Progress on progress: how (not to) measure prosperity
It is well known that the Gross Domestic Product (GDP) is ill-suited as a measure of material living standards. Efforts to find alternatives abound, ranging from surveys of well-being to a multiplicity of measures that lump together various indicators of prosperity into one overall score. We have contributed to this debate through the LSE Growth Commission report published earlier in the year.
Last week this somewhat crowded and competitive arena welcomed a new member to much fanfare – the ‘Social Progress Index’ (SPI), published by a Washington-based think-tank. It made headlines the fact that the UK came second amongst a sample of fifty countries, outstripping the United States, Germany, France and Japan for overall progress in providing for ‘basic human needs,’ laying the ‘foundations for individuals to enhance and sustain wellbeing’ and offering ‘opportunity for all citizens to reach their full potential.’
The team of US economists (Harvard and MIT) that produced the index make bold claims about the distinctiveness of their creation. This index, they argue, is different in that it does not allow for standard, uni-dimensional economic indicators like income, but focuses instead on a broad range of social indicators that seek to capture what ‘really matters to people’s lives,’ such as ‘do I have a roof over my head?’ and ‘do I have enough food to eat?’ It is different because it draws on ‘the most granular outcome elements available given our current understanding from diverse literatures in economics, sociology, history, political science, and social psychology.’ And in packing these elements into an overall score, the authors were ‘agnostic’ in that they ‘did not judge any one of them to have an a priori higher weighting than any other.’
So, is the UK really at or near the top in terms of what ‘really matters?’ Is this index truly of a different breed, capturing aspects of social progress which so far have escaped the radar of other measures? The short answer is no.
As it stands, the SPI index suffers from a number of important conceptual and methodological problems. The top three in my list:
• Creating an artificial dividing line between social and economic indicators. It is not by coincidence or mistake that the words ‘social’ and ‘economic’ are often found together. Economics, properly defined, is an inquiry into social prosperity, so it is should not be surprising that its empirical ingredients (income, for example) focus on precisely the same sorts of things that the SPI index purports to measure – what “really matters” to people’s living standards. To exclude ‘economic’ indicators on the grounds that they are not ‘social’ indicators is a contradiction in terms;
• Using ‘prevalence of tuberculosis’ and ‘access to piped water’ as measures of basic human needs; ‘primary and secondary school enrolment’ and ‘deaths from HIV’ as foundations of well-being; and ‘freedom of speech’ and ‘freedom of assembly/association’ as measures of opportunity, is likely to bear fruits in the context of developing economies, where these types of issues ‘really matter.’ In the case of mature economies (like the UK), where these issues matter a lot less, this approach becomes mostly pointless, struggling to find any significant variation between countries.
• Eschewing difficult questions about the relative importance ascribed to the different components of the index (or, in the technical jargon, the weighting scheme used) is tempting, but ultimately misguided. To argue that, by weighting each component equally, one is being ‘agnostic’ and not making judgements about their relative importance, is to divert attention from the assumptions/judgements that one does – in this case to weight each component equally. It does not resolve the problem. It simply dodges it.
Efforts to move beyond GDP in measuring socio-economic prosperity are a praiseworthy endeavour. Our work with the London School of Economics suggested that a useful way of moving the debate forward would be to supplement GDP releases with information on median household income – a better way of capturing the living standards experienced by the bulk of society.
Developing composite indexes of the SPI type is another approach that may offer novel insights about how living standards are changing and how policies can help to support improvements. Yet, for these measures to be credible, they need to be strongly grounded on sound theoretical frameworks, as opposed to being simple exercises in exploratory empirics. The SPI index may well develop into a credible measure and help improve the measurement of social prosperity. But with the index’s current configuration, that remains just an aspiration.