Last week, the Planning Commission filed an affidavit in the Supreme Court updating the official poverty line to Rs 965 per month in urban areas and Rs 781 in rural areas. This works out to Rs 32 and and Rs 26 per day, respectively. The perceived inadequacy of these figures has led to widespread discussion and criticism in the media. In light of the controversy, it may be worth looking at where the numbers come from in the first place. Two Measures of the BPL Population The official poverty line is determined by the Planning Commission, on the basis of data provided by the National Sample Survey Organisation (NSSO). NSSO data is based on a survey of consumer expenditure which takes place every five years.  The most recent Planning Commission poverty estimates are for the year 2004-05. In addition to Planning Commission efforts to determine the poverty line, the Ministry of Rural Development has conducted a BPL Census in 1992, 1997, 2002, and 2011 to identify poor households. The BPL Census is used to target families for assistance through various schemes of the central government. The 2011 BPL Census is being conducted along with a caste census, and is dubbed the Socio-Economic & Caste Census (SECC) 2011. Details on the methodology of SECC 2011 are available in this short Ministry of Rural Development circular. Planning Commission Methodology Rural and urban poverty lines were first defined in 1973-74 in terms of Per Capita Total Expenditure (PCTE). Consumption is measured in terms of a collection of goods and services known as reference Poverty Line Baskets (PLB). These PLB were determined separately for urban and rural areas and based on a per-day calorie intake of 2400 (rural) and 2100 (urban), each containing items such as food, clothing, fuel, rent, conveyance and entertainment, among others. The official poverty line is the national average expenditure per person incurred to obtain the goods in the PLB. Since 1973-74, prices for goods in the PLB have been periodically adjusted over time and across states to deduce the official poverty line. Uniform Reference Period (URP) vs Mixed Reference Period (MRP) Until 1993-94, consumption information collected by the NSSO was based on the Uniform Reference Period (URP), which measured consumption across a 30-day recall period. That is, survey respondents were asked about their consumption  in the previous 30 days. From 1999-2000 onwards, the NSSO switched to a method known as the Mixed Reference Period (MRP). The MRP measures consumption of five low-frequency items (clothing, footwear, durables, education and institutional health expenditure) over the previous year, and all other items over the previous 30 days. That is to say, for the five items, survey respondents are asked about consumption in the previous one year. For the remaining items, they are asked about consumption in the previous 30 days. Tendulkar Committee Report In 2009, the Tendulkar Committee Report suggested several changes to the way poverty is measured.  First, it recommended a shift away from basing the PLB in caloric intake and towards target nutritional outcomes instead. Second, it recommended that a uniform PLB be used for both rural and urban areas. In addition, it recommended a change in the way prices are adjusted, and called for an explicit provision in the PLB to account for private expenditure in health and education. For these reasons, the Tendulkar estimate of poverty for the years 1993-94 and 2004-05 is higher than the official estimate, regardless of whether one looks at URP or MRP figures. For example, while the official 1993-94 All-India poverty figure is 36% (URP), applying the Tendulkar methodology yields a rate of 45.3%. Similarly, the official 2004-05 poverty rate is 21.8% (MRP) or 27.5% (URP), while applying the the Tendulkar methodology brings the number to 37.2%. A Planning Commission table of poverty rates by state comparing the two methodologies by is available here.