4.2.2 Proxy Advisors

The literature on proxy advisors can be roughly divided into two strands, one of which looks at the quality of voting recommendations and governance ratings. In terms of governance ratings, Daines et al. (2010) examine whether proxy advisors' governance ratings provide useful information for shareholders; their results suggest they do not. One explanation for this may be conflicts of interest caused by the fact that ISS also issues advice to listed companies as to how to optimize their governance structure (Belinfanti 2009), although it should be noted that ISS has gone at great length over the recent years to develop an objective framework for rating governance structures.1 In terras of proxy voting recommendations, Alexander et al. (2009) study the information content of recommendations issued during proxy contests and find that these appear to certify the extent to which dissidents can add value. Finally, Choi et al. (2009) examine how proxy advisors make their recommendations and find that they differ substantially from each other in the factors that affect their recommendations and the relative weight of these factors.

Another strand of the literature on proxy advisors attempts to measure their influence. Bethel & Gillan (2002) find that ISS recommendations unfavourable to management are associated with 13.6% to 20.6% fewer votes cast in favour of management, depending on proposal type. Cai et al. (2009) control for performance and governance characteristics, and find that a negative ISS recommendation reduces the average percentage of votes FOR management by 20.7%. However, Choi et al. (2010), too, control for firrn-specific factors and find that the impact of an ISS recommendation ranges only from 6% to 13%. Finally, Alexander et al. (2010) conduct event studies of announcements of proxy voting recommendations and find that such recommendations are good statistical predictors of outcomes.

Two studies that are closely related to our study are Choi, Fisch & Kahan (2011) and Cotter et al. (2009). The latter analyze mutual fund voting data from 2003-2008, and find that mutual funds vote consistently with ISS recommendations more often than with management recommendations, both on non-routine management proposals and shareholder proposals. Choi, Fisch & Kahan (2011) analyze the voting behaviour of mutual funds in uncontested director elections. They examine the extent to which funds rely on shortcuts to reduce the transaction costs associated with voting, and view a fund as employing a `primary' ISS shortcut if the fund votes in accordance with the ISS recommendation in 99.5% or more of its votes, and a `secondary' shortcut if the fund otherwise rarely deviates from ISS recommendations. Their results show that 10 of the 127 fund clusters in their sample employ the primary ISS shortcut, and another 26 employ the secondary shortcut. However, funds that follow ISS more than 99% of the time manage just over 3% of assets in the sample; funds that follow ISS more than 95% of the time manage 19% of assets in the sample, with funds managing the remainder deviating more often. The difference between their study and the present study is that they study a large sample of funds, but only look at votes in uncontested director elections; by contrast, our sample is small, but we look at votes on a range of issues, which enables us to examine differences in the propensity to deviate from proxy voting recommendations across these issues. Moreover, in analyzing factors that potentially affect voting behaviour we do not only look at exogenous factors such as portfolio firm performance, but also at endogenous factors such as the size of the fund's stake in the portfolio company and whether the portfolio firm is a foreign or domestic firm from the fund's point of view.

Three recent surveys among institutional investors shed further light on the role of proxy advisors. About half of the respondents to a survey by McCahery et al. (2010) claim they do not retain proxy advice, and a majority of those who do, claim they merely use it to determine their own position. Paape & Lachotzki (2010) report that 56% of the respondents to their survey retain proxy advice. Respondents to this survey consider verification of the accuracy of voting recommendations to be important, and they rate the "degree of deviation" from their proxy advisors' voting recommendation at 1.7 (on a scale from 1-10). Finally, according to a recent survey conducted by the Australian Institute of Company Directors (2011), only 22% of the funds responding to the survey consider proxy advisors influential; 54% consider them somewhat influential; and 11% consider them not influential.