Modern behavioral science rests on a framework Daniel Kahneman built with Amos Tversky over three decades of experiments. Kahneman won the 2002 Nobel in economics, and Tversky, who died in 1996, would certainly have shared it. Kahneman set the framework out for the general reader in Thinking, Fast and Slow (2011). It's a metaphor, and Kahneman is explicit that there aren't literally two systems in the brain. It's still the most useful metaphor an investor can carry.
How do System 1 and System 2 differ?
System 1 is fast, automatic, intuitive, effortless, emotional, and always on. It's the part of you that reads the word on a billboard without choosing to, that knows 2 + 2, that detects hostility in a voice, that recognizes a friend's face, that jumps at a sudden noise. It's also what feels a jolt of fear when a stock gaps down. It runs on pattern-matching and association. It's superb at what it's good at, and you can't switch it off. It also has no sense of its own limits. It produces an answer to almost any question instantly and hands it up with full confidence, whether or not the question is one it can actually answer.
System 2 is slow, deliberate, effortful, logical, and lazy. It's the part that computes 17 × 24, that fills out a tax form, that checks the validity of a complex argument, that holds several considerations in mind at once. It can do the disciplined reasoning good analysis requires. But it's metabolically expensive and tires easily, so it defaults to the path of least resistance. That laziness is the hinge of the whole theory. System 2 usually endorses the answer System 1 has already produced, without doing the work to check it. Kahneman calls System 2 "the lazy controller." We feel like the reasoning System 2, the careful analyst. Most of the time we're System 1 with a System 2 press secretary that rationalizes the snap judgment after the fact.
Advertisement
Which System 1 features do the most damage?
Three features of System 1 do most of the damage in investing. They come back again and again in the bias catalog, so it's worth naming them now:
- WYSIATI, "What You See Is All There Is." System 1 builds the best possible story from the information immediately available and doesn't register the information it lacks. It never asks "what am I missing?" It works only with what's in front of it, and its confidence depends on the coherence of that story while ignoring the amount or quality of the evidence. This is the root of overconfidence, narrative fallacy, and the failure to ask what isn't being shown. A polished investor presentation is a WYSIATI weapon. It fills your field of view with a coherent bull story, and System 1 won't spontaneously notice the segment note, the related-party disclosure, or the bear case that isn't on the slide.
- Substitution. Give System 1 a hard question, such as "Is this a good investment at this price over the next five years?", and it swaps in an easier one it can answer: "Do I like this company / this CEO / this product / this story?" Then it answers that one, without noticing the switch. The difficult valuation question has become an easy affect question, and the easy answer arrives with the confidence the hard one would deserve.
- Affect and cognitive ease. Things that are familiar, fluent, easy to process, and repeated feel truer and safer, even when they aren't. A stock you've heard about many times, a story told smoothly, a chart that has been going up all generate cognitive ease, and System 1 reads that ease as a signal of quality and safety. So familiarity gets mistaken for knowledge, and a well-told narrative beats a spreadsheet.
Where do investing errors come from?
So where do investing errors live? In the gap between the instant System 1 answer and the checking that System 2 fails to do. The sequence is almost always the same. You look at an idea, and within a second or two System 1 has already decided ("great business, I like this" or "falling knife, stay away"). It decides on the basis of a coherent surface story (WYSIATI), an affect heuristic (you like the product) and cognitive ease (you've heard of it, the chart looks good). Everything you do afterward, the "research," the model, the reading, is then at grave risk of being System 2 building the case for the conclusion System 1 already reached. That's confirmation bias wearing the costume of due diligence. Having intuitions isn't the error. Expert intuition is real and valuable, and Kahneman studied that too. The error is mistaking a System 1 output for a System 2 conclusion, feeling that you've analyzed when you've merely rationalized.
How do you fix a System 1 error?
The diagnosis dictates the cure, and it's the intellectual foundation for the whole behavioral edge. You can't fix a System 1 error by trying harder. Trying harder is a System 2 activity, and System 2 isn't the part that made the error. Worse, System 2's main instinct is to defend System 1's answer. The only reliable fix is structural: an external process that forces System 2 to engage on the specific questions System 1 is skipping. A written checklist with the line "state the bear case in its strongest form" forces the disconfirmation System 1 won't volunteer. A decision journal that demands "what would change my mind, written before the outcome" forces the falsifiable prediction System 1 avoids. A required cooling-off period puts time between the System 1 impulse and the irreversible act.
So when you ask "how do I become less biased?", the answer is never "be more rational." It's always "build a machine that makes rationality the path of least resistance." Hold that thought. Everything else rests on it.