While income-based poverty represents a key factor influencing well-being and societal progress, there is a broader set of deprivations—relating to health, education and basic standards of living—that affect the lives and livelihoods of individuals and families. These directly affect people’s ability to break out of poverty.

Many countries now measure multidimensional poverty alongside income-based poverty. The former is considered a broader definition of poverty. Under this broader definition of poverty, many more people come into view as the global Multidimensional Poverty Index (MPI), a person is identified as multidimensionally poor (or MPI poor) if they have been deprived in at least one-third of the weighted MPI indicators. In other words, a person is MPI poor if their weighted deprivation score is equal to or higher than the poverty cutoff of 33.33%.

In this summary blog, we show the multidimensional poverty index in Lahore, Pakistan by using data from the USAID’s Demographic and Health Surveys (DHS), our work done in Sustainable Development Goals (SDG) Tech Lab in collaboration with United Nations Development Program (UNDP), United Nations Population Fund (UNFPA) and Intelligent Machines Lab which is supported by the Google ResearchScholarAward.

Our findings present the extent of deprivation that exists not only in the 13 clusters identified by DHS in Lahore city but also the 35 sampled slums included in the 4km (2.5 miles) radius of every respective Cluster. This comparison can help us in designing policy that will have a constructive impact on upgrading interventions for different sectors across the city and consequently a better guide for budget allocation

The following sections elaborates on our study area, data, methodology, results and analysis.

Lahore

According to Pakistan’s most recent Census (Pakistan Bureau of Statistics, 2017), 36% of Pakistan’s population lives in urban areas. The largest province in terms of population is Punjab; it accounts for 52.9% of the population. Punjab is the second most urbanized province in Pakistan where 36.71% of the population resides in cities. Punjab’s capital is Lahore that hosts nearly 15% of the province’s population, making it the most populous city of the province.

Lahore’s population is growing at 4.22% per annum and is expected to be home to 42.46 million people by 2050. The city spans 1,772 km² (District at a Glance Lahore | Pakistan Bureau of Statistics, n.d.). Within Punjab, Lahore is the most urbanized city of the province hosting almost a third of Punjab’s urban population (State of Pakistan Cities Report 2018 | UN-Habitat). This makes Lahore our preferred area of study. Population density is estimated at around 6,300 persons per km² and the average household size is 6.3 persons, which is quite high compared to cities of similar size Lahore Population 2021 (Demographics, Maps, Graphs).

So how do we estimate multidimensional poverty?

In line with Alkire & Foster (2011), we estimate multidimensional poverty by combining the incidence of poverty with its intensity.  After the poverty threshold identification step, we aggregate across individuals to obtain the incidence of poverty or headcount ratio (H), which represents the proportion of poor people. We then compute the intensity of poverty (A), representing the average number of weighted deprivations experienced by the poor. Finally, we compute the adjusted poverty headcount ratio (M0) or MPI by combining H and A in a multiplicative form (MPI = H x A) (Alkire & Foster, 2011).

Dimensions of Multidimensional Poverty Index (MPI)

A key step in the development of the MPI is to decide the structure of the measure—that is, the dimensions and indicators that together measure poverty in the country.

We take three core dimensions: education, health, and living standard that sums up to a total of 15 indicators, out of which three belong to education, four to health, and eight to living standards. The indicators use a nested weight structure: equal weights across dimensions and an equal weight for each indicator within a dimension.

Table 1 provides a summary of the dimensions, indicators, thresholds and weights used in the MPI.

Data Sources

Pakistan Demographic and Health Surveys (DHS).

The MPI relies on different datasets that are publicly available and comparable to most developing countries. For our analysis, we have opted for the USAID-funded Demographic and Health Surveys (DHS). DHS is a major source of population and health data for the poorest countries in the world and provides high-quality and detailed data on a wide range of monitoring and impact evaluation indicators in the areas of population, health, and nutrition.

The primary sampling units in the DHS are “clusters.” Cluster size can vary but contains a number of households within a geographic area who participated in the survey. Since the data included in DHS is personal and potentially sensitive, the DHS maintains the confidentiality of the respondents by shifting the spatial coordinates of the cluster in the published data (Burgert et al., 2013).

DHS Clusters in Pakistan

The spatial coordinates for rural locations are displaced by 0–5 km in any direction. Additionally, a small fraction of coordinates, 1 percent, are randomly shifted up to 10 km. For urban locations, the displacement is up to 2 km only. DHS recommends that researchers average any environmental data over a 5–10 km buffer around each DHS rural cluster with the specific community falling somewhere within the disc around each point (Perez-Heydrich et al., 2015).

Figure 1 DHS Clusters in Pakistan from the 2017-18 PDHS (Sampling Methodology)

The USAID’s Demographic and Health Surveys (DHS) conducted in Lahore, Pakistan identified a total of 13 clusters for the survey. Table 2 in our study presents the Multidimensional Poverty Index (MPI) for each of these clusters.

Table 2 DHS Clusters in Lahore and their estimated MPI

Comparing Slums with DHS Cluster

Along with analyzing poverty in the DHS clusters in Lahore, we also present our findings from slums. We show the extent of deprivation in slums that are a part of the cluster which we are analyzing. Data from slums were collected in March 2019 under a study we conducted in collaboration with UNDP Pakistan and UNFPA (Aftab et al, 2019). This is primary data and we have brought it as close as possible to the DHS questionnaire.

Slums in Lahore

‘Slum’ has become a term to describe a wide range of low-income settlements and/or poor human living conditions. UN-HABITAT attempts a common definition by defining slums as households that lack one or more of the following indicators: durable housing of permanent nature, sufficient living space, easy access to safe water, access to adequate sanitation, and security of tenure (United Nations Human Settlements Programme (UN- Habitat), 2008).

