Covid-19: Forecasting and Modelling

Aaron Bell Excerpts

Tuesday 18th January 2022 Westminster Hall

[Aaron Bell (Newcastle-under-Lyme) (Con)]

It is a pleasure to see you in the Chair, Sir Edward, and to follow all my hon. Friends, who I note have usually been in a different Lobby from me on most coronavirus measures. I am sure the Minister will be grateful to have somebody speaking from the Government Benches who has been supporting the Government on coronavirus throughout.

However, I too have issues with modelling, which is why I chose to speak in today’s debate. I have more sympathy with modelling, and I will be offering some sort of partial defence and explanation of it in my remarks, because before I was an MP, I was a modeller myself—a software engineer. I wrote in Visual Basic.NET, which is nice and simple: engineers can see what the code does. I worked for bet365, and I used to write models that worked out the chance of somebody winning a tennis match, a team winning a baseball game, or whatever. I had some advantages that Neil Ferguson and these models do not have, in that there are many tennis matches, and I could repeat the model again and again and calibrate it. If I got my model wrong, there were people out there who would tell me that it was wrong by beating me and winning money off me, so my models got better and better.

The problem we have with covid is that we cannot repeat that exercise—there is no counterfactual. We have heard the phrase “marking your own homework”.

[Bob Seely]

I am deeply impressed by all this stuff— I do not quite understand what my hon. Friends are talking about, but it sounds fantastic. However, there is a counterfactual. The counterfactual is when people say, “We are not going to follow the lockdown,” and hey presto! we do not get 3,000 or 5,000 deaths a day and all the people who predicted that are proved wrong. There is a counterfactual called real life.

[Aaron Bell]

I thank my hon. Friend for his point, and I accept it, but the problem is that none of these models model changes in human behaviour. We discussed this issue during our debate on the measures that we brought in before Christmas, and as I said at the time, the reality was that people were not going to the pub, the supermarket or anything because they were changing their behaviour in the face of the virus. If the models do not take that into account, they cannot know where the peak will be. The models show what would happen if nobody changed their behaviour at all, but of course, the reality is that people do. We have not got good enough at modelling that, because we do not know exactly how people change their behaviour.

As a tangential point, behavioural science has had a really bad pandemic. We were told that people would not stand for lockdowns, but—to the chagrin, I am sure, of many of my hon. Friends—people did stand for them. Looking at the polling, they were incredibly popular: they were incredibly damaging, as colleagues have said, but people were prepared to live with lockdowns for longer than the scientists thought they would. There was initially an attempt to time the lockdown, because people would not last for that long. In reality, that is not what happened, so behavioural science also has a lot to answer for as a result of the pandemic.

I think that models still have value. My biggest concern arising from the experience of the pandemic is the bad parameters that have gone into those models at times—I will refer to two particular examples.

The time when I was nearest to following my colleagues into the Lobby was the extension to freedom day in June, because on that day we had a session of the Science and Technology Committee, which has taken excellent evidence throughout; it has a session on reproducibility in science tomorrow, where we will also look at this sort of thing. On the day of that vote, I was questioning Susan Hopkins and we were considering vaccine effectiveness. Public Health England had just produced figures showing that the actual effectiveness against hospitalisation of the Pfizer vaccine was 96%, yet the model that we were being asked to rely on for the vote that day said it was 89%. Now, 89 to 96 may not sound like a huge difference, but it is the difference between 4% of people going to hospital and 11%, which is three times higher. It was ludicrous that that data was available on that day but had not yet been plugged into the models. As I said to my hon. Friend the Member for Penistone and Stocksbridge (Miriam Cates), that was one of the reasons that I said in the Chamber that the case was getting weaker and weaker, and that if the Government tried to push it back any further, I would join my colleagues in the Lobby on the next occasion.

The other case is with omicron. Just before Christmas, we had these models that basically assumed that omicron was as severe as delta. We already had some evidence from South Africa that it was not, and since then we have discovered that it was even better than we thought. That feeds into what my hon. Friend was saying about the total number of people who are susceptible. The fact that omicron has peaked early is not because people have changed their behaviour but because the susceptible population was not as big as we thought: more people had been exposed, more people have had asymptomatic disease. There are all those sorts of problems there.

More philosophically, my models when I worked for a bookmaker were about probabilities. Too often we focus on a single line and too often that has been the so-called worst-case scenario. Well, the worst-case scenario is very black indeed at all times, but Governments cannot work purely on a worst-case scenario; they have to come up with a reasonable percentile to work with, whether it is 95% or 90%. Obviously, it must be tempered by how bad the scenario would be for the country. The precautionary principle is important and we should take measures to protect against scenarios that have only a 5% chance of happening or indeed a 2% chance, but we should do that only if the insurance price that we pay––the premium for doing that––is worth paying. That comes down to the fact that not many economic models have been plugged in, as my hon. Friend the Member for Wycombe (Mr Baker) has repeatedly said in the Chamber and elsewhere throughout.

Any Government must try to predict the course of a pandemic to make sensible plans and I believe that the best tool for that is still modelling, but we must learn the lessons of this pandemic. We must learn from shortcomings such as the failure to understand human behaviour properly, the failure to make code open source so that other people can interrogate a model and change the parameters, and the failure to enter the right parameters and update the model at the moment politicians are being asked to vote on it. For all those reasons, I am grateful for today’s debate and look forward to hearing the Opposition spokespeople and the Minister. I thank my hon. Friend the Member for Wycombe for today’s debate.