require further
all index providers at the end of March 2013. From the investigation. The next section provides a review of liter-
remaining three stocks, Emaar Properties and Royal Dutch ature on the performance of Islamic equity investments.
Shell comply with MSCI and FTSE criteria while Intertek
Group Plc complies with S&P and DJ criteria. It is further
evident that market-value-based Shari’ah screens are more Literature Review
likely and more often to screen out equities over time as
compared with total-assets-based Shari’ah screens. The performance of Islamic equity funds has received
In terms of the tolerance level of financial ratios, considerable attention over the past decade. However, the
thresholds are determined based on qeyas. Obaidullah empirical findings on the comparative performance of
(2005) criticizes that historical rulings/events used as the Islamic equity investment funds (IEIFs) based on Shari’ah
precedence for screening of stocks differs considerably and screening principles are mixed. It is documented that IEIFs
may be out of the context.8 There are some differences due to their adherence to religious faith perform better than
among tolerance levels such as FTSE screening criterion conventional funds during an economic downturn (Alam
allows higher thresholds for quick assets up to 50 % while and Rajjaque 2010; Abdullah et al. 2007; Ashraf 2013).
all others allow only to maximum of 33 %. This resilience in performance is attributed to the selection
of securities in funds based on the screening criteria
8
Usmani (2002) and Obaidullah (2005) provide a detailed discussion applied (Ashraf 2013; Gangi and Trotta 2013). However,
on the rationale for financial screening criteria. there is evidence that contradicts these findings whereby it

123
 Author's personal copy
214 D. Ashraf

is found that IEIFs lag behind benchmark indices irre- varying Shari’ah standards on the portfolio composition of
spective of market conditions (Hayat and Kraeussl 2011; Islamic portfolios drawn from the constituent list of S&P
Nainggolan 2011). Elfakhani et al. (2007) suggest that the 500. They found that the application of a different Shari’ah
outperformance of IEIFs depends on the measure and screening criterion to the same set of equities may lead to
benchmark used for the performance evaluation. It also significantly different constituents in a portfolio. Rahman
depends on the time period chosen for performance et al. (2010) compared the screening criteria for level of
evaluation. debt and level of liquidity between the Kuala Lumpur
Due to the existence of trading costs, managements’ Stock Exchange Islamic Index (KLSESI) and the Dow
stock selection and market timing skills, the impact of Jones Islamic Market (DJIM) Index. Their findings suggest
Shari’ah screens on the performance of IEIFs is unclear. that KLSESI does not use both criteria as a measure during
Furthermore, the empirical literature on the performance of the screening process and if both criteria are applied con-
IEIFs often applies Jensen’s alpha, the intercept of CAPM currently, only 198 out of 565 Shari’ah compliant equities
regression to measure the abnormal performance of a fund. listed on KLSESI would pass the Shari’ah screening cri-
Fama and French (2004) suggest that funds focusing on teria. In terms of impact on the return performance of
small, low beta or value stocks tend to produce positive Islamic equity investment, Derigs and Marzban (2009)
significant abnormal returns relative to their benchmarks compared the return performance of Shari’ah compliant
without any special talent from fund managers. Since IEIFs portfolios drawn from S&P 500 based on six different
focus on lower leverage stocks that may be small stocks standards and found that Shari’ah screening standards
and/or low beta stocks, the outperformance results are applying a market capitalization for financial screening
questionable. This implies that IEIFs are not necessarily a perform better than those that apply book value of total
better tool to assess the performance of Islamic equity assets. However, empirical evidence on performance
investments both in terms of assessing management skills deviation based on varying Shari’ah screening criteria is
or the impact of Shari’ah screens. Schroder (2007) sug- missing for wider geographic regions and longer time
gests that the use of a screened index rather than a screened periods especially the period of the recent global financial
fund for performance comparison isolate the impact of crisis. This study attempts to fill this gap by providing
screening on the performance from that of trading costs and empirical evidence on the return performance deviation of
management skills. IEIs following different Shari’ah screening criteria by
Most of the empirical literature on the performance of using longer time series data with robust methodology. The
IEIs suggests that IEIs perform at least as good as con- next section presents the empirical estimation methodology
ventional indices without any loss of diversification (Guyot used for the evaluation of IEIs’ performance.
2011; Hakim and Rashidian 2004; Kok et al. 2009; Girard
and Hassan 2008; Albaity and Ahmad 2008; Walkshäusl
and Lobe 2012). Hassan and Girard (2011) found that the Islamic Stock Index Evaluation Methodology
performance of IEIs is time period sensitive while when
comparing the performance of DJIM indices versus con- The basic model to measure the relative performance of
ventional indices. Others found that the IEIs underperform IEIs is within the context of the standard CAPM and has
compared to their conventional counterpart when the cap- the following specifications:
ital markets were generally falling (Hussein and Omran Rit ¼ ai þ bij Rjt þ eit for t ¼ 1; 2; . . .; T; i ¼ 1; 2; . . .; I
2005; Hussein 2004). However Ashraf and Mohammad
(2014) and Adamsson et al. (2014), while applying more and j ¼ 1; 2; . . .; J: ð1Þ
robust empirical methodology on recent market data, found In Eq. (1), an investor selects a Shari’ah compliant port-
that relatively better performance of IEIs especially during folio (IEI) at time, t - 1 that produces a return, Rit at time
the falling capital markets stem from the lower levels of t. Where Rjt is the similar return at time = t of benchmark
relative riskiness of IEIs with their benchmarks. j from where Shari’ah compliant portfolio draws its equi-
In most of the studies discussed above, authors have ties. Both Rit and Rjt are in excess of risk free rate of
focused on single Shari’ah screening criteria by analyzing return.10 The constant term ai is Jensen’s alpha and mea-
the performance of IEIs on relatively older time series sures the performance of IEI once the return has been
data.9 Derigs and Marzban (2008) analyzed the impact of
9 10
See for example Hassan and Girard (2011) Dow Jones Islamic One may argue that Muslim investors are prohibited from
Market Indices (DJIM) covering a time period from 1996 to 2005, investing or borrowing on interest and hence would not borrow or
Guyot (2011) DJIM from 1999 to 2008, Girard and Hassan (2008) lend at a risk free rate. The author agrees with the preposition
FTSE from 1998 to 2006, Hussain (2006) FTSE and Dow Jones from however, the intent of the paper is to address a wider audience
1996 to 2004, Hussein (2004) from 1996 to 2003. therefore a return excess over the risk free rate is used in this study.

123
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 215

adjusted to benchmark j. bij ¼ covðRit ; Rjt Þ=varðRjt Þ is non- IEIs from different index providers. The SUR model esti-
diversifiable risk and can be used to measure the relative mates the parameters of all equations simultaneously, so
riskiness of IEI, i as compared with benchmark j. An IEI that the parameters of each single equation also takes the
with bij [ 1 indicates the relative higher riskiness of IEI, information provided by the other equations into account.
i as compared with benchmark j. While bij \ 1 or bij = 1 This results in greater efficiency of the parameter estimates
indicates that an IEI has relatively less risk or the same risk because additional information is used to describe the
as compared with their benchmarks, respectively. system (Hodgson et al. 2000; Schroder 2007). Equation (1)
The null hypothesis is that there is no difference in the can be rewritten as a system of equations:
return performance of IEIs and their equity benchmarks. If R1t ¼ a1 þ b1j R1t þ e1t ; . . .; Rit ¼ ai þ bij Rjt þ eit
IEIs do not perform any differently than their respective for t ¼ 1; 2; . . .; T; i ¼ 1; 2; 3. . .; I and j ¼ 1; 2; 3. . .; J:
benchmarks then ai should not be statistically different than
zero and bij should be equal to 1. Schroder (2007) suggests ð2Þ
testing coefficients individually and jointly under null Under the SUR model, dependent variables (Rit) are
hypotheses H01: ai = 0, H02: bij = 1 and H03: (ai = 0 and allowed to have different sets of explanatory variables (Rjt).
bij = 1) by using the spanning test based on Huberman and Capital markets witnessed a severe downfall in late 2008
Kandel (1987). These tests of coefficients are performed to and early 2009. To capture the impact of this market crash,
assess whether IEIs can be replicated by their benchmarks. a dichotomous variable, CRSHt is used with all empirical
If null hypotheses are not rejected, it would imply that the estimations that takes the value of one for the period
screens used by IEIs do not affect the performance and August 2008–March 2009 and zero otherwise. The next
produce similar risk-return payoffs as the benchmark section presents the data sources and descriptive statistics
index. followed by estimation results of the single and multi-
equation framework.

