Post-Kahneman Decision Theory: What Behavioral Economics Missed

The field of behavioral economics has spent three decades proving that humans are irrational—and in doing so, it has systematically overlooked the rationality of irrationality itself.

Kahneman and Tversky's heuristics-and-biases framework was revolutionary because it showed that human judgment deviates from rational choice theory in predictable, measurable ways. Anchoring, availability bias, loss aversion, framing effects—these became the vocabulary of how we understand decision-making. But the framework contains a hidden assumption that has calcified into dogma: that these deviations are mistakes. Errors to be corrected, biases to be debiased, systematic failures to be engineered away.

This is precisely backwards.

The problem is not that behavioral economics discovered irrationality. The problem is that it never asked whether the patterns it discovered might be adaptive. A bias is only a bias if you're measuring against the wrong criterion. Loss aversion looks like a failure when you're optimizing for expected value maximization across infinite time horizons. It looks like wisdom when you're trying not to go bankrupt before next quarter. Anchoring looks like a cognitive flaw when you assume people should ignore all available information. It looks like information integration when you recognize that anchors often contain real signal.

What behavioral economics missed is that human decision-making evolved in environments where the cost of certain errors was catastrophic. A hunter-gatherer who treated a rustle in the grass as random noise rather than a potential predator paid with their life. The bias toward false positives—the availability heuristic applied to threats—was not a bug. It was the operating system.

The real insight is not that we're irrational. It's that we're rational in ways that don't map onto the utility functions economists prefer to measure. We optimize for robustness, not expected value. We optimize for regret minimization, not outcome maximization. We optimize for social standing, not individual consumption. These are not deviations from rationality. They are different rationalities, calibrated to different problems.

This matters because the behavioral economics framework has become the intellectual foundation for a particular kind of intervention: the nudge. If people are systematically biased, then small changes to choice architecture can push them toward "better" decisions. Opt-out pension schemes. Calorie labeling. Default options. The logic is seductive: we're not restricting choice, just correcting for bias.

But this approach assumes that the person designing the choice architecture knows what "better" means better than the person making the choice does. It assumes that the deviation from rational choice theory is a deviation from the person's own interests. Often, it is not.

Consider the post-purchase reinforcement of benefits. When someone buys a product, they experience a moment of vulnerability—what if I made the wrong choice? The reinforcement of benefits after purchase isn't a manipulation that exploits bias. It's information that resolves uncertainty in a way that aligns with the buyer's own interests. They wanted the product. The reinforcement confirms that judgment was sound. This isn't debiasing. It's decision support.

The next generation of decision theory needs to move beyond the assumption that human judgment is a corrupted version of rational choice theory. It needs to ask: what problem is this decision-making pattern solving? What environment was it designed for? What does it optimize for?

Only then can we distinguish between biases that genuinely harm people's interests and decision patterns that serve them well—even if they violate the axioms of expected utility theory.

The irony is that Kahneman himself understood this. His later work on System 1 and System 2 hinted at it. But the field he created moved in the opposite direction, toward ever-more-sophisticated debiasing techniques. It's time to reverse course. The real frontier in decision science is not correcting human judgment. It's understanding what human judgment is actually trying to do.