How the poverty line and poverty rate are calculated
he World Bank is preparing a new Poverty Assessment for the Consultative Group Meeting to be held in March 2006. World Bank newsletter interviews Tim Conway, Poverty Specialist at the World Bank’s Cambodia Office, about poverty rates, poverty lines and related indicators.
Q: How is the poverty line calculated?
A: To calculate a “poverty line” requires household consumption data from a representative sample survey. For Cambodia, the latest data comes from the Cambodia Socio-Economic Survey (CSES) 2004. Just under 15,000 households from throughout the country were asked questions about consumption and other important aspects of living standards. A similar survey was conducted in 1993/4.
From the consumption data collected in the CSES, the poverty line in Cambodia is calculated in the same way as in other countries. This involves first working out what food someone in Cambodia needs to consume to meet their basic nutritional needs (2,100 calories a day). This is the food poverty line.
Using data on household consumption collected in a sample survey (the CSES 2004), an analyst then finds how much households on the food poverty line spend on essential non-food goods and services (e.g. clothing, shelter, health care). These basic food and non-food allowances are added together, and give us the poverty line.
Q: What is the poverty line in Cambodia?
A: Separate poverty lines were identified for people living in Phnom Penh, other towns, and rural areas, reflecting different consumption patterns in these three areas. When these are averaged, the national poverty line for 2004 is approximately 1,826 Riels per person per day (or 9,130 riels per day for a family of five).
At 2004 exchange rates, this is about $0.45 per person per day (or $2.25 per day for a family of five). About 80 percent of this is food; 20 percent is for non-food basic needs (clothes, housing, etc.). This reflects the reality of what households actually spend on food and non-food items, as measured in the survey.
Q: What is the poverty rate in Cambodia now? What was it ten years ago?
A: The poverty rate – that is, the percentage of the population living under the poverty line – is estimated at 35 percent for 2004. It is highest in rural areas and lowest in Phnom Penh.
It is harder to know precisely what the poverty rate was ten years ago, because the 1993/4 household survey did not cover the whole country: it was not safe for the survey teams to visit the outlying areas where the Khmer Rouge were still active. Almost all of the urban population were included in the sampling frame in 1993/4, but only 65 percent of the rural population were represented in the sample.
However, we can compare the rate of change in poverty between 1993/4 and 2004 in those parts of the country that were accessible and covered in 1993/4. There the poverty rate fell from 39 percent in 1993/4 to 28 percent in 2004. If we apply this rate of poverty reduction to the whole country, we project back to estimate that poverty in Cambodia in 1993/4 was 47 percent
(or somewhere between 45 percent and 50 percent). This means that the poverty rate has fallen by between 10 percent and 15 percent over the ten years between the 1993/4 and 2004 surveys.
This makes sense, given what we know about rates of economic growth over the last decade and the pattern of distribution of that growth. It is also broadly confirmed by other indicators of improving living standards. In the 2004 survey, indicators such as quality of housing, ownership of consumer durables such as bicycles and radios, and access to primary schooling have all improved relative to 1993/4.
Q: How is the poverty rate calculated?
A: The poverty rate (sometimes called the poverty headcount or poverty incidence) is calculated by comparing the distribution of household per capita consumption, collected during a random sample survey, to the poverty line. When we say the poverty rate in Cambodia in 2004 was 35percent, we mean that 35 percent of the individuals surveyed in the 2004 CSES came from households with per capita consumption less than the poverty line. If the survey is well designed (and the CSES 2004 was), then we can generalize to say that 35 percent of the Cambodian population is under the poverty line.
Q: Is the World Bank proposing a lower poverty line for Cambodia than the international standard?
A: Categorically not. The poverty line and the comparison of survey data on household consumption against this line to calculate the poverty rate in Cambodia are calculated in the same way as in other countries. In dollar terms, national poverty lines in neighboring countries are the same or slightly lower than Cambodia’s. Thus Cambodia’s national poverty line at 1,826 riels per person per day in 2004 is roughly $0.45. In Vietnam in the 1998 survey the poverty line was 4,900 dong per person per day (or $0.35); in the Lao survey in 2002/3, the national poverty line was 3,065 kip per person per day ($0.26).
Because the “normal” poverty line used internationally is often referred to as the “dollar-a-day” poverty line and the national poverty line is set at $0.45 (1,826 riels) per day, some people feel that the national line is unfair to Cambodians.
