Month: December 2017

Enrolment Rates Are Climbing. So What Explains the Sorry State of India’s Education Sector?

[This article was published by on 21st December 2017. It was written in collaboration with Aniruddha Ghosh, a classmate from LSE. To read the article on, please click here . The article has also been posted on the LSE South Asia Blog and has been featured at Qrius]

Economic research shows that interventions aimed at improving cognitive skills rather than mere enrolment rates are required to boost economic growth.

Despite high enrolment ratios of over 96% in the last eight years, improvement in reading outcomes and arithmetic ability continues to be low. Credit: World Bank/Curt Carnemark/Flickr (CC BY-NC-ND 2.0)

Despite high enrolment ratios of over 96% in the last eight years, improvement in reading outcomes and arithmetic ability continues to be low. Credit: World Bank/Curt Carnemark/Flickr (CC BY-NC-ND 2.0)

Come January 2018, the Annual Status of Education Report (ASER) 2017 will be up for discussion among policy experts. Based on household-based surveys that cover children in the age group 3-16 across almost all rural districts of India, ASER provides estimates of children’s schooling status and their ability to do basic reading and arithmetic tasks.

ASER’s 2016 report observed that despite high enrolment ratios of over 96% in the last eight years, improvement in reading outcomes and arithmetic ability continues to be low. Moreover, a large proportion of students in both government and private schools continue to be below ‘grade level’. Grade level means that a student can deal with what is expected of her in that grade.

The ASER 2017 is targeted to look ‘beyond basics’: the age group between 14-18, primarily those outside the Right to Education ambit. There is a significant dearth of information in this regard and therefore, ASER 2017 will be a critical information asset to assess India’s madhyamik shiksha scenario. The government’s flagship Rashtriya Madhyamik Shiksha Abhiyan (RMSA) launched in 2009 and re-booted in 2013 as RMSA-Integrated has not been much of a success in India’s secondary education scene. In light of these observations, it is likely that ASER 2017 will throw up systematic issues that have continued to plague our secondary education.

ASER 2016 and other previous reports

Let’s take a step back and illustrate how our performance has been in the age group 3-16. Here is a snapshot of ASER’s previous reports on the state of India’s early grade development.

Figure 1: Percentage of children not enrolled in school. The trend for enrollment has been on the rise. Source: ASER 2016 Report

Figure 1: Percentage of children not enrolled in school. The trend for enrolment has been on the rise. Source: ASER 2016

Figure 2: Children in class III who are at ‘Grade Level’ 2008-2016. The lack of commensurate ability with grade level is evident. Source: ASER 2016

Figure 2: Children in class III who are at ‘Grade Level’ 2008-2016. The lack of commensurate ability with grade level is evident. Source: ASER 2016

Fig 3: Percentage of students who can at least do subtraction in Grade III. Large variation across states. In 2016, the number of children at grade level ranged from 50% in Himachal Pradesh to 10% in Uttar Pradesh. Source: ASER 2016 Report

Enrolment levels have been high for the 6-14 age group, and around 96% since 2009 onwards.

The percentage of children in grade 3 who are able to read at least a grade 1 level text has improved marginally – from 40% in 2014 to 42.5% in 2016.

Percentage of grade 3 children who could do two digit subtraction has marginally improved from 25% in 2014 to 27% in 2016. This has been the first year since 2010 when there has been an upward trend observed in arithmetic ability.

However, trends over time point to a dismal outlook. Figure 4 demonstrates the ability of grade 4 children in successive cohorts to read and do basic arithmetic. One can see a downward trend in the ability of successive cohorts.

Figure 4: The graphs show the performance of three cohorts from class IV to class VIII. The graph on the left shows the percentage of students who can do division; the graph on the right shows the percentage of students who can read a class II 2 level text. Source: ASER 2016

Figure 4: The graphs show the performance of three cohorts from class IV to class VIII. The graph on the left shows the percentage of students who can do division; the graph on the right shows the percentage of students who can read a class II 2 level text. Source: ASER 2016

Economics, enrolment rates and cognitive ability

One of the most influential studies to look at the relationship between education and economic growth was professor Lance Pritchett’s  Where has all the Education Gone? His 2001 study found no significant positive relationship between educational attainment and economic growth.