Risk-Return Payoff of the Index Providers Using

the SUR Model Data Sources and Descriptive Statistics

A common feature of published work concerning index This study investigates the risk adjusted performance of 29
performance is the use of single-equation CAPM models. IEIs following different Shari’ah screening standards,
In this approach, either one equation is used to assess the constructed and published by four different index service
performance of all indices or several independent equations providers. These providers are MSCI, DJ, FTSE and S&P.
are estimated separately by using ordinary least squares To assess the performance of IEIs, the benchmark indices
(OLS) and the estimated parameters are used to predict the for each IEI has been selected that best reflects the
under or over performance. However, the performance of investment universe of that IEI and would be a preferred
an individual index can be correlated with other indices due benchmark for any investor who is not following the
to the economic environment, selection of securities and/or Shari’ah screening of equities. The sample consists of end-
factors not specifically incorporated in the model and share of-the month index level data from the base date in US
a common error structure with non-zero covariance, often dollars for all IEIs and their conventional benchmark with
referred to as contemporaneous correlations. Hence, OLS the launch date of September 30, 2007 or earlier11 from
estimates from Eq. (1) will be inefficient from a statistical December 2000 to May 2012 (inclusive).12 The launch date
point of view since it ignores information about the con- is chosen to capture the impact of extreme shortfalls in
temporaneous cross equation correlation of error terms security values at the peak of the global financial crisis in
(Judge et al. 1985). The contemporaneous error in case of September 2008. Data for DJ, FTSE and S&P is obtained
IEIs may arise due to the application of heterogeneous
Shari’ah screening process adopted by index providers
11
where IEIs use different Shari’ah screening standards, Market data for IEIs is available from the base date. The time
period between the base date and effective launch date is used for
different timing and frequency of rebalancing. Other pos-
back testing. Since index providers structure IEIs based on the same
sible reasons may include the possible investment universe methodology and Shari’ah screening criteria, there is no sample
coverage problem where index providers may choose a selection bias or survivorship bias during this time. Hence using
different investment universe and investment horizon for market data from the base date provide further depth to the analysis.
12
the performance evaluation. To capture the impact for Shari’ah screening, monthly data is
preferred due to the review frequency of IEIs’ constituents for
Zellner’s (1962) ‘‘Seemingly Unrelated Regression’’
compliance status. Among the index providers, only S&P review
model takes into account the contemporaneous correlations constituent list monthly while all other index providers review
and allows for differences in the relative performance of constituent lists quarterly.

123
 Author's personal copy
216 D. Ashraf

from DataStream, while performance data for MSCI indi- In terms of Shari’ah screening criteria based return; it is
ces is downloaded from the MSCI Index Service website evident that indices that use trailing market value of equity
([Link] For a further (DJ and S&P) as a divisor for financial ratio calculation
robustness check, data for the additional factors is down- perform worse than that of IEIs using total assets as a
loaded from the website of Professor French. Since most of divisor. In the case of DJ Islamic market indices, 9 out of
the IEIs in the present study are international, using addi- 10 IEIs lagged their benchmark indices, while MSCI and
tional factors from global portfolios is more appropriate as FTSE indices perform considerably better than their
compared with country specific additional factors such as benchmarks. This clear divide in the performance suggests
the US stock market based portfolios. Additional factor further investigation.
data from global portfolios is downloaded as proposed in
Fama and French (2012).13
The IEIs cover different global, regional and country Empirical Results
specific indices. Of the 29 indices in the sample, seven
indices provide global coverage, three indices provide The empirical results of the CAPM and hypotheses test
equity performance of European stocks, seven indices [H01: ai = 0, H02: bij = 1 and H03: (ai = 0 and bij = 1)]
cover the Asian markets (of which four focused on Asia for each IEI with reference to the selected benchmark
Pacific), and 11 indices contain stocks of single countries: developed in Eq. (2) is estimated using the Newey and
USA, Japan, Canada, and Malaysia. To assess the perfor- West (1987) method that correct the variance and
mance of each index, excess return has been calculated by covariance matrix of residuals for both autocorrelation
subtracting the 3-month US T-bills rate from the continu- and heteroskedasticity. Table 2 report the estimation
ously compounding returns of each index at the end of results in Panels A to C for individual IEIs from MSCI,
month. DJ, and FTSE, S&P, respectively. It is evident from these
Table 1 provides detailed information about the indices results that there is no IEI that reports abnormal returns in
and some descriptive statistics i.e., mean excess returns for the form of Jensen’s alpha, except for MSCI-Europe, DJ-
IEIs and their respective benchmarks along with the Sharpe US Large, DJ-Canada and DJ-World. Jensen’s alpha is
ratio. The Sharpe ratio is commonly used to rank portfolios positive and slightly significant for MSCI-Europe but
in terms of their risk-return tradeoff. Mathematically, it can negative and significant for DJ-US Large, DJ-Canada and
be written as: DJ-World suggesting that IEIs from these markets fol-
Ri  R f lowing DJ Shari’ah screening guidelines pickup equities
Sharpe ratio ¼ ; ð3Þ that cause the IEIs to lag their benchmark. While in the
ri
case of systematic risk, all IEIs reported significant bij
where (Ri - Rf) is the annualized excess return of index coefficient at a 1 % level suggesting that IEIs’ returns can
i over risk free rate (3-month T-bill rates) and ri is the be explained relative to their benchmark indices. For ten
standard deviation of the return of the index. out of 29 IEIs, bij coefficient is greater than unity sug-
The comparison of mean excess returns and risk adjus- gesting these IEIs are carrying a systematic risk greater
ted performance of IEIs with conventional benchmarks than the benchmark risk. For the remaining IEIs bij
shows that in 18 out of 29 indices, IEIs perform better than coefficient is approximately equal to one except for two
their conventional benchmarks on an annual return basis as indices—FTSE-Developed and S&P-Japan, where bij
well as on a risk adjusted basis through the Sharpe ratio. coefficient is significantly below unity indicating a lower
Among the notables, FTSE-Bursa has produced an excess systematic risk as compared with the benchmark index.
return of 4.07 % relative to the benchmark index on an These results indicate the appropriateness of the bench-
annual basis followed by S&P-Japan and FTSE-Mult with mark used for empirical analysis.
excess annual returns of 3.11 and 2.45 %, respectively. The coefficient for the variable of capital markets
Among the worst performing IEIs are DJ-World, DJ-Can- crash, CRSHt is positive and significant for those IEIs
ada and DJ-US Large trailing behind their benchmarks by investing in equities from the global markets (DJ-World)
4.48, 3.67 and 1.86 % annually, respectively. and developed Western markets (MSCI-Americas, DJ-US
Large, DJ-Canada, and DJ-US). This suggests that IEIs
13
The global portfolio and additional factors consist of stock returns provided a hedging opportunity to the investors of these
from 23 countries: Australia, Austria, Belgium, Canada, Denmark, markets especially, IEIs following DJ Shari’ah guide-
Finland, France, Germany, Greece, Hong Kong, Ireland, Italy, Japan, lines. This finding is in line with Ashraf (2013) and
Netherlands, New Zealand, Norway, Portugal, Singapore, Spain,
Hoepner et al. (2011), who both found a positive return
Switzerland, Sweden, United Kingdom, and the United States. These
factors can be downloaded from: [Link] associated with Islamic investments during bearish mar-
pages/faculty/[Link]/Data_Library/int_index_port_formed.html. ket trends.

123
 Table 1 Performance characteristics of IEIs versus conventional benchmark indices
ID Pneumonic Name of Islamic index Benchmark for Islamic index Base Launch Mean Mean SD SD Sharpe Sharpe
IEI BM IEI BM IEI BM

1 MSCI-America MSCI AC Americas Islamic Standard AC Americas Standard 05/2002 07/2007 2.02 1.94 19.77 21.24 0.10 0.09
2 MSCI-Asia MSCI AC Asia Islamic Standard AC Asia Standard 05/2002 07/2007 2.31 2.29 26.54 24.86 0.09 0.09
3 MSCI-AP MSCI AC Asia Pacific Islamic Standard AC Asia Pacific Standard 05/2002 07/2007 3.32 3.03 27.14 25.63 0.12 0.12
4 MECI-Europe MSCI AC Europe Islamic Standard AC Europe Standard 05/2002 07/2007 2.52 0.38 24.49 26.63 0.10 0.01
5 MSCI-Far-East MSCI AC Far East Islamic Standard AC Far East Standard 05/2002 07/2007 1.93 1.86 25.83 24.15 0.07 0.08
6 MSCI-Pacific MSCI AC Pacific Islamic Standard AC Pacific Standard 05/2002 07/2007 2.99 2.66 26.53 25.01 0.11 0.10
7 MSCI-World MSCI ACWI Islamic Standard ACWI Standard 05/2002 07/2007 1.98 1.41 21.40 22.35 0.09 0.06
8 MSCI-GD MSCI Golden Dragon Islamic Standard Golden Dragon Standard 05/2002 07/2007 7.72 6.77 33.45 33.64 0.23 0.20
Does Shari’ah Screening Cause Abnormal Returns?

