Distributive policy

Poverty Survey Data and Broad Policy Directions

There must be engagement with survey data, but realities on the ground must shape programmatic interventions

There must be engagement with survey data, but realities on the ground must shape programmatic interventions

Based on the multidimensional poverty measure, the poverty rate (headcount ratio) in Tamil Nadu fell from 4.89% in 2015-2016 to 1.57% in 2020-21, based on fourth and fifth rounds of data from the National Family Health Survey (NFHS). Is it too good to believe? Maybe. Academics have questioned the quality of the NFHS data for various reasons, based on the previous four rounds of NFHS databases. Such questions can also be raised in relation to the NFHS 5 database. But first, let’s explore poverty statistics derived from NFHS 5 data using the multidimensional poverty measure suggested by NITI Aayog and its indicators of poverty. political intervention. After that, we will raise questions about the quality of NFHS data with the aim of using it with caution and improving data quality with the future in mind.

On the MPI

NITI Aayog, armed with quite a large sample of survey data from NFHS 4 (with over six million households in India), estimated the Multidimensional Poverty Index (MPI) and released the benchmark report in 2021 The rationale for MPI was derived from the concept that poverty is the result of simultaneous deprivations in multiple functions such as health, education, and standard of living. NITI Aayog identified 12 indicators in these three sectors and calculated the weighted average of deprivations in each of these 12 indicators for all men and women surveyed in NFHS 4. If an individual’s overall weighted deprivation score was greater than 0.33, it was considered multidimensionally poor. .

The non-poor may also be deprived of some of these indicators, but not so much to be classified as multidimensionally poor. The proportion of the population with a deprivation score greater than 0.33 out of the total population is defined as the poverty rate or headcount rate. The authors estimated the MPI and its components for Tamil Nadu using NFHS 5 and compared it with the estimates based on NFHS 4 provided by NITI Aayog.

Another interesting aspect of this approach is the estimation of the intensity of poverty. This is the weighted average deprivation score of multidimensionally poor people. For example, the intensity of poverty in Tamil Nadu fell from 39.97% to 38.78% during this period, indicating that the summary measure of the multiple deprivations of the poor declined only slightly over the of these five years, and must be highlighted for policy guidance.

The MPI is a product of the ratio of the number of inhabitants and the intensity of poverty. Tamil Nadu’s MPI fell from 0.020 to 0.006. This sharp drop in the MPI is largely due to a larger drop in the count ratio relative to the intensity of poverty. This gives us a clue that any further decline in MPI in Tamil Nadu should only occur by addressing all dimensions of poverty and significantly reducing its intensity across the state.

Direction of intervention

The deprivation estimate also indicates that the overall population that was identified as disadvantaged in most indicators individually is higher than the population identified as multidimensionally poor. This reiterates once again the fact that people may be severely deprived in a few functions, but may not be multidimensionally poor. This adds another aspect of public policy intervention, i.e. poverty alleviation in Tamil Nadu should not only be multidimensional but also universal. Only this approach can address deprivations in all indicators. It will also surely and squarely reduce the intensity of poverty in Tamil Nadu.

Read also | TN’s success in reducing poverty

Statistically, the headcount ratio and poverty intensity can be calculated for each district and separated by gender, rural and urban, and other dimensions. Therefore, the usefulness of the MPI and its components is enormous in terms of understanding poverty in its totality as well as the granular details that are essential for sectoral and spatial policies and programmatic interventions. The strength of IPM as a data-driven public policy instrument depends on the quality of survey data, namely NFHS data.

NFHS Data Quality

The quality of survey data has been widely debated in academia. The National Sample Survey Organization (NSSO) sample surveys have been the subject of debate among economists and statisticians, both in terms of sampling and non-sampling errors, from its earliest days. in the 1950s. Following several review reports on NSSO methodologies, NSSO attempted to improve sampling design and reduce non-sampling errors, especially with regard to relates to the reminder periods for providing household consumption expenditure. All of this is well documented.