He wrote:

“In the decades since 1960, nearly all developing economies have already seen education attainment grow rapidly. The cross-national data show, however, that on average, education attainment contributed much less to growth than would have been expected in the standard augmented Solow model.”

He proposed several possible explanations for this phenomenon, the most important of which being the classic quantity-quality argument: Quality of educational attainment was so low that despite high educational attainment there wasn’t a significant increase in cognitive skills and human capital.

As outlined above, the very idea that improved schooling (through the lens of school attainment rates) – which has often been a cornerstone of most interventionist strategies – will raise economic well-being has often been discounted by economists. Eric Hanushek and Ludger Woessmann in their authoritative paper, ‘The Role of Cognitive Skills in Economic Development (2008)‘ find strong evidence that the cognitive skills of the population – rather than school attainment rates – are powerful determinants of earnings, economic growth and income distribution.

Their work premises itself on a rather simple question: whether education is the steering force or merely one of the several factors that are correlated with more fundamental development forces, say cognitive skills when it comes to economic growth. Hanushek-Woessmann armed with strong econometric evidence argue that cognitive skills overwhelmingly outnumbers schooling attainment rates when it comes to influencing economic growth. Moreover, the effect of cognitive skills on economic growth is larger in developing countries than in the developed ones.


Building on the works of Pritchett, and Hanushek and Woessmann, a recent research paper titled Does One Size Fit? The Impact of Cognitive Skills of Economic Growth by professor Nadir Altinok of the University of Lorraine and professor Abdurrahman Aydemir of Sabanci University studies the difference in the impact of cognitive skills on economic growth between developing and developed economies.

In addition to the findings mentioned above, the heterogeneous effects of cognitive skills vis-a-vis the income levels of the economies is rather stark: the magnitude of the effect of cognitive skills is about 60% higher for low-income countries compared to high-income countries, and this more than doubles when low total-factor productivity (TFP) countries are compared to high TFP countries. From a policy perspective, this encourages the view that the promotion of education policies that focus on the quality of education has especially larger payoffs in the least developed regions through the productivity channel. To sum it up, economic research overwhelmingly supports the idea of interventions aimed at improving cognitive skills rather than mere enrolment rates.

To cite a success story, Colombia has made impressive progress towards universal enrolment in basic education and at the same time has raised learning outcomes. A lot of this accrues to the flexible ‘new school’ model, commonly known as Escuela Nueva.

Escuela Nueva accepts multigrade teaching as an unavoidable condition in small schools of rural areas. It encourages to develop special materials and teaching methods for multigrade teaching. The academic achievement of the students in Escuela Nueva has been consistently higher than in urban schools. There are plenty of cross-country experiences to learn from and we should actively explore and replicate them here.

India’s scenario

Sarva Shiksha Abhiyan (SSA) was launched in 2000 to spread the availability of universal elementary education across India. Under SSA, commendable progress has been made in increasing enrolment rates; as well as providing basic infrastructures such as classrooms, water, toilets and boundary walls to all schools. Yet, what is the scenario when it comes to learning outcomes?

India made its debut in the Programme for International Student Aptitude (PISA) test in 2009 with 16,000 students from 400 schools across Himachal Pradesh and Tamil Nadu. While China – also a first-timer in 2009 – stormed into the number one position with Shanghai schools topping in math and science, India was at a paltry 72nd among the 74 participating countries. Since then, India has boycotted the PISA rankings citing ‘methodological differences’ but it plans to return to the rankings fold in 2021. Coupled with ASER’s findings over the years, this paints a sorry state of India’s primary education sector.

Now that enrolment rates are high, we would need to look for innovative interventions to improve learning-based outcomes in India. One of the recommendations of the ASER report, which the government has been focussing on in the last few years, has been early-childhood care and nurture, especially for children in the 0-3 age group. In fact, ASER’s study on three states (Andhra Pradesh, Assam and Rajasthan) found a positive and significant relationship between early childhood care and nurture and early grade learning outcomes. Although issues do remain in implementation, this avenue holds promise to further improve outcomes in India.

Additionally, in India, there are two interesting policy interventions that are in the process of being rolled out: same language subtitling (SLS) and outcomes fund for the education sector.

Conceived by Professor Brij Kothari of IIM Ahmedabad, SLS is the concept of subtitling existing Bollywood film songs on television. Kothari’s research estimates a 9% increase in the number of functional readers who watch TV programmes with SLS within a period of two years.

