Against Kind Informed Voters

In a democracy, voters pick which candidates become government officials. In this context, we usually admire those whose efforts help voters to become better informed, and to vote more altruistically. After all, we generally praise altruism. And we see how much harm could from ignorant voters who have no idea what candidates might do, which actions would have what effects, and who would be helped or hurt by them.

However, such efforts are actually less laudable than they seem. The reason is that selfish and ignorant voters can form a quite effective democratic system, and, relative to this system, efforts to induce more altruistic and informed voters are mainly ways for some voters to gain at the expense of others. Let me explain.

First, quite selfish and ignorant voters can form the quite effective democratic system of retrospective voting. At election time, each voter need only ask themselves how good their life seems overall now, percentile rank that situation relative to their expectations when the current incumbent took office, and have a chance of re-electing that incumbent which is increasing in that percentile rank. This is a reasonable strategy by which a selfish ignorant voter can infer how much they’d gain from an incumbent staying in office, and it also gives incumbents good incentives to set policy to increase a weighted average of voter life-goodness self-estimates. Which seems pretty close to an ideal policy outcome.

Yes, this fails when incumbents could more easily steal from citizens than work to get re-elected, this doesn’t induce voter goodness that voters can’t see, this doesn’t tell voters which challenger to pick when rejecting an incumbent, voters may not remember their before-incumbent expectations, and this doesn’t give incumbents incentives to make public investments when voters cannot immediately see their value, and which take longer than incumbent terms to realize. Furthermore, the simplest version of this voter strategy doesn’t distinguish outcomes that particular incumbents can greatly influence from those that they can hardly influence. 

Even so, this is overall quite a good system, and democratic details can be designed to promote it. Such as via two party systems that make the challenger obvious, via preferring price-scored private investments to public investments, via reminding voters of their prior expectations, and via centralized power and responsibility, or dividing responsibilities according to areas of life with easily distinguishable outcomes. 

Second, relative to this simple retrospective voting benchmark, efforts to induce more altruistic voters, and more informed voters, can both be seen mainly as ways to help some voters to gain at the expense of others. And if that is their main effect, such efforts can be seen as wasteful “rent-seeking” that doesn’t improve the system as a whole.

To see this, consider first an incumbent who has limited policy budgets to spend (not just money) to attract voters. In seeking re-election, such an incumbent should ask themselves, for each type of spending, how much a given spending amount will induce each voter to increase their chance of helping them win the election. This effect is the product of (a) how much that spending would improve each voter’s life, times (b) how likely that voter is to notice a given gain in the noise of everything else happening in their life, times (c) the degree to which they will credit the incumbent for that change, times (d) the chance a given noticed and credited life change will induce them to change their vote, (e) times the chance that voter will be pivotal in the election. (Note that this analysis applies also prospective as well as retrospective voting.) With diminishing returns to many of these factors, incumbents should set alternative spending levels to give equal added chances of helping them win the election. 

In the simplest extreme retrospective voting scenario, voters have similar minimal levels for most of these parameters. They are similarly able to notice when their lives get better, they have similar low chances of being pivotal, and they know little about which particular policies were caused by which incumbents or induced which changes in their lives. In this context, one way for a voter to induce an incumbent to slant more spending in their direction, and away from other voters, is to increase their chance of noticing some incumbent-induced change, and of crediting them for it. And then informing the incumbent of their increased awareness levels. Yes, such efforts may make the election slightly more informed overall, for example by cutting the chances of re-electing a bad incumbent. But their main effect will be to slant spending more toward better informed voters, and to raise over all spending on voter information. 

Another way for voters to induce incumbents to slant spending in their direction is to form voting coalitions who vote together either to support or reject an incumbent. By voting together, such coalitions greatly increase their chance of being pivotal in the election. Which should increase in the weight that incumbents give to spending on members’ behalf. And a simple way to construct and interpret such coalitions is via altruism: instead of just voting based on personal life outcomes, coalition members vote based on outcomes for the coalition as a whole. Intermediate levels of altruism, where voters put some weight on others but more weigh on themselves, produce intermediate chances of their being pivotal in an election. 

Yes, if some voters put high altruistic weight on others who instead vote selfishly, that later group can gain at the expense of the first. Which is why voting coalitions monitor each other to ensure that they put similar weight on each other. This is why politics is such a social and moral activity; political allies must watch out for and punish free-riders. All of which induces yet more wasteful rent-seeking efforts, in addition to efforts required to learn voting inclinations of other coalition members.

Thus while simple selfish ignorant retrospective voting is an admirably effective democratic system, democracy tempts each voter to instead devote more efforts to becoming more informed and altruistic than they would be in this simple system. This not only wastes energy due to such efforts, but it also produces far more unequal policy outcomes. As voters vary greatly in their abilities to form coalitions and become better informed, incumbents put very different weights on different voters in their policy choices. Such voter weights may well vary more than do voter income levels. “One person one vote” does not at all ensure egalitarian policy outcomes. 

Perhaps there are no good ways to prevent these strategies in admirable democratic systems. (Though I doubt this.) But we can at least stop admiring and praising wasteful efforts to induce more voter altruism and information. They likely make for worse, not better, democratic outcomes.