Absolute poverty: when necessity displaces desire

How many poor people are there in the world?  Where are they located? Are economic development and globalisation increasing or decreasing their number?  If so, which countries have been most affected? To answer these questions, we need a metric to measure poverty.

There already exists a widely recognized metric for measuring poverty–the World Bank’s famous $-a-day poverty line.  However, this measure does not account for differences in basic needs across countries. For example, the clothing, housing, and heating needs for a poor person living in a cold northern country are very different from what is needed for a poor resident of a warm southern country, and the prices of these goods vary across developed and undeveloped nations.

The World Bank’s $-a-day poverty live does not account for differences in basic needs across countries

One shoe does not fit all – the subsistence level of income differs across countries. In my paper, ‘Absolute Poverty: When Necessity Displaces Desire,” I propose a new method of measuring poverty lines that takes account of these differences.

The World Bank’s poverty line is based upon the minimum nutritional, clothing, and housing needs for people living in the world’s 15 poorest countries.  It was first established in 1991 and has been increased several times as prices have risen.  In the original World Bank study, the average of these poorest poverty lines was $1 per day–hence the famous line.  The investigation was redone in 2005 with a larger sample of countries, and the average of the 15 poorest poverty lines equaled $1.25 per day, which became the Bank’s line for 2005.  In 2011 the WBPL reached $1.90 per day.  These 15 poverty lines, updated for inflation, are still the basis of the WBPL.

An unavoidable feature of the World Bank’s statistical approach is that its poverty line is dominated by the experience of sub-Saharan Africa since that is where the poorest countries are located.  Indeed, 13 of the 15 countries in the 2005 sample of poorest countries lie in south of the Sahara, so the WPBL inevitably incorporates that region’s culinary practices, its clothing norms (which reflect its tropical climate), and the region’s real estate market where housing is inexpensive compared to many other parts of the world.

The World Bank poverty lines account for differences in prices across countries, but there is no allowance for differences in needs. In my paper, I explore an alternative approach to setting a poverty line that specifies a reference budget defined in terms of the basic needs for each country.  The classical economists had such a standard in mind with their concept of the subsistence wage.  Malthus, Ricardo, and Marx thought that demographic and technological evolution would keep wages at this low level.  In modern parlance this standard is called absolute or extreme poverty.

The Basic Needs Poverty Line

Basic needs include food and non-food goods like clothing, fuel, and housing.  I compute the poverty line for twenty countries using data from 2011. The countries in the sample range from very poor countries south of the Sahara to the USA, UK, and France.  Many Asian countries including China and India are included.  My line comes closest to matching the WBPL for the sub-Saharan countries.  In other cases, I compute higher lines reflecting differences in dietary regimes, climate, and real estate markets.

To estimate basic needs budgets, prices are needed.  My main source is the data collected for the 2011 round of the International Comparison Project.  This ambitious effort, which is coordinated by the World Bank, compiles the prices of thousands of precisely defined goods and services for 180 countries.  The principal objective of this exercise is to produce the purchasing power parity exchange rates used to compare GDP across countries. The household consumption expenditure component of that exercise is used by the World Bank to find the local currency equivalent of its poverty line for every country.

The Basic Needs Poverty Line is the sum of the expenditures on three categories of basic needs, food, non-food goods, and housing.

The Basic Needs Poverty Line is the sum of the expenditures on three categories of basic needs, food, non-food goods, and housing


The food basket for each country used to calculate the BNPL is based upon the ‘diet problem,’ first formulated and solved (almost correctly) by George Stigler in 1945.  The diet problem is based upon a mathematical technique know as linear programming.The problem begins with a potentially long list of foods, their prices, and their nutritional compositions.  Foods are selected to minimize the cost of the diet, so long as it meets a list of nutritional requirements. Stigler’s model included requirements for nine nutrients and used prices for the USA in 1939 and 1945.  The solution for 1939 was 168 kg of wheat flour, 129 kg of dried navy beans, 23 kg of evaporated milk, 50 kg of cabbage, and 10 kg of spinach.  He regarded this as impractical as nutritional advice for Americans since it was at such variance to their norms.  I agree that this approach does not describe the behaviour of rich people very well, which includes Americans in 1939, but it is a different story for poor people in poor countries.  For them, survival is the issue and many features of their behaviour are, in fact, described by the solution to the diet problem.  Stigler’s solution for 1945 is not far from the solutions that I compute for poor countries in 2011, and the broad outlines of the diet are consistent with the eating habits of many poor people in poor countries.

A big problem in setting up the diet problem is choosing the nutritional requirements.  I proceed by specifying models with increasingly stringent requirements.  The simplest specification has only a calorie requirement; the most comprehensive includes 15 nutrient requirements.  The most satisfactory model is the one I term the ‘Basic’ model since it comes closest to matching actual consumption patterns, and since it provides enough nutrition for healthy living.  It includes requirements for calories, protein, and fat as well as six vitamins and minerals to prevent anaemia, beri-beri, pellagra, and scurvy.

