The National Food Security Bill (NFSB), the recent government effort to ensure food security, intends to cover 75% of the rural and 50% of the urban population. Yet, this Bill chose to create a hierarchy of ‘priority, general and excluded’ households; with differential food entitlements ranging from high to nil. In our view, the Bill, instead of solving the problem of exclusion and inclusion error, has given ways to augment such problems. Decades of research has shown that it is a tedious task to identify vulnerable households through subjective methods, and the BPL-socioeconomic and caste census carries many such elements.
Studies have shown large exclusion and inclusion errors in the country’s public distribution system. For example, Jalan and Murgai estimate an undercoverage of 49% for 2002; and Andhra Pradesh, Gujarat, Haryana, Himachal Pradesh, Karnataka, Kerala, Punjab, Rajasthan and Tamil Nadu recording over 60% of undercoverage. Another study by Shariff and Ganga has found an exclusion error of 58% in rural and 54% in urban areas for 2005; and for Delhi (NCT), Shariff and Bajpai estimate 65% exclusion error for 2005.
The SECC scoring method is susceptible to high levels of non-sampling errors during the survey. Since the characteristics needed for scoring are a priori known to the respondents, high misreporting is expected reflecting the local social and caste equations. There are other issues in scoring and aggregation; for example, each indicator is assigned an equal weight assuming that all indicators wield the same impact on poverty conditions.
Households reporting ‘less than one square meal per day for most of the year’ and ‘non-ownership of any of the listed consumer durables’ are assigned the same weight; similarly, scoring method presumes ‘artisan’ households to be always better off than ‘subsistence cultivators’.
Further, the methodology is uniform across states; the unirrigated land in a high-rainfall state has the probability to get the same score as unirrigated land in a desert state. The methodology ignores many such ground realities, especially large social and economic inequalities that exist based on caste and religious affiliation.
There is no consensus on the number of households to be included in the priority list at state level and which variables will be utilised to fix the cutoff at state and further at district level. In an attempt to find a solution, an alternative proposal is described below: identify the high food-stress regions and districts across the country, a composite index of ‘spread of agriculture’, ‘agricultural (cereal) productivity’, ‘irrigation intensity’, ‘presence of agricultural markets’ and ‘social practices’ that influence local-area food production and markets.
Given the proportionately low share of agricultural in the country’s GDP, a good number of districts will qualify for universal coverage under the NFSB. The remaining districts, which will be mostly found in Punjab, Haryana, western Uttar Pradesh and parts of Andhra Pradesh, can devise their own universal food security policies, including the PDS.
The poverty and food stress in urban areas are expected to be low due to relatively higher wage rates and easier access to a number of social services. A major urban poverty problem is the exclusion of migrant households, and the ongoing universal ID project can be effectively used to include such households in urban welfare programmes. Thus, overall, about a third of the urban households would be requiring supply of sustenance through the food Bill..