Here’s yet another post from my project on character-based voting (CBV). It’s the first of three posts on CBV and leadership effects, and one of many on CBV.

As I’ve said in previous posts, “character-based voting” is voting for or against a political candidate on the basis of what the voter takes to be his traits of character. That contrasts with “policy-based voting,” which is voting for a candidate based on the expected consequences of the policies the voter expects the candidate to pass.

In The Ethics of Voting, Jason Brennan asserts that we shouldn’t vote for a candidate on the basis of the candidate’s character unless we have good reason to believe that the candidate’s having that character “is likely to produce” good quality “governance” (elsewhere glossed as “policy”). Here’s the operative sentence from the book:

So character-based voting is acceptable only insofar as it is a proxy to the quality of the governance a candidate is likely to produce (p. 84).

In the absence of any account of “quality of governance,” we could in principle assume that possession of good character was constitutive of good quality of governance,  meaning that the sheer expression of certain politically-relevant character traits would exemplify good governance simply by virtue of being expressed by a candidate in office. If so, given a specification of the traits constitutive of good governance, character would be guaranteed to be a proxy for good governance. So, for instance, if honesty is constitutive of good governance, then the propensity for honest-dealings-in-office would be guaranteed to be a proxy for good governance. If so, it makes no sense to say (as Brennan does) that CBV is “largely wrongful voting.” On the contrary, on this interpretation, CBV would practically be unavoidable.

Now, maybe it overstates things to say that character traits literally constitute good governance. If so, we might ratchet things down to the weaker claim that some traits of character are instrumentally valuable to good governance. If honesty doesn’t do it for you on instrumentalist grounds, maybe competence will. If competence doesn’t, maybe something else will. But as long as something does, and that something is a character trait, there is a case to be made for CBV.

Here’s another tack to take, however. The “quality of governance a candidate is likely to produce” is ambiguous as between “the quality of governance that the candidate is likely to produce by his own direct authorship,” and “the quality of governance that the candidate is likely to produce, whether by his own direct authorship, or through secondary consequences, intended or unintended, that arise indirectly through his being in office.” I’m inclined to think that Brennan means the first rather than the second of these, but the second could in principle be read back into his claim, and can’t be dismissed out of hand.*

Take a political candidate, Jones. Suppose that Jones’s moral character lends implicit moral support to other people besides Jones, so that Jones’s having that character, and expressing it in a particularly forceful and effective way, gives moral confidence to others besides Jones. Suppose further that those others wouldn’t have had that moral confidence had Jones not assumed office. Now suppose that Jones’s character is deeply flawed or immoral. Suppose further that Jones has both a very strong personality and a very prominent social role, so that Jones functions as a role model and morale-booster for lots of immoral people.

Now suppose that some of these people are in positions of power, not positions directly subordinate to Jones, or directly connected to Jones’s authority, but in some sense autonomous of Jones’s authority, in domains at some appreciable remove from that authority. Think for instance of the relationship between the President of the United States and a small business owner or head of a household inspired or emboldened by him, each with final authority over the business or the household and its members. Or think of the relationship between the President and a police chief, a sheriff, or a mayor. For that matter, though the relationship is somewhat less autonomous, think of the relationship between the President of the United States and Congress, or the President of the United States and the candidates he supports in an election. Though the President “authors” legislation by signing or vetoing it, he doesn’t literally write the laws that Congress writes, and doesn’t literally cast the votes that are cast for the candidates he supports. But he strongly influences both processes so that he might be said to be indirectly responsible for what arises from them.

Suppose now that the people in question are in a position to put Jones’s ethos into practice in the autonomous domains of authority that they control.  Now suppose that lots of these people enact deeply immoral policies at their own initiative, inspired by but not explicitly ordered or even explicitly advised or prescribed by Jones. Suppose, finally, that we can say with justification that these Jones-inspired policies really are policies of Jonesian inspiration; they genuinely exemplify Jones’s distinctive ethos without being literally authored, produced, or demanded (or prescribed, etc.) by Jones. If so, Jones would enjoy concentric circles of indirect policy influence well beyond the direct lines of policy influence he exercised. He would also enjoy plausible deniability for any of the influence he exercised beyond his direct control, where attributions of moral responsibility, though warranted and politically relevant, would be very hard to establish beyond a reasonable doubt.

If this happened, Jones’s moral character would function as an important factor in the production of immoral policies without its literally having to be the case that Jones  directly produced the policies. Setting aside the very real epistemic difficulty of tracing the effects back to Jones, it seems to me that it would be a mistake to vote for someone like Jones, assuming that there was a better alternative than him. Put differently, Jones’s bad character would at least be an electoral strike against him–a consideration of some weight against voting for him, even if Jones’s character produced no unjust policies literally of his own making. Other things being equal, if Jones differs from Smith in having the kind of character that produces bad secondary consequences, you should vote for Smith. How much weight to give this consideration, and how to establish when the relevant effect obtains, is a topic for another post or two. My point is simply that it has some weight.

In the next post, I’ll discuss a small handful of cases, from Hitler to Trump. In the third (and probably last) post, I’ll discuss epistemic issues that arise in thinking about leadership effects in a political context.

*Social scientists call these “leadership effects.” I’ve gone back and forth in this post between using the phrase “leadership effects” and “secondary consequences,” and for present purposes regard the two as synonymous. I hesitate a bit to use “leadership effects,” despite its greater currency in social science, because I don’t want to be tied to social scientists’ conception of the relevant phenomenon, and suspect that I have a broader conception of it than they do.