 |
By Rajiv Shah
A recent Planning Commission-sponsored study, “Refining State Level Comparisons in India”, by Pranjul Bhandari, who works as economist at the Office of the Chief Economic Advisor in the Ministry of Finance, Government of India, has found that a “refined” analysis of the performance across 21 major states suggests that Gujarat ranks 16th in health, 14th in education and 11th in infrastructure. The analysis is already creating flutter against the backdrop of the claim by those favouring Gujarat’s neoliberal model, that state’s alleged poor performance in the social sector is more an outcome of “fudged” figures, dished out by the Government India’s different ministries.
The “refined” analysis for comparing states is considered a new and a more scientific methodology, adopted in order to find out how well states perform given the availability of resources at their disposal. It differs from the “raw” analysis, under which Gujarat ranked 12th in health, 10th in education and 6th in infrastructure among 21 major states.
For the sake of convenience, Bhandari has sought to confine her analyze across just three critical sectors – health, education and infrastructure. Explaining her new methodology, she says, “Our ‘raw’ results conform with the already well-established findings of several other studies that states such as Kerala are amongst the best performing while the so-called BIMARU states (Bihar, MP, Rajasthan and UP) are laggards. While this is true on an absolute level, it does not reveal the performance conditional on state level factors”, the economist insists, pointing towards the need to “refine this analysis”.
To provide a refined analysis, the economist seeks to “control” the three indices for per capita consumption to “put the states on a level playing field and for gauging how well the states have used available resources”, to quote her. She says, “Our ‘refined’ analysis throws up rankings which are quite different from the ‘raw’ analysis. For instance, we find clear differentiation between the BIMARU states – while Orissa, Bihar and Chhattisgarh are amongst the best performers, Uttarakhand, Rajasthan and Jharkhand are amongst the worst. While the performance of Himachal Pradesh has been most impressive, Gujarat is amongst the worst on health, Maharashtra on infrastructure, and Haryana on both.”
A recent Planning Commission-sponsored study, “Refining State Level Comparisons in India”, by Pranjul Bhandari, who works as economist at the Office of the Chief Economic Advisor in the Ministry of Finance, Government of India, has found that a “refined” analysis of the performance across 21 major states suggests that Gujarat ranks 16th in health, 14th in education and 11th in infrastructure. The analysis is already creating flutter against the backdrop of the claim by those favouring Gujarat’s neoliberal model, that state’s alleged poor performance in the social sector is more an outcome of “fudged” figures, dished out by the Government India’s different ministries.
The “refined” analysis for comparing states is considered a new and a more scientific methodology, adopted in order to find out how well states perform given the availability of resources at their disposal. It differs from the “raw” analysis, under which Gujarat ranked 12th in health, 10th in education and 6th in infrastructure among 21 major states.
For the sake of convenience, Bhandari has sought to confine her analyze across just three critical sectors – health, education and infrastructure. Explaining her new methodology, she says, “Our ‘raw’ results conform with the already well-established findings of several other studies that states such as Kerala are amongst the best performing while the so-called BIMARU states (Bihar, MP, Rajasthan and UP) are laggards. While this is true on an absolute level, it does not reveal the performance conditional on state level factors”, the economist insists, pointing towards the need to “refine this analysis”.
 |
The economist says, “The comparative performance of individual states has become an important area of research for a number of reasons. Given the well-known regional disparities in India, a study of parts (i.e. the states) becomes important if the sum of parts (i.e. the country) needs to progress in a balanced way. Also, a study of states throws up successful experiments and examples which can be replicated or adapted by other states. Issues at the state level are increasingly dictating election outcomes both at the centre and the states, making this study important for the political class as well.”
She further says, each of the sectors she has chosen is “complex”, and “given the sheer size of resources needed for scale up, each of these three needs effort from both the public and private sectors. The public sector for instance not only needs to provide resources, but also create a policy environment conducive for scale-up.” Given this framework, she adds, “We try to analyze the long term performance of states in the provision of health and education services as well as infrastructure. We rank the states and gauge if performance across the three sectors are correlated or divergent. We compare states for both absolute performance as well as for performance after controlling for consumption levels.”
