Explainable Governance

My once bumper sticker: “Question Authority, But Raise Your Hand First”.

In families and small groups, we can usually challenge and question our leaders. When they declare an official policy, we can often ask “why?”, and get a moderately coherent response. If we notice inconsistencies between explanations for related policies, we can point them out, and pressure leaders to reduce them. And to a limited extent, we can challenge such explanations, giving counter arguments to official reasons, and offering reasons for alternate policies.

Of course sometimes busy or exasperated parents, and other authorities, retreat to “because I said so”. But if they go there too often or quickly, we think less of them, share that opinion via gossip, and undermine their authority and tenure.

However, as our social systems get larger, we tend to lose this crucial human option, to see and challenge justifications for the policies we live under. Oh sure, some justifications are offered, but such things tend to be more rare, shallow, inconsistent, and unresponsive to criticism.

Today when someone tells you can’t do that, there’s a rule, and you ask why, you mostly get shoulder shugs or vague platitudes. When you hear reasons, there usually seems little point in pointing out contradictions. As a result, you feel disrespected, and these rules feel less legitimate. All of which probably contributes to our living under a less justified and coherent total set of policies.

Can we do better? I can imagine a legal requirement that all laws and agency rules have an explicit justification text, but I doubt that would create substantially better justifications than we see now.

I’ve recently tried to think about how to make our systems of governance better at offering persuasive justifications. To explain my ideas, I will make some simplifying assumptions. Not because I’m sure I need them, but because they seem to make this initial concept exploration easier.

First, let us assume that each policy has a text description, and applies to a subset of the space of policy applications. Each policy also has a text justification, and an author. Assume none of these subsets overlap partially; that is, if there’s any overlap, then one is a strict subset of the other. For each such set, we can (somehow) create a rough dollar estimate of the annual value of good policy in that set. These dollar values add up in the obvious way across sets.

My idea has two levels. The first level just tries to create good policy justifications for fixed policies, while the second allows policies to be changed to get better justifications. Let’s start with the first level.

Let the author of the current justification for a policy be continually paid X% (1%?) of the estimated value of its policy set (minus the values for subset policies). At anytime, anyone can challenge that author in a justification court, by offering an alternate justification, and by paying for a jury trial. At the trial, both sides can make arguments beyond the texts of their justifications. If jurors decide that this is a better justification, then that becomes the official justification, and its author now gets paid for its value.

When policies change, policy makers either offer a justification, or its an empty one that should be easy to beat by a challenger. To allow better targeting of justifications, I’d let challengers offer a justification for a strict subset of an existing policy. If the jury likes it, that would become the official justification for that subset, and its author would be paid the value for that subset. To help deal with inconsistency, I’d also let challengers offer a new justification to replace (the union of) any set of existing justifications. The challenger can argue that these different prior justifications are inconsistent, and that this new justification is overall more coherent, and better.

At least that’s my simple first-cut design. I can imagine doing better via betting markets on who would win if a jury were invoked. I can also imagine starting with a small jury and then moving to larger juries only after small jury wins, and only changing the policy justification after a large enough jury win. But for now these issues seem like distractions from our main ones concerns, so I’ll set them aside for now.

My first level proposal seems to create incentives to make policy justifications that ordinary people would accept. At least to the extent that there’s any reasonable way to justify such policies. But what if the policies are just stupid and incoherent, or at least seem so to ordinary people?

My second level tries to address this by moving further in the direction of governance by jury. Now allow challengers to also specify new policies, as well as new justifications for those new policies. As with my first level proposal, they can do this not only for particular existing policies, but also for sets of such policies, and for strict subsets. Juries are now empowered to approve such new policies along with their justifications.

As before, we might be able to improve on this via betting markets, small then larger juries, etc., but as before it seems premature to go into those details. For now, the main question must be: does this whole approach make sense? Would it be good to make policy more coherent and justified, in the eyes of ordinary citizens, even if this may come at the expense of making it less coherent and justified in the eyes of elites who might otherwise decide such things? To whom exactly should policy seem justified, if anyone?