9 MSCI-ZH MSCI Zhong Hua Islamic Standard Zhong Hua Standard 05/2002 07/2007 9.29 9.55 34.20 35.12 0.26 0.26
10 DJ-AP DJ Islamic Asia/Pacific Index DJTM Asia/Pacific $ Index 12/1995 05/1999 0.32 0.99 28.40 25.66 0.01 0.04
11 DJ-US large DJ Islamic US Large Cap Index DJ US Large Cap Total Stock Mkt Index 12/1995 05/1999 -0.60 1.26 21.26 22.37 -0.03 0.06
12 DJ-US mid DJ Islamic US Mid Cap Index DJ US Mid Cap Total Stock Mkt Index 12/1995 05/1999 6.87 6.96 27.88 26.22 0.25 0.27
13 DJ-US small DJ Islamic US Small Cap Index DJ US Small Cap Total Stock Mkt Index 12/1995 05/1999 6.67 6.81 25.40 26.99 0.27 0.26
14 DJ-Europe DJ Islamic Europe Index DJIM Europe Index 12/1995 05/1999 6.10 6.13 28.78 28.67 0.21 0.21
15 DJ-Euro DJ Islamic Euro Index DJTM Euro $ Index 12/1995 05/1999 0.09 -0.27 27.51 28.95 0.00 -0.01
16 DJ-Canada DJ Islamic Canadian Index DJTM Canada $ Index 12/1995 05/1999 4.58 8.25 37.24 29.94 0.12 0.28
17 DJ-World DJ Islamic Market World Index DJ Global Index 12/1995 05/1999 1.66 6.13 21.90 25.17 0.08 0.24
18 DJ-Japan DJ Islamic Japan Index DJTM Japan $ Index 12/1995 05/1999 -4.60 -3.30 21.58 21.60 -0.22 -0.15
19 DJ-US DJ Islamic US Index DJ US Total Stock Market Index 12/1995 05/1999 0.77 1.94 22.05 22.96 0.04 0.09
Author's personal copy

20 FTSE-Malaysia FTSE Bursa Malaysia Shariah $ Index FTSE Bursa Malaysia EMAS $ Index 12/1995 02/2007 20.56 16.48 40.72 36.90 0.49 0.44
21 FTSE-World FTSE Shariah All World Index FTSE All World $ Index 12/1986 09/2003 -1.42 -2.87 32.59 35.92 -0.04 -0.08
22 FTSE-Developed FTSE Shariah Developed Index FTSE GWA Developed $ Index 12/1986 09/2003 -1.17 -1.09 29.97 53.91 -0.04 -0.02
23 FTSE-Japan FTSE Shariah Japan 100 $ Index FTSE W Japan $ :L Index 12/1986 09/2003 -4.51 -5.44 23.88 21.45 -0.19 -0.26
24 FTSE-Mult FTSE Shariah Mult 150 Index FTSE Multinationals ($) Index 09/1999 09/2003 1.71 -0.74 28.01 34.52 0.06 -0.02
25 FTSE-USA FTSE Shariah USA $ index FTSE US $ :A Index 09/1999 09/2003 2.40 1.30 28.79 33.98 0.09 0.04
26 FTSE-GWA FTSE GWA Shariah Dev $ Index FTSE GWA All-World $ Index 09/1999 09/2007 3.46 3.12 31.90 38.70 0.11 0.08
27 FTSE-AP FTSE Shariah Dev Asia Pacific Index FTSE Dev Asia Pac $:A Index 09/1999 09/2003 -7.36 -7.69 31.56 32.67 -0.24 -0.24
28 S&P-US S&P 500 Shariah $ Index S&P 500 Composite - Aggregate Price 12/2000 12/2006 3.57 2.10 35.64 33.81 0.10 0.06
29 S&P-Japan S&P Japan 500 Shariah $ Index S&P JAPAN 500 INDEX 12/2000 12/2006 0.33 -2.78 19.91 22.12 0.02 -0.12
217

123
 Table 2 Regression results from single equation standard CAPM model and test of coefficients
218

ID 1 ID 2 ID 3 ID 4 ID 5 ID 6 ID 7 ID 8 ID 9

123
Panel A: MSCI
bij 0.9517*** 1.0053*** 1.0284*** 0.9203*** 1.0072*** 1.0315*** 0.9468*** 0.9522*** 0.9712***
(0.0270) (0.0221) (0.0190) (0.0216) (0.0230) (0.0195) (0.0178) (0.0142) (0.0161)
CRSHt 0.0106** 0.0052 0.0071 0.0089* 0.0054 0.0073 0.0010 -0.0013 0.0023
(0.0053) (0.0079) (0.0057) (0.0047) (0.0078) (0.0054) (0.0026) (0.0040) (0.0059)
ai 0.0003 -0.0003 -0.0002 0.0019* -0.0003 -0.0001 0.0010 0.0013 0.0004
(0.0009) (0.0008) (0.0008) (0.0011) (0.0009) (0.0008) (0.0008) (0.0008) (0.0010)
F-statistic 892.96 1174.45 1606.58 1215.8 1115.91 1567.76 1414.41 2299.38 1830.64
H01: ai = 0 0.14 0.15 0.04 2.92* 0.10 0.03 1.56 2.41 0.14
H02: bij = 1 3.20* 0.06 2.24 13.66*** 0.10 2.61 8.92*** 11.35*** 3.20*
H03: (ai = 0 and bij = 1) 1.61 0.11 1.14 7.43*** 0.11 1.33 7.45** 5.99*** 1.60
ID 10 ID 11 ID 12 ID 13 ID 14 ID 15 ID 16 ID 17 ID 18 ID 19

Panel B: DJ
bij 1.0359*** 0.9820*** 1.0906*** 1.0643*** 1.0163*** 0.9453*** 1.3075*** 0.8623*** 0.9215*** 1.0119***
(0.0269) (0.0409) (0.0499) (0.0464) (0.0081) (0.0282) (0.0582) (0.0252) (0.0374) (0.0405)
CRSHt 0.0027 0.0207** 0.0135 0.0077 0.0013 -0.0015 0.0423** 0.0124** -0.0034 0.0185**
(0.0088) (0.0079) (0.0139) (0.0132) (0.0014) (0.0095) (0.0184) (0.0050) (0.0072) (0.0085)
ai -0.0012 -0.0021* -0.0010 -0.0002 -0.0001 0.0001 -0.0067** -0.0037*** -0.0016 -0.0014
(0.0012) (0.0011) (0.0018) (0.0018) (0.0003) (0.0016) (0.0027) (0.0013) (0.0015) (0.0011)
F-statistic 798.64 391.21 306.35 411.51 23617.01 736.69 272.15 617.06 355.88 405.09
H01: ai = 0 0.94 3.47* 0.33 0.01 0.03 0.01 6.31** 8.01*** 1.10 1.57
H02: bij = 1 1.78 0.19 3.29* 1.92 4.04 3.77* 27.94*** 29.78*** 4.40** 0.09
Author's personal copy

H03: (ai = 0 and bij = 1) 1.46 2.34 1.66 0.96 2.06 1.92 13.99*** 30.21*** 3.44** 0.79

ID 20 ID 21 ID 22 ID 23 ID 24 ID 25 ID 26 ID 27 ID 28 ID 29

Panel C: FTSE S&P

bij 0.9893*** 0.9384*** 0.6197*** 0.8619*** 0.9130*** 0.8525*** 0.8383*** 0.9855*** 0.9263*** 0.7667***
(0.0397) (0.0187) (0.0459) (0.0581) (0.0231) (0.0300) (0.0283) (0.0148) (0.1380) (0.0446)
CRSHt 0.0029 0.0031 -0.0242 -0.0117 0.0037 -0.0010 0.0026 -0.0003 -0.0059 -0.0024
(0.0047) (0.0067) (0.0190) (0.0098) (0.0046) (0.0073) (0.0095) (0.0066) (0.0180) (0.0137)
ai 0.0009 0.0011 0.0035 0.0020 0.0005 0.0014 0.0019 0.0009 0.0021 0.0025
(0.0026) (0.0015) (0.0030) (0.0023) (0.0014) (0.0016) (0.0021) (0.0011) (0.0033) (0.0021)
F-statistic 465.15 1266.35 155.96 242.61 788.18 430.61 493.94 2220.29 56.38 186.79
H01: ai = 0 0.13 0.55 1.38 0.77 0.12 0.72 0.75 0.65 0.41 1.40
D. Ashraf
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 219