She further says, each of the sectors she has chosen is “complex”, and “given the sheer size of resources needed for scale up, each of these three needs effort from both the public and private sectors. The public sector for instance not only needs to provide resources, but also create a policy environment conducive for scale-up.” Given this framework, she adds, “We try to analyze the long term performance of states in the provision of health and education services as well as infrastructure. We rank the states and gauge if performance across the three sectors are correlated or divergent. We compare states for both absolute performance as well as for performance after controlling for consumption levels.”
 |
Explaining her “refined” analysis, the economist says, the performance after controlling for consumption levels “can be associated with governance – how well the resources at the state’s disposal have been used for progress in the critical sectors of health, education and infrastructure… What we do here is to control for per capita consumption before analyzing or ranking performance. This puts the states on a level playing field before comparisons are made. For instance, Bihar’s underperformance on many fronts could partly be explained by lower resources at its disposal which makes it difficult for the state to invest more on health and education. Our analysis controls for this factor while evaluating the state’s performance in delivering key services.”
Raw comparison of states: For raw analysis, the economist makes three separate indices (for health, education and infrastructure), each of which combine several widely used and publicly available variables that are available across states. In all, 21 major states are covered. For health, the economist uses both input (e.g. immunization) and output (e.g. infant mortality rate) variables. For education, she uses variables which reflect both the quantity (e.g. net enrolment rate) as well as quality (e.g. reading level for enrolled children). As for infrastructure, it is divided across sectors, such as agriculture, electricity and transportation, to ensure that the main sectors are included.
The raw analysis entails the following steps – first, the economist gets a complete data set of all the variables across the 21 states. Then, the data are stratified and a ranking is given in accordance with performance. For instance, in health ranking, higher institutional deliveries are better and the data is left as is. But higher infant mortality rate is worse, therefore the inverse of IMR is taken.
In the raw analysis following rankings are observed:In the first tier are Kerala, Goa, Himachal, Punjab, Tamil Nadu, Maharashtra and Haryana, who are found to be the best performers. However, performance of Maharashtra in infrastructure and that of Haryana in health is found to be markedly poor.
The second tier states comprise West Bengal, Uttarakhand, Karnataka, Andhra, Gujarat, J&K and Orissa. They are considered “medium performers”. Orissa stands out for worse performance on infrastructure, compared to its performance in health and education.
The third tier states comprising Rajasthan, Assam, MP, Chhattisgarh, UP, Bihar and Jharkhand are the laggards, mostly comprising of the BIMARU states.
“The rank correlation between the three indices is high, ranging from 81% to 88%, implying similarities in performance across health, education and infrastructure”, the economist says, adding, “Of the three correlations, the one between health and education is the highest. The rank correlation between each of the three indices and monthly per capita consumption expenditure (MPCE; source: NSSO, 2009/10) is also high, ranging between 80% and 87%. While these are simple associations and not causal relations, they suggest that higher growth and income are associated with better health, education and infrastructure status. “
Raw comparison of states: For raw analysis, the economist makes three separate indices (for health, education and infrastructure), each of which combine several widely used and publicly available variables that are available across states. In all, 21 major states are covered. For health, the economist uses both input (e.g. immunization) and output (e.g. infant mortality rate) variables. For education, she uses variables which reflect both the quantity (e.g. net enrolment rate) as well as quality (e.g. reading level for enrolled children). As for infrastructure, it is divided across sectors, such as agriculture, electricity and transportation, to ensure that the main sectors are included.
The raw analysis entails the following steps – first, the economist gets a complete data set of all the variables across the 21 states. Then, the data are stratified and a ranking is given in accordance with performance. For instance, in health ranking, higher institutional deliveries are better and the data is left as is. But higher infant mortality rate is worse, therefore the inverse of IMR is taken.
In the raw analysis following rankings are observed:In the first tier are Kerala, Goa, Himachal, Punjab, Tamil Nadu, Maharashtra and Haryana, who are found to be the best performers. However, performance of Maharashtra in infrastructure and that of Haryana in health is found to be markedly poor.
The second tier states comprise West Bengal, Uttarakhand, Karnataka, Andhra, Gujarat, J&K and Orissa. They are considered “medium performers”. Orissa stands out for worse performance on infrastructure, compared to its performance in health and education.
The third tier states comprising Rajasthan, Assam, MP, Chhattisgarh, UP, Bihar and Jharkhand are the laggards, mostly comprising of the BIMARU states.
“The rank correlation between the three indices is high, ranging from 81% to 88%, implying similarities in performance across health, education and infrastructure”, the economist says, adding, “Of the three correlations, the one between health and education is the highest. The rank correlation between each of the three indices and monthly per capita consumption expenditure (MPCE; source: NSSO, 2009/10) is also high, ranging between 80% and 87%. While these are simple associations and not causal relations, they suggest that higher growth and income are associated with better health, education and infrastructure status. “
 |
Refined comparison of states: The economist believes that while the “analysis above is insightful, it only reiterates the well known fact that states like Kerala have done well on health and education, while the BIMARU states have been laggards. States with lower resources at their disposal are likely to underperform.” Hence the need for a refined analysis, “by creating a level playing field before comparing states”.
