Suppose that there's a trade-off between profitability and CSR mandates. Why might this be the case? Any CSR initiatives that increased profitability would have already been adopted under normal profit-seeking strategies.
Managers want to think themselves good people and enjoy doing CSR things. They get approbation from others for doing so. They'll get in trouble with shareholders if profitability drops too much, so they won't do the nuttiest things. But they'll take some CSR as consumption.
NBER points to some relevant evidence. From the abstract of Cheng, Hong and Shue's new paper:
We find support for two key predictions of an agency theory of unproductive corporate social responsibility. First, increasing managerial ownership decreases measures of firm goodness. We use the 2003 Dividend Tax Cut to increase after-tax insider ownership. Firms with moderate levels of insider ownership cut goodness by more than firms with low levels (where the tax cut has no effect) and high levels (where agency is less of an issue). Second, increasing monitoring reduces corporate goodness. A regression discontinuity design of close votes around the 50% cut-off finds that passage of shareholder governance proposals leads to slower growth in goodness.In short, the more skin that decision-makers have in the game, the less they go for CSR policies. And the more that those with skin in the game can monitor the decision-makers, the less they go for CSR policies. It's slack in the principal-agent relationship that allows managerial consumption of "do-good" benefits.
Sometimes, we need Larry the Liquidator.