Estimates of Eq. (1) using monthly data of IEIs and their benchmarks from December 2000 to May 2012 (inclusive). The last three rows in each panel report the value of v2-test statistics for the
To further understand these results, test of coefficients

27.37***
14.35***
ID 29 have been performed using the Wald test14 as described in
Schroder (2007) and Judge et al. (1985) for three null
hypotheses as described above under H01–H03, and repor-
ted in the last three rows underneath each IEI in Table 2.
Test statistics for Jensen’s alpha under the null hypothesis
ID 28

H01: ai = 0, cannot be rejected for 25 IEIs. The coefficient

0.29
0.26

test under the null hypothesis H02: bij = 1 for same sys-
tematic risk (bij) is rejected for 15 IEIs suggesting that
these IEIs exhibit a different risk/return performance as
compared with their benchmark returns. The joint
ID 27

0.96
0.70

hypothesis of spanning cannot be rejected for 16 IEIs,

suggesting that the performance of these IEIs is similar to
that of their respective benchmark indices. For the
32.62***
16.39***

remaining 13 IEIs, the spanning hypothesis is rejected at a

ID 26

significance level of 10 % or lower suggesting that these

IEIs exhibit a different risk-return payoff than conventional
benchmark indices.
Furthermore, the test results show that there are only
24.11***
12.42***

three IEIs—MSCI-Europe, DJ-Canada, and DJ-World,

ID 25

where all three null hypotheses are rejected at the mini-

mum 10 % significance level. This suggests that these IEIs
have produced an alpha significantly different from zero
and a beta significantly different from unity independently
14.23***
7.34**
ID 24

as well as jointly. For the remaining IEIs, in the absence of

significant Jensen’s alpha and statistically significant beta
coefficients, it can be argued that any over (under) per-
formance of IEIs is due to higher (lower) assumed sys-
stated null hypotheses. Standard errors of estimated coefficients are reported in parenthesis
*** Statistically different from zero, 1 % level; two tail test, ** 5 % level; * 10 % level

tematic risk in the form of bij coefficient.

5.64**
ID 23

3.00*

In summary, the above results indicate that most of the

IEIs exhibit a return performance significantly different
from their benchmark indices. However this deviation does
not stem from abnormal returns as indicated by the insig-
68.77***
35.92***

nificant Jensen’s alpha coefficient. A significant bij coef-

ID 22

ficient in most of the IEIs indicates that the majority of the

IEIs only deviate from their benchmark in terms of sys-
tematic risk. This finding suggests that an investor in an IEI
10.81***

can expect at least a similar risk-return payoff as of the

5.50***
ID 21

benchmark.
Estimation results, by using the multi-equation frame-
work as developed in Eq. (3), are reported in Tables 3 and
Table 4 for the whole sample and by individual index
providers, respectively. Before presenting estimation
ID 20

0.07
0.09

results, it is pertinent to know whether the SUR model is

appropriate for estimation. The Breusch and Pagan (1979)
test for independence was performed for separate equations
H03: (ai = 0 and bij = 1)

of the system in this study for the overall group and for the
individual index providers with the null hypothesis of zero
Table 2 continued

correlation in the error term. The SUR model is appropriate

if the null hypothesis is rejected. Null hypothesis of no
H02: bij = 1

14
Wald tests are approximately v2-distribution with n degrees of
freedom and 2n in case of the spanning test.

123
 Table 3 Regression results of the multi-equation framework using the SUR model across all IEIs
220

ID 1 ID 2 ID 3 ID 4 ID 5 ID 6 ID 7 ID 8 ID 9

123
Panel A: MSCI
bij 0.941*** 1.03*** 1.041*** 0.927*** 1.038*** 1.049*** 0.981*** 0.939*** 0.972***
(0.016) (0.010) (0.009) (0.012) (0.011) (0.010) (0.011) (0.011) (0.013)
CRSHt 0.012** 0.005 0.008** 0.011** 0.006 0.008** 0.002 -0.002 0.004
(0.005) (0.004) (0.004) (0.006) (0.004) (0.004) (0.003) (0.004) (0.005)
ai -0.002 0.001 0.000 0.000 0.001 0.000 0.000 0.002 -0.001
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
R2 0.96 0.98 0.99 0.97 0.98 0.99 0.98 0.99 0.98
ID 10 ID 11 ID 12 ID 13 ID 14 ID 15 ID 16 ID 17 ID 18 ID 19

Panel B: DJ
bij 1.013*** 0.837*** 0.907*** 0.897*** 1.021*** 0.880*** 1.134*** 0.883*** 0.884*** 0.857***
(0.019) (0.015) (0.017) (0.013) (0.006) (0.017) (0.027) (0.014) (0.020) (0.014)
CRSHt -0.003 0.003 -0.012* -0.016*** 0.002 -0.015* 0.018 0.007 -0.011* 0.000
(0.006) (0.005) (0.007) (0.005) (0.002) (0.008) (0.012) (0.004) (0.006) (0.005)
ai 0.002 0.000 0.003 0.004** 0.000 0.005* -0.004 0.004*** 0.003 0.001
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
R2 0.97 0.97 0.97 0.98 1.00 0.96 0.95 0.97 0.94 0.97

ID 20 ID 21 ID 22 ID 23 ID 24 ID 25 ID 26 ID 27 ID 28 ID 29

Panel C: FTSE S&P

bij 0.923*** 0.902*** 0.616*** 0.866*** 0.874*** 0.808*** 0.789*** 0.965*** 0.848*** 0.664***
(0.022) (0.012) (0.025) (0.024) (0.014) (0.017) (0.015) (0.014) (0.051) (0.038)
Author's personal copy

CRSHt 0.002 0.000 -0.026** -0.012 0.000 -0.006 -0.003 -0.002 -0.014 -0.014
(0.006) (0.005) (0.011) (0.008) (0.004) (0.006) (0.007) (0.005) (0.014) (0.011)
ai -0.002 0.001 0.004 0.002 0.000 0.001 0.002 0.001 0.003 0.005
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.01) (0.00)
R2 0.95 0.98 0.88 0.90 0.98 0.96 0.96 0.98 0.78 0.79
BP test N = 40 2098.74***
H01: ai = 0 N = 29 89.73***
H02: bij = 1 N = 29 1334.29***
H03: (ai = 0 and bij = 1) N = 58 1433.43***

Estimates of Eq. (2) using monthly data of IEIs and their benchmarks from December 2000 to May 2012 (inclusive) across individual index providers based on Shari’ah screening criteria adopted.
The Breusch-Pagan test of independence reports a v2-test statistic with N degree of freedom and null hypothesis that residuals from all equations are independent. A rejection of null hypothesis
confirms the appropriateness of using a multi-equation framework. The last three rows in each panel report v2-test statistics with N degrees of freedom for the stated null hypotheses. Standard
errors of estimated coefficients are reported in parenthesis
D. Ashraf

*** Statistically different from zero, 1 % level; two tail test, ** 5 % level; * 10 % level
 Table 4 Regression results of the multi-equation framework using the SUR model across IEIs providers
ID 1 ID 2 ID 3 ID 4 ID 5 ID 6 ID 7 ID 8 ID 9

Panel A: MSCI
bij 0.943*** 0.993*** 1.006*** 0.914*** 0.997*** 1.010*** 0.944*** 0.945*** 0.967***
(0.017) (0.013) (0.012) (0.015) (0.013) (0.012) (0.015) (0.011) (0.012)
CRSHt 0.010** 0.005 0.006* 0.008* 0.005 0.006* 0.001 -0.002 0.002
(0.004) (0.004) (0.004) (0.005) (0.004) (0.004) (0.004) (0.004) (0.004)
ai 0.000 0.000 0.000 0.002* 0.000 0.000 0.001 0.001 0.000
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
R2 0.95 0.97 0.98 0.96 0.97 0.98 0.97 0.98
BP test N = 36 892.04***
H01: ai = 0 N=9 10.55
H02: bij = 1 N=9 104.34***
Does Shari’ah Screening Cause Abnormal Returns?