This is not correct. The full description of the international poverty line is “$1 per capita per day, 1990 PPP.” The PPP stands for “Purchasing Power Parity”. That means, the $1 in the poverty line is the purchasing power of $1 in the United States in 1990. When adjusted cost-of-living differences between countries, it comes out as 1,382 riels per person per day in Cambodia in 2004. To put it another way, what you could buy (and just survive on) for $1 a day in the United
States in 1990 – defined as the poverty line – you could buy (and just about survive on) for 1,382 riels per day in Cambodia in 2004. The international “dollar-a day” poverty line is actually much lower than Cambodia’s national poverty line (see Q2), and results in a lower estimated poverty rate (see Q3).
Q: What is the poverty rate for Cambodia according to the international (“dollar-a-day”) poverty line?
A: In October the World Bank revised the dollar-a-day poverty estimates for East Asia. According to this, 19 percent of the Cambodian population were below the dollar-a-day rate in 2004, down from an estimated 29
Q: Why do we have two poverty lines, and so two poverty rates? Why are they different?
A: The national poverty line used to calculate the national poverty rate in 1993/4, and the updating of this line (to allow for inflation) in 2004, have been produced specifically for Cambodia, but follow accepted standard international practice (see FAQ 1 above).
The international (dollar-a-day 1990 PPP) poverty line was adopted in 2000 by the international community – that is, all the Governments of the world, plus the World Bank, the UN agencies and other international bodies – as a common base, applicable across all countries, against which to measure progress in reducing global poverty. The first of the Millennium Development Goals (MDGs) commits the Governments of the world to “eradicate extreme poverty and hunger”: the first target under this goal, and the best-known MDG, is to “halve, between
1990 and 2015, the proportion of people who live on less than $1 a day (PPP).”
Discussions about poverty in Cambodia – who is poor, where the poor live, which groups are getting poorer or richer, whether inequality is becoming greater or less – should all be based on the national poverty line, which is designed specifically for Cambodia.
Q: Do other countries have this same difference between two poverty lines, and so two poverty rates?
A: Yes. For example, in Vietnam in 1998 the international, “dollar-a-day” poverty rate was estimated at 16.4 percent. The poverty rate according to the national poverty line of 4,900 dong per person per day ($0. 35), based on the 1998 Vietnam Living Standards Survey, was 37 percent. In Laos the international, “dollar-a-day” estimate in 2002 was 28.1percent; in 2003 it was 25.8 percent. The poverty rate according to the national poverty line (3,065 kip per person per day, or $0.26), measured in the third Lao Economic and Consumption Survey (LECS III) was 30.7 percent.
Q: Why did the World Bank revise down its estimate of dollar-a-day poverty for Cambodia in 2004 from 42 percent in April to 18.5 percent in October?
In April 2005 the Bank included an estimate of 42 percent dollar-a-day poverty in Cambodia for 2004 in a publication called the East Asia Update. In October 2005, in the East Asia Update, the Bank revised this 2004 dollar-a-day estimate significantly down, to 19 percent.
Although dramatic, this revision is simply an attempt to improve the accuracy of the dollar-a day estimates on the basis of new and better information. The critical factor that underpins the revision is that in April 2004 we were still working on projections of Cambodian household consumption based on the last available year (1997) in which there was a household survey.
By October we had new data on actual household consumption from the 2004 CSES (entering, checking, and analyzing the survey data took time, so the information was only available in June). The October estimates were based on this actual (surveyed) consumption data; the previous estimate was based on a forward projection of the 1997 data.
Q: Is the rate of poverty reduction seen in Cambodia over the last ten years good, bad or average?
A: If poverty in Cambodia fell by between 1 percent and 1.5 percent per annum between 1993/4 and 2004 (from 45-50 percent to 35 percent), this is relatively good. However, it also needs to be put in context. Cambodia in 1993/4 was just coming out of over 25 years of civil conflict. It was also emerging from isolation, having been almost entirely cut off from international trade, investment and aid for two decades.
Since the 1990s Cambodia has benefited from a significant influx of investment and trade (and aid) that has helped to lift growth rates to a high average of 7 percent per year. The achievements in rapid growth and relatively modest poverty reduction between 1993/4 and 2004 need to be understood in large part as a one-off “peace dividend”: as the country was starting from a very low economic base, initial rates of improvement were likely to be rapid. Some countries emerging from similar twentyyear legacies of conflict have, with good policies in place, experienced higher rates of poverty reduction (e.g. 2.4 percent per year in Uganda between 1992 and 2000 surveys; or 2.5 percent per year in Mozambique between surveys in 1997 and 2003).
The challenge for the Government now is to find new, sustainable sources of growth which can continue to drive a rise in living standards for the poor; and to combine this with improved delivery of basic services such as healthcare and education. The challenge for donors is to help the Government do this with aid that is more coordinated and effective and better targeted to poverty reduction.
(Source: The World Bank Cambodia Newsletter, Volume 3, Number 12, December 2005)