4.4.3 Proposal Type

Previous studies have shown that the voting behaviour of institutional investors varies across proposal types (see e.g. Ashraf & Jayaraman 2006, Iliev et al. 2010), and that the association between ISS recommendations AGAINST management and the percentage of fewer votes cast FOR management varies across proposal types (Bethel & Gillan 2002). As Table 3 (Panel A) shows, the voting behaviour of the funds in our sample also varies across proposal types. For the median fund, the rate of deviation by proposal type ranges from as low as 0.3% to as high as 15.3%.

Naturally, some proposals, such as merger proposals, potentially have more significant value implications than others, and accordingly are more likely to impact on portfolio performance than others. We hypothesize that the more significant the potential value implications of a proposal, the more resources the fund devotes to verifying the accuracy of the voting recommendation on that proposal, and the greater the likelihood (at the margin) that the fund reaches a different conclusion than its proxy advisor.1 To test this hypothesis, in Panel B we divide the sample into votes on routine matters, which generally are unlikely to have significant value implications, and votes on non-routine matters (such as mergers), which potentially have significant value implications. The results, visualized below, show that the median fund deviates from proxy voting recommendations regarding routine matters only 0.6% of the time versus 1.7% of the time for non-routine matters, consistent with our hypothesis.

For the full sample, we obtain similar results. As Panel C shows, the average deviation rate for routine matters is 4.8% versus 6.5% for non-routine matters, and the difference between the means is statistically significant (p-value<1%).