H03: (ai = 0 and bij = 1) N = 18 114.36***

ID 10 ID 11 ID 12 ID 13 ID 14 ID 15 ID 16 ID 17 ID 18 ID 19

Panel B: DJ
bij 1.013*** 0.880*** 0.921*** 0.920*** 1.015*** 0.895*** 1.110*** 0.847*** 0.946*** 0.895***
(0.020) (0.021) (0.020) (0.016) (0.006) (0.021) (0.032) (0.020) (0.027) (0.019)
CRSHt -0.002 0.008 -0.010* -0.013** 0.001 -0.011 0.011 0.008* -0.004 0.004
(0.005) (0.005) (0.006) (0.005) (0.001) (0.007) (0.011) (0.005) (0.006) (0.005)
ai 0.001 0.001 0.004** 0.003** 0.000 0.003* 0.000 0.000 0.001 0.002
(0.001) (0.001) (0.001) (0.001) 0.000 (0.002) (0.003) (0.001) (0.002) (0.001)
R2 0.96 0.96 0.96 0.98 1.00 0.95 0.93 0.95 0.91 0.96
BP test N = 45 540.39***
Author's personal copy

H01: ai = 0 N = 10 17.61*
H02: bij = 1 N = 10 164.19***
H03: (ai = 0 and bij = 1) N = 20 184.47***

ID 20 ID 21 ID 22 ID 23 ID 24 ID 25 ID 26 ID 27 ID 28 ID 29

Panel C: FTSE S&P

bij 0.940*** 0.885*** 0.576*** 0.788*** 0.858*** 0.803*** 0.773*** 0.922*** 0.868*** 0.678***
(0.032) (0.018) (0.034) (0.041) (0.018) (0.024) (0.023) (0.020) (0.083) (0.058)
CRSHt 0.004 -0.002 -0.031*** -0.018** -0.002 -0.006 -0.005 -0.006 -0.012 -0.013
(0.007) (0.005) (0.011) (0.009) (0.005) (0.006) (0.007) (0.005) (0.015) (0.012)
ai -0.003 0.001 0.005 0.002 0.000 0.002 0.002 0.001 0.003 0.005
(0.002) (0.002) (0.003) (0.003) (0.001) (0.002) (0.002) (0.002) (0.005) (0.004)
221

123
 Author's personal copy
222 D. Ashraf

31.23***
33.11***

adopted. The Breusch-Pagan test of independence reports a v2-test statistic with N degree of freedom and null hypothesis that residuals from all equations are independent. A rejection of null
hypothesis confirms the appropriateness of using a multi-equation framework. The last three rows in each panel report v2-test statistics with N degrees of freedom for the stated null hypotheses.
Estimates of Eq. (2) using monthly data of IEIs and their benchmarks from December 2000 to May 2012 (inclusive) across individual index providers based on Shari’ah screening criteria
correlation is rejected at a significance level of 1 % for the

3.544*
ID 29 overall sample and for each of the providers, indicating the
0.79

1.67
presence of cross-equation contemporaneous correlation.
Thus, the set of equations developed to estimate the per-
formance of IEIs is a seemingly unrelated system. There-
fore, the Zellner (1962) estimation approach (SUR) is
N=1
N=2
N=2
N=4
ID 28

0.79

appropriate for this purpose.

Table 3 reports the estimation results for performance of
the overall sample of IEIs as a group using the multi-equation
framework. These results are qualitatively similar to that of
ID 27

the single equation model as reported in Table 2 for the

0.97

overall sample. bij coefficient for 20 IEIs is smaller than that

of the single equation model. For the remaining IEIs the
upward change in the magnitude of bij coefficient is negli-
gible. In case of the multi-equation framework, there is evi-
ID 26

0.95

dence of positive and slightly significant alpha for DJ-US

Small, DJ-Euro and DJ-World. The application of SUR
model results in a clear divide in case of CRSHt, IEIs from
MSCI that uses the book value of total assets as a divisor
ID 25

report with a positive and significant coefficient for four IEIs.

0.96

While another four IEIs, adhering to the DJ guideline of using

market value of equity reports a negative and significant
coefficient of CRSHt. A closer look at these results suggest
that negative CRSHt is associated with the more volatile small
ID 24

0.98

and mid-cap markets of the US, Euro zone and Japan.

In terms of coefficient tests, the null hypotheses of zero
abnormal return, same systematic risk individually and
jointly are rejected at 10 % level of significance or better.
*** Statistically different from zero, 1 % level; two tail test, ** 5 % level; * 10 % level
ID 23

These results suggest that IEIs, as a group, not only provide

0.89

a significantly different systematic risk level as compared

with conventional indices but also offer some stock selec-
tion benefit. The significant abnormal returns coefficient
warrants further investigation to see the source of abnormal
ID 22

returns by individual index providers. Table 4 provides the

0.88

Standard errors of estimated coefficients are reported in parenthesis

estimation results for performance of IEIs based on unique

screening criterion adopted by individual index providers.
The most notable difference in the single equation model
338.44***

213.63***
219.67***

and the SUR model is the general shrinkage of bij coeffi-

ID 21

0.97

5.88

cients, where all of the IEIs report bij coefficient smaller

than the single equation model except for DJ-Japan which
support the application of multi equation framework.
Panel A of Table 4 reports the estimation results for
N = 28

N = 14
N=7
N=7
ID 20

MSCI-IEIs showing insignificant Jensen’s alpha for all IEIs

0.95

except for a positive and slightly significant Jenson’s alpha

for MSCI-Europe. bij coefficient among all MSCI-IEIs is
smaller than single equation framework and significant at
H03: (ai = 0 and bij = 1)

the 1 % level. CRSHt is significant for four IEIs: MSCI-

Americas, MSCI-Asia Pacific, MSCI-Europe and MSCI-
Table 4 continued

Pacific suggesting hedging benefit during the peak of glo-

bal financial crisis. The joint test of Jensen’s alpha for all
H02: bij = 1
H01: ai = 0

MSCI-IEIs as a group is not rejected suggesting that IEIs

BP test

following Shari’ah screening criteria adopted by MSCI do

R2

not offer any abnormal return as a group. The joint

123
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 223

hypothesis of bij coefficient equal to one and the spanning benefit as compared with their benchmark. However, the
test of zero abnormal return are rejected at the 1 % sig- abnormal return is associated with small and mid-cap IEIs
nificance level. Combining these results, further confirms and those also reflect relatively higher losses during the
that IEIs following MSCI based Shari’ah screening criteria global financial crisis. On the other hand, IEIs using any
on average, exhibit a lower systematic risk as compared other screening criteria do not offer any stock selection
with their conventional counterparts with no abnormal benefit for the similar universe of equities. However, the
returns, but do offer some hedging benefit within specific clear benefit of investing in IEIs would be lower systematic
regional markets. risk with a similar risk-return payoff as conventional
The empirical results based on the SUR model devel- indices. These results are in line with the findings of Ashraf
oped from Eq. (3) for IEIs following Shari’ah screening and Mohammad (2014), Hassan and Girard (2011) and
criteria adopted by DJ are reported in Panel B of Table 4. Guyot (2011) that performance of Islamic indices do not
The major difference between the overall performance of differ significantly from that of conventional indices.
DJ and MSCI IEIs is a significant Jensen’s alpha in case of
DJ-US small, DJ-US mid and DJ-Euro. There is a general
shrinkage in bij coefficient as compared with single equa- Robustness Checks
tion model except for DJ-Japan where it increased slightly.
The coefficient of CRSHt is slightly significant for three In order to better understand the return performance of
IEIs but with opposite signs. For the two IEIs with positive diversified portfolios, Fama and French (1993) propose
and significant alpha: DJ-US small, DJ-US mid, CRSHt is size and value versus growth factors in addition to the
negative and slightly significant indicating the cost asso- market risk premium in CAPM model. The inclusion of
ciated with abnormal return. The test of coefficient for zero additional factors helps to isolate the impact on return
abnormal return for DJ-IEIs as a group is rejected at 10 % performance if a manager is following a style strategy.
significance. Null hypotheses for same systematic risk and Carhart (1997) proposed an additional momentum factor to
spanning are rejected at the 1 % significance level. Overall, Fama and French (1993) factors to capture the persistence
test results suggests that DJ-IEIs deviates significantly from in return performance of diversified portfolios. This
that of their benchmark indices and do offer some abnor- inclusion of additional factors in the standard CAPM
mal returns in some regional markets. However, this must model is expected to improve the estimation of alpha.
be interpreted carefully due to lower statistical significance Pastor and Stambaugh (2002) suggest that the use of non-
and negative association of CRSHt with individual IEI benchmark assets in a benchmark model such as the CAPM
having positive alpha. helps in estimating the alpha that tend to exhibit less var-
Empirical estimations for IEIs based on Shari’ah iation across different specification of benchmarks. They
screening criteria adopted by FTSE, and S&P are reported in further argue that in the extreme event when a benchmark
Panel C. These results are similar to the results from MSCI- model completely estimates the other passive asset, the
IEIs in Panel A of Table 4 and suggest that the relative under estimate of alpha is the same regardless of additional
(over) performance of IEIs stem from relative riskiness of subsets of assets in the estimation.
IEIs with respect to the benchmark indices. For IEIs from As a robustness test of IEIs performance, additional
FTSE and S&P; Jensen’s alpha is not significantly different factors are included for empirical estimations. The inclu-
from zero. CRSHt is negative and significant in case of FTSE- sion of the small cap factor, the growth-value factor, and
World and FTSE-Developed indicating higher losses during momentum factor transform Eq. (3) into a system of
the peak of financial crisis. This suggests that a relaxed tol- n equations is as follows:
erance level as in the case of FTSE, do not offer the hedging
benefit usually found with Islamic investments. Null R1t ¼ a1 þ b1j R1t þ b2 SMBt þ b3 HMLt þ b4 WMLt þ e1t ; . . .;
hypothesis of zero abnormal returns cannot be rejected for Rit ¼ ai þ bij Rjt þ b2 SMBt þ b3 HMLt þ b4 WMLt þ þ eit
both providers. On the other hand the test for bij coefficient for t ¼ 1; 2; . . .. . .; T; i ¼ 1; 2; 3. . .:; I
and spanning test are rejected at the 1 % significance level
and j ¼ 1; 2; 3. . .; J ð4Þ
for both index providers suggesting that any performance
deviation among these IEIs are associated with level of where SMBt is the difference between the returns of small
systematic risk that IEIs assume. stocks and big stocks on diversified portfolios. HMLt is the
The empirical results from the SUR model confirm that difference between the returns of value (high book-to-
different versions of Shari’ah screening criteria adopted by market) stocks and growth (low book-to-market) stocks on
individual index providers may result in performance diversified portfolios of stocks, and WMLt is the difference
deviation. Only those IEIs following screening criteria between the returns on diversified portfolios of the winners
adopted by DJ as a group may offer some stock selection and losers over the past year in month t. Table 5 presents

