Trends in the Literature on Socially Responsible Investment: Looking for the Keys Under the Lamppost

In this paper, we use online search engines and archive collections to examine the popularity of socially responsible investing (SRI) in newspapers and academic journals. A simple content analysis suggests that most of the papers on SRI focus on financial performance. This profusion of research is somewhat puzzling as most of the studies used roughly the same methodology and obtained very similar results. So, why so many studies on SRI financial performance? We argue that the academic literature on SRI is mostly data-driven: the famous “Looking for the Keys Under the Lamppost” syndrome. The question of the financial performance of the SRI funds is certainly relevant, but maybe too much attention has been paid to this issue, whereas more research is needed on a conceptual and theoretical ground, in particular the aspirations of SRI investors, the relationship between regulation and SRI, as well as the assessment of extra-financial performances.


Introduction
Data, especially hard data, have become the lifeblood of social sciences. Kim et al. (2006) report that, in the early 1970s, 77% of the most highly cited articles in economics were theoretical and only 11% empirical. In the late 1990s, proportions were almost reversed: 60% were empirical and only 11% theoretical. Technical progress in collecting and processing quantitative information has obviously fueled the growth of empirical studies and we can be delighted with this. Nevertheless, this is not without some methodological issues.
From Anscombe (1973) to Angrist & Pischke (2010), scholars have consistently cautioned against the misuse of data. Most of the critics focus on econometric issues (multicollinearity, simultaneity, endogeneity, structural breaks, to name a few) and data mining -pros and cons.
Several articles are also concerned about how empirical evidence is interpreted and how data are used as a rhetorical tool (see Leamer 1983, Griliches 1985, McCloskey & Ziliak 1996, Shugan 2002, Kennedy 2002, Teixeira 2007, or Hoover & Siegler, 2008. In this paper, we illustrate some of the concerns about the use of data in academic research with the literature on socially responsible investing. In particular, we aim to draw attention on the extent to which research is data-driven. Socially responsible investing (SRI), also known as ethical investing, refers to the integration of environmental, social and corporate governance considerations (ESG) into the investment process. Responsible investors generally avoid investment in the so-called "sin stocks" (alcohol, gambling, tobacco, weapons, etc.), while they favor firms engaged in best practices with respect to environmental sustainability, labor conditions and community relations; they are also likely to promote shareholder engagement. Since the financial collapse caused by the subprime turmoil, SRI is sometimes even considered as an answer to the moral crisis of capitalism. In any case, SRI enjoys a large consensus: to conciliate finance and sustainable development. Doing well by doing good: the intention is undoubtedly admirable, though to such an extent that it may result in wishful thinking.
Although it is a relatively new concept, SRI has been the subject of a good deal of research. In this study, we provide quantitative evidence of a craze for SRI in the newspapers and the academic journals. We use online search engines and archive collections to examine the popularity of SRI in the public debate. We also rely on a content analysis of articles (in newspapers, books, academic journals and online sources) that deal with SRI. Our content analysis shows mainly that most of these articles focus on the performance of SRI, while few of them -and the portion declines over time -are concerned with ethics, altruism or moral values. Moreover, those studies on the performance of SRI are amongst the most influential papers, if we stick to the number of citations. This profusion of academic research on SRI financial performance is somewhat puzzling. Most of the studies used roughly the same methodology and obtained very similar results. So, why so many studies on financial performance? In this study, we argue that the main explanation is that the academic literature on SRI is mostly data-driven. Researchers 1 have been victim of the famous LKUL syndrome -"Looking for the Keys Under the Lamppost". 2 The question of the financial performance of the SRI funds is relevant, but maybe too much attention has been paid to this issue. The risk is that data-driven research can oust some fundamental, but hard, topics. Of course, nobody is able to mark out the most important topics. However, it can reasonably be said that ethical investing raises conceptual and theoretical issues which have not been fully addressed.