Comparisons of the diets implied by the basic model to actual consumption show a fairly good correspondence for poor countries but not for rich countries, as Stigler noted.  A poor person in a poor country generally eats a mainly vegetarian diet that is dominated by the common grain of that country along with some vegetables or fruit, small amounts of oil or butter and fish or milk.  This is the kind of diet predicted by the diet problem with the basic requirements, and it has much in common with Stigler’s solution (although his lacks oil or butter because he did not include a fat requirement, and it has too many beans because of some of his vitamin requirements).

The calculated basic needs diets do respond to local prices.  Comparability across countries is obtained by using the same requirements everywhere.  I have set nutritional requirements at levels recommended by the medical profession, which means that they are somewhat above the levels that would trigger deficiency diseases.  This allows a limited scope for people to exercise choice by including some sugar, for instance, in their consumption at the expense of a nutrient level required by the model.  But the scope for such behaviour is limited: It is the nature of poverty that necessity displaces desire.

Non-food goods

The United Nations Sustainable Development Goals require the World Bank to use its poverty line to track the number of poor people in every country–not just the developing countries that it has been focusing on.   It is imperative, therefore, to define the poverty line in a way that allows clothing and heating, for instance, to vary with the temperature.  The clothing and fuel needed to survive the winter in Niger are not enough to survive in New York.

It is imperative to define the poverty line in a way that allows clothing and heating to vary with temperature

I have defined requirements for clothing, foot ware, bedding, fuel, and lighting using household expenditure surveys for industrial workers in St Petersburg in 1907/8 and Bombay in 1921/2.  These workers were indubitably poor and the cities lie at opposite ends of the climate spectrum.  Low wage workers in St Petersburg consumed three times as much clothing, bedding, and foot ware as their India counterparts (including 27 times more foot ware!).  Energy consumed for lighting in St Petersburg was double Bombay consumption, while fuel consumption for cooking and heating was four times greater.  Values for the cities in my sample were scaled between those for Bombay and St Petersburg according to temperature.  One might argue that modern life requires more non-food goods and services, but the aim here is to calculate an austere poverty line.


The Bombay expenditure survey reported on the housing of the workers and typical arrangement was that a family (on average close to four people) lived in one room of approximately 12 square meters.  This suggests a requirement of 3 square metres of space per person.  Indeed, this number recurs in other surveys of slum housing in poor countries–although poor people living in ‘over crowded’ housing in rich countries often have more space–so I have adopted it as the minimal requirement.

To obtain an estimate of housing expenditures, the rental price of housing is needed.  Rental costs for ‘traditional’ housing with some modern conveniences are reported in the ICP for six less developed countries in my sample, but it was necessary to determine rents for the remaining countries from a variety of sources.  The aim was to find the rent of low quality accommodation in large cities.  Rents ranged from $0.50 per square metre per month in poor countries to about $25 in London, Paris, and New York.  Since even in rich countries rural accommodation can be very cheap, the problem (which the World Bank will have to face) of establishing a ‘nationally representative rent’ will be more challenging in rich countries than in poor.

The Basic Needs Poverty Line is the sum of the food, non-food, and housing expenses.


Amongst the less developed countries, food amounts to 65% of spending, non-foods to 27% and housing to 9%.  In the USA, UK, and France, food was 24%, non-foods 23%, and housing a staggering 53%.

The BNPL is defined in terms of local currencies, but it can be converted to dollars for comparison with the WBPL.  In sub-Saharan Africa, the BNPL is close to the WBPL, but in north Africa and most of Asia, the BNPL is $2.50-$3.00–-considerably above the WBPL value of $1.90.  In the rich countries, the BNPL approaches $4 per day.

The upshot is that in much of Asia, the BNPL indicates that there are a substantially larger number of poor people than indicated by the WBPL.

The BNPL indicates that there are a substantially larger number of poor people in Asia than the World Bank suggests

While economic growth has been impressive in Asia and raised the incomes of many people, many of those who have escaped poverty using the $1.90 WBPL are reclassified as poor when the higher BNPL is used.  The BNPL indicates that there are about 50% more poor people in the world than the WBPL implies.

Another surprising result is that the BNPL indicates there are millions of people living in absolute poverty even in the richest countries–0.6% of the population in France, 1% in the UK, and 1.5% in the USA.


Many aid programs devoted to reducing poverty allocate their resources based in part upon the number of people living in absolute poverty in various countries, so it’s important to have the most accurate estimates possible. Accurate estimates are also needed to measure the success of these programs and to document changes in absolute poverty over time.

My paper improves upon previous measures by allowing basic needs for food, housing, and fuel to vary across countries in a way that is consistent with the optimizing behaviour predicted by economic models, and the new measure does a good job of capturing the spending patterns of the poorest of the poor.  It also indicates there are substantially more poor people than we thought in Asia and in rich countries like the UK and the US giving us new insight into the extent and severity of the poverty problem.

This article summarizes ‘Absolute poverty: when necessity displaces desire‘ by Robert Allen, in the American Economic Review.