top of page

How far can we rely on the predictive judgements made by professionals?

On the evening of January the 4th 2021, British students heard an announcement that under normal conditions would be like music to their ears. Boris Johnson, the British prime minister at this point declared that schools would shut down temporarily and everyone should stay at home for a second time. Another lockdown! This decision, like many others, was made to prevent the Covid-19’s impact on the country. But was this the right decision? Or, more interestingly, how big is the role that psychological variability can play in a politician’s judgement? The question is whether Mr. Johnson would have come to his conclusion at different times that same day?

We like to believe that those making principled decisions should not be prone to inconsistency within their judgements. Unfortunately, one of the many dangers of politics are evaluative errors, where executives are often faced with situations in which they must make informed choices with significant stakes. This fundamental issue undermines how much trust we can assign to the people allocated the authority to make judgements on behalf of the general public. The COVID-19 pandemic is an excellent example of this issue, as the pressure put on policymakers to move forward and make decisions based on uncertainty was immense. The problem with choices made under such novel circumstances is that politicians do not tend to have comparable past experiences readily available to provide quantifiable scales to guide their judgements.

For the first question, we can only speculate. Attempted evaluation of the effectiveness of implemented policies occurs in hindsight, after the policy (and its consequences) have occurred , and even then, it is difficult to isolate each cause and its effects. Even though the pandemic hit various countries at around the same time in roughly the same manner, countries reacted very differently, which proves that there exists some variability in decision-making where the outcome is yet unknown.

However, the second question is far easier to address from a statistical point of view. Evidence suggests, for example, that decision makers tend to be more pessimistic when they are hungry, meaning that if Boris had waited until after dinner before deciding whether to close the schools, policy could have been significantly different. We can make assumptions about the consistency of judgements based on the mental processes we engage in, where there is a very low probability for them to be precisely the same at any given time. The mind is a measuring instrument that is unpredictable and inevitably leads to inconsistency in human-made conclusions due to its susceptibility to varying situational factors such as mood, fatigue, or social reinforcement.

When defining political judgement as a form of measurement suggesting cognitive assessments rather than simple opinions, it implies that these judgements aim to get as close to an actual value (i.e., an optimal solution) as possible. Consequently, two different conclusions based on the same set of data statistically cannot both be correct.

A kind of lottery is involved when asking the question of which answer will be proposed by the same individual. Why is that?

This is where an unwanted statistical variable comes into play. This error is called “Noise” and was highlighted by the so-called father of behavioural economics, Daniel Kahneman. Because politics involves singular problem-solving rather than repeated analysis of the same case, this phenomenon in policymakers is not easy to spot. When observing professionals making recurrent strategic decisions, by contrast, one can notice that the same person will often come to different conclusions when evaluating the same problem twice. For example, in the United States the chance that the same judge will assign asylum to an applicant reduces by 19% if the individual's hearing follows two successful ones (Ramji-Nogales et al., 2007). This sounds bizarre; one would expect them to have an accurate picture of a case at both times if the data stays identical and not be susceptible to influences like streak aversion.

Over the long run, there is arguably cause for comfort; the positive and negative deviations from the actual outcome could cancel each other out. However, as the decisions that policymakers make are not repeated, and the institutions involved don’t make multiple evaluations of the same case, every faulty conclusion is simply added to the pile. So, when Mr. Johnson drastically changed policies during the pandemic, we cannot know which one hit the nail on the head. The only thing we can know is that each approach that wasn't the optimal solution would have added to the total sum of “Noise”.

The argument that politicians can claim to have at least some expertise in their profession, and should be sufficiently knowledgeable to accurately seek a solution in a single decision-making situation, has merit. However, because of the informal nature of the types of judgements they must make i.e. decisions that are not bound to a rigid set of rules but are guided by individual experience and subjectivity, space for the noise problem is created.

There is a difference between wanted diversity of thought and unwanted, irrational variability. The former is welcome in competitive situations, where the best judgement is rewarded. Political opinion must vary among people to some extent as it provides space for debate. Variability is wanted in contention but a crucial disability in analysing facts and data.

However, this article is concerned with assessments that are primarily irrespective of personal preference but the fact that even this single individual will show inconsistency within their judgement regardless of their ideology.

We should also realise that political decisions are rarely made in isolation but are more likely to occur under group conditions. Intuitively one would think that this might ease the problem of having one individual randomly assign the solutions to a one-shot question. However, factors such as who speaks first, the sceptic or the optimist, might sway the opinions of the whole group towards one polarised end instead of delivering a well-balanced analysis. For example, research has shown that when Juries have to allocate monetary fines to punish a company for product liability cases, their final value turned out to be more extreme if this group's median member favoured this standpoint (Kahneman et al., 1998). This shows that even when introducing multiple individuals to a problem, there is no guaranteed diversity in perspectives when trying to find an optimal solution.

Furthermore politicians, just like any other human being, are also susceptible to a concept called Naïve Realism, which is the propensity to believe that our perception of the world reflects it precisely as it is. Therefore, we tend to be satisfied with a single interpretation of an issue and rarely seek alternative explanations. It is also a matter of environmental reinforcement as politicians share a common language and establish rules about considerations that matter in making a judgement. They establish a certain fluency and ease by solving familiar issues. Therefore, when a complex problem such as the pandemic is approached in a reductionist manner, politicians might fall into the trap of unconsciously agreeing with their past selves and applying the same problem-solving techniques that they are confident with, even though each novel case should invite original judgements.

Independently of Noise, pre-established behavioural economic theories have illustrated how cognitive errors can lead economic agents to come to systematic faulty conclusions, which are undesirable but at least can be foreseen. Biases provide a causal explanation that can help us find solutions to avoid such mistakes. Nevertheless, the concept of Noise is more complicated as it can only predict that there is some sort of variability in judgement but cannot quantify the extent of these errors. The only attempt one can make is to ensure that decisions are made only under conditions where our cognitive state is at its optimal, such as in the morning rather than at night, next to encouraging a healthy amount of doubt concerning our general overestimation of the likeliness that our individual judgement must be closer to an actual value as any other. Maybe it was not for nothing that Boris Johnson earned the nickname Mr. U-Turn during the pandemic. It could indicate that there is already some awareness about inconsistencies in decision-making, even in the political sphere.


bottom of page