SRI in the news and the academic debate
In the last decade, socially responsible investing has earned its spurs. In the financial markets nowadays, everybody knows what the acronym SRI stands for -albeit some disagreements on the exact meaning of the "S" (socially or sustainable?) -and it seems that SRI has come of 1 Including the two authors of the present paper, Capelle-Blancard and Monjon (2011). 2 A woman sees a man looking for his keys under a lamppost outside a tunnel. She proposes to help him and she asks where he lost them. He lost them inside the tunnel. The woman looks puzzled. -"Then why are you looking for them under the lamppost?" -"Because I see much better here" he answers. age. A sure sign of this trend is the success of the United Nations Principles for Responsible Investment (PRI). As of March 2010, there were more than 700 signatories worldwide representing about US$ 18 trillion in assets under management.
Outstanding amounts managed according to SRI guidelines are usually considered as an indicator of its attractiveness (while there is often a tendency to over-emphasize the size or the growth of the SRI market; see Capelle-Blancard and Monjon, 2010). But it is also interesting to monitor directly the way the debate on SRI evolves. To do so, we propose to quantify SRI's prevalence in the public and the academic debates. Then, we examine the changes in the terminology associated with SRI.

Methodology
To proxy the popularity of SRI, we simply count the number of occurrences of the concept in newspapers and academic journals (as well as in books and on the web, albeit to a lesser extent). The count was conducted with automatic requests using various internet search engines. We chose to consider the following key wordings: "socially responsible investing", "socially responsible investment", "ethical investment" or "ethical investing".
Newspaper articles are collected from Dow Jones Factiva. This software covers all major newspapers and publications in the world; that is, more than 10,000 news sources including major publications such as The Wall Street Journal, The Financial Times, etc. We search for academic articles using online archive collections provided by ScienceDirect-Elsevier, Wiley Interscience, SpringerLink, and Jstor. These platforms give access to the most influential academic journals worldwide. We search in the full-text (all fields), but restrict our query to journal articles in Business, Economics, and Finance; we also avoid double-counting. Lastly, we use the Google search engine to run queries on the web, as well as Google Books and Google Scholar (as of May 2010). Results obtained with Google are used mainly in this section and hereafter only as robustness check.
Over the period 1982-2009 3 , we identify approximately 513,000 webpages, 27,500 newspaper articles and 673 academic journal articles 4 which include the phrases "socially responsible investing" or "socially responsible investment" or "ethical investment" or "ethical investing".
We also obtained 28,200 results in Google Books and 11,000 results in Google Scholar.
Clearly, one may argue that our results overestimate the number of articles genuinely devoted to SRI, but it is not crucial to our purpose as we are mostly interested by the trends. However, if we want to draw comparisons, we need to take sample size effects into account. 5 To do so, we regularize the number of articles mentioning SRI with the number of articles mentioning the words "investing" or "investment". The purpose is to proxy the percentage of articles dealing with SRI amongst the bulk of articles about investing/investment. Thus, about 25,000,000 newspapers articles and 150,000 academic journal articles were identified including the words "investing" or "investment" over the whole period. Hence, our ratios for SRI newspapers articles and SRI academic articles are equal to 0.11% and 0.44%, respectively. The suitable ratios are equal to 0.19% for webpages, 0.29% for Google Books and 0.81% for Google Scholar. Results are presented  etc. 5 What we call the sample size effects are twofold. First, we need to compare the different sources of information. But web audience is larger than newspapers audience, which is itself larger than academic audience; consequently assessment might be distorted. Second, the number of articles available each year in the archives is growing, which might skew inter-temporal comparisons.
Note: The number of SRI articles is obtained by running the following query: "socially responsible investing" or "socially responsible investment" or "ethical investment" or "ethical investing" for the period 1982-2009. The percentage is computed as the number of SRI articles divided by the number of articles with the words "investing" or "investment". Web articles are collected from Google. Books occurrences are collected from Google Books. Newspaper articles are collected from Factiva (Dow Jones). Academic articles are collected either from a) Google Scholar or b) ScienceDirect-Elsevier, Wiley Interscience, SpringerLink and Jstor (only journal articles are considered).
As an opening remark, we can say that scholars seems relatively more concerned by SRI issues than others, as testified by the highest percentage obtained with academic journals or Google Scholar (differences compared to the others are highly significant). However, this is somewhat unexpected. Two features make SRI particularly attractive to journalists: its novelty and the fact that it offers unusual angles to talk about financial products (Winnet and Lewis, 2000). Such features may also be appealing for scholars, but in a much lower extent a priori. To explain the large number of SRI articles in academic journals, we need to further examine their content.

