In a recent Times article, Juliet Samuel argues that a "new modelling army is leading us astray", and that, "be it policy for lockdown, net zero or immigration, it must be ministers and not scientists or analysts who lead the way". It may be too late for that, especially when models used are often hidden from public view.
The article was mainly about the pandemic, but also looked at the proliferation of models, modelling and modellers across our lives. Models are a key ingredient in the quantification and future-focusing of climate change. The IPPC uses them to game a huge set of 'what ifs'. It usually ends up by setting out a number of possible scenarios ranging from the worst-case down to the relatively benign. Critics say that too many pressure groups focus on the worst-case prophecies in order to maximise their influence, with ministers not usually knowing any better than to accept them.
Samuel wrote:
"This is a problem beyond pandemic management. It is rife across government. Treasury economic and fiscal models have systematically starved our economy of investment. Highly speculative climate models and net zero accounting (which ignores emissions driven offshore) have driven us to adopt a completely unworkable set of energy policies that have loaded up costs while having no material effect on global emissions.
In immigration policy, economic models that focus on labour costs, productivity and wage growth have eclipsed qualitative arguments about social cohesion, values and cultural change. Never mind, say the models, if veterans can no longer sell poppies for Remembrance Sunday without being harassed. Just think of the marginal productivity gains.
It’s easy to see why politicians have surrendered to models. Most are not experts in the fields in which they are making decisions. Expert modellers have invested years in honing and promoting their methods and bear no responsibility for the decisions made. If ignored or defied, critics hungry for ammunition to use against ministers will deploy them with relish. For ministers, there is little to gain and everything to lose by trusting their instincts or moral compass above the conclusions of an apparently sophisticated model.
But in the end such models are fickle friends. They might be useful tools with which to consider scenarios, devise specific strategies or test one’s beliefs, but they cannot take the place of judgment calls and values. They cannot decide what is right or wrong or give us a true measure of the human soul. For that purpose, God help us, we elect leaders. And when they fail to lead, we eject them, models be damned."
.........................................
My title comes from an iconic Flanders & Swann song [Friendly Duet] in which there's a reference to the Profumo affaire in the early 1960s when fashion models moonlighting as call girls led a government minister (and at least one peer of the realm), to stray. F&S concluded that "such models of friendship are precious and rare though the friendship of models is not!"
Just so. You can listen to it here.
Responses
It is refreshing to see this post about modeling. While computer generated models have a place in helping to guide potential decision making, but they are rife with danger when they are too readily believed as infallible. Once only has to look at weather forecasting, with a long history of fallibility, to understand that 'chaos factors' always undermine conclusions. Allowing technocrats to 'put all our policy eggs in one basket' because they fit a current belief is more catastrophic than many of the worse case scaenarios.