4.5 Do Customized Voting Policies Matter?

In this section, we explore the possibility that institutional investors receive voting recommendations that reflect their stated preferences. This may be the case, first, when the proxy advisor has taken into account the investor's preference regarding the information to be collected in preparing the voting recommendation. Indeed, a prior study has determined that a major component of ISS's influence stems from its role as an information agent, aggregating factors that its clients consider important (Choi et al. 2010). Second, voting recommendations can be said to reflect investors' preferences when the proxy advisor has taken their preference into account regarding the governance standard against which the relevant information is evaluated. There are various ways in which investors can signal their governance preferences to their proxy advisor. First, ISS enables investors to provide input on its general voting policy (which varies per region) prior to the start of the proxy season.1 The two funds in our sample that retain proxy advice from ISS have indicated that they have both provided such input, but of course it is difficult to measure the extent to which this input has resulted in changes to ISS 's voting policy. Two other ways to provide input are by informally discussing a specific proposal with the proxy advisor before the proxy advisor issues its voting recommendation, and by informally discussing the proxy advisor's services from time to time. Some funds in our sample have indicated that such discussions occasionally take place, but again it is difficult to measure impact. This is different, however, for the fourth and potentially most significant way to provide input: through a customized voting policy.

Proxy advisors generally issue recommendations on the basis of their own general voting policy, but provide their clients with the opportunity to customize that voting policy so that it accurately reflects the client's governance preferences; the proxy advisor then issues recommendations to this client on the basis of its customized voting policy. Survey evidence suggests that many institutional investors make use of this opportunity (Paape & Lachotzki 2002), and two of the funds in our sample have done the same. This raises the question, do customized voting policies matter? In this section, we try to answer this question by measuring the frequency of differences between recommendations based on the proxy advisor's general voting policy and recommendations based on the customized voting policy of the funds in our sample.

Table 5 (Panel A) shows the results for the two funds in question (Fund 1 and Fund 3). In 2010, the customized voting policy resulted in different voting recommendations 0.8% of the time for Fund 1 and 15% of the time for Fund 3. This suggests that the significance of customized voting policies varies per fund; some funds may put a lot of effort into customizing the voting policy, whereas others may put only limited effort in (or fmd that the proxy advisor's general voting policy already reflects their govemance preferences). As a robustness check, we also look at whether the fund tends to deviate less often from the voting recommendations based on the customized voting policy than from the voting recommendations based on the general voting policy, as one would expect. The results reported in Panel B confirrn this.

Interestingly, because proxy advisors provide many of their clients with a logistics platform to execute votes, they can observe the voting behaviour of their clients, and thus can also observe systematic deviations by their clients from these voting recommendations. Such systematic deviations signal a difference between the proxy advisor's govemance standards and the govemance preferences of its clients. The same is true when clients' customized voting policies systematically depart from one or more govemance standards reflected in the proxy advisor's general voting policy. Given that proxy advisors are forprofit organizations that generally aim to provide services that satisfy their clients' demands, we may expect proxy advisors to 'kam' about their clients' preferences from observing their voting behaviour, and adapt their general voting policies to these preferences. Over time, then, there should be a decreasing need for institutional investors to deviate from the voting recommendations based on these policies.