The growth of the SRI literature
In this subsection, we provide quantitative evidence of the prevalence of the SRI concept in the public debate over the period 1982-2009. Our queries confirm the craze for SRI, both in relative and absolute terms (see Figure 1).  1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000  2001  2002  2003  2004  2005  2006  2007  2008  2009 # Academic Articles -LHS Ratio Academic Articles (x 100,000) -RHS Note: The number of SRI articles is obtained by running the following query: "socially responsible investing" or "socially responsible investment" or "ethical investment" or "ethical investing". The ratio is computed as the number of SRI articles divided by the number of articles with the words "investing" or "investment" and multiplied by 100,000. Newspaper articles are collected from Factiva (Dow Jones). Academic articles are collected from ScienceDirect-Elsevier, Wiley Interscience, SpringerLink and Jstor (only journal articles are considered).

A simple content analysis
The number of papers dealing with SRI is growing. But what do they talk about? To try to identify the most favored topics, the points of view, possibly the prejudice, etc. is not an easy task. Besides, this kind of analysis is rather uncommon in economics. 6 This is partly due to the subjective nature of such analysis. It is a priori non-formal, hardly quantifiable, and consequently barely refutable. In this section, we try to overcome these challenges by carrying out a simple quantitative content analysis of our large sample of SRI articles. By doing so, we aim to contribute to a better understanding of the opinions and attitudes of journalists and scholars vis-à-vis SRI.
Content analysis is an increasingly important research tool in social sciences, but not very common in economics. This is a set of techniques used to determine the presence of certain words or concepts within a collection of texts. Formally, content analysis can be defined as a "technique for making inferences by objectively and systematically identifying specified characteristics of messages" (Holsti 1969). 7 In this section, we first provide evidence of the growing reluctance to use the qualifier "ethical" to the benefit of the less connoted term "socially responsible". Then, we discuss the current drift towards "sustainability". Finally, we attempt to identify the main topics currently addressed in the SRI articles.

Ethical vs. Socially Responsible
The qualifiers "ethical" or "socially responsible" are sometimes considered as perfectly synonymous, including in the dedicated academic journals where one might expect that the concepts are clearly defined. Thus, for instance, according to Schueth (2003: 189): "the terms social investing, socially responsible investing, ethical investing, socially aware investing, socially conscious investing, green investing, value-based investing, and mission-based or mission-related investing all refer to the same general process and are often used interchangeably". Likewise, Hellsten & Mallin (2006: 393) "use the terms 'ethical investments' and 'socially responsible investments' interchangeably". Until now, we do not distinguish between the two terms. But clearly, the terminology is not neutral.
As stated by Sparkes & Cowton (2004: 46), "as time has passed the term 'ethical investment' has increasingly been replaced by that of 'socially responsible investment' (…)".
Accordingly, we posit a progressive decline in the use of the term "ethics" within the SRI debate. To test formally this hypothesis, we search within our sample of SRI articles (reminder: 27,500 newspaper articles and 673 academic journal articles), the percentage of articles including the word "ethics". Results are presented in Figure 2.

Figure 2. Ethics and SRI
Percentage of articles about SRI which mention the term "ethics" Note: The number of articles is obtained by running the following query: ("socially responsible investing" or "socially responsible investment" or "ethical investment" or "ethical investing") and "ethics". Newspaper articles are collected from Factiva (Dow Jones This argument has been precisely documented by Sparkes (2001) and Sparkes & Cowton (2004). The most common acceptation today is that ethical investing refers to negative screening (that is, the exclusion of "sin stocks"), while socially responsible investing encompasses the integration of the ESG criteria, the best-in-class approach and shareholder activism, which appear to be at the cutting-edge of SRI.
The previous argument is eventually relevant to understand the attitude of journalists who convey investors' messages, but it does not fully explain why scholars themselves are reluctant to use the word "ethics". One possible interpretation lies in the way research in economics has developed since centuries. Since Bernard de Mandeville and "The Fable of the Bees" published in 1705, economics has been gradually emancipated from moral sciences.
Economic research is deeply affected by this initial demarcation: for some, it is nothing less than the very justification for its scientific status. Those who endorsed this position considered that the relevance of economics is purely technical. This approach is regularly and vigorously criticized (Sen 1987, Hausman & McPherson 1993), but it is still dominant, albeit in less extreme versions. Therefore, if we suppose that researchers have strong incentives to comply with their peers who consider, more or less, that economists should not be interested by moral questions, it is clear that researchers will prefer to talk about SRI, rather than ethical investment.

