Developing an activism rating
Yesterday I blogged about FocusLists, which we’ll be releasing on ProxyDemocracy with our (re-)launch slated for next week. As I explained in that post, FocusLists are an attempt to let users select the proposals they think are important and rank mutual funds on those proposals, much as political advocacy groups have done for a long time with the voting records members of Congress. It’s an old approach, but I believe we’re the first to create a public tool so that anyone can make her own list and share the results.
In conjunction with that, I’ve worked on developing a ProxyDemocracy activism rating of our own, to provide a different approach to the same problem. The goal was to come up with a broad-based measure of activism that reflects experts’ sense of what the most important issues on proxy ballots are. As with FocusLists, I hoped that our activism rating would quickly convey information that would be useful to an investor comparing funds’ voting records. Unlike FocusLists, which can reflect the particular interests of individual users (for example, resolutions on climate change disclosure), our ratings would try to capture a general consensus view of what it means to have an activist voting record. My hope is that FocusLists will be particularly useful because they can be created about any topic and are associated with a particular viewpoint, while our ratings will be useful as a broad, impartial, easily understandable measure of activism.
Roughly speaking, the rating system I’ve come up with measures how often a fund votes against the voting recommendation of management. The simplest way to produce this would be to calculate the percentage of all votes cast by a fund that went against management’s recommendation. (In fact we used this in the beta version of the site we released in December 2007.) The problem with this approach is that it weights the different types of proposals according to how frequently they arise at shareholder meetings, whereas most observers of proxy voting would not assess things in that way. For example, the vast majority of votes are on director nominations — about 10 times as many as any other proposal type — so the score calculated based on all votes would be dominated by how the fund voted on director nominations. Since it’s clear that director nominations are far more common than they are “important” (to both proxy voting experts and, I believe, the investors PD is trying to reach), we need a different approach.
I decided to think about what the important issues really are, and devised four categories that seem to be of roughly equal importance to people who care about proxy voting:
- director nominations: who represents shareholders
- executive compensation: how and how much the execs get paid
- corporate governance: the internal rules that structure decision-making
- corporate impact: relationships between the firm and its employees, customers, the world.
So our overall activism score is the average of the activism scores from these four issue areas. One way of putting it is that our activism score measures the percentage of votes a fund casts against management, where the votes are weighted to reflect the equal importance of the four categories above. Another way of putting it is that our activism score is an estimate of how many proposals the fund would vote against, on average, if it were faced with 100 proposals, 25 of which were chosen from each of the above four categories.
Ultimately our ratings are only good if they summarize the information our users want to know about these voting records. I’m not sure if we’re there yet, and (especially once we’ve launched the new site and you can see the ratings) I look forward to getting your feedback.
What I’m particularly worried about with this score is whether we can communicate useful information without editorializing about how much activism is the “right” amount. If ProxyDemocracy took the position that a mutual fund should always vote against management, then this would be easier: I could give our most activist fund, which has a score of 76, a C (or an A, if I want to highlight its leadership) and give most of the other funds an F (the median score in our database is around 20). Since we were all constantly receiving grades for about 15 years of our lives, we would easily understand a rating system like that. But I don’t believe that the most activist fund is the best, and I don’t think ProxyDemocracy should take that position either. So instead I’m using a rating system that is difficult to interpret as a grade and therefore requires the user to come up with her own idea of what the right level of activism is. I don’t know if that’s asking too much. I’m trying to help by showing the funds’ ratings in relation to each other (by reporting percentiles and displaying little histograms) but I guess we’ll just have to see whether this makes sense to people or not once we release the new vote profiles.