A Toolbox for Rationality

I have no time, no energy, don’t care enough and am stressed out of my mind. Are you surprised my decision wasn’t optimal?!

We make many decisions a day. As a result, we attribute a lot more time and cognitive effort into some decisions than we do into others. We don’t have unlimited time and resources to find out the best option and choose it. To deal with this we have learned certain rules of thumb. These rules tend to be somewhat correct, because otherwise we would have either updated or discarded them. But, they are not applicable to each and every situation. As a result, continuously using these rules of thumb makes us prone to mistakes. It happens. We tried our best with the (limited) resources we had.

As a result, we are irrational according to the strict guidelines posed by neo-classical economists. We fail to maximise our utility by making the optimal and therefore rational decision. Because we have such constraints on our time and resources, we don’t maximise, we satisfice. We are just aiming to find an option that satisfies all our criteria, and then move on with our lives instead of building an algorithm that finds us the optimum. Is not wanting to spend half a month deciding whether to eat the chocolate pudding or not, more rational than spending all that time reaching your optimal result? At least, that is the argument proposed by bounded rationality.

Bounded rationality was proposed by Herbert A. Simon, using the analogy of a pair of scissors. One blade represents the cognitive limitations of the human mind and the other blade represents the structures of the environment. He was trying to illustrate how minds compensate for limited resources by exploiting known structural regularity in the environment.

If that analogy did not really hit home with you, don’t worry. Not to take credit away from Simon, but he’s not the person who made bounded rationality famous. That honour falls mainly to Gerd Gigerenzer. Gigerenzer was of the opinion that decision theorists have not really adhered to Simon's original ideas. Rather, they have considered how decisions were crippled by limitations to rationality, or have modeled how people might cope with their inability to optimize. Gigerenzer rejected the idea of bounded rationality being a limitation and proposed that the simple rules of thumb we have are optimal in themselves. In his most famous paper, which we will now dive into, he argued that simple heuristics often lead to better decisions than theoretically optimal procedures.

In the 2002 paper, Gigerenzer and Selten outline what is now widely known as the adaptive toolbox, a seemingly easier to understand analogy. The rules of thumb were dubbed mental shortcuts, or just simply heuristics. They argue that heuristics were better adapted to the decision-making environment, and as a result better than general-purpose algorithms. When I say better, I am referring to a double-whammy of excellence: heuristics are faster as they use less cognitive resources and in contradiction to economic theory: more accurate.

So what tools do we have in our toolbox? The three ultimate ones are rules of searching, stopping and decision. Searching rules focus on what you actually want to find in your environment. They identify the available choice set. We use heuristics for this. For example, when looking for your next holiday destination. Instead of carefully weighing each and every benefit and cost of the destination we take the most important characteristic: pearl white beaches. If we employ a Take-The-Best heuristic we will simply look for the best pearl white beach in the world, no other questions asked about temperatures, prices, accessibility etc. We have booked our fights and hotel and have called it a day.

Stopping rules focus on when to stop searching. A heuristic for this is to stop as soon as the next alternative is not better than the option we currently have in our minds. Another way to stop a search is to set a limited time frame in which you can come up with alternatives. Emotions are great for stopping search as well. Instead of deciding every day, making use of an extensive cost-benefit analysis, whether we want to stay with our partners and/or children, we look at them and spend time with them. Their presence in our mind invokes feelings of love. As a result we stay. That is all it really takes. The opposite holds as well. If your partner inspires feelings of disgust, most people call quit and move on.

Lastly, decision rules. Once we have decided to stop searching, a decision must be made regarding the choice set we have accumulated through our search. In economics, it is now time to implement a multiple linear regression model to value all the cues of the choice options against each other, having pre-determined which weight to attribute to each cue, and choose the one having the highest value. Is this too complicated when you are trying to pick out what to eat for dinner tonight? Definitely. Most people employ heuristics for this as well. I have already mentioned the Take-The-Best heuristic which focusses only on one cue and maximises that one. Another one is known to most people as a habit: on Tuesday we eat pancakes, on Wednesday we eat potatoes, fish and salad. Really does solve a lot of planning issues.

Another interesting point is that the adaptive toolbox does not care much for transitivity, or consistency in choice and judgement. Because bounded rationality focusses so much on what is needed in a certain situation, it accepts variability. So don’t get confused by the habit example. Bounded rationality recognises that in certain situations, consistency is key. In others it is unimportant, and sometimes it can be a massive disadvantage. An example provided in the 2002 paper is that of relationships. Trust, communication and fairness are integral parts of cooperative relationships. But when taking a different kind of relationship, let’s say a competitive relationship, our judgement and behaviour should not follow the same (consistent) structure. We have not become intransitive, we have adapted towards what the situations demands from us. We are ecologically rational.

Making use of the adaptive toolbox, or any type of heuristic sounds simple. And it is. Which is exactly why it has gotten such a bad rep. Because we have a heuristic in our mind saying: the more complicated something is, the better/more accurate it is. As mentioned before, not all our heuristics are accurate all the time, but they do a pretty decent job.

The learning of inaccurate heuristics is something we will look into in the next article, in which I will discuss the heuristic that assumes that price is directly impacting quality.

References Gigerenzer, G., & Selten, R. (Eds.). (2002). Bounded rationality: The adaptive toolbox. MIT press.