123
 Table 5 Robustness test: multivariate regression results of the multi-equation framework after including additional factors: small-cap, value-growth, and momentum within the single factor
224

CAPM model across IEIs providers

ID 1 ID 2 ID 3 ID 4 ID 5 ID 6 ID 7 ID 8 ID 9

123
Panel A: MSCI
bij 0.991*** 0.994*** 1.010*** 0.954*** 0.999*** 1.015*** 0.937*** 0.942*** 0.972***
(0.016) (0.014) (0.013) (0.014) (0.015) (0.014) (0.015) (0.011) (0.013)
CRSHt 0.008** 0.004 0.005 0.007* 0.004 0.005 0.002 -0.002 0.002
(0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004)
SMBt 0.022 -0.032 -0.015 0.032 -0.038 -0.019 -0.037 0.044 -0.021
(0.045) (0.055) (0.049) (0.053) (0.057) (0.051) (0.049) (0.050) (0.057)
HMLt -0.266*** -0.030 -0.057 -0.183*** -0.026 -0.055 0.068 -0.023 0.002
(0.049) (0.058) (0.052) (0.058) (0.060) (0.053) (0.056) (0.054) (0.062)
WMLt 0.099*** -0.025 -0.006 0.127*** -0.024 -0.006 -0.009 -0.011 0.032
(0.020) (0.023) (0.020) (0.024) (0.023) (0.021) (0.020) (0.021) (0.024)
ai 0.001 0.000 0.000 0.002** 0.000 0.000 0.001 0.001 0.000
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
R2 0.97 0.97 0.98 0.97 0.97 0.98 0.97 0.98 0.98
BP test N = 36 874.17***
H01: ai = 0 N=9 9.110
H02: bij = 1 N=9 83.02***
H03: (ai = 0 and bij = 1) N = 18 91.42***
ID 10 ID 11 ID 12 ID 13 ID 14 ID 15 ID 16 ID 17 ID 18 ID 19

Panel B: DJ
bij 1.015*** 0.897*** 0.918*** 0.921*** 1.014*** 0.893*** 1.106*** 0.831*** 0.944*** 0.906***
Author's personal copy

(0.021) (0.021) (0.021) (0.017) (0.006) (0.022) (0.034) (0.019) (0.029) (0.019)
CRSHt -0.003 0.011** -0.011* -0.012** 0 -0.014* 0.011 0.003 -0.006 0.006
(0.006) (0.005) (0.007) (0.005) (0.001) (0.008) (0.011) (0.004) (0.006) (0.005)
SMBt -0.022 -0.139* -0.003 0.011 0.018 0.128 0.136 0.249*** -0.035 -0.079
(0.088) (0.074) (0.100) (0.084) (0.022) (0.116) (0.175) (0.067) (0.105) (0.074)
HMLt -0.077 0.069 -0.012 0.030 -0.028 -0.154 0.049 -0.161** -0.082 0.064
(0.086) (0.074) (0.098) (0.082) (0.022) (0.117) (0.171) (0.064) (0.101) (0.073)
WMLt -0.047 0.035 -0.036 -0.001 0.005 0.042 -0.013 0.026 -0.030 0.028
(0.034) (0.029) (0.038) (0.032) (0.009) (0.046) (0.067) (0.026) (0.039) (0.029)
ai 0.002 0.000 0.004** 0.003** 0.000 0.003* 0.000 0.000 0.001 0.001
(0.001) (0.001) (0.002) (0.001) 0.000 (0.002) (0.003) (0.001) (0.002) (0.001)
R2 0.98 0.96 0.96 0.96 0.98 0.99 0.95 0.93 0.97 0.91
D. Ashraf
 Table 5 continued
ID 10 ID 11 ID 12 ID 13 ID 14 ID 15 ID 16 ID 17 ID 18 ID 19

BP test N = 45 572.08***
H01: ai = 0 N = 10 17.48**
H02: bij = 1 N = 10 164.39***
H03: (ai = 0 and bij = 1) N = 20 186.00***

ID 20 ID 21 ID 22 ID 23 ID 24 ID 25 ID 26 ID 27 ID 28 ID 29

Panel C: FTSE S&P

bij 0.929*** 0.873*** 0.574*** 0.737*** 0.856*** 0.796*** 0.767*** 0.889*** 0.857 0.647
(0.033) (0.018) (0.035) (0.043) (0.018) (0.024) (0.023) (0.020) (0.087) (0.066)
CRSHt 0.002 -0.004 -0.029*** -0.022** -0.001 -0.007 -0.006 -0.007 -0.017 -0.018
(0.007) (0.005) (0.011) (0.009) (0.005) (0.006) (0.007) (0.005) (0.015) (0.012)
Does Shari’ah Screening Cause Abnormal Returns?

SMBt 0.250* 0.294*** 0.32 0.498*** 0.202** 0.168 0.365*** 0.367*** 0.091 0.379
(0.143) (0.102) (0.227) (0.189) (0.092) (0.123) (0.141) (0.098) (0.314) (0.280)
HMLt -0.048 0.021 0.151 -0.064 0.071 0.049 0.026 0.095 -0.280 -0.114
(0.118) (0.086) (0.187) (0.148) (0.077) (0.104) (0.118) (0.079) (0.245) (0.205)
WMLt 0.004 0.028 0.101 -0.056 0.047 0.021 0.061 -0.021 -0.062 0.038
(0.044) (0.032) (0.071) (0.055) (0.029) (0.039) (0.044) (0.029) (0.090) (0.076)
ai -0.003 0.001 0.005 0.001 0.000 0.001 0.002 0.000 0.002 0.004
-0.002 -0.002 -0.003 -0.003 -0.001 -0.002 -0.002 -0.001 -0.005 -0.004
2
R 0.95 0.98 0.89 0.91 0.98 0.96 0.96 0.95 0.79 0.81
BP test N = 28 337.83*** N=1 3.45*
H01: ai = 0 N=7 7.330 N=2 1.230
Author's personal copy

H02: bij = 1 N=7 211.55*** N=2 29.27***

H03: (ai = 0 and bij = 1) N = 14 217.76*** N=4 32.06***

Estimates of Eq. (4) using monthly data of IEIs and their benchmarks from December 2000 to May 2012 (inclusive) across individual index providers based on Shari’ah screening criteria
adopted. The Breusch–Pagan test of independence reports a v2-test statistic with N degree of freedom and null hypothesis that residuals from all equations are independent. A rejection of null
hypothesis confirms the appropriateness of using a multi-equation framework. The last three rows in each panel report v2-test statistics with N degrees of freedom for the stated null hypotheses.
Standard errors of estimated coefficients are reported in parenthesis
*** Statistically different from zero, 1 % level; two tail test, ** 5 % level; * 10 % level
225

