Thinking in terms of incentives

One of the key takeaways for me from reading Thomas Sowell's Basic Economics is that we should always think in terms of incentives when evaluating policies. A commonly-cited example is rent controls, but a local example that I observed firsthand was face mask prices during Covid in Kenya.

I remember in early 2020 right at the beginning of Covid, face masks were retailing in Kenya for hundreds of shillings (and over Ksh 1,000 for the N95 type masks). Capping the prices to ensure a "fair" price would have made the situation worse and fortunately policy makers didn't try to do this. Instead, the private sector was incentivised by the high demand to manufacture and distribute more masks. Supply drastically increased, and in a short time masks were being hawked on every street for between Ksh 10-30.

To quote Sowell:

"Among other things, this illustrates the crucial importance of making a distinction between intentions and consequences. Economic policies need to be analyzed in terms of the incentives they create, rather than the hopes that inspired them."

I think our tendency to focus on immediate stressors and consequences instead of the second-order effects of incentives is an example of evolutionary mismatch; our hunter-gatherer brains evolved to survive in much simpler environments and societies. As a result we end up having lots of disagreements and many harmful policies.

An unfortunate consequence of ignoring second-order effects is that politicians are incentivized to implement short-sighted policies that might bring immediate relief (e.g. price caps on goods or services) but that lead to disaster in the long run.

To make matters worse, economic cycles and political cycles often don't coincide. So a politician could enact policies whose first-order effects are pleasant, finish their term, and have the next politician deal with the fallout from the negative second-order effects of the very same policies!

I don't have a silver bullet solution to this but I do think we'd be better off if we looked at things less from the point of view of "Let's do this and that to solve this problem" - assuming that people are like static pieces on a chess board - and more from the point of view of "If we do this, what behavior will it likely induce and what is the likely outcome of that behavior"?