The performance of socially responsible equity mutual funds: Evidence from Sweden

• Author: Florinda Silva,Maria Ceu Cortez,Carlos Leite,Christopher Adcock
• DOI: http://doi.org/10.1111/beer.12174
• Date: 01 April 2018
• Published date: 01 April 2018

ORIGINAL ARTICLE

The performance of socially responsible equity mutual funds:

Evidence from Sweden

Carlos Leite

1

|

Maria Ceu Cortez

2

|

Florinda Silva

2

|

Christopher Adcock

2,3

1

School of Economics and Management,

University of Minho, Braga, Portugal

2

NIPE –School of Economics and

Management, University of Minho, Braga,

Portugal

3

School of Finance and Management,

SOAS –University of London, UK

Correspondence

Florinda Silva, NIPE –School of Economics

and Management, University of Minho,

Gualtar, 4710-057 Braga, Portugal.

Email: fsilva@eeg.uminho.pt

Funding information

Fundaç~

ao para a Ci^

encia e a Tecnologia,

Grant/Award Number: POCI-01-0145-

FEDER-006683; National Funding;

The ERDF

This paper presents a comprehensive analysis of socially responsible (SR) funds in Sweden by

assessing fund managers’abilities and performances across different market states. These issues

are analyzed at the aggregate and individual fund levels. The paper also presents several new sta-

tistical tests that allow more precise inferences about differences in performance and the

variability in fundreturns arising from differentbenchmarks. In general, SR and conventional funds

perform similarly to the market. At the aggregate level, SR funds investing in Sweden andEurope

perform similarly to conventional funds,while those investing globally tend to underperform. This

underperformance seems to be linked with poor selectivity abilities of global SR fund managers.

For individual funds, the performance of both types of funds is more similar. Most funds perform

similarly in crisis periods compared to non-crisis periods. Overall, our results are consistent with a

mature market for SRinvesting and support the view that the similar performanceof SR and con-

ventional funds is associated with the mainstreaming of SR investment in Sweden. These findings

encourage SR investing both by socially conscious investors, who wish to align their social values

with their investment decisions, as well as by conventional investors,who will not be penalized by

investing in these funds. We also call attention to the difficulties investors face when trying to

identify fundswith high social standards, considering that thereis scarce information on the extent

to which each fund(SR or conventional) holds stocksthat comply with ethical andsocial criteria.

1

|

INTRODUCTION

Socially responsible (SR) investments have experienced remarkable

growth in recentdecades. Within this segment ofthe financial industry,

the number of SR mutual funds have increased substantially. These

funds use screens to select companies that meet certain social criteria

or to exclude those that are involved in undesirable activities or prac-

tices. The emergence of such funds and their growing importance in

the industry has attracted the interest of academics and raised some

pertinent questions, mainly relatedto the impact that considering social

criteria has on portfolio performance. Theoretically, there are argu-

ments in favor of overperformance, underperformance, and even neu-

tral performance of SR funds relative to their conventional peers. At

the empirical level,most studies conclude that bothtypes of funds per-

form similarly(e.g., Capelle-Blancard & Monjon, 2012;Revelli & Viviani,

2015).

In this paper, we extend the international evidence and contribute

to the field by carrying out a comprehensive analysis of SR equity

mutual fund performance in Sweden.Considering extant evidence that

SR funds are more sensitive to conventional benchmarks than SR

benchmarks, we also explore this issue by implementing a number of

statistical tests that are either rarely used in fund performance studies

or which to the best of our knowledge are new. Empirically, the pur-

pose of this paper is threefold. First, the main objective is to compare

the performance of Swedish SR funds with that of conventional funds

by using robust models that consider time-varying risk and perform-

ance. Second, considering that the screening process might restrict or

enhance opportunities for different manager skills, we compare timing

and selectivity abilities of both types of funds. Third, we investigate

whether there are differences in performance over different market

states. This analysis is motivated bythe fact that most studies evaluate

SR funds over a specific time period without comparing performance

across differentmarket states. To the best of our knowledge, thisis the

first study on Swedish SR funds that analyzes whether there are statisti-

cally significant differences in individual fund performance across crisis

and non-crisis periods. Finally, all these performance issues are analyzed

at both the aggregate and individual fund levels. Most studies on SR funds

only report fund performance at the aggregate level, even though analysis

of individual fund performance is both more relevant and more useful for

investors. Furthermore, evidence of outperformance/underperformance

108

|

V

C2017 JohnWiley & Sons Ltd wileyonlinelibrary.com/journal/beer BusinessEthics: A Eur Rev. 2018;27:1 08–126.

Received:17 October 2016

|

Revised: 22 August2017

|

Accepted:17 September 2017

DOI: 10.1111/beer.12174

at the individual fund level may cancel out at the aggregate level or evi-

dence of outperformance/underperformance at the aggregate level may

be driven by a few individual funds.