123
 Author's personal copy
226 D. Ashraf

the empirical results based on Eq. (4) using the SUR Table 6 Robustness check: Regression results of the multi-equation
model. Similar to Eq. (3), estimation results and hypothe- framework using the SUR model on fund of fund portfolios across
ses tests: ai = 0, bij = 1 and spanning test ai = 0 and index providers
bij = 1, are presented based on individual index providers. MSCI DJ FTSE S&P
The results reported in Table 5 are very similar to the
bij 1.034*** 0.970*** 0.825*** 0.836***
results reported in Table 4. There is very little change in
(0.022) (0.029) (0.023) (0.058)
the size of bij coefficient with no change in the statistical
CRSHt 0.012*** 0.003 -0.004 -0.006
significance. There is no change in signs or significance of
(0.004) (0.007) (0.005) (0.011)
abnormal return as reflected by alpha after including
ai -0.001 0.000 0.000 0.003
additional factors in the SUR model. The CRSHt is sig-
(0.001) (0.002) (0.002) (0.003)
nificant in fewer IEIs after including additional factors with
R2 0.98 0.97 0.98 0.79
no change in signs. This confirms the findings from the
SUR model as reported in Table 4. BP test N=6 10.78*
The impact of additional factors on the return perfor- H01: ai = 0 N=4 1.90
mance of IEIs is generally insignificant with some excep- H02: bij = 1 N=4 60.91***
tions. The size factor (SMBt) is positive and significant for H03: (ai = 0 and N=8 65.12***
bij = 1)
all FTSE IEIs except for FTSE-Word and FTSE-Mult. This
can be attributed to relaxed screening criteria whereby Estimates of Eq. (2) using monthly data for value-weighted portfolios
FTSE allows a higher tolerance level for quick assets as of IEIs and their benchmarks from December 2000 to May 2012
(inclusive) across individual index providers based on Shari’ah
compared with other index providers that might cause screening criteria adopted. The Breusch–Pagan test of independence
smaller firms to be included in the FTSE IEIs as compared reports a v2-test statistic with N degree of freedom and null hypoth-
with others. HMLt and WMLt are mostly insignificant for esis that residuals from all equations are independent. A rejection of
all index providers suggesting that IEIs do not follow a null hypothesis confirms the appropriateness of using a multi-equation
framework. The last three rows in each panel report v2-test statistics
specific style or momentum strategy. with N degrees of freedom for the stated null hypotheses. Standard
The estimation results based on the test of coefficients as errors of estimated coefficients are reported in parenthesis
reported in Table 5 are the same as previously reported in *** Statistically different from zero, 1 % level; two tail test,
Table 4. The null hypothesis for ai coefficient cannot be ** 5 % level; * 10 % level
rejected for any index provider except for DJ IEIs. This
further confirms that IEIs following screening criteria systematic risk and spanning test are rejected at 1 % sig-
based on the market-value of equity as a divisor for nificance level suggesting that IEI providers as a group do
quantitative screening may exhibit some stock selection not offer any abnormal returns and return deviation, if any,
benefit. However, null hypotheses regarding the spanning stems from the relative riskiness of IEIs relative to the
test and test for bij coefficient are rejected at the 5 % sig- benchmark indices.
nificance level or better. These results again indicate the
existence of a different risk return payoff of IEIs and
conventional benchmarks. However, this difference in risk Summary and Conclusion
return payoff arises from the relative riskiness of IEIs with
respect to their benchmark indices. Shari’ah compliant equity investments have opened an
To further confirm the differences on the risk-return outlet for faith-adhering Muslims to benefits from capital
performance of IEIs, eight value weighted portfolios (four markets without compromising their religious beliefs. For
IEIs and four conventional benchmark portfolios) are an equity to obtain Shari’ah compliance status, its major
structured by including all indices from each provider. The sources of revenue should not come from non-permissible
resulting portfolios are similar to a fund of funds-like activities such as dealing in lending money on interest,
portfolio for each index data provider. Table 6 presents the pork or pork products, alcohol, speculation whether from
estimation results for fund-of-fund portfolios using the gambling, derivatives or insurance related activities. In
SUR model. Test of coefficient for three null hypotheses: addition, equities have to comply with financial ratio based
H01: ai = 0, H02: bij = 1 and H03: (ai = 0 and bij = 1) are screening criteria of financial leverage, cash holding and
obtained in support of the empirical estimations. Similar to account receivables. In practice, there are several different
previous findings bij is significant for all providers at 1 %. divisors used to calculate the financial ratios required to
CRSHt is positive and significant for MSCI. While alpha is attain the compliance status. These include total assets,
not significant for any of the providers. In terms of test of trailing market value of equity ranging from 12 months to
coefficients, null hypothesis for zero abnormal returns 36 months. Aside from divisor differences, tolerance levels
cannot be rejected while the null hypotheses of same for impermissible activities are not the same. These

123
 Author's personal copy
Does Shari’ah Screening Cause Abnormal Returns? 227

different financial screening criteria may result in com- more frequent rebalancing wherein under-performing stocks
pletely different portfolios even if selected from the same are replaced before reaching the trough in the downturn of
pool of assets. Aside from diverse portfolio compositions the economy. These finding are in line with recent studies
and varying return performance, investment portfolios including, Hassan and Girard (2011) for DJIM indices ver-
based on trailing market value of equity may require higher sus conventional indices from 1999 to 2006, Guyot (2011)
monitoring and rebalancing costs to maintain the Shari’ah for DJIM from 1999 to 2008 and Albaity and Ahmad (2008)
compliance status. for Kuala Lumpur Syariah Index versus Kuala Lumpur
Islamic equity investment has grown exponentially over Composite Index from 1999 to 2005, Derigs and Marzban
the past decade. Recent empirical literature suggests that (2009) for S&P 500 and Adamsson et al. (2014).
IEFs performed relatively better than conventional funds in The findings of this paper are of interest to policy
the recent past. However, performance comparison of makers, practitioners and the general investing public.
Shari’ah complaint mutual funds with conventional funds Investors in IEFs are not detrimentally affected for their
lacks the rigor due to trading costs and active management faith-adherence and can expect a risk adjusted return not
by fund managers. To isolate the impact of Shari’ah dissimilar to that of a conventional benchmark index.
screenings, use of indices for performance comparison is Furthermore, the screening criteria used by index providers
preferred over investment funds as it makes the screen do not affect the return performance of IEIs. Findings of
related performance attribution more robust since it does this study are relevant in performance measurement of fund
not require adjustment for transaction costs, management managers. Use of appropriate Islamic equity benchmarks
fees and/or timing abilities of fund managers. for the performance measurement of IEFs based on
This study provides a comparison of risk-return char- screening criteria adopted by IEIs provides more relevant
acteristics of IEIs based on different Shari’ah screening information about the management skills of fund managers.
criteria versus conventional benchmarks by utilizing a Further work needs to be done in streamlining Shari’ah
longer time series and cross sectional data for IEIs con- screening standards to avoid confusion among the investing
structed by four major index providers: MSCI, DJ, FTSE public. Both policy makers and fund managers can benefit
and S&P. Each of these index providers uses a different by adhering to a unified criterion without any significant
Shari’ah screening criterion especially for financial factors loss to return performance. This will help to attract more
for inclusion of equities in the index. MSCI and FTSE use faith-adhering investors wishing to invest in capital mar-
the book-value of total assets for calculating financial ratios kets. The managers of conventional funds can also benefit
while DJ and S&P use the trailing market value of equity by holding a passively managed portfolio of investments
for the calculation of a similar ratio. This study investigates adhering to Shari’ah principles.
whether different screening criteria result in a different This study has investigated the relationship between the
risk-return payoff by using index-level data for Islamic as screening criteria and the performance of IEIs versus
well as conventional indices for the period December 2000 conventional indices with no consideration to transaction
to May 2012. costs and/or liquidity issues. However, return for a passive
The empirical findings are obtained using several dif- portfolio that often requires rebalancing can diverge con-
ferent performance measurement models in a single equa- siderably from that of a more passive portfolio due to
tion as well as in a multi-equation framework. Conventional transaction costs and liquidity risk. A portfolio following
indices are matched as a benchmark for IEIs from where Shari’ah screening based on the market-value of equity
IEIs normally draws its equities over a longer time period requires more often rebalancing and thus incurs higher
than any other previously published studies in this area. The trading costs as compared with a portfolio following book-
use of a multi-equation framework is one of the major value of total assets based criteria and thus may lead to
contributions of this paper. A multi-equation framework lower performance after these adjustments. This warrants
using the SUR model is applied to groups of IEIs based on further investigation and provides an avenue for future
distinct screening criteria used by each index provider. The research.
multi-equation framework results in a better estimation of
coefficients and further testing of coefficients. Empirical Acknowledgments I thank Dr. Barbara Marie L’Huillier, Professor
Ashfaque Hassan Khan, and participants of IRTI Knowledge Sharing
results under both frameworks suggest that IEIs, on average, Session, 20th Annual Conference of the Multinational Finance
do not produce any abnormal returns except for small and Society and NUST Business School workshop on Islamic finance for
mid-cap IEIs from the US following quantitative Shari’ah helpful comments. The author is grateful to the editor and two
screening criteria adopted by DJ. However, those IEIs anonymous reviewers for their helpful comments and suggestions to
improve the earlier version of this paper. The views expressed in this
reflecting positive abnormal return reported steeper declines paper are those of the author and do not necessarily reflect the views
in returns during the peak of the financial crisis. The positive of the Islamic Research and Training Institute or the Islamic Devel-
abnormal return associated with DJ IEIs may be attributed to opment Bank Group.

