Accounting for Growth Effects of Age Structure Transition through Public Education Expenditure: New Macroeconomic Evidence from India

Abstract

This article quantifies the growth effects of age structure transition through current public education expenditure. Using the National Transfer Accounts’ (NTA) First Demographic Dividend (FDD) model, growth effects are accounted by the impact of current public education expenditure on economic support ratio (ESR) and labour productivity through human capital investments. The results offer new macroeconomic evidence. Age structure transition reduces the education dependency ratio (EDR) by all levels of education but the highest in the elementary education. This impacts on a long-term decline in enrolment in elementary education where the current gross enrolment ratio (GER) is close to 100 per cent and a decline in current public education expenditure. Other things being equal, the resultant potential savings, or the availability of extra budgetary resources, is a new way of financing the investment requirements for secondary and higher education with the aim of increasing national economic growth through human capital investments. In particular, growth effects are shown to be positive, higher and longer up to 2050, if the current public education spending is reallocated more for the secondary and higher education. Surprisingly, growth effects are explained less by the ESR than labour productivity. This justifies a higher human capital investment to enhance labour productivity for attainment of higher economic growth. The afore- mentioned macroeconomic framework, results and implications are of general relevance for other developing countries in South Asia and elsewhere in the world.

JEL Classification: E65, H52, J11

Keywords

Age structure transition public education expenditure human capital investment economic growth National Transfer Accounts demographic dividend

Introduction

India’s education sector is remarkable by its institutional networks and enrolment of students. Government of India (2016) shows that, by March 2015, India had 15.17 million schools up to secondary education (Class IX to Class XII) with 84 per cent belonging to elementary education (Class 1 to Class VIII); 760 universities and 38,498 colleges in higher (post-secondary) education and 12,276 stand-alone institutions (e.g., diploma-level institutions). Total enrolment of students in the elementary, secondary and higher education institutions was equal to 293 million with 67.31 per cent of students (197.67 million) in elementary education: 21.04 per cent of students (61.80 million) in secondary education and 11.65 per cent of students (34.21 million) in higher education. At the same time, gross enrolment ratio (GER) varied across the levels of education: 95.90 per cent for the elementary education, 64.83 per cent for the secondary education and 24.30 per cent for the higher education.¹ Further, 76 per cent of elementary schools, 37 per cent of total secondary schools and 80 per cent of universities were in the public sector (Government of India, 2014; Mehta, 2013). Thus, public spending or expenditure is important for education in general and for elementary and higher education in particular. In fact, a higher public spending is essential to invest more on human capital formation and for improvements of India’s ranking in Human Development Index (HDI) and Human Capital Index (HCI) as they are related to the education index or pillar. For instance, India was ranked 130th country among the 188 countries by the value of HDI (0.609) in the UNDP’s Human Development Report 2015 (UNDP, 2015). As compared to the value of Health Index (0.738) and Income Index (0.605), the value of Education Index was lowest at 0.505. On the other hand, of the 122 countries, India was placed at the 78th rank in the overall HCI and 63rd rank in education pillar by the World Economic Forum’s HCI (World Economic Forum, 2013). Thus, there is a greater need to spend more on the education sector, now and in future, to improve India’s international ranking through higher human capital investments.²

In economic terms, public spending on education aims at the creation of physical and/or human capital by educational institutions which produce and provision the educational services. Students are the key stakeholders by their direct consumption of the educational services, as provided by the institutions. A decline in enrolment of pre-secondary education in public institutions may result in less current and future public expenditure due to a fall in the number of student-consumers. This is particularly relevant to the school-going population (comprising young population in age group 6–13 years) in elementary (or pre- secondary) education because the GER in elementary education is close to 100 per cent. Thus, a long-term decline in population in the school-going age may reduce the size of school enrolment and per student expenditure if the total budget for the education sector continues to grow and the quality of school education remains at a benchmark year. This process may free up budgetary resources at the school education, or lead to the availability of extra budgetary resources or contribute to potential savings, which are allocable for human capital investments in secondary and higher education to achieve higher economic growth. Thus, demographic or age structure transition is an important determinant of current and future public spending by levels of education and its resultant growth effects through human capital investments.³

This article aims at economic analyses of the empirical linkages between demographic transition, public spending on education and economic growth. This analysis is important to answer many policy relevant questions for India, such as, the following. What is the nature and magnitude of public education expenditure by age? How do age-specific expenditures vary across levels of education and over time? How does age structure impact on public education expenditure through the education dependency ratio (EDR)? Will age structure transition result in potential savings in elementary education? If yes, can those savings be a new instrument of financing the investment requirements at secondary and higher education? What are the growth effects of such public education expenditure through human capital investments and how to quantify them? Plausible answers to these questions are essential to offer empirical evidence on how public educational spending responds to demographic changes and to draw implications for economic growth through human capital investments.