Similarly, a large outcome-based fund for education is all set to launch in India in early 2018. Touted as one of the first and largest funds for social enterprises, the fund would invest in education providers to work with government-run schools to deliver outcomes. There could be a variety of outcomes like early childhood interventions, retention of girl students, learning in primary schools and employability of students after high school. The fund is being launched by the Global Social Impact Investment Steering Group, an organisation comprising 13 member countries (including India), with a focus on channeling global social impact investment. Outcome fund based models are actively being employed by nations across the globe to fund social projects and has the potential to deliver the necessary outcomes.

While we wait for ASER’s 2017 findings, much of the theory and evidence that we have strongly suggests that raising enrolment rates hasn’t been enough to push our growth frontiers. The hope is that the required stakeholders will aim, plan and push for innovative interventions that encourage student achievements.

Finite Populations: How do we think about them?

Consider how a normal regression output on STATA looks.

Sample Regression Output


We find that along with the coefficients, we get other values like standard errors, p values, confidence intervals, etc. So what is the purpose of these stats?

The standard assumption in stat-econ literature is as follows: i)the observed sample is a random representative sample from the entire population, ii) we are interested in the true population parameters.

What does this mean? We assume that we pick a random representative sample from a population. After running a regression, we get a particular coefficient on all the x’s. We assume that in repeated sampling, we would get a distribution of coefficients for the x’s. In fact, one of the fundamental properties of a good estimator is unbiasedness or that the Expected Value of the coefficient from the sampling distribution is equal to the true value, i.e. the true coefficient is equal to the expected value of the coefficient in repeated sampling. Although we rarely do repeated sampling in practice, we do check for the significance of the coefficients with a null hypothesis that they are equal to zero. Thus, in a normal regression framework, the population has the true coefficients and is quite distinct from the samples which have a distribution of coefficients.

So what happens when we do have the entire population?

This is a tricky question, as we usually always assume that the population is infinite and we can never have the entire population. Thus, in cases of finite population, like the 28 states of India, it becomes a little bit tricky as to how one should interpret the regression output. I will attempt to summarise two popular views in the stat-econ literature in this regard.


One view says that if you have the entire population, the p values and t statistics are irrelevant.  The finite population is considered to be a fixed set of elements. The coefficients derived are the true relationship since you have the entire population and not a sample of the population. Thus, the concepts of hypothesis testing and significance become meaningless for an entire population. This is because the stats and tests are only relevant for a sample, and not the entire population. The caveat here is that the model has to be correctly specified.

Super Population / Underlying Process

This view states that the observations in the population are simply a sample from an infinite population. For example, if we’re looking at the 28 states in India, we can interpret the set of observations at one time period as a sample; and the set of observations across all time as the infinite super population.

This becomes important, especially if you want to make inferences not just about relationships today but also if you want the relationships to hold for similar groups in the future.

Another related but slightly different view would be that the observed outcomes in the population are the products of an underlying process. In that regard, what standard errors and other test stats capture become relevant. According to Abadie et al (2014) “there are for each unit missing potential outcomes for the treatment levels the unit was not exposed to.”

On a similar vein Wallis and Robers (1956) claim, “the totality of numbers that would result from indefinitely many repetitions of the same process of selecting objects, measuring or classifying them, and recording results.”

In these regard, test stats become important in understanding whether the coefficients generated are true or have been generated as a result of a chance outcome.



  1. Abadie A, Athey S, Imbens G, Wooldridge J (2014), Finite Population and Causal Standard Errors, NBER Working Paper Series.
    Available at:
  2. Asali M (2012), Can I make a regression model with the whole population? [Msg 1], ResearchGate.
    Message posted:
  3. Frick R (1998), Interpreting statistical testing: Process propensity, not population and random sampling, Behavioral Research Methods, Instruments, & Computers. Vol 30 (3), pp: 527-535
  4. Hartley H, Sielker Jr R (1975), A “Super Population-Viewpoint” for finite population sampling, Biometrics, Vol. 31 (2) , pp: 411-422
  5. Hidiroglou, Michael Arsene (1974), Estimation of regression parameters for finite populations , Iowa State University Digital Repository.
    Available at:
  6. January (2013), How to report data for an entire population? [Msg 1], CrossValidated.
    Message posted to:



© 2019 Sujan Bandyopadhyay

Theme by Anders NorenUp ↑