As the prime minister prepares to respond to his MPs' demands to publish a cost-benefit analysis of the tier restrictions, Giles Wilkes argues that it would not improve decision-making – or public confidence in the decisions taken
There is a growing clamour for the government to release a cost-benefit analysis (CBA) of its lockdown decision-making. This can be heard mainly from a band of rebellious Tory backbenchers, grown increasingly unhappy with the prime minister’s decision to introduce a second full national lockdown a month ago, and now chafing at a tiering system that feels as harsh as what went before.
Superficially, nothing could sound more reasonable. The powers taken by this government are as extensive as any since the Emergency Powers Act (1939) which, to paraphrase, allowed the government to do whatever it wanted to win the war. Since the passage of the Coronavirus Bill, the economy has endured its sharpest recession since a frosty year in 1709. Education ground to a halt, as did many non-Covid health services. No one can doubt that there have been costs to the Covid measures – can these not be weighed against the benefit of a lower infection rate?
On our podcast, the IfG’s director Bronwen Maddox called me “no stranger to cost-benefit analysis” and asked what I thought. My answer was immediately to turn to the serious difficulties of carrying out such an exercise.
I write this as someone who has guiltily tried to build my own “SIR” models of the spread of the infection, inspired by the example of Ian Mulheirn at the Tony Blair Institute. Any such attempt must pin down several great unknowns. We do not know how much people will comply with measures, whether they are smart in how they shun riskier ‘super-spreading’ activities, or how different types of people react to having a different vulnerability to the disease. You can build a model that roughly apes the course of Covid so far – I found that this was possible, if you just assume that people have short memories and increase their risky behaviour as the salience of high death rates slips from their mind. But this gave me no confidence that I had a tool for predicting what might happen next.
Of course, there is much that we do know. In particular, we can see that lockdowns reduce infections – as I write, there has been welcome news on a fall in the “R” rate of the virus. This can be contrasted with the fast-growing numbers in the United States, where efforts to stop the spread have been lax in many states (South Dakota, with a population one seventy-fifth of the UK, has suffered 400 fatalities in November alone). But we cannot know the counterfactual with any accuracy, and the experience of Downing Street briefings is that attempts to set out the dire consequences of failing to act risks undermining the credibility of the forecasters. People’s behaviour will almost certainly render them inaccurate.
Then there is the challenge of calculating economic harm. Perhaps it is possible to generate an all-encompassing model of the economy fuelled by people’s interactions, which manages to capture both their propensity to spread the disease, but also how such contacts generate GDP. Even then, however, it is fair to ask how much the immediate fall in GDP should weigh, compared to the medium-term scarring effects of permanently lost jobs and shuttered businesses. How should we change our verdict if the government efficiently ‘shields’ businesses through policies like the Coronavirus Job Retention Scheme? How to value the longer-term damage from lost months of education for school children – or the cost of a parent incapacitated by Long Covid? These are irreducible value judgements.
Contra Homer Simpson's mantra, something being hard is a poor reason not to try. But what worries me more about calls for a CBA of Covid is how it would mislead the public about how decisions are really taken. Government is riddled with CBA, and analysis is taken very seriously. I remember, for example, the analysis the Treasury carried out in 2013 of the effect cutting corporate tax (higher business investment of up to 4.5 per cent). Every new regulation was meant to be subject to a Regulatory Impact Assessment, measuring how much it might cost business, scoring every department on the burden it created.
Yet in no case can I recall a decision made because the CBA said so. The truth is that (as with the corporate tax cuts) the analysis often followed a decision, rather than led it. Even if a Treasury analysis generated a higher cost from imposing a lockdown than the benefit, this would not oblige any politician to salute and obey. In particular, models can produce an average expected result, but what keeps a prime minister up at night is the small risk of something truly disastrous – like a spike in infection overwhelming the NHS, the mortality rate sky-rocketing, the public in a state of panic, and the state’s hand forced by the course of chaotic events now totally out of its control. That kind of risk is ultimately a matter of judgement for the prime minister, not something a brilliantly-formulated spreadsheet can handle.
Measures to control coronavirus remind me of other difficult and controversial decisions, in particular the austerity measures of ten years ago. Then, too, politicians said they were doing something difficult now to forestall worse outcomes that they could not prove would happen (and might not have). I remember worrying that the creation of the Office for Budget Responsibility (OBR) might be undermined by politicians trying to enlist an authority figure on their side of a political argument. But I was wrong, in large part because the OBR was never called upon to do anything as controversial as an objective analysis of whether austerity was worth it. Instead, it produces reams of useful analyses of the effect of individual measures – and honest discussions of the uncertainty inherent in the task it performs. It has enhanced the quality of public fiscal debate, without impinging on the political decision-making.
There is no doubt that we could have better public information about Covid-19 and its management. But the emphasis should be on smaller matters than whether to lock down or not: for example, the effect of different-length quarantine periods, the relative riskiness of a busy bar or restaurant compared to a quiet café, the role (or not) of schools in spreading the disease, and the effect of mass-testing such as Liverpool has recently undergone.
The pandemic is the most unanticipated crisis in UK government history, and the decisions the government has made have reflected the learning curve up which it has been scrambling. Something as definitive as a cost-benefit analysis of the lockdown would prove a political distraction, and is unlikely to improve either the decision-making or the public’s confidence in it.