Conflicts of Interest Among Proxy Advisors

The Importance of Proxy Advisors

Proxy advisors play a critical role in corporate governance. One important way in which shareholders exert governance – ensure that executives act in shareholders’ interest, rather than their own interest – is through voting. Votes can be for particular directors (who monitor executives), for or against a say-on-pay vote, or for or against manager or shareholder proposals on issues such as corporate social responsibility, strategy, payout policy, charitable contributions, or corporate governance itself (e.g. shareholder rights).

However, many investors may be uninformed about how to vote, either due to lack of expertise in corporate governance (as may be the case for some, but not all, retail investors) or lack of incentives to become informed due to their small stakes. Even a large institutional investor may only hold a small stake in any one particular company, due to being spread thinly across thousands of companies. These concerns are particularly strong given the rapid growth of index funds who are especially diversified.

Proxy advisors, such as Institutional Shareholder Services (“ISS”) and Glass-Lewis are a potential solution to this problem. Just as credit ratings agencies guide investors on how much to charge for lending to a company, and credit and equity research reports guide investors on whether to buy or sell a company’s debt or equity, proxy advisors can guide an investor on how to vote. Even though some investors may not blindly follow a proxy advisor’s recommendation, their recommendations remain very influential. Exxon Mobil recently stated that “proxy advisors hold a position of unparalleled influence” and estimated that “between 20-25% of the votes cast at Exxon Mobil’s most recent annual meeting were voted automatically in accordance with proxy advisor recommendations.” This is backed up by large-scale causal evidence – investors are significantly more likely to vote for (against) a resolution if ISS recommends voting for (against).

Thus, it is critical that proxy advisor recommendations are unbiased. However, evidence suggests that credit rating agencies may be biased, and the Global Analyst Research Settlement was in response to concerns of equity analyst bias. A forthcoming paper in Management Science by Tao Li of Warwick Business School, based on his PhD dissertation at Columbia Business School, suggests that conflicts may also exist with proxy advisors.

Why Might Conflicts Exist?

What’s the source of such a conflict? Tao studies ISS, who sells not only proxy voting services to investors but also consulting services to firms who are considering management-sponsored proposals. Thus, there is a potential concern that ISS may recommend that investors vote for management if it is a consulting client – either as a “quid pro quo” for commissioning consulting services, or to show that its consulting has succeeded in designing management-sponsored proposals that shareholders are willing to approve. Indeed, the Ohio Public Employees Retirement System dropped ISS’s services in the mid-2000s, commenting that “the thing that tipped us was [ISS’s] actual or perceived conflicts due to the corporate consulting.”

Documenting Potential Conflicts

How did Tao study whether these conflicts actually existed? Simply showing that ISS is more likely to recommend votes for management when also provides consulting services does not suggest bias, because it could be that the consulting genuinely improved proposal quality. So, Tao looks at the entry into the proxy advisory market by Glass Lewis, a major competitor providing no consulting services. The hypothesis is that competition disciplines ISS from issuing biased recommendations, just as it does so with equity analysts.

But, simply showing that ISS recommends “against” more often when Glass Lewis enters will still not suggest that ISS was previously biased. It could be that managers are starting to make more egregious proposals, which both causes ISS to recommend “against” and also encourages Glass Lewis to enter, since investors need more voting guidance. Thus, Tao compares firms that are potential clients of ISS (proxied by large firms) – where conflicts exist – with firms that are not potential ISS clients because they are too small. This “difference-in-differences” analysis controls for time trends. Using a medical analogy, large firms are “treated” firms, where any bias is reduced by Glass Lewis’s entry, and small firms are “placebo” firms – Glass Lewis’s entry is irrelevant, since there were few conflicts to begin with.

Indeed, Tao finds that, when Glass Lewis initially covers a new firm, ISS becomes tougher towards potential clients than non-clients. Also supporting a “conflict” explanation, the effect is stronger for more complex votes where it is easier to be biased (e.g. on proposals related to governance and executive compensation) rather than “no-brainer” votes where bias might appear blatant (e.g. uncontested director elections). For contested director elections, the effect is stronger for first-time director nominations than director reappointments, again potentially because the former are more complex and thus bias was easier before Glass Lewis’s entry.

Moreover, any potential bias has real outcomes – it is important. Tao compares proposals with potentially biased recommendations from ISS that pass by a narrow margin to those that fail by a narrow margin (focusing on narrow margins zones in on proposals where any potential bias was particularly important). Firms with proposals that narrowly pass subsequently outperform those that narrowly fail. These effects are small, but Tao finds that executives at these firms have higher abnormal pay, higher growth in pay, and more cash payments. These results suggest that any potential bias allows managers to extract value at the expense of shareholders.


Tao is careful not to over-interpret his results. While suggestive of conflicts, they do not prove bias. The published paper does not use data on ISS’s actual client base, and can only proxy for it using firm size. (Tao, like any institutional client that purchases ISS’s voting recommendations, was able to check potential conflicts of interest using ISS’s actual client base, but was not allowed to write any papers based on the client list.) Moreover, even if ISS’s recommendations were biased, they may still be informative – a biased recommendation may still be better than no recommendation, and so proxy advisory firms play an important role in compensation. But the result do suggest that, just as investors use credit ratings and analyst reports to guide their investment decisions but do not blindly follow them, investors should similarly use proxy recommendations as an important input to their votes but it is critical that they also do their own analysis. If investors do not have sufficient “skin-in-the-game” to do their own analysis, policymakers should encourage them to take large stakes in firms – potentially by increasing the threshold (currently 5% in the US and EU, and 3% in the UK) at which they need to disclose their holdings. My conversations with a very large, but very diversified, institutional investor revealed that they virtually automatically follow ISS in their voting, which may not lead to the best outcomes.