123
 Author's personal copy
228 D. Ashraf

References Online, Accessed October 17, 2012, from [Link]

com/english/wp-content/uploads/2009/09/Risk-Return-of-Islamic-
Abdullah, F., Hassan, T., & Mohamad, S. (2007). Investigation of [Link].
performance of Malaysian Islamic unit trust funds. Managerial Hassan, K., & Girard, E. (2011). Faith-based ethical investing: The
Finance, 33(2), 142–153. case of Dow Jones Islamic Indexes. Working Paper 2011-WP-
Adamsson, H., Bouslah, K., & Hoepner, A. G. F. (2014). An Islamic 05. Network Financial Institute, Indiana State University,
Equity Premium Puzzle. Retrieved April 29, 2014, from http:// Indiana, April 2011.
[Link]/abstract=2390347/. Hayat, R., & Kraeussl, R. (2011). Risk and return characteristics of
Alam, N., & Rajjaque, M. (2010). Shariah-compliant equities: Islamic equity funds. Emerging Markets Review, 12, 189–203.
Empirical evaluation of performance in the European market Hodgson, D. J., & Linton, O. (2000). Vorkink. K.: Testing the capital
during credit crunch. Journal of Financial Services Marketing, asset pricing model efficiently under elliptical symmetry: A
15(3), 228–240. semiparametric approach. Journal of Applied Econometrics,
Albaity, M., & Ahmad, R. (2008). Performance of Syariah and 17(6), 617–639.
Composite Indices: Evidence from Bursa Malaysia. Asian Hoepner, A. G., Rammal, H. G., & Rezec, M. (2011). Islamic mutual
Academy of Management Journal of Accounting and Finance, funds’ financial performance and international investment style:
4(1), 23–43. Evidence from 20 countries. The European Journal of Finance,
Ashraf, D. (2013). Performance evaluation of Islamic mutual funds 17(9–10), 828–850.
relative to conventional funds: Empirical evidence from Saudi Huberman, G., & Kandel, S. (1987). Mean-variance spanning.
Arabia. International Journal of Islamic and Middle Eastern and Journal of Finance, 42, 873–888.
Management, 6(2), 105–121. Hussein, K. A. (2004). Ethical Investment: Empirical evidence from
Ashraf, D., & Mohammad, N. (2014). Matching perception with the FTSE Islamic Indices. Islamic Economic Studies, 12(1), 21–40.
reality—Performance of Islamic equity investments. Pacific- Hussein, K., & Omran, M. (2005). Ethical investment revisited:
Basin Finance Journal, 28, 175–189. doi:10.1016/[Link].2013. Evidence from Dow Jones Islamic Indexes. Journal of Investing,
12.005. 12(2), 105–124.
Breusch, T. S., & Pagan, R. A. (1979). A simple test for heterosced- Hussain, K. A. (2006). Islamic investment: Evidence from Dow Jones
asticity and random coefficient variation. Econometrica, 47(5), and FTSE Indices. In Proceedings of 6th international confer-
1287–1294. ence on Islamic Economics and Finance (Vol. 1, pp. 387–401).
Carhart, M. M. (1997). On Persistence in Mutual Fund Performance. Judge, G. G., Griffiths, W. E., Hill, R. C., Lütkepohl, H., & Lee, T. C.
Journal of Finance, LI, I(1), 57–82. (1985). The theory and practice of econometrics (2nd ed.). New
Derigs, U., & Marzban, S. (2008). Review and analysis of current York: Wiley.
Shariah-compliant equity screening practices. International Kok, S., Giorgioni, G., & Laws, J. (2009). Performance of Shariah-
Journal of Islamic and Middle Eastern and Management, 1(4), compliant indices in London and N.Y. stock markets and their
285–303. potential for diversification. International Journal of Monetary
Derigs, U., & Marzban, S. (2009). New strategies and a new paradigm Economics and Finance, 2(3–4), 398–408.
for Shariah-compliant portfolio optimization. Journal of Banking Nainggolan, Y. A. (2011). Taking a leap of faith: Are investors left
& Finance, 33(6), 1166–1176. short changed. PhD Thesis, Queens University of Technology,
Elfakhani, S. M., Hassan, K., & Sidani, Y. M. (2007). Islamic mutual Brisbane, Australia.
funds. In K. Hassan & M. K. Lewis (Eds.), Handbook of Islamic Newey, W., & West, K. (1987). Hypothesis testing with efficient
Banking (pp. 253–273). Cheltenham: Edward Elgar. method of moments estimation. International Economic Review,
Ernst & Young. (2011). Islamic Funds & Investment Report 2011. 28(3), 777–787.
Achieving growth in challenging times. Accessed December 20, Obaidullah, M. (2005). Islamic financial services. Jeddah: Scientific
2011, from [Link] Publishing Center, King AbdulAziz University.
Fama, E. F., & French, K. R. (1993). Common risk factors in the Pastor, L., & Stambaugh, R. F. (2002). Mutual fund performance and
returns on stocks and bonds. Journal of Financial Economics, seemingly unrelated assets. Journal of Financial Economics,
33(1), 3–56. 63(3), 315–349.
Fama, E. F., & French, K. R. (2004). The capital asset pricing model: Rahman, A. A., Yahya, M. A., & Nasir, M. H. M. (2010). Islamic
Theory and evidence. Journal of Economic Perspectives, 18(3), norms for stock screening—A comparison between the Kuala
25–46. Lumpur Stock Exchange Islamic index and the Dow Jones
Fama, E. F., & French, K. R. (2012). Size, value, and momentum in Islamic Market Index. International Journal of Islamic and
international stock returns. Journal of Financial Economics, Middle Eastern Finance and Management, 3(3), 228–240.
105(3), 457–472. Schroder, M. (2007). Is there a difference? The performance
Gangi, F., & Trotta, C. (2013). The ethical finance as a response to the characteristics of SRI equity indices. Journal of Business
financial crises: An empirical survey of European SRFs perfor- Finance & Accounting, 34(1–2), 331–348.
mance. Journal of Management and Governance,. doi:10.1007/ Shanmugam, B., & Zahari, Z. R. (2009). A primer of Islamic Finance.
s10997-013-9264-7. Charlottesville, VA: The Research Foundation of CFA Institute.
Girard, E., & Hassan, M. K. (2008). Is there a cost to faith-based Usmani, M. I. A. (2002). Meezan bank’s guide to Islamic banking.
investing: Evidence from FTSE Islamic indices. The Journal of Karachi: Darul-Ishaat.
Investing, 17(4), 112–121. Walkshäusl, C., & Lobe, S. (2012). Islamic investing. Review of
Guyot, A. (2011). Efficiency and dynamics of Islamic investment: Financial Economics, 21(2), 53–62.
Evidence of geopolitical effects on Dow Jones Islamic Market Zellner, A. (1962). An efficient method of estimating seemingly
Indexes. Emerging Markets Finance & Trade, 47(6), 24–45. unrelated regression and tests for aggregation bias. Journal of the
Hakim, S., & Rashidian, M. (2004). Risk & return of Islamic Stock American Statistical Association, 57(298), 348–368.
Market Indexes. In Forum Annual Conference in Sharjah, UAE.

123