Measuring the Cost of Socially Responsible Investing

Correctly understanding Kritzman’s argument

Kritzman was at some pains to make clear that he does not necessarily disagree with the social objectives of those who would have portfolios managed in a socially responsible fashion. He takes seriously, in particular, the environmental consequences of reliance on fossil fuels. His concern, rather, is that advocates of SRI acknowledge the costs that he claims the portfolios would incur, and then make an informed decision of whether these costs are justified by the ends. As he wrote in an op-ed essay in the Chronicle of Higher Education:

His investment argument is controversial, which is why he made clear his position favoring the ethical objectives of socially responsible investing, clearer in his talk even than in the preceding statement. He said that at a number of recent speaking occasions, he has seen at firsthand how his paper has riled some readers and listeners, including Bill McKibben, who has been a prominent spokesman for the movement to divest endowments of the stocks of fossil fuel companies, and with whom he participated in a panel at Middlebury College. (You can see McKibben’s comment on Kritzman’s essay in the Chronicle of Higher Education by following the link above.) Kritzman explained that many of the complaints against his and Adler’s paper are predicated on a misconception.

Kritzman began his argument by clarifying the difference between SRI and active management. The distinction is critical to his argument. He adopted Langbein and Posner’s definition of SRI, which he quoted: It is “excluding the securities of otherwise attractive companies from an investor’s portfolio because the companies are judged to be socially irresponsible, and including the securities of certain otherwise unattractive companies because they are judged to be behaving in a socially laudable way.”

So, Kritzman said, if an investor invests in socially responsible companies only because he expects them to perform better than the market as a whole, and doesn’t invest in socially irresponsible companies because he expects them to underperform, then what he is doing isn’t distinctively SRI; he’s just engaging in normal active (as distinct from passive) management. Active managers, by definition, invest in the stocks of companies that they expect to have superior investment performance and avoid companies that they expect to underperform. If, in contrast, the investor invests in socially responsible companies that he expects to underperform the market and does not invest in socially irresponsible companies that he believes will outperform the market (and even if he expects that, on average, good companies will outperform bad companies), then he is engaging in socially responsible investing.

“You cannot have it both ways,” he said, meaning that you can’t claim to be carrying out an intention to be socially responsible in your investing if you are only picking companies that you believe will have the best returns. The issue of cost arises for socially responsible (active) investors because they are forgoing better returns that they believe they could otherwise have achieved, and these lost returns are a cost. These costs may be negligible, he said, before presenting his computations, but there has to be a cost.

Kritzman’s argument has no bearing on socially responsible passive investing, because he is not claiming that the broad market, excluding socially irresponsible companies, will underperform a broader market index that includes them. It might or it might not, but that is more an empirical issue, and the methodology of his demonstration cannot address that question. A failure to recognize this distinction between active and passive socially responsible investing is the principal misconception that underlies many of the criticisms of his paper. (And it is why the article that McKibben referenced in his comment, cited above, does not vitiate Kritzman’s argument.)

In his paper, Kritzman made no allusion to his opinion, stated elsewhere, that active management through security selection is generally a waste of effort, because of the efficiency of markets. As he stated in a recent guest editorial in the Financial Analysts Journal, “there is … overwhelming empirical support for [Samuelson’s] thesis of micro-efficiency’ (which is efficiency at the level of individual securities, as distinct from the pricing of asset classes).² In light of this, Kritzman’s discussion of the cost of SRI, although he talked of it as real, might be understood from his perspective as being very much hypothetical. If active management can’t produce superior returns, then it follows from his argument that socially responsible investing won’t incur a cost. He conceded this only during the question-and-answer session, when a listener familiar with his writings raised the matter.

Kritzman pointed out yet another misconception about his argument. Some have taken him to task because he has said that by restricting the universe from which the manager can select investments, the manager loses the possibility of acquiring stocks with superior returns. Their objection is that (except in the case of passive investors), managers draw from restricted universes all the time, because they choose not to invest in, say, Nigerian stocks or Asian currencies. Kritzman countered that this misrepresents his usage of the word “restrict.” Active managers almost always restrict their universes to groups of stocks in which they have skill. He is not arguing that that incurs a cost. The cost is incurred when the universe of stocks in which the manager has skill is restricted.

