The Institute for Fiscal Studies (IFS) recently published a paper claiming that ‘households receiving the winter fuel payment are almost 14 times as likely to spend the money on fuel than would have been the case had their incomes been increased in other ways’. They said this would come as a result of how the benefit is labelled: it’s because it’s called ‘winter fuel’ that people spend it disproportionately more on heating their homes in winter –an example of the fashionable ‘nudge’ theory (i.e. behavioural economics) in practice. For once, the IFS may be wrong.
The IFS economists are very competent, and this paper has been covered in the national media, so the possibility that they may have got it wrong took me by surprise.
Statistical regression analysis investigates whether one variable depends on a number of factors. Needless to say, there are always more potential factors influencing the variable under study than the number any researcher can get hold of. Provided none of the aspects excluded from the analysis are highly relevant, the results would be valid nonetheless. Instead, if some crucial factors were left out, then you cannot say anything about what’s being looked into –preferably, you would not even start crunching the numbers.
The IFS investigated the change in expenditure on fuel as a proportion of total household budget following the receipt of the winter fuel payment benefit. Through various statistical techniques they came to the conclusion above. But –and this is the killer ‘but’- the winter fuel allowance is automatically paid… at the onset of winter. Would people spend the extra £200-£300 in heating their homes if the allowance were paid in, say, June? Hardly so, even if it were re-labelled ‘the extremely polar conditions mother of all winters payment’. The IFS did not take into account one decisive factor: the time of year the payment is given. True, no-one can disentangle the likely influence of labelling and timing on how the winter fuel allowance is spent –but the consequence of this impossibility is that the existence of a labelling effect is far from established.
We, at Age UK, welcome the fact that recipients of the winter fuel payment spend more on fuel, of course, but this is beside the point: this extra spending may not come as a consequence of the word ‘fuel’ but of the allowance being paid when people need to heat their houses most –it’d be more the ‘winter’ side of things than the ‘fuel’. Hence, it’d be very premature to draw the conclusion that benefit recipients could be nudged towards consuming this or that by merely labelling and re-labelling allowances. It might work, it might not. The IFS paper didn’t prove it.
In some countries in Latin America, the annual salaries are divided into 13 payments –not because there are thirteen months, just in case you wonder. The 13th instalment –known as ‘aguinaldo’ in Spanish- corresponds to a payment given a couple of weeks before Christmas. What do people do with their ‘aguinaldo’? They spend it disproportionately on Christmas gifts, dos and dinners, and the vacations they tend to take shortly after. I remember one year in Argentina the government decided to postpone the payment of the ‘aguinaldo’ until the beginning of January. That year, Christmas dinners lacked some of the usual trimmings and retailers saw their seasonal sales plummet, but the following month there was more pocket money to spend by the seaside than usual.
Behavioural economics provides many a useful insight. How benefits are labelled may still be relevant –though I doubt that a ‘five-a-day allowance’ would do much to increase the consumption of fruit and vegetables- but as the potential effects of labelling and timing could not be told apart, there’s precious little we can say about the effect of the label on how older people spend their winter fuel allowance. The method had one flaw. Perhaps, the conclusion was flawed too.