She explaining the “refined” analysis method, she says, “We adjust the three indices created for monthly per capita consumption (MPCE). Although gross domestic product (GDP) per capita and consumption per capita broadly measure the same thing and are tightly correlated (with a correlation coefficient of 90%), consumption has the benefits of reflecting the actual purchasing power and including income generated from outside the state (i.e. inter state remittances).” Hence, the economist calculates “state-wise MPCE by taking a population weighted average of rural and urban MPCE for each state. Population statistics are taken from the Census 2011, and rural and urban MPCE from NSSO 2009/10.”
She further explains, ”MPCE works well for health and education as both are household decisions to a large extent. While it could be argued that GDP per capita should be used for infrastructure, we continue to use MPCE because … using MPCE for each of the three sectors is important for doing a comparable analysis.”
The ‘refined’ analysis throws up the following observations:
She explaining the “refined” analysis method, she says, “We adjust the three indices created for monthly per capita consumption (MPCE). Although gross domestic product (GDP) per capita and consumption per capita broadly measure the same thing and are tightly correlated (with a correlation coefficient of 90%), consumption has the benefits of reflecting the actual purchasing power and including income generated from outside the state (i.e. inter state remittances).” Hence, the economist calculates “state-wise MPCE by taking a population weighted average of rural and urban MPCE for each state. Population statistics are taken from the Census 2011, and rural and urban MPCE from NSSO 2009/10.”
She further explains, ”MPCE works well for health and education as both are household decisions to a large extent. While it could be argued that GDP per capita should be used for infrastructure, we continue to use MPCE because … using MPCE for each of the three sectors is important for doing a comparable analysis.”
The ‘refined’ analysis throws up the following observations:
Good performers – Himachal Pradesh, Kerala, Orissa, Tamil Nadu and Bihar have been the best performers across all the three sectors. West Bengal and Chattisgarh have also been amongst the best off states.
Laggards – Uttarakhand, Rajasthan, J&K and Jharkhand have been laggards across all the three sectors.
Average performers – The remaining middle ranking states have varied performance. Goa, Punjab and Karnataka have done well in health and infrastructure, but underperformed in education. On the other hand, Haryana, Andhra, Gujarat, Assam, MP, UP and Maharashtra have each underperformed in two of the three sectors have been analyzed.
The economist says, “We also rank the states across health, education and infrastructure based on the residuals. The rank correlations between them have fallen to the 25% to 50% range (46% between health and education; 25% between education and infrastructure; 50% between infrastructure and health) compared to the 80% to 87% range in the raw analysis. This was expected given that we have now controlled for consumption which could have been directly or indirectly driving some of the similarities in rankings in the raw analysis.”
Clearly, the rankings of many states change when the indices are refined. “Thus, Bihar, Orissa and Chattisgarh have risen sharply in rankings across all the three sectors. Relative ranking of Jharkhand has also improved but it remains a laggard state. Haryana and Uttarakhand have fallen in rankings across all the three sectors. Gujarat, Punjab and Maharashtra have also slipped in ranks in the refined analysis”, the economist concludes.
Laggards – Uttarakhand, Rajasthan, J&K and Jharkhand have been laggards across all the three sectors.
Average performers – The remaining middle ranking states have varied performance. Goa, Punjab and Karnataka have done well in health and infrastructure, but underperformed in education. On the other hand, Haryana, Andhra, Gujarat, Assam, MP, UP and Maharashtra have each underperformed in two of the three sectors have been analyzed.
The economist says, “We also rank the states across health, education and infrastructure based on the residuals. The rank correlations between them have fallen to the 25% to 50% range (46% between health and education; 25% between education and infrastructure; 50% between infrastructure and health) compared to the 80% to 87% range in the raw analysis. This was expected given that we have now controlled for consumption which could have been directly or indirectly driving some of the similarities in rankings in the raw analysis.”
Clearly, the rankings of many states change when the indices are refined. “Thus, Bihar, Orissa and Chattisgarh have risen sharply in rankings across all the three sectors. Relative ranking of Jharkhand has also improved but it remains a laggard state. Haryana and Uttarakhand have fallen in rankings across all the three sectors. Gujarat, Punjab and Maharashtra have also slipped in ranks in the refined analysis”, the economist concludes.
Comments