Inconsistent Thoughts on Time-Inconsistency Problems

Since the financial crisis, many have argued that we need to restructure the incentives that bankers face by extending their time horizons. See, e.g., this post by Brad DeLong. In its best manifestation it's not a principle-agent argument, but rather the opposite: shareholders incentivize management to maximize short run profit at the expense of long run stability. Bankers respond to those incentives by maximizing short run profit, sacrificing long run stability in the process if necessary, and reap large rewards for doing so. All fine and good, except this creates systemic risk, which is a negative social externality. If taxpayers are going to give the banking sector an implicit explicit bailout guarantee, then we should be able to legislate the way that banks behave -- including how bankers are compensated -- in order to bring private interests in the banking sector more in line with the public interests of society to not have to pay for their mistakes. This, in effect, means limiting the input of shareholders and management.

The wisdom of this is not apparent to me, since the "public interest" in the run up to the recent crisis seemed to be extending credit to everyone everywhere, which is why policy and practice shifted in that direction, but right now I'm more interested in extending the logic to other contexts to see where it leads. Let's assume that the public is purely interested in maximizing long run stability over short run profits. Is this a laudable goal?

The same logic could be used to argue against democracy. Democratic leaders also must balance short run incentives against long run interests. After all, if a politician loses an election it doesn't matter what her long run policy preferences are, since she won't be in office to enact them. Political scientists don't agree about much, especially across subfields, but I imagine that we could get a large majority to agree that politicians privilege short run constraints over long run outcomes, especially in democracies. So we often see spikes in government spending before elections, even if this causes fiscal problems in the longer run. Greece We often see leaders fail to slash inflation near elections, despite the adverse consequences of inflation on long run economic health, because the resulting economic decline will hurt their re-election chances. Episodes of trade protectionism often follow a similar logic, where long run inefficiencies are entrenched because of short run political incentives. I know that Dr. Oatley is developing an argument along these lines with regards to American macroeconomic policy; he can elaborate if he likes.

Yet I rarely hear calls to limit or abandon democracy. True, we often pay lip-service to checks and balances, but generally not for the reasons I've described. (And even that is subject to preference; witness the "Abolish the Senate" movement.) This begs a question: Why would we demand something of the private sector that we wouldn't even consider for the public sector? Is there an implicit positive model of politics that explains the separation? Or a normative model of justice? I'm having difficulty seeing it. To me it looks more like cognitive dissonance.

Some have called for CEO compensation to be dependent on the performance of a firm over the medium run. Suppose we did the same with laws? Any significant change in policy must first be subject to a 5-10 year review process. If it hasn't been revoked in that period, then it becomes law. Obviously we'd have to exempt some policy areas, like natural security, from such a long lag. But the debate over energy policy or health care policy would look a lot different if lawmakers weren't subject to immediate pressures that track the election cycle.

This is probably a really bad idea, and I'm not actually endorsing it. (For evidence of how bad it is, here's Thomas Friedman endorsing its logical extreme: technocratic authoritarianism.) But I am thinking about it. After all, we try to depoliticize some aspects of social management, like monetary policy, for precisely this reason.

The point is to think more seriously about why we want to shift incentives along time horizons in some situations but not others. What is the justification for treating some issue areas differently from others?