By Alex Edmans, London Business School; Doron Levit, University of Washington; and Jan Schneemeier, Indiana University



Responsible investing – the practice of incorporating environmental, social, and governance (ESG) factors into investment decisions – is becoming increasingly mainstream. For example, when the PRI launched in 2006, it had 63 signatories managing a total of US$6.5 trillion. By the end of 2021, this had grown to 4,375 signatories, representing US$121 trillion.

Blanket exclusion of brown industries – such as tobacco, gambling, and fossil fuels – is often seen as the purest and most effective form of responsible investing. Divesting starves these industries of capital, the argument goes, preventing them from creating further harm. Accordingly, practitioners and the public hold investors accountable for their holdings of brown firms and call for them to commit to their divestment.

But divestment does not actually deprive a company of capital. An investor can only sell if someone else buys. Divestment is often compared to customer boycotts, but this is a false analogy: boycotts do deprive a company of revenue if no-one else steps in. A more nuanced argument is that divestment does not defund a company immediately but lowers the share price and makes it harder to sell shares in the future.

The limits of blanket exclusion

Even if this is the case, our new paper, Socially Responsible Divestment, shows that driving down the stock price as much as possible may not be the optimal strategy. We show that tilting (leaning away from a brown sector but being willing to hold the most sustainable companies in that sector) may be more effective than exclusion (shunning the sector outright).

Central to our results is the observation that brown companies can take corrective actions that minimise the harm they create – for example, fossil fuel firms can develop renewable energy. The problem with blanket exclusion is that it provides no incentives for such companies to implement corrective actions: even if it invests in renewables, a fossil fuel company will still be classed as a fossil fuel company and be divested. In contrast, since a tilting investor is willing to buy a fossil fuel firm if it is best-in-class, this motivates the firm to ensure that it is indeed best-in-class.

There is a close analogy with executive pay. Many investors support linking pay to ESG performance to provide CEOs with incentives to improve sustainability. However, blanket exclusion means that their capital flows are independent of ESG performance and provides no incentives to bolster it.

We build a model in which responsible investment affects social welfare through both channels: it lowers the stock price and affects incentives. There is a single firm that emits negative externalities. The firm’s manager can take a corrective action that reduces both externalities and firm value. The firm also raises capital that it uses to fund an expansion, increasing both firm value and externalities.

We show that divestment is most effective at starving a company of capital and hindering expansion, but tilting is more powerful at inducing the corrective action. The optimal strategy is a trade-off between these two forces and involves tilting if the corrective action is particularly effective at reducing the externality, because this consideration dominates the trade-off. This result suggests that exclusion may be optimal for industries such as controversial weapons, where it is relatively difficult to reduce the harm produced. In contrast, tilting may be preferred for fossil fuels, where firm managers can take corrective actions.

Unobservable corrective actions

In the core model, the corrective action is publicly observable, so the investor can commit to a tilting strategy that rewards a brown company from being best-in-class. We extend the model to the case in which the investor cannot observe the corrective action but only an imperfect signal such as an ESG rating, and so she can only base her trades on the rating. Even if the firm’s manager takes the corrective action, it may not be reflected in the rating, so the investor does not buy any shares and the firm is not rewarded for this action. Thus, to provide sufficient incentives to reform, the investor must promise greater purchases upon a positive ESG rating to compensate for the fact that a reforming firm will not always enjoy a rating uplift. These greater purchases allow the firm to expand even more and create even more harm, so tilting may no longer be optimal.

Many responsible investors claim to go beyond ESG ratings and conduct their own analysis. By doing their own research, they can find out whether the company has taken the corrective action and buy shares if it has. However, doing so may lead to the investor being accused of greenwashing – buying a brown company even though it has not reformed in the eyes of the market, which only observes ratings. If the responsible investor were to suffer a sufficiently large reputational cost from buying a lowly-rated company, then she will not do so. This reduces her incentives to do her own research, and indeed to implement the tilting strategy in the first place.

Implications for practitioners

Our model has important implications for how investment managers should practise responsible investing, and for how their clients and the public should evaluate them. It is common to assess the responsible investment practices of an investment manager through their holdings of green stocks and accuse those with brown firms of greenwashing. Such simplistic evaluations are damaging. They discourage responsible investors from tilting strategies, which may change behaviour more effectively than exclusion. Such evaluations also discourage investors from gathering private information on whether companies have taken corrective actions – if they have, but their actions are not yet publicly observable, the investor can’t buy shares as she may be accused of greenwashing. She will follow ESG ratings rather than doing her own research.

Our concerns about the effectiveness of divestment are quite different from the common counterargument that “divestment is bad because you can’t engage”. While true, some investors rarely engage to begin with – their expertise may be stock selection rather than engagement, or they lack the substantial financial resources needed. For example, Engine No. 1 needed to spend US$30 million in its campaign to elect three climate-friendly directors onto Exxon’s board; many investors may not have such resources. Even for investors who rarely engage, our research shows that exclusion may not be the most effective divestment strategy.


This paper won the PRI Award for Outstanding Research.


the presentation of the research at the PRI Academic Network Week 2022.


the presentation of the research and the award at PRI in Person & Online 2022.


The PRI’s academic blog aims to bring investors insights from the latest academic research on responsible investment. It is written by academic guest contributors. Blog authors write in their individual capacity – posts do not necessarily represent a PRI view.