When Anthony Atkinson was seventeen and finished with boarding school, in 1962, he quit his job at IBM in London to spend nine months working at a hospital for the poor in Hamburg, Germany, as a kind of British Peace Corps volunteer. Working for IBM had been interesting, he told his friend Nicholas Stern in an interview: “there were still valves in computers in those days.” But the effect of the demanding work in the hospital was more profound. . .
Returning to his studies, he joined the Child Poverty Action Group. The Poor and the Poorest, by Brian Abel-Smith and Peter Townsend (1965), came out that Christmas Eve and made it clear that dire poverty still existed in post-war Britain. Three years before, Michael Harrington had published The Other America, the first broadside in what eventually became the “war on poverty.”
Strongly influenced by James Meade’s austere little monograph Efficiency, Equality, and the Ownership of Property, Atkinson switched from mathematics to economics at Churchill College, Cambridge University. In 1970, in the Journal of Economic Theory, then just beginning its transformative run, he published “The Measurement of Inequality.” Thereafter came a book and half a dozen articles almost every year for nearly half a century.
Atkinson died earlier this year, on New Year’s Day, a deeply beloved figure among applied economists. Among many other things, he was the scholar more responsible than any other for the first of the Sustainable Development Goals adopted by the United Nations in 2016: “Goal 1.1: by 2030, eradicate extreme poverty, for all people, everywhere, currently measured as people living on $1.25 a day.” An autobiographical interview conducted by Stern appeared in the lead position of this year’s Annual Review of Economics, an honor previously accorded only Kenneth Arrow and Paul Samuelson.
More or less the last thing Atkinson wrote was as head of Monitoring Global Poverty: Report of the Commission on Global Poverty, for the World Bank, which appeared earlier this year. The original $1 a day figure (in local currency equivalents), which had risen to $1.25 in 2005 dollars, was by now $1.90 a day. Atkinson was well aware of the deficiencies of the methods by which that figure was arrived at, but, from a practical standpoint, he reasoned that it was better to stick with the original yardstick than to try change it midstream. “The $1.90 line has acquired an independent political status,” he wrote.
Perhaps not. In Absolute Poverty: When Necessity Replaces Desire, the lead article in the December American Economic Review, Robert C. Allen, a distinguished economic historian, takes up the question. Originally conceived for developing countries, the World Bank poverty line can be profoundly misleading, he says.
For one thing, because of an accident of history, it is “not valid outside of the tropics,” he says. For another, it makes no attempt to assess the difference between “food” and “non-food” spending. Neither varies with climate. How far would a tropical allowance for fuel and clothing go in surviving a Russian winter?
Allen’s approach implies much more “absolute poverty” than the World Bank measure, especially in Asia. It has the added virtue of revealing that millions of persons are living in that condition in rich countries around the world – especially the US and the UK.
It was in 1991 that a trio of World Bank economists collected poverty lines that had been established in 33 countries, ranging from poor to rich, converted them to dollars using purchasing power parity exchange rates, and plotted the results against per capita consumption. The minimum standard of living in the six poorest countries clustered around a $1 a day. That became the measure of absolute poverty, especially after their finding was buttressed by a second study of 15 poor countries in 2009.
Problems with the method were pointed out over the years. For example, Angus Deaton, of Princeton University, noted in 2010 that, while India had been in the original sample of poor countries, it had grown so much in fifteen years that it was no longer included when the second study was made in 2005. The Indian poverty line was very low in 1991, so leaving it out of the new sample raised the average World Bank poverty level. That meant that measured poverty in India increased dramatically in the second survey, despite all the growth that had occurred.
A historian of the British industrial revolution, Allen, 70, came to the problem of poverty measurement from historians’ longstanding tradition of assembling basic baskets of necessary goods with which to measure the purchasing power of workers’ wages. In Poverty Lines in History, Theory, and Current International Practice, a discussion paper at Oxford, where Allen taught at Nuffield College for many years, he compared historians’ straightforward practices in comparing wages in widely differing countries and climate with the World Bank’s statistical approach. He concluded.
Perhaps the World Bank can learn a lesson from historians and settle on an explicit definition of poverty that can be applied across space and over time. Historical research indicates that this is practical. The benefits in terms of transparency and intelligibility would be large.
In the new article, Allen shows how. Using the mathematics of linear programming, techniques discovered during World War II, Allen calculates a “food security line,” the cheapest possible diet required to keep an adult alive, with certain small allowances permitted for local substitutions and preferences. Economist George Stigler, of the University of Chicago, had reasoned his way to a close approximation of the same during the war, using trial and error.
The truly poor do the same out of experience. Absolute poverty means that your life is governed by linear programing, says Allen, not by standard consumer theory. You eat to live. Necessity displaces desire. Inevitably there is a little “wiggle-room” in the diets of the absolute poor: a small amount of sugar, a favorite spice, food prepared for festivals. But for those on the margin of survival, linear programming pretty much describes the choices made.
He augments his “Basic Needs” measure with two other headings – nonfood goods, and rented housing – and, by the end of the exercise, has described a method much more clearly related, not just to survival but to well-being. Many conceptual problems with the World Bank dollar-denominated yardstick disappear as well. Atkinson recognized as much, Allen says; he just didn’t think it was a good idea to move the goalposts in the middle of the match.
It’s a long road ahead. Allen, today Global Distinguished Professor of New York University Abu Dhabi as well as Nuffield College, doesn’t have as much time to proselytize the historians’ approach as Atkinson did his. But the two are brothers under the skin. Measurement economists will take up Allen’s methods. The picture of poverty around the world will come into sharper focus. Safety nets will be improved.
And on a day on which Ron Lieber, of The New York Times, reported on yet another little-noticed provision of the new US tax act – the one that allows families to use tax-exempt savings accounts to pay private school tuitions from elementary school instead of just college bills – Atkinson surely would have welcomed Allen’s diversion from the ordinary on another Christmas Eve.