Quantifying the cost of SRI

Kritzman and Adler used Monte Carlo simulations both to validate their argument and to try to quantify the cost of socially responsible active management. As Kritzman said, you can convince yourself from first principles that there is a cost, but to get a sense of the size of the cost, and how it varies with manager skill, and with the size of the portfolio, and with the dispersion of returns within the universe, you have to run experiments. In outline, their method depends on modeling the returns to portfolios of very large numbers of stocks (as would be held by institutional portfolios), assuming normal distributions of returns within the universes, which are randomized, and, over different simulations, assuming different levels of analyst foresight, all, for the sake of realism, fairly low. The simulations do not use actual security returns, only these hypothetical returns, which are based upon the cross-sectional standard deviations of returns of the various universes they tested, the S&P 500, MSCI EAFE, and others. They then restricted the universes randomly, to model the restrictions that would arise from the constraints of SRI, assuming that 10%, 20%, and 30% of the securities were eliminated. And they ran these simulations over spans of five, ten, and twenty years. They ran 10,000 simulations each time.

Their results demonstrated that, the greater the restriction, the greater the cost to the portfolio, the greater the skill of the manager, the greater the cost, and the greater the dispersion of returns within the universe, the greater the cost.

The following table presents a subset of their results, assuming a portfolio with a beginning value of $1 billion, an average annual market return of 8.0%, and the ability of the portfolio manager to correctly rank 52% of the stocks in advance by their subsequent returns.

Multi-Year Cost of Socially Responsible Investing:

100-Stock Portfolios, 52% Correct Choices, 20% Excluded

Kritzman said that there have been a number of studies claiming that there is no cost, or only a very low cost, to SRI. His objection to these studies, often adduced in opposition to his argument, is that they rely on historical data, and so reflect just the particular period of the study, which can’t be taken as representative of the future. That, he said, is like saying, “Last year, on July 15th, it rained, so therefore I won’t plan any outdoor activities today.” He and Adler were using simulations to model what could happen, not what did happen.

The costs in the table are very large, and Kritzman backed off them to some extent under questioning at the end of his talk, when his interlocutors objected to some of the assumptions that underlay the simulations, for example, the large proportion of the securities excluded through restriction. At that point, Kritzman emphasized that he and Adler had presented a methodology, and he said that anyone could re-run the analysis using assumptions that they considered more realistic.

Another point that was raised not so much to object to his argument but rather to temper its interpretation was that the costs that Kritzman and Adler computed are expected costs. They arise from probabilistic distributions of future results. But stocks will have only one future between any two dates, and just as in some studies based historical data, there was no cost to SRI restrictions, so, too, it may turn out that there will be no cost in the future. Or the costs could be greater.

Kritzman also conceded that he was not looking at how SRI restrictions would alter the risk exposures of a portfolio, a matter that many analysts consider to be of much greater interest than the costs. But he countered that this was beside the point he was making.

Three options for SRI investors

Kritzman concluded his presentation by saying that, in light of his and Adler’s results, those who advocate SRI should consider three options.

The first option is to proceed as SRI advocates wish, and restrict the investment universe. By not investing in the bad companies, you might raise their cost of capital, and thereby induce them to reform. He said, though, that because all of these endowments together are small in comparison with the total market capitalization of the bad companies, he guessed that this would be unlikely. Alternatively, drawing attention to a company’s bad behavior might influence consumers to avoid its products and services and thereby induce it to reform. Or the attention might, through moral suasion, induce the company to change. He said, though, that he thought that divestment is largely symbolic, which might by itself be good.

The second option is not to restrict the investment universe. Continue to permit investment in these bad companies, and rather than incur the cost of restricting the universe, deploy the saved money however you like, which might include improving the carbon footprint of the university.

The third option is also not to restrict the investment universe, but instead to use the money that is saved by investing it toward environmental ends, such as the exploitation of alternative energy sources, or the development of policies to curb the consumption of fossil fuels, or maybe even shareholder activism to alter bad company behavior.

Adam Jared Apt, CFA, is a financial advisor and the owner of Peabody River Asset Management, based in Cambridge, MA.

Read more articles by Adam Jared Apt