This article uses the accounting methodology of National Transfer Accounts (NTA) which is a macroeconomic framework for introduction of age into National Income and Product Accounts (NIPA). As individuals pass through their lifecycle from young to youth, youth to working age and from working age to old age, both production and consumption changes create deficits (consumption exceeds production) or surplus (consumption is less than production) at each age. As an accounting framework, NTA aims at (a) quantifying the nature and magnitude of these economic lifecycle changes and (b) developing the public and private institutional mechanisms by which deficits are financed by surplus generated during the working ages through age reallocations in terms of transfers and asset-based reallocations. These aims are accomplished by developing a conceptual framework for calculation of inter-age flows by the age profiles of consumption, production and age reallocations. This framework is the basis for construction of the Flow Account of NTA, consistent with the National Income Identity in NIPA. The Flow Account gives accounting relationships through inter-age flows (i.e., inflows and outflows) of all macroeconomic variables for an accounting year in monetary terms and at national level of aggregation.⁴ Using this methodology, (a) age profile of public education expenditure is calculated and impact of age structure transition on public education expenditure is forecast over the period 2005–2050 and (b) sources of changes in public education expenditure over the period 2005–2050 are decomposed by interactive and isolated effects of age structure transition using the methodology of Lee and Mason (2014). Further, using the NTA-based growth model (Lee & Mason, 2011b), growth effects of public education spending are quantified through First Demographic Dividend (FDD) model over the period 2005–2050. In particular, growth effects are calculated by the impact of public education expenditure on human capital investments. This approach is useful to distinguish the impact of human capital investments on economic support ratio (ESR) and labour productivity. These forward-looking analyses are aimed to provide with unambiguous answers to the research questions in this article and to draw useful policy implications for India in particular and other South Asian developing countries in general.

Rest of the article is organized as follows. The second section presents a brief review of related literature. The third section describes India’s age structure transition and its relevance for analysis of public expenditure by levels of education. The next section presents the empirical frameworks for linking between the age structure transition, public education expenditure and economic growth. Variables and data descriptions are given in the fifth section. Analyses of results are presented in the sixth section. Major conclusions and implications are summarized in the last section.

Review of Related Literature

There exists an extensive literature on description, calculation/estimation and analyses of nature, extent and impact of the sources of financing of higher educational services at the national and state levels in India. This literature includes Tilak and Varghese (1991), Tilak (1993), Narayana (2006, 2008), Prakash (2007), Agarwal (2009), Ernst and Young (2013) and Rani (2014). These studies do not focus on linkages between the public education spending, demographic transition and economic growth, and hence, do not provide with answers to the earlier questions. Are there studies which establish this linkage, and answer the research questions, outside the literature on India’s public financing of education? This is answered further by a review of literature in and outside India.

Demographic Change and Public Education Expenditure

Empirical studies which focus on impact of demographic changes on public education spending include Poterba (1997, 1998) for the USA, Grob and Wolter (2005) for Switzerland, Lee and Mason (2014) for 40 Asia and Pacific countries, Marchionni and Alejo (2015) and Gragnolati and Troiano (2015) for Argentina, and Miller and Castanheira (2013) for Brazil. In general, this literature shows the negative impact of demographic changes (e.g., age structure transition, population ageing and decline in fertility) on per pupil public education expenditure and distinguishes the sources for this impact through a decomposition analysis. For instance, Poterba (1997) used a panel data of 48 states from 1961 to 1991 to estimate the determinants of per student public spending on K-12 education (age group 5–17 years) in the USA. The estimation results showed, among others, that an increase in share of elderly population has a negative impact on per pupil educational spending and variations in school-age population do not result in proportionate changes in educational spending. Poterba (1998) analysed the available evidence from different studies to explain a lower per pupil expenditure in public schools in states with a higher proportion of elderly population in the USA. This explanation is mainly driven by the self-interest of elderly generation (e.g., public-funded old-age social security and long-term health care needs) which is best served by keeping expenditure on public schools at lower levels.⁵ Grob and Wolter’s (2005) used a panel of the Swiss Cantons from 1990 to 2002 and found that the public education system exhibited little elasticity in adjusting to changes in the school-age population, and that the share of the elderly population had a significant negative influence on the willingness to spend on public education. Lee and Mason (2014) focused on the impact of demographic changes, among others, on public education spending of the Central Government in 40 Asia and Pacific countries. Trends in government expenditure on education as a percentage of GDP over the period 1995–2005 or 2010 and projections up to 2050 are analysed for each country. Overall, the results show a decline in public spending on education in all Asian and Pacific countries due to decline in fertility. Further, Lee and Mason (2014) decomposed the sources of change in public spending on education over the period 2010–2050 by age structure, age-specific spending levels and interaction of age structure and age-specific spending levels. In all countries, changes in population age structure showed negative effects, changes in spending levels showed positive effects, and interaction effects were negligible for 18 countries and negative for the rest. However, these effects on public education spending varied in magnitude across countries and regions in Asia.

Marchionni and Alejo (2015) focused on impact of age structure transition on the basic education (comprises children in early childhood, primary and secondary levels from age 3 to 17 years) in Argentina. Expenditure on the basic education as a percentage of GDP is decomposed by three sources: benefit-generosity ratio (or per student expenditure as a percentage of average GDP per potential worker), GER and EDR. This decomposition approach is useful to quantify the impact of a targeted increase in public education expenditure by the sources and education levels. Accordingly, desired policy changes can be identified including demographic factors which impact through the sources, either directly or indirectly. Further, as demographic changes free up more resources per student, new policy options are opened to shift resources towards quality, expansion of coverage, fostering inclusion and reducing overage rates. This result adds a new perspective on the linkage between demographic change and public education expenditure.

Gragnolati and Troiano (2015) and Miller and Castanheira (2013) analyse, among others, the implications of demographic transition on education (i.e., school education) through a decomposition of the share of public education expenditure as a percentage of GDP by demographic component (measured by population at each age relative to the working-age population) and economic/policy component (measured by average annual public benefits received at each age relative to GDP per working-age adult). This decomposition is useful to understand the importance of age structure in translating social effort (aggregate spending) into individual welfare (spending and measuring benefits relative to GDP per working-age adult.

It should be emphasized that impact of demographic transition on public education expenