Skip to main content
Open this photo in gallery:


Jillian Horton is a Canadian internist and writer. Her memoir, We Are All Perfectly Fine, will be published this month.

Have you ever stayed far too long in a relationship that was clearly failing? Or kept a piece of ugly furniture you really wanted to get rid of, still dragging it with you every time switched apartments? If you have, you’ll understand at least one of the reasons most of Canada’s premiers have struggled to make appropriate shifts in public health strategy at critical junctures in our long struggle to contain COVID-19.

On Aug. 7, just five days before Manitoba’s ill-fated reopening, I wrote about the concept of situational awareness, and how it helps us avert disasters before they happen. Countries that have displayed high and consistent levels of situational awareness have been rewarded for their early and aggressive pre-emptive actions. Their lockdowns are limited or over, their GDPs have taken the smallest hits, and they remain ready to take swift and definitive action as soon as single cases or small clusters appear. COVID-19 follows a pattern of predictable, exponential growth – in that one regard, it practically comes with its own crystal ball. So why have most provincial governments continued to make the same mundane stumbles over and over? As with the lack of situational awareness, I think some of those shocking failures can be explained by the frequent occurrence of a second thinking problem: a cognitive error known as the sunk costs fallacy.

Richard Thaler, winner of the 2017 Nobel Prize in economics, first described the sunk costs fallacy as the tendency for us to make decisions about present actions based on our past investment of money, time or energy – whether or not that investment has been fruitful. In this framework, instead of 20/20 hindsight, our view of the present is actually clouded by what our choices in the past have cost us – despite the fact that our investment cannot be recouped – keeping us locked in, even when our horse is clearly flailing. In medicine, the sunk costs fallacy often manifests as sticking to a particular diagnosis or management plan even when it isn’t working. In a sense, our thinking is distorted by our subconscious attachment to what we have already invested. So we sink deeper into the hole we’re digging, unaware of the fact that we’re sinking at all.

With the situation so desperate in most of the country, some people might say, “Isn’t the lockdown itself the sunk cost?” No, the sunk cost is the too-late, too-light, half-baked lockdowns that have allowed COVID-19 to spread beyond the point of control in most of our provinces, with the notable exceptions of Nova Scotia and PEI. The rest of our lockdowns have generally been associated with high loss of life because of our premiers’ attachments to their original choices, including slow closings, premature openings and delayed reactions to rising case counts. And because they had invested so much of their political and leadership capital in those decisions, they struggled to separate the cost of correcting those errors with what they had already lost. The irony, of course, is that this attachment made everyone’s losses even worse.

How can our leaders avoid being unduly influenced by the sunk costs fallacy? They can start by simply being aware of what it is, and how it tends to influence their decisions. When I’m managing a complex medical case, especially when things aren’t going well for my patient, I know it’s crucial to stop and ask myself: Are my choices today clouded by yesterday’s failures, or do they have internal validity right now? Another strategy is using the cognitive framing that is sometimes called “beginner’s mind.” A useful question from a beginner’s mindset is: If I were encountering this situation today, for the first time, what would I decide to do? This simple exercise can liberate us from the outsized influence of sunk costs, allowing us to see things in a different light.

Many premiers will fret that the political cost of changing course is untenable. But the view that a leader shouldn’t change course when their strategy is failing always spells trouble. Do you want to be a passenger on the plane where a pilot who believes they should never override their flight plan, even if the weather changes? The safest teams in medicine, aviation and other high-stakes industries allow for acknowledgment of mistakes and real-time adjustments. In fact, in medicine, the best-led teams report making the most mistakes. That’s because those teams are the ones where people can admit their errors without fear of retribution or criticism – or, notably, fear of not being re-elected.

Finally, immunization against sunk-cost losses involves leaders tasking people around them with the job of pointing out what the leader might be missing. This is particularly important in any setting where loyalty means unbridled support for the boss, no matter what happens to be coming out of the boss’s mouth. Sycophants don’t speak freely, and squeaky wheels are more likely to be replaced than greased. The end result is a dangerous form of groupthink that puts us all at risk.

Only a fool would argue that our premiers face easy decisions right now. Businesses are facing utter devastation, Canadians are sick and dying, and citizens who have been following public health rules are struggling with pandemic fatigue. But too frequently, our premiers have made these problems worse by falling into the trap of repeating an endless loop of the same mistakes, and ultimately sinking deeper into the mire. The sunk costs fallacy tells us that previous errors are destined to script our choices in the future – unless we are willing to cut our losses, admit our previous failures and start again. The best doctors do it all the time, and so do the best leaders. And that is the good thing about calling a spade a spade. At least you can start using it to dig yourself out.

Keep your Opinions sharp and informed. Get the Opinion newsletter. Sign up today.

Your Globe

Build your personal news feed

Follow topics related to this article:

Check Following for new articles

Interact with The Globe