The Semantics of Responsible Investing: An Ongoing Debate
We focus previously on the semantic shift from "ethical investment" towards "socially responsible investment". However, the terminology -and so the significance -continue to change: the disputes now are on the exact meaning of the "S" (Socially or Sustainable?), not to mention, its suitability. To sum up, we can give a last -but very symptomatic -example of the semantic drift by considering how things evolved in Australia in the last ten years. In 1999, some Australian professionals founded the Ethical Investment Association. In 2002, they launched their first annual survey entitled "Socially Responsible Investment in Australia", while maintaining the name of their association. 9 But in 2007, the terms "ethical investment" and "socially responsible investment" were finally discarded in favor of "responsible investment": the association is now called the Responsible Investment Association Australasia. 8 According to the US SIF CEO, Lisa Woll, the name-change process was embarked "with trepidation" (see the interview by Mark Nicholls, Environmental Finance, June 22, 2011). 9 It should be noted that Sparkes & Cowton (2004) considered that Australia was an exception, because they "generally kept the older usage" -that is, the term "ethical investment".

What is in question in SRI articles?
If papers on SRI do not deal with ethics, what do they talk about? To put it differently, within SRI articles, what are the most common topics addressed by journalists and scholars? To answer this question, we search for the existence 10 of a certain number of concepts within our collection of texts on SRI in newspapers and academic journals.
Most quantitative content analysis uses predefined word categories based on specialized dictionaries (such as Harvard or Lasswell) to limit subjectivity. But our approach is different since we are looking for concepts dedicated to SRI; consequently, we need to consider specific terms. So, we proceed in two steps. First, we list a set of words likely to be representative of the concepts relevant to the SRI debate. Then, we identify the percentage of SRI articles including those words.
Fifteen words have been chosen which can be categorized in four ad-hoc groups. The first group is about personal values: "altruism", "sacrifice", "moral", "religion". The second group deals with financial characteristics: "performance", "diversification". The third group refers to the SRI strategy: "best-in-class", "filter" or "screen", "activism", "stakeholder". The fourth group involves some proxies for the environmental, social and governance (ESG) factors: "South Africa", "human rights", "climate", "sustainable", "corporate governance". 11,12 Results are documented in Table 2. 10 An alternative approach would have been to count the number of times each concept appears. Such an approach might be more indicative (most of the time). However, it is not applicable with a very broad collection of texts, as it is the case in this study. 11 Several attempts were necessary to draw up this list. We based our choice on two criteria: relevance, but also the number of occurrences. In particular, the words "pro-social" or "community" are too seldom used to be incorporated in the analysis. 12 One may argue again that our method overestimates the number of studies truly related to a specific concept. But, as the bias is likely the same for all the words included in our list, we believe that it is not a decisive issue. The main striking result is the predominance of the term "performance" in all kinds of publications. Amongst the terms on the list, it is the most used -far beyond the others. In newspapers, this term appears in one third of the articles. In academic journals, it was mentioned in half of the papers over the period 1982-1999, and this percentage has increased to almost three quarters over the period 2000-2009.
Such a high proportion of appearances for the term "performance" (at least, relative to the others terms) is somewhat surprising. However, it confirms the qualitative analysis provided by Winnett & Lewis (2000) on the coverage of ethical investment (they focus on the UK broadsheet Sunday press over the period [1994][1995]. According to them, a representative press article on SRI begins with some definitions, followed by a presentation of the different strategies and it typically ends with a comparison of performances. In truth, we expected that journalists would devote more to the issue of performance than scholars. Indeed, as stated by Winnett & Lewis (2000), league tables are the mainstay of financial journalism; they foster the hope that gains can be realized by using relevant insights and by careful screening. However, unlike "popular" models, "economic" models deny the belief that it is possible to "beat the market". In this regard, the fact that scholars attach even more importance than journalists to fund performance is quite remarkable.
