AI and It’s Implications for Income Distribution and Unemployment (2017) by Anton Korienk and Joseph E. Stiglitz. Background paper for NBER conference “The Economics of AI”.
Technological Progress and Welfare: A Taxonomy
First-best scenarios are those in which we assume all markets are perfect. I’m not sure what utility there is in discussing such impossible states as these, but the authors seem to position them as a baseline ideal we can compare all other, viable, scenarios. Of note, the authors highlight that in a first-best world, all individuals would have access to perfect insurance against all such technological boons. For instance, if some folks made out better than others, redistribution would emerge naturally, balancing everything out.
Not everyone is equally hip to the frontiers of innovation. Some will get there first and make more money than others. In a world where all spoils would be effectively redistributed, all parties would be aligned on the promise of innovation, since no one would be left in the lurch. This, as we know, is not how things work, so what other scenarios might we consider?
Perfect markets ex-post and no costs of redistribution is a “second-best” scenario in which the market does not have perfect insurance all parties (as we assume in first-best), yet there is no cost to redistribution so no harm no foul. In this world, it is in the best interest of innovators to effectively redistribuate all spoils to their workers (which make up the majority of the planet and workforce), to ensure all keep rowing together towards the horizon of technological progress without rupturous dissent. This, as we know, is not been someting we’ve gotten right in tech, or in society at large.
Perfect markets but costly redistribution presents a scenario where we may experience significant opposition to technological progress, as the cost of redistribution will prevent it from happening altogether, which means groups will oppose it tooth and nail, as I believe they should. Why should they get on board with progress that does not take their economic reality into consideration? So, the authors ask how costly is redistribution in practice? That’s a great question! How do we even calculate that? And, say we demonstrate that the cost is negligible, do we have precedent for it taking place and the infrastructure needed to ensure it does? Hmmmmm…
Imperfect markets are world in which improving the functioning of one market may deteriorate overall welfare. Market imperfections “include all the market arrangements that differ from the Arrow-Debreu ‘optimal’ benchmark, the conditions which ensure the Pareto efficiency of the market”. Of note, “the possiblity of achieving a Pareto improvement depends on how broadly we define the classes of individuals which are affected by an innovation”, for instance “workers” vs. “entrepreneurs”.