The dialogue about the performance of socially responsible investments is full of mixed messages and mixed motivations. Recent developments and data on corporate governance are likely to bring on a fresh wave of intellectual dissonance.
Corporate governance has been kind of a stepchild for social investors. Although attention has been paid, and the most egregious violators are usually excluded from SRI portfolios, I would argue it has not been a central criterion for inclusion or exclusion - activist social investors have been more interested in things like South Africa, climate change, and Darfur, while less-activist ones have often delegated corporate governance due diligence to their money managers. Still, everyone says governance is important.
It doesn't help that we don't have a good consensus definition of corporate governance. ISS, Morningstar, and others provide rating services, but there are major differences in both definitions and ratings of individual companies. CalPERS has its own take, which is different again - it has managed to both add value and annoy Warren Buffett, another mixed message for returns-focused investors.
But governance has special significance because it's one of the social or quasi-social variables where there's been academic support for the idea of a performance benefit. It makes sense conceptually, and there's data. The best data so far cam from Gompers et al, which showed that companies with corporate governance they didn't like underperformed materially in the 1990s. For the most part, the Gompers governance metrics focused on takeover defenses - the more anti-takeover defenses, the lower your rating. This raised the question of what we were measuring because large acquirors (the firms most likely to have strong anti-takeover defenses) underperformed materially in the 1990s. But in the 2000s acquirors have done well, as takeover premiums became smaller and investors rejected big deals resembling the ugly stories of the 90s.
And simultaneously, a study by Empirical Research Partners ('Does Corporate Governance Matter?')finds that companies with high scores on the 'Gompers index' of corporate governance have significantly underperformed low-rated companies since 2000. Empirical is a strong sell-side quant shop, and I have no reason to doubt their results. Author Michael Goldstein (formerly a strategist for Sanford Bernstein) writes:
"There was, over the entire 15-year period, some payoff to good governance in holding periods beyond a year, but it wasn't impressive... We also considered whether the governance score could help us distinguish winners from losers among the best- and worst-ranked stocks in our core model and its earnings quality module. We found thre was an advantage conveyed by paying attention to governance when avoiding losers but only in the 1990s and not thereafter."
This is unpleasant news for governance advocates who have ridden the Gompers data a long way. Goldstein goes on to argue that governance shouldn't matter much for stock prices since it's easily observable and tends to change slowly. I disagree there - I don't know that we've perfected the art of measuring it, but one of the few points on which Ben Graham and Amy Domini would both agree that investing with a management you can't trust is a fool's game.