According to official figures, the urban population in Pakistan is 36.4% with an estimated increase of 118 million people by 2030 (Pakistan | UN-Habitat, n.d.). If no concrete framework for action is carried out to address the problem, more than 50 per cent of the population of major cities will live in slums and become victims of these “spatial poverty traps”(Bird et al., 2002).

We have previously used remote sensing (Rehman et al., 2022) to distinguish slums from their surroundings because of the differences in spectral characteristics and as a result, help narrow areas for the socio-economic surveys to compute MPI. As a result, we have collected primary data for 35 slums in Lahore encompassing 348 households for the estimation of MPI in 2019. Table 3 in our blog presents the MPI values for each of the sampled slums.

Table 3 Sample slums in Lahore with their respective MPI, and its decomposition with respect to three dimensions.

* No deprived individuals in by-groups 18 & 30.

Analysis

The comparison of MPI between DHS clusters in Lahore and sampled slums provides insight into the extent of multidimensional poverty within different areas of the city. The Multidimensional Poverty Index (MPI) serves as an indicator of multiple deprivations in living conditions, including health, education, and standard of living. The higher the MPI value, the higher the level of deprivation in the area.

DHS Clusters:

The DHS clusters’ MPI values range from 0.086 (HV001-153) to 0.447 (HV001-151). The average MPI value for the DHS clusters is approximately 0.295. This indicates a moderate level of multidimensional poverty across the 13 clusters.

Sampled Slums:

In the sampled slums, MPI values range from 0 (L18 and L30) to 0.673 (L3). The average MPI value for the sampled slums is approximately 0.255. This suggests that the sampled slums, on average, experience a slightly lower level of multidimensional poverty compared to the DHS clusters.

Differences between DHS Clusters and Sampled Slums:

Variation in MPI values: The sampled slums show a wider range of MPI values, with some slums having no deprived individuals (L18 and L30), while others show high levels of deprivation (L3). This indicates that the living conditions in slums can vary greatly, with some slums having better living conditions than certain DHS clusters. In contrast, the MPI values of the DHS clusters are more consistent, with no extreme highs or lows.

Average MPI values: While the average MPI value for sampled slums is slightly lower than that of the DHS clusters, it is essential to note that slums represent a small, specific subset of the urban population. The DHS clusters provide a broader representation of the general population in Lahore, and thus may more accurately reflect the overall living conditions in the city.

Difference in Methodology: The MPI values for the DHS clusters were obtained from the USAID’s Demographic and Health Surveys, while the slum MPI values were collected through a separate study conducted in collaboration with UNDP Pakistan and UNFPA. The difference in data collection methods and survey instruments may have influenced the resulting MPI values.

Conclusion

In conclusion, the comparison of MPI values between DHS clusters and sampled slums reveals disparities in living conditions within Lahore. The slums exhibit a wide range of MPI values, suggesting significant variation in poverty levels among them. The DHS clusters, representing a broader cross-section of Lahore’s population, show a more consistent range of MPI values, with a slightly higher average level of multidimensional poverty. Further research and targeted interventions are necessary to address the specific needs of both the DHS clusters and the slums in order to improve overall living conditions and reduce poverty in Lahore.

Bibliography

Aftab, I., Ali, M., Akbar,S., Islam,N., Rehman, M. F. U. (2019). Mapping  and measuring Multidimensional poverty index in Lahore, Hyderabad and Peshawar. Unpublished Manuscript

Alkire, S., & Foster, J. (2011). Counting and multidimensional poverty measurement. PUBEC Journal of Public Economics, 95(7–8), 476–487.

Bird, K., Hulme, D., Shepherd, A., & Moore, K. (2002). Chronic Poverty and Remote Rural Areas (SSRN Scholarly Paper ID 1754490). Social Science Research Network. https://doi.org/10.2139/ssrn.1754490

Block Wise Provisional Summary Results of 6th Population & Housing Census-2017 [As on January 03, 2018] | Pakistan Bureau of Statistics. (n.d.). Retrieved October 31, 2019, from http://www.pbs.gov.pk/content/block-wise-provisional-summary-results-6th-population-housing-census-2017-january-03-2018

Burgert, C. R., Colston, J., Roy, T., Zachary, B., ICF International (Firm), United States, & Agency for International Development. (2013). Geographic displacement procedure and georeferenced data release policy for the Demographic and Health Surveys.

District at a glance Lahore | Pakistan Bureau of Statistics. (n.d.). Retrieved October 31, 2019, from http://www.pbs.gov.pk/content/district-glance-lahore

Lahore Population 2021 (Demographics, Maps, Graphs). (n.d.). Retrieved May 29, 2021, from https://worldpopulationreview.com/world-cities/lahore-population

Pakistan | UN-Habitat. (n.d.). Retrieved April 18, 2022, from https://unhabitat.org/pakistan

Perez-Heydrich, C., Warren, Joshua L., Burgert, C. R., Emch, M. E., & SpringerLink (Online service). (2015). Influence of Demographic and Health Survey Point Displacements on Raster-Based Analyses. https://doi.org/10.1007/s40980-015-0013-1

Rehman, M. F. U., Aftab, I., Sultani, W., & Ali, M. (2022). Mapping Temporary Slums From Satellite Imagery Using a Semi-Supervised Approach. IEEE Geoscience and Remote Sensing Letters, 19, 1–5. https://doi.org/10.1109/LGRS.2022.3180162

State of Pakistan Cities report 2018 | UN-Habitat. (n.d.). Retrieved September 24, 2020, from https://unhabitat.org/state-of-pakistan-cities-report-2018

United Nations Human Settlements Programme (UN- Habitat). (2008). State of the world’s cities 2010/2011: Bridging the urban divide. Earthscan