What is counterfactual thinking?

Counterfactual thinking is all about asking “what would happen if I didn’t do x?” It is key to evaluating the impact of our actions. The impact of an action x is the difference between the world where I take action x, and the world where I don’t. For example, the impact of me throwing a stone into the water, is the difference between the world where I do throw the stone, and the world where I do not. This sounds very obvious. So what’s the big deal?

Suppose a business needs a £1,000 loan to stop it going bust. I’m considering loaning them the money. What is the impact of me loaning it to them? Well at first I might think the impact is stopping the business going bust! But what if I ask specifically “what would happen if I didn’t loan them the money?”. Well I’m not the only lender in town! They could quite possibly get a loan from a bank or from someone else. So in fact, they might not have gone out of business anyway and my impact is smaller than I think it is. In fact, I also have to think about whether the bank that would have lent them the money would lend it to someone else instead!

Another example is scientific discovery. In 1861 Louis Pasteur created the germ theory, realising that germs caused disease. What if he had not made this discovery? Well it’s likely that someone else would have eventually made the discovery. So the impact is not the difference between a world where no one ever knows about germs, and a world where they do. It’s the difference between the world where people find out about germs in 1861 versus a world where they find out in 1881. Still a big impact of course, but perhaps not what we would first expect.

A key point is choosing the correct counterfactual. Suppose you’re interested in becoming a doctor in the UK. One question you might ask is: What is the impact of being a doctor? Well suppose a man comes into hospital after an accident and needs emergency treatment. The doctor treats him and the man makes a full recovery. So is the impact of becoming a doctor at least saving one life? Well, compared to the scenario where the doctor refused to treat the man, maybe the man did die. Or perhaps another doctor stepped in and saved him. But there probably is a big difference between these two worlds. But this isn’t the correct alternative world. We’re considering choosing to pursue a career as a doctor. So the counterfactual is you not becoming a doctor at all in the first place. In the UK, many people apply for medical roles and get rejected. So it’s likely that someone else takes the doctor role you would have taken and treats that man, probably about as well as you would have. The difference between these two worlds is not that big, and so maybe the counterfactual impact of becoming a doctor is not l as large as we would expect! (This even is a simplification though. You have to think about what job you would do if you weren’t a doctor, and what job the person who would replace you as a doctor would have done instead, and the people they displaced and so on and so on!)

One type of counterfactual people like to consider is historical counterfactuals. What if x hadn’t happened? These are very tricky to consider. The world is incredibly complex and chaotic. Predicting different histories or the future is incredibly hard. But even before then we often run into a problem. Suppose we ask “What if World War Two hadn’t happened? What would the world be like?” The answer to this question is “it depends!”. The question isn’t specific enough. World War Two may not have happened because Hitler had a heart attack in 1934 and the Nazis pursued a less violent foreign policy. World War Two may not have happened because an asteroid hit the Earth in 1938 and killed all of humanity. We have to be very specific in how the counterfactual world is different. This is very hard. And the complexity of history makes it even harder, because I have no idea what I would have to change to make World War Two not happen.

Counterfactual thinking sounds obvious but we often make mistakes when trying to. It can be very hard and complicated, but that doesn’t mean it’s not the correct way to evaluate impact!