In addition to the empirical analyses described above, this paper

makes a number of methodological contributions. These are motivated

by the classic paperof Roll (1978), who first drew attention to thefact

that the assessment of portfolio performance can vary markedly

depending on the choice of benchmark.Notwithstanding the analytical

results in Roll’s paper, the use of formal statistical tests to assess such

differences is rare. The methods described in Grinblatt and Titman

(1994) are an exception; it is more usual for papers on fund perform-

ance measurementto comment on differences. Such differences, when

they occur, may be due solely to sampling variation. Formal statistical

tests allow more precise inferences to be made about differences in

fund performance.This aspect of performance measurement is particu-

larly relevant for the assessment of SR funds, as the choice of an

appropriate benchmark is an important issue. Several studies (e.g.,Reh-

man, Zhang, Uppal,Cullinan, & Naseem, 2016) findno significant differ-

ences between the performance of SR and conventional indices.

Nevertheless, there is evidence that SR funds are more exposed to

conventional than to SR indices (e.g., Bauer, Koedijk, & Otten, 2005;

Bauer, Otten, & Rad, 2006; Cortez, Silva, & Areal, 2009, 2012), moti-

vating the debate on whether SR funds should beevaluated relative to

a socially restricted index or a conventional one. Equally important is

the comparison of performance of an SR fund with a comparable con-

ventional portf olio. In this paper, we employ f our tests. First , we apply

a procedure due to Hotelling (1940), which tests theequality of corre-

lation of two alternative independent variables with a specified

dependent variable. This test, published over 70 years ago, is of

obvious relevance when there are two different indices (SR and con-

ventional) used in applications of the market model. However, Hotel-

ling’s test is not well known in finance. Second, we employ a

development of Hotelling’s procedure (referred to as the Beta test in

the rest of the paper),which tests the equality of the betasin two sim-

ple ordinary least squares (OLS) models with a common dependent

variable. When applied together, these two tests allow a joint assess-

ment of (a) the extent to which each index accounts for the variability

in fund returns and (b) differences in the exposure of a fund to each

index. Third, a similar procedure is also applied to test for the equality

of alphas (Alpha test). As estimated values of alpha are widely used to

measure fundperformance, the results of the testinform the important

issue of benchmark choice. Last, we are concerned with the extent to

which different indices may be used to account for the variability in

both SR and conventional funds. We present a procedure to test the

equality of two adjusted R

2

s in separate regressi on models. For large

samples, this results in an Ftest. To the best of our knowledge, these

tests have not been usedin the fund performance literature beforebut

are of obvious relevance. In this particular study, the tests often serve

to confirm results that may be inferred by inspection. This would not

necessarilybe the case for other studies.

Sweden has long been at the forefront of SR investments with all

large institutions having formal SR policies (Eurosif, 2014) and public

policies playingan active role in promoting corporate social responsible

practices (Steurer, Martinuzzi, & Margula, 2012). It is considered a

mature market for SR investments (Rietz, 2016), having experienced

exponential growth in recent years, and representing more than 674

billion euros in2013. Additionally, in terms of the numberof SR mutual

funds, Sweden is oneof the leading European markets for equityfunds

(Basso & Funari, 2014).

1

Thefactthatcorporatesocialresponsibility

(CSR) issues are at the center of Scandinavian investors’and institu-

tions’concerns and that SR investing has developed from a niche to a

mainstream approach to investing (Bengtsson, 2008; Nilsson, 2008;

Vidaver-Cohen & Brønn, 2015) also motivates this study. There are

additional features of the Swedish market that make it of interest.

Swedish SR funds follow investment approaches that have cometo be

known as the “Scandinavian model”(Jensen, 2016). Accordingly,

Swedish SR funds use mainly norms-based screening (Rietz, 2016),

which involves excluding companies that violate international norms

and standards. The Scandinavian model isalso characterized by mutual

funds holding diversified portfolios rather than focusing on a particular

theme (e.g., environment) or dimension of CSR (Jensen, 2016). Despite

these particularities of the Swedish SRmarket and research suggesting

country and cultural idiosyncratic contexts in SR investing (Bengtsson,

2008), to the best of our knowledge, only a few studies have included

Swedish SR funds. None of these studies perform an in-depth investi-

gation of the performance aspectsof Swedish funds mentioned above.

This paper is organized into six sections. In the second section, we

present a briefreview of the literature in this area.The third and fourth

sections describethe methods and the data, respectively. The empirical

results are presented and discussedin the fifth section. In the last sec-

tion, we draw the main conclusions.

2

|

LITERATURE REVIEW

Recent years have witnessed an increasinginterest in companies’social

responsibility practices. Following some well-known corporate ethical,

accounting, and environmental scandals, there is an increased general

awareness that companies’socially irresponsible or unethical behavior

has damaging consequences to the wellbeing of society. Consequently,

investors arenow more concerned with incorporatingcorporate behav-

ior selection screens in the investment decision process. A major issue

in this process is whether considering social criteria impacts portfolio

performance.

There are contradictory arguments on the impact of considering

social screens in the financial performance of mutual funds. Following

Markowitz (1952),imposing social filters on the portfolio selection pro-

cess would result in a suboptimal portfolio and, therefore, in a less

attractive risk-return tradeoff. In contrast, another school of thought

argues that socially screened portfolios may benefit from improved

financial performance. This perspective is supported by studies that

document that SR companies perform better compared to those that

are less SR (e.g.,Kempf & Osthoff, 2007; Statman & Glushkov, 2009).

2

The underlying argument is that SR companies benefit from competi-

tive advantages and have a better quality of management (Frynas &

Yamahaki, 2016), besides experiencing lower risks (G

omez-Bezares,

LEITE ET AL.

|

109