The AI Will Vote the Shares

We talked in November about proxy voting at US public companies. My basic point was that, if you own a diversified portfolio of stocks, you will be asked to vote on hundreds of different propositions — annual elections of directors, non-binding social and environmental proposals, the occasional merger — each year. Almost none of these votes will matter to you economically, and your vote is vanishingly unlikely to matter to any of the companies. In the abstract, in the aggregate, shareholder voting is the foundation of corporate democracy, a way to make sure that managers are answerable to the ultimate owners of their companies. But in almost every particular case it’s a waste of time.

I suggested that this tension — shareholder votes are important in the aggregate, but not to most actual shareholders — is most easily solved by outsourcing proxy voting to specialists. It would be insane for millions of individual investors to read hundreds of proxy statements to decide how to vote on hundreds of different questions. But if, like, eight people each read a few hundred proxy statements and decided how to vote, and everyone else paid those people a small fee to vote their shares for them, the problem would be solved. I mean, there would be other problems. (Would those eight people actually vote the shares in ways that align with the interests of the ultimate owners of the companies? Would they have their own agendas? Would the proxy-voting profession attract people whose views on proxy voting are different from those of most shareholders?) But, again, most of these votes don’t matter very much to most of the shareholders. A cheap outsourced system is probably good enough.