Philippe Aurain, chief executive and CIO at Federis, lists four reasons why the usual definition of SRI is unusually slippery.

Considering the usual definition of socially responsible investment (SRI) as "taking into account, in a systematic way, extra financial data concerning issuers in order to select securities for financial portfolio management", we have to recognise that we are left with a very general principle.

That definition doesn't tell us anything about how to deal with the data, or, more importantly, what is the objective of SRI. For these reasons, I tend to consider SRI as a fuzzy concept that even the same people would give a different meaning depending on time and circumstances.

There are three main issues when considering SRI: defining objectives, aligning financial strategies with theses objectives and understanding the impact of these choices.

Let's begin with the first issue: defining the objective. The point is that there is not one single objective in investors' mind when they consider SRI, there are at least two. And, unfortunately, the two of them are independent, which makes it very difficult to attain them simultaneously.

The first kind of objective, that we call identity-driven objective (IDO), reflects the idea that investors are keen to invest in assets embedding values or externalities matching their own values, political views or reputational concerns. Typical IDO filters are arms, GMO, pornography, games or nuclear industry exclusions. They could include 'social' approaches like 'community welfare', 'gender parity' or 'employee welfare' and so on. This kind of objective is, of course, totally legitimate, as investors are the most appropriate agents for determining what matters to them.

This said, they should also recognise that there is absolutely no link between their preferences and the expected returns of securities that meet these criteria. The performance of excluded sectors is random as confirmed by most studies. The only reason why the exclusion could introduce a negative bias on performance would be if a sufficient proportion of investors excluded the same sectors for extended periods of time, thus impacting their performance. As investors have different identities and values, they tend to be diversified in the way they filter the investment universe on an identity-driven objective, which leads to a lack of correlation between sectors' exclusions and their performances.

The second kind of objective, which we call Financial SRI (FSRI), is designed to add value by incorporating extra financial filters. The idea is that taking into account the information that is usually ignored by traditional financial analysts gives a competitive edge in security selection. Of course, to be efficient, FSRI has to be defined in a way consistent with the objective, meaning that an extra financial filter should not be chosen depending on investors' preferences but rather based on its measured capacity to provide meaningful insights for portfolio management – in other words, a better risk/reward profile. This also implies that FSRI should be supported by quantitative research of 'material' impact on the investment process.

Consequently, once these two objectives are defined, the investor is confronted with the sad truth that they are indeed 'independent'. That doesn't mean that IDO will systematically underperform but rather that its performance is quite unpredictable and cannot be used as a foundation of an investment process seeking performance.

Unfortunately, many investors do not recognise the independence of these objectives and therefore expect better returns from securities that respect SRI criteria as a general rule. It is crucially important that, first, investors separate the two kinds of objectives, and second, that they design specific investment strategies that extract value from extra financial data in order to fulfil the second objective. And this is a difficult task as the relationship between SRI criteria and performance and risk is difficult to evidence. If the definition 'identity-driven objective' should not be outsourced to an asset manager, the financially driven is by nature part of asset managers' jobs in their daily quest for value.

Finally, the investment management industry will have to improve the performance reporting process. Current SRI reporting standards do not provide any distinction between the two kinds of objectives. Moreover, the reports usually don't even report any specific extra-financial data, disclosing only a global score that tells nothing about the issuers' SRI behaviour but rather gives a relative assessment based on an aggregated rating.

To conclude, we can claim that SRI is definitely a fuzzy concept for four reasons. First, each investor has his own definition of IDO, and, hence, SRI cannot be defined as "content" but rather a "family of process". Second, investors do not make a distinction in their investment objectives between IDO and FSRI, which leads to false return expectations. Third, investors do not design investment management processes according to these different objectives. And fourth, the reporting standards currently in use do not reflect the different natures of the two objectives.

Philippe Aurain is chief executive and CIO at Federis GA and a former financial director at France's FRR.