Thinking, Fast and Slow: System 1 and Sum-like Variables

I apologize for how incoherent the following will be. The idea is a fresh one, and I’m still struggling to articulate it.


Something that sparked some of the first original thoughts I’ve had while reading Thinking, Fast and Slow came about when Daniel Kahneman was covering how System 1 deals with valuation of joint products. In short, economics assumes that prices are sum-like in nature, essentially that the value of a stack of goods is determined by summing the individual prices (not account for things like bulk discounts).  In short, again not accounting for these other factors, the value of a good with 5 more of something than another good will cost 5 more (forget the absence of units if you would).  An alternative, and an important alternative at that, would be that the price of the goods is determined by averaging the price of the individual constituents.  In the case above, a good with 5 more will not cost 5 more but somewhere around 5 over the number of the goods (the additional product is averaged over the units).

Here’s why this distinction matters. System 1, the “thinking fast” system, cannot deal with sum-like variables, making thinking fast a difficult thing to do when creating economic models. But since the average consumer doesn’t use their System 2 frequently while at the grocery store, it might be valuable to introduce System 1 thought into some System 2 economic models. As current models of supply and demand are dependent on utility functions that assume sum-like evaluations of utility, they lapse in functionality when System 1 interferes with the rational consumer model. In order to account for this lapse, something that behavioral economics focuses on as a field, it might be possible to reorient the way we calculate utility such that utility is an averaging variable instead of a sum-like one. What exactly that looks like mathematically is something I’ve yet to apprehend.

This entry was posted in Uncategorized. Bookmark the permalink.

2 Responses to Thinking, Fast and Slow: System 1 and Sum-like Variables

  1. Karolina says:

    Hello,i accidently found your article, and am currently doing project in the university about “Thinking fast and slow”. But i only have two chapters of the whole book, and right now i’m struggling with understanding what is the “sum-like variable” exactly? If you find time during upcoming two days, and could answer that, i’d very grateful!
    Best wishes-

    • pwilliams says:

      I hope I can help you understand. I’ll do my best, although it’s not my forte so I know I’ll be weaker than Daniel Kahneman. The reason we call sum-like variables sum-like is because of the way they are valued. The example that Kahneman focuses on in Thinking, Fast and Slow is the dish sets. What it means for a dish set to be sum-like is that each additional unit included in the package should raise the value of that package by approximately the value of that particular unit. Instead, our minds value these dishes on average, rather than by adding up each dish. I’m not going to have space to unpack it in the same way Kahneman does, but basically, we are treating variables as sum-like when more quantity means more value (at about the per unit cost) instead of valuing more quantity by evaluating the dishes as a set. Each unit should add value to, not take away from, the whole.

Leave a Reply

Your email address will not be published. Required fields are marked *