Demographers such as K. Srinivasan, S. Irudaya Rajan and KS James have written several articles on non-sampling errors in different NFHS data sets. They tested, for example, arbitrariness in reporting age of death, differences in data quality between educated and uneducated respondents, data quality based on differences in time needed to complete a survey among of different types of households, etc. All of this has serious implications. for health data such as fertility and mortality rates. A market-based approach to deciding the data collection process is also criticized by demographers.

The authors performed another type of quality control for the NFHS 5 data for Tamil Nadu. For example, in Tamil Nadu, NFHS data was collected over two periods: 8,382 households (30%) during the pre-pandemic period and 19,547 households (70%) during the post-lockdown period, totaling 27,929 households for the state. . Data collected from 19,547 households in the post-lockdown period is expected to reflect the impact of the first wave of the COVID-19 pandemic. Let’s compare pregnant women and their age distribution in the two time periods to get some insight. The proportion of pregnant women under 19 was 18:82; those aged 19-21 was 25:75 compared to the proportion of 32:68 for pregnant women over 21. The pandemic has led to an increase in pregnancies among women under 21, more so among teenage girls. The number of deaths per 1,000 households surveyed rose from 118.23 to 135.01 – this is clear evidence of the impact of the pandemic.

The authors estimated count ratios for the 12 indicators and found that these ratios were lower in the post-lockdown period than in the pre-pandemic period, leading to the conclusion that after lockdown, deprivation in several functionings was lower, implying a lower poverty rate as well as the intensity of poverty. In particular, deprivation in terms of nutrition and maternal health has decreased, and school enrollment and attendance have increased in the post-lockdown period.

Substitution of dry rations for hot meals in midday meal programs and strong pressures on hospitals to manage COVID-19 cases are expected to increase nutritional and maternal health deprivation in the post-lockdown period, unlike to the decline in nutritional deprivation and maternal health in the post-pandemic period that we derived from this database. Tamil Nadu is known for increasing enrollment and reducing dropout rate year after year; therefore, the increase in deprivation in terms of schooling should raise questions. Regarding school attendance, we do not know how parents interpreted school attendance during the long period of school closures in the period of confinement. Therefore, combined survey data from two different time periods separated by a major pandemic should be approached with caution when interpreting statistics derived from the entire database.

Assuming that the survey data comes from a single period, it is normal to compare the results of survey data on specific indicators, with programmatic data derived from official records. Some claim that the deprivation indicators in terms of drinking water and sanitation are at a higher level in Tamil Nadu than the claims made by the respective state government departments. Such problems are common in survey data. For example, consumption expenditure on food grains derived from ONSS data would not agree with the estimate of food consumption, according to the System of National Accounts.

Data use and quality

The quality of survey data has always been a contentious issue in academic and policy debates for a variety of (well-founded) reasons. However, this has not prevented academics and policy makers from inferring policy guidance, as such data at a reasonably aggregated level (e.g. state level) should be useful. As mentioned earlier, in Tamil Nadu, the sharp drop in count ratio and a marginal drop in poverty intensity in NFHS 5 compared to NFHS 4 cannot be ignored. From this, we can deduce that to reduce the intensity of poverty, we need to tackle the deprivations of the whole population, i.e. there should be a one-size-fits-all approach instead of a targeted approach to address it.

The survey data only gives us general guidance on policies while programmatic interventions need to be tailored to the realities on the ground. At the same time, an ongoing engagement with survey data in terms of improving sample design and quality of responses must be maintained. Analyzing the data and finding the incongruity of inferences from different databases on a question would help improve data collection systems. Let’s continue to use survey data both to draw policy conclusions (with caution) and also to help improve data quality.

R. Srinivasan is a member of the Tamil Nadu State Planning Commission. S. Raja Sethu Durai is Professor of Economics at Hyderabad University. Opinions expressed are personal