Back in 2018, two economists made a £1,000 bet about future child poverty figures for the UK and I had the job of deciding who would win.
Jonathan Portes of King’s College London wagered that harsh social security policies would raise the headline poverty rate from 30 per cent of children in 2016-17 to more than 37 per cent by 2021-22. His economic model was predicting a rise to more than 41 per cent. Christopher Snowdon of the Institute of Economic Affairs took the other side of the bet, saying he’d seen similar forecasts far too often and they never reflected reality.
My role was as an independent arbiter if there was a disagreement over the bet’s outcome. There were various clauses that would nullify the bet — for example, if ministers did not implement the policies they had promised. But they did, and the child poverty rate for 2021-22 was published last Thursday and showed a rate essentially unchanged at 29 per cent. To my relief, my judgment was not required. Portes conceded and paid up.
Both Portes and Snowdon have outlined what they learnt from their big bet. In my view, there are four important lessons for social policy, how we talk about poverty and the complicated forecasts that arise from economic models.
First, we need to understand why Portes got it wrong. The child poverty measure in question was the proportion of children living in households with incomes after housing costs below 60 per cent of the median. The problem for Portes was not a government U-turn and greater generosity to the poor, nor more employment in response to stingy benefits, but a failure to foresee that median incomes could grow so slowly.
Social security levels barely increased over the period, but roughly kept up with incomes rather than falling behind — this meant that the headline relative poverty measure was left unchanged. The failure of Portes’ bet had nothing to do with poverty and everything to do with a much more disappointing period of general economic performance than he expected.
Second, it follows that the results demonstrate that the headline relative measures abuse the word “poverty”. Measured poverty tends to rise in good times because real median incomes increase faster and more people fall below a threshold linked to the median. They fall in bad times when true poverty is increasing.
Deprivation exists in many communities in the UK and genuine destitution in some. Tom Clark of the Joseph Rowntree Foundation highlights stories of families suffering on incomes below the breadline in his new book, Broke. The numbers involved, unable to feed their kids properly, are almost certainly rising, but account for far fewer than 30 per cent of children.
We shouldn’t rely on a flawed inequality measure to describe the plight of those worst off in the UK.
Third, Portes’ lost bet should not stop economists trying to predict the consequences of state policy. But they need to be more transparent about what can go wrong with forecasts. The past decade has seen many respected economic institutions, including the Institute for Fiscal Studies and the Resolution Foundation, making similarly erroneous predictions about rising income inequality and poverty levels.
The Office for Budget Responsibility is alone in regularly highlighting the errors it has made in its forecasts and publicly listing the lessons learnt. This should be normal practice.
Finally and most importantly, what really matters for genuine poverty reduction is economic growth. It improves employment prospects for the poorest households while also providing more funds for the government to redistribute without making others worse off.
Of course, addressing genuine poverty is more complicated than generating economic growth, but, truth be told, not that much more complicated.