Can Socially Responsible Investments Be Compatible with Financial Performance? A Meta‐analysis

• DOI: http://doi.org/10.1111/ajfs.12244
• Date: 01 February 2019
• Author: Chang‐Soo Kim
• Published date: 01 February 2019

Can Socially Responsible Investments Be

Compatible with Financial Performance?

A Meta-analysis*

Chang-Soo Kim**

College of Government and Business, Yonsei University, Republic of Korea

Received 29 March 2018; Accepted 22 November 2018

Abstract

This paper examines whether socially responsible investment (SRI) outperforms conven-

tional investments by meta-analysis. The result shows that weighted average effect size is

not significantly different from zero, suggesting that SRI performance is not different from

conventional investments. Meta-regression results regarding sampling issues indicate that

economic crisis, control group, investment universe, screening procedure, and mutual

funds are important determinants of effect sizes, and publication year, author type, and

control group are important factors of the absolute value of effect sizes. Regarding

methodological issues, risk adjustment, weighting scheme, data refinement, benchmark

model, and matching procedure are significant factors determining the absolute value of

effect sizes.

Keywords Effect size; Meta-analysis; Meta-regression; Random-effect model; Socially

responsible investment

JEL Classification: G11, G32, M14

1. Introduction

Researchers and investors have long been curious about whether socially respon-

sible investment (SRI) funds and portfolios financially outperform conventional

ones. Early studies date back to the 1970s (Moskowitz, 1972; Alexander and

*This paper was supported by the Ministry of Education of the Republic of Korea and the

National Research Foundation of Korea (NRF-2016S1A5B8925203). This work was supported

by Fulbright Senior Researcher Scholarship.

**Corresponding author: Department of Business Administration, Yonsei University, Yonsei-

dae-gil 1, Heungup-myon, Wonju, Gangwon-do 26493, Republic of Korea. Tel: +82-33-760-

2331, Fax: +82-33-760-2918, email: kimc@yonsei.ac.kr.

Asia-Pacific Journal of Financial Studies (2019) 48, 30–64 doi:10.1111/ajfs.12244

30 ©2019 Korean Securities Association

Buchholz, 1978), and a significant amount of research has been accumulated so

far.

1

However, existing evidence shows mixed results. For example, Auer (2016)

shows that investors can do financially well while doing social good based on

environmental, social, and governance ratings data for the European market.

However, Renneboog et al. (2008a) report that the SRI funds of France, Ireland,

Sweden, and Japan significantly underperform conventional funds by 4–7% per

annum during the period 1991–2003. In addition, there are many other papers

that report no difference in performance between SRI and conventional funds

(Hamilton et al., 1993; Kreander et al., 2005; Kempf and Osthoff, 2008; Hoepner

and Schopohl, 2018).

It is imperative to determine the overall influence of SRI on investment perfor-

mance, because if the impact is positive, investors will voluntarily participate in

socially beneficial investments and government does not have to intervene in the

market to promote social goals. If the impact is negative, however, investors have

to sacrifice their investment benefits to pursue their social preferences and the pub-

lic sector has to intervene in one way or another to encourage SRI activities. The

direction and magnitude of SRI influence can also have a profound implication for

financial industries with respect to product development, asset management, fidu-

ciary duty, performance evaluation, and so forth. Understanding the impact of SRI

is also important to academic research. As Combs et al. (2011) point out, without a

proper synthesis of existing literature more fruitful further research is difficult.

Despite this, there have been few attempts to synthesize existing literature on

SRI and even these have not been successful. For example, Rathner (2013) lacks

representativeness due to a small sample size and incomplete methodology. Revelli

and Viviani (2015) performed a meta-analysis based on 85 studies. However, they

fail to draw a sharp conclusion due to ambiguity caused by pooling several coun-

tries in one sample. L

opez-Arceiz et al. (2018) show that cultu re is an important

determinant of SRI influence, so combining countries with different cultural back-

grounds will prevent researchers from obtaining a correct picture of SRI.

2

More

importantly, they analyze the impact of moderators only on effect size (hereafter,

ES) ignoring the impact on the absolute value of effect size (hereafter, AES ),

1

SRI is different from corporate social responsibility (CSR). CSR focuses on corporate social

activities and papers on CSR examine whether firms can improve financial performance by

engaging in socially desirable activities. SRI focuses on investment strategies that employ vari-

ous screening procedures (e.g., positive, negative, and best-in-the-class) based on so-called

ESG (environmental, social, and governance) criteria and papers on SRI compare the invest-

ment performance of these portfolio strategies with conventional ones.

2

Moreover, their study seems to have multiple errors. For example, they include Smith

(1996), which is an event study, even though they explicitly mention that they do not include

event studies in the text. Gregory et al. (1997) have a Hedges’ g value of 0.24 in their Appen-

dix, but the sign should be negative since the ethical funds performed worse than the non-

ethical funds.

Meta-analysis on SRI

©2019 Korean Securities Association 31

although many valuable conjectures can be made regarding AES. This point is espe-

cially important in the current situation where positive and negative effects coexist,

since focusing only on the level of ESs will average out positive and negativ e

impacts leading researchers to underestimate the impact of SRI. To our knowledge,

our paper is the first that investigates the influence of moderators on both ESs and

AESs.

Another reason for the lack of sound summary papers is the diversity and

complexity of existing studies with regard to samples, methodologies, perfor-

mance measures, investment universe, benchmarks, etc. A meta-analysis is a very

useful scientific tool to synthesize a wide variety of heterogeneous papers. It cal-

culates ES that can be interpreted as a standardized difference between perfor-

mances of SRI and conventional portfolios. By standardizing performance

differences, whatever the performance measures, a common measure of difference

between the two types of portfolios is generated, making synthesis possible across

very diverse papers.

Given the situation in SRI research, this article attempts to achieve the fol-

lowing goals. First, we provide a sound synthesis of existing literature which

shows an overall influence of SRI on investment performance in terms of sign

and magnitude through a meta-analysis. Then, we examine a variety of determi-

nants of SRI influence, since it is conceivable that the magnitude and direction

of SRI effects may have a systematic relationship with sample characteristics and

research methodologies. The analyses on moderators will help us to produce

interesting implications for future research and practice. When performing the

analysis, we carefully explain our meta-analysis and data collection process to

ensure replicability.

This study is organized as follows. In the next section, we develop hypotheses

regarding the overall impact of SRI and moderators that can have a systematic rela-

tionship with ESs and AESs. In Section 3, we explain the meta-analysis and the con-

cept of ES and show definitions of moderators used for hypothesis testing. In

Section 4, we present our sampling process in detail, since sampling is crucial in

performing and interpreting a meta-analysis. Section 5 reports the meta-analysis

results, and Section 6 presents conclusions and discussions.

2. Hypotheses

Our initial research interest is on the average impact of SRI on investment perfor-

mance. Thus, we first propose a hypothesis on the overall ES and then proceed to

establish hypotheses for moderators that may have a systematic relationship with

ESs and AESs. We attempt to employ moderators to reflect research characteristics

as extensively as possible based on theory and practice. We split the moderators

into two broad categories: sampling and methodology. Sampling relates to data col-

lection, and methodology is about analysis tools.

C.-S. Kim

32 ©2019 Korean Securities Association