This blog is based on an article in the Journal of Social Policy. Click here to access the article.
In the global North, comparisons of poverty rates across countries, time and demographic groups rely largely on indicators of which households have low relative incomes – such as those below 60% of the national median. But such household-based analysis depends on crucial assumptions about the relative costs faced by different household types. You need more income to escape poverty if you’re a couple with three children than a single person. But how much more? Assumed relativities are incorporated into ‘equivalence’ scales used to adjust household incomes before comparing them.
These equivalence scales matter, particularly to which groups show the highest poverty rates, influencing policy priorities. Yet their values are based on a range of abstract and contested evidence and analysis, often combined in an arbitrary fashion. New analysis based on more direct estimates of the minimum needs of various household types shows that the prevailing equivalence scales are not just inaccurate, but systematically so. In particular, they consistently underestimate the relative cost of children.
The scales currently in use have their origins in an economic approach to estimating equivalence, based on how different kinds of households distribute spending between essential and discretionary items as they become better off. Rising income is associated with more spending on discretionary things like holidays and proportionately less on essentials like food. So economic models assume that two different households (such as a single person and a couple with two children) have the same living standard if both spend the same proportions on discretionary items. Observing the difference between the average incomes of larger families and smaller families who reach such a common ‘equivalent’ standard allows the scales to be constructed.
Unfortunately, technical applications of this method are widely disputed by economists, and the scales they produce also suffer from two underlying weaknesses. One concerns the robustness of the assumption that, across different types of household with very different types of need, an equivalent living standard can be inferred solely by how much spending is on ‘discretionary’ items. The other is that the prevailing current scale being used (the ‘OECD modified’ scale) was derived over 25 years ago, drawing on a combination of studies and some largely arbitrary judgements about which weightings to use.
A newer strand of research looks at different household needs much more directly. The Minimum Income Standard (MIS), a research technique developed in the UK and currently operating in nine countries, identifies minimum household budgets by listing the actual goods and services required to reach a an acceptable standard of living. Judgements about what this entails are made by groups of ordinary citizens, and items costed by researchers.
As an example of how this method informs equivalence calculations can be seen in citizens’ judgements about minimum transport needs, in urban parts of the UK outside London. They say that you can manage life without a car if you do not have children, but families need a basic car. This has two kinds of implication for relative transport costs. It increases the degree to which they increase when you have children. It also means that adult transport costs are about double for a couple than a single for those without children (two people rather than one using public transport), but a lower ratio once they start a family (because one car can serve two adults). A single parent gets the rough end of both these aspects: they need a car to get their children places, but don’t economise by sharing it between two adults.
A comparison of four countries with comparable results using the MIS method produces some telling findings. Many features of their results vary significantly across countries, influenced by differences in living patterns, in the extent of public provision and in relative prices of different kinds of goods. However, what is striking is that in all the countries in this comparison, the cost of children relative to adults is shown to be greater than assumed by the main official equivalence scale. They also predominantly suggest that the scale underestimates the cost of singles compared to couples (ie how much couples save through economies of scale). The equivalence scales also take no account of concessions that bring down costs to certain groups, like free bus passes and prescriptions for UK pensioners.
The combination of these factors can cause big distortions. For example, in the UK, the equivalence scale applying to incomes net of housing costs suggests that life costs slightly less for a lone parent with two children than for a pensioner couple; yet the MIS results show that the minimum income required for the lone parent family is about a third higher than for the pensioners. As result, in official statistics, just under twice as many individuals on low income are shown to live in lone parent families than in pensioner couples, whereas according to the MIS research, it is actually over four times as many.
These findings are hard for policy makers and poverty analysts to ignore, given their size and the consistent direction of the bias, across countries. Yet it would also be difficult to use this evidence simply to construct alternative equivalence scales. Not only do relative costs for different groups vary considerably across countries, but they can vary across time with precisely the kinds of policies designed to respond to poverty, such as subsidising certain goods and services for low income groups. A single, inflexible equivalence scale will never capture these changes. Instead, ongoing research looking directly at minimum costs and the things that influence them needs to be taken into account alongside the standard income poverty indicators, to inform policymakers about which groups are having the toughest time.
About the author
Donald Hirsch is Director of the Centre for Research in Social Policy at Loughborough University.