Those of us who are friendly to decentralized market processes and are skeptical of centralized intervention in the market often speak about unintended consequences of intervention. The idea of unintended consequences is important, but it also comes in a few different degrees and forms worth unpacking. For now, I want to elaborate and expand on the varieties of unintended consequences – I’ll discuss some implications of these distinctions in a future post.

I’ve often heard the terms “unintended consequences” and “unanticipated consequences” used interchangeably, but there is an important distinction between these two phrases. Not all unintended consequences are unanticipated. It’s possible to anticipate an outcome which you don’t necessarily intend or even desire, but still find the action worth carrying out. An easy example of this is known and predictable side effects of medication. When a doctor informs a patient that some drug will treat their condition but is also known to cause drowsiness as a side effect, the drowsiness is an unintended but still anticipated consequence of the medical intervention. The doctor does not intend to make the patient drowsy, but nonetheless still anticipates drowsiness will occur.

Unanticipated consequences are exactly what they sound like – a result which was neither intended nor predicted. In this framing, unanticipated consequences are a subset of unintended consequences. Not all unintended consequences are unanticipated, but all unanticipated consequences are unintended. One example of unanticipated consequences comes from a program putting bounties on cobras in India. The British government wanted to reduce the number of cobras, and so decided to pay people for every cobra they killed. Seems reasonable, right? But the policymakers didn’t anticipate how people would react. Many people simply began to breed cobras in large numbers, in order to kill them and turn in their skins for money. Eventually, the British government realized what was happening and terminated the program. This in turn led the snake breeders to release their now worthless breeding stock. As a result, the cobra population actually increased. This consequence was both unintended and unanticipated.

Even within the framework of unanticipated consequences, a further distinction deserves to be made – between predictable and unpredictable unanticipated consequences. Just because an outcome was unanticipated (and thus unintended) doesn’t necessarily imply that it was unpredictable. If you’ve ever said to yourself “I should have seen that coming” when something catches you off guard, you’re acknowledging that whatever just happened was in principle predictable, even if it was unanticipated in practice. One could argue that the outcome of the snake bounty program was unanticipated but was still predictable, if the architects of that program had thought more clearly about the incentive structure they were creating. However, it’s worth pointing out it’s much easier to retrodict an explanation for why an outcome should have been anticipated than it is to predict that outcome in advance. Hindsight bias probably makes many instances of what was unpredictable at the time seem predictable in retrospect – not in every case, but at least some of the time.

To all this, I would add one further meta level distinction – the predictability of predictability. Was the fact that an outcome was unpredictable itself something that could have been predicted? A clear example of predictable unpredictability would be a Powerball drawing – you can’t predict what any given drawing will be, but you also know the result can’t be predicted. By contrast, an outcome can be called radically unpredictable when the unpredictability of the result was not itself predictable. Black swan events, as described by Nassim Taleb, are events which are radically unpredictable. Not only can you not see them coming, but you don’t even know that you can’t see them coming.

To sum up: when we act, we do so to bring about a particular result. Some results of our actions are unintended but still anticipated. Other times, there will be results which are both unintended and unanticipated. Among this result set, some of the unanticipated results were predictable in principle, while others could not have been predicted. And among the results that were unpredictable, some of them will have been predictably unpredictable, while others would be radically unpredictable.


Kevin Corcoran is a Marine Corps veteran and a consultant in healthcare economics and analytics and holds a Bachelor of Science in Economics from George Mason University.