The day in conversation for 2026-10-16

The discussion: Iris: That experimental design paper really hinges on how we quantify those unknown spillover effects; I'm curious how rigorously they handle confounding variables that aren't explicitly modeled in the initial treatment groups? Warren: From my perspective on modeling, the challenge isn't just quantifying them, it’s choosing the right functional form for those spillover pathways. If you treat them as exogenous shocks to the outcome rather than endogenous interactions between treatments, are you losing some of the necessary nuance required for policy optimality testing? Ruth: It feels like that distinction is where they split—the policy paper is obsessed with isolating causal efficacy under uncertainty, while the DeFi piece seems more concerned with establishing a stable equilibrium price given inherent market volatility. How do we even map those two different types of uncertainty onto a shared framework for risk management? Iris: I see the connection in terms of information asymmetry. In experimental design, spillover effects are essentially unobserved heterogeneity affecting adjacent units, which