Corporate Social Responsibility and Imperfect Regulatory Oversight: Theory and Evidence from Greenhouse Gas Emissions Disclosures

September 2020 marked the 50th anniversary of Milton Friedman’s famous New York Times Magazine article, in which he summarized and expanded on an earlier argument that “there is one and only one social responsibility of business—to use its resources and engage in activities designed to increase its profits so long as it stays within the rules of the game” (Milton Friedman, Capitalism and Freedom, p.133). This argument was controversial then and remains so today. In his view, it is the regulator’s job to ensure the appropriate behavior of profit-maximizing firms by setting proper rules and regulations, rather than firms’ responsibility to determine and implement their notion of socially responsible behavior. Implicit in this argument is the assumption that the government or regulatory body is able to ensure appropriate behavior by firms. But what happens if the regulator is unable to monitor firm behavior effectively? What if the regulator doesn’t know whether the firm is staying within the rules of the game?

To put the question slightly differently, would a self-interested, profit-maximizing entrepreneur ever find it optimal to engage in corporate social responsibility (CSR) when regulatory oversight was imperfect? More generally, how is the quality of oversight connected to the prevalence of CSR? How does this change how we view Friedman’s admonition?

We take up these questions in Corporate Social Responsibility and Imperfect Regulatory Oversight: Theory and Evidence from Greenhouse Gas Emissions Disclosures, where we develop and test a simple economic model in which a self-interested entrepreneur, a socially responsible worker and a regulator interact. The entrepreneur operates a technology that unavoidably imposes negative externalities (for example, pollution) on the rest of society, but can select an action that affects the degree of pollution that occurs. The regulator may choose to monitor the firm and set a regulatory ceiling that balances the firm’s profits against the social costs of these negative externalities; or may forfeit oversight altogether if the benefits from regulation are more than offset by the monitoring costs.

The entrepreneur has two main decisions to make. First, she decides whether to ignore the environmental consequences of her actions and manage her firm herself as a pure profit-maximizing organization or instead to hire the responsible worker to manage her firm as a purpose-driven organization. The second decision is what action to take, whether she implements the action herself or elicits it from the responsible worker, depending on the organizational form she chooses. This action choice may involve complying with the regulation and selecting an action at or below the regulatory ceiling; or it may involve “cheating” by selecting an action above the ceiling, anticipating that non-compliance will only be caught with some probability that reflects the quality of regulatory oversight.

Two central results emerge from the theoretical analysis. First, hiring the responsible worker to manage the firm as a purpose-driven organization is always strictly optimal for the entrepreneur. The key is that in equilibrium the entrepreneur can invariably commit to operate her firm in a CSR-free manner. This commitment allows her to hire the responsible worker, let him enjoy extra utility by selecting an action that produces fewer externalities than the action that would be selected in a pure profit-maximizing organization, and then capture this surplus through lower wages.

Of course, in our model workers are simply a convenient embodiment of the social preferences toward environmental action that affect the entrepreneur’s optimal action. The same basic message would apply if the entrepreneur were interacting with other stakeholders (e.g. investors, suppliers), taking into consideration social preferences for the negative externalities associated with production. The key insight is that the profit motive combined with the bargaining power to extract rents causes the self-interested, agnostic entrepreneur to internalize pro-social preferences. In effect, by pursuing the very profit motive for which Friedman advocated, the entrepreneur becomes pro-social herself, creating the purpose-driven organization in the process.

The second key result concerns the relation between regulatory oversight and CSR. When the regulator possesses the ability to monitor firm compliance with sufficient effectiveness, she can ensure that the firm complies with regulatory standards that approach the socially optimal level. In an environment with strict and enforceable standards, the optimal choice for the purpose-driven organization is to exactly comply with the regulatory ceiling. Thus in this case the firm precisely follows Friedman’s dictum: it acts within the law but at the limit of what the law allows—it does not engage in CSR. In a world with sufficiently effective regulatory oversight, CSR would hardly exist.

In contrast, when the regulator’s monitoring technology is insufficiently effective, she must adopt—in order to ensure firm compliance—a regulatory threshold that is so lax that the benefits from oversight are outweighed by the costs of monitoring. In these instances the effective regulatory ceiling vanishes, and the entrepreneur has the option to elicit the pure profit-maximizing action from the worker in the purpose-driven organization. Instead the entrepreneur elicits an action strictly lower than the pure profit-maximizing action, in order to extract rents from the responsible worker through lower wages, as discussed above. Indeed in those cases the purpose-driven organization does engage in CSR through self-regulation, producing strictly fewer externalities than pure profit-maximization would require. Thus, CSR emerges when the firm’s profit motive causes it to have a comparative advantage over the regulator in reducing negative externalities.

The key prediction of our model is that all else equal, the effectiveness of oversight should be negatively related to the level of CSR that firms adopt. To evaluate the prediction, we develop two sets of empirical tests based on plausibly exogenous variation in regulatory oversight. Our first set of tests follows Philipp Krüger’s 2015 working paper, “Climate Change and Firm Valuation: Evidence from a Quasi-Natural Experiment” and exploits a shock in UK reporting standards surrounding greenhouse gas (GHG) emissions. In 2012, the UK government imposed a mandatory GHG emissions disclosure policy on all public firms operating in the UK. This corresponds to an increase in oversight in our environment. Using a difference-in-difference strategy in which we compare UK with similar non-UK firms before and after the policy ruling, we find that after the mandatory disclosure policy, UK firms on average had lower corporate social responsibility ratings compared to firms from the other 15 European countries, which did not have a mandatory disclosure policy in place. The negative and significant effect of mandatory disclosure on CSR is consistent with the main prediction of our model, and robust to a large number of robustness checks.

Our second set of tests uses US data and relies on the fact that changes in the degree of outsourcing across industries have differentially changed regulators’ monitoring ability. The central argument is that it may be more difficult for a regulator to monitor firms that offshore a large fraction of their activities; and that because these firms face less oversight they may choose higher levels of CSR. Here again we find evidence consistent with the model: our results show a positive relationship between industry-level foreign outsourcing and firms’ CSR activities.

Thus, in sum, the evidence from changing greenhouse gas emissions standards supports the predictions of the model, and points to CSR as a strategic policy that firms adopt when a regulatory vacuum is created by imperfect regulatory oversight.

The complete paper is available for download here.