I don’t have an article or statistic to point to, just a question: what is the threshold at which a betting market outcome becomes endogenous to the existence of the betting market?
Polymarket recenttly removed its market for a nuclear detonation. The implication is straightforward: if a market exists for an outcome that a singular individual (or small group) can make manifest, then as the market becomes takes on greater volume, the maximal reward for independently producing the wagered upon outcome increases. This generates a testable hypothesis: does the predicted possibility of a positive outcome increase with market volume?
If and when the answer is yes, market volume has a positive causal relationship with market outcomes, then the welfare proposition of the market existing comes into question. I don’t care about more accurately predicting nuclear detonation if the market yielding the prediction is increasing it’s probability of occurring.
Now, do I think betting markets are increasing the probability of a nuclear explosion? Eh….probably not, or at the very least the effect should be quite modest. But there are lots of events on prediction markets that raise this possibility without wading into the waters of apocalypse-adjacent outcomes. As I discussed previously, the endogeneity of sports outcomes to betting markets is threatening the integrity of professional competition. Many of those obscure sports wagers don’t seem like thick markets, but relative to a 20 years where such wagers didn’t even exist, they are positively rippling with volume.
In a world where unscrupulous individuals are betting big on their ability to extract enough rents in public life before the world catches up with them, we would be wise not that add more profit channels for corruptuon than we are capable of credibly monitoring.
Economist Writing Every Day 