To corroborate the primacy of the issue of performance in studies on SRI, we also rely on a manual count of the academic articles that specifically examine the financial performance of SRI indexes or SRI mutual funds. The conclusion is clear-cut: in a companion paper, we list more than fifty academic papers that specifically examine this issue (Identifying reference To have such a number of academic articles devoted to SRI financial performance is somewhat puzzling. Of course, this is an important issue given the fiduciary responsibility of the fund managers. Of course, replication of published empirical results is a fundamental pillar of the scientific method (Dewald et al. 1986, Hamermesh 2007. But these arguments are not sufficient.
Recall first that the rationale of SRI funds is to go beyond the financial aspects. Accordingly, it is surprising that the main question in the academic literature is whether it pays.
Furthermore, most of the studies used roughly the same methodology 15 and obtained very similar results on (seemingly) different samples. They almost unanimously show that the financial performance of the SRI funds is not significantly different relative to their conventional peers or relative to a benchmark index. Besides, this empirical result is widely expected. 16 The availability of the data and the easiness of access, together with the opportunity to make use of econometric tests clearly play a major role in the choice of questions that are dealt with by scholars. Obviously, it is almost impossible to measure the extent of such behavior.
Notwithstanding, we assert that the most plausible explanation to the high number of papers on financial performance of SRI funds is that this strand of research is mostly data-driven. 17 Researches on SRI have been victim of the famous LKUL syndrome -"Looking for the Keys Under the Lamppost": a general tendency to over-analyze data. Following Kimball (1957), Kennedy (2002), in its ten commandments of applied econometrics, called it a Type III error: to produce the right answer to a wrong question. Truly, the question of the financial performance of the SRI funds is not out of place. But maybe economists have paid too much attention to this issue.
Aside from the question of financial performance, we also document four notable trends concerning the topics addressed in our set of SRI articles (overall, the trends are the same in newspapers and academic journals). 18 First, it is clear that the words connected with personal values ("altruism", "sacrifice", "moral", "religion") are used less and less frequently, in the 15 They used the CAPM or a multifactor model to assess the risk-adjusted return of SRI funds. 16 And yet, newspapers articles persistently present contradictory evidence based on specific case studies. For examples of "selective use of evidence", see Winnett & Lewis (2000: 333). 17 Again, while Hoepner & McMillan (2009) use a different methodological approach, they reach a similar conclusion: "many influential analyses of responsible investment's process appear focused on aspects with sufficient data for statistical analysis, which might supply more developmental or exploratory research tasks with insufficient attention from prospective influential authors". 18 To investigate how the percentages significantly evolved over time, we use the following Z-statistics: Z = (p 1 -p 2 ) / √ [p 1 (1 -p 1 )/n 1 + p 2 (1 -p 2 )/n 2 ] where p 1 and p 2 are the sample proportions and n 1 and n 2 are the sample sizes corresponding to the proportions. This test is especially informative when the samples are large and the percentages are far from zero. The asymptotic distribution of the Z statistic is standard normal under the null hypothesis of no change in the percentage. same manner as the word "ethics". Second, there is a growing use of the term "best-in-class" in the newspapers, together with a lesser use of the words "filter" or "screen". This was anticipated since it reflects a change in the asset managers' practices. Third, the term "stakeholder" appears to be very popular amongst scholars: almost two third of the academic papers examined mention this concept (the second most-used in our list). Fourth, along with a renewal of societal issues, SRI papers mention less frequently the boycott of South Africa and instead cite human rights, climate change, and sustainable development problems with increased frequency. The topic of corporate governance is also more and more prevalent.

Conclusion
"Not everything that counts can be counted, and not everything that can be counted counts" William Bruce Cameron (1963: 13). 19 Advocates of SRI usually consider that good financial performance is likely to promote SRI (the "green-is-more-profitable" argument). But there is eventually an element of delusion here. Moreover, economic incentives can have many perverse effects, in particular that of discouraging "pro-social" behavior (Bénabou & Tirole, 2010). Altruism, reputation or selfesteem can be the powerful motives which lead people to be socially responsible investors.
Actually, few studies examine the aspirations of SRI investors, but they represent a valuable source of information. 20 Likewise, studies on the relationship between SRI and regulation, as well as empirical evidence on extra-financial performances seem to be promising avenues, to name a few.
Overall, the question of the financial performance of the SRI funds is certainly relevant, but maybe too much attention has been paid to this issue, whereas more research is needed on a conceptual and theoretical ground. The motto of W.B. Cameron, often endorsed by the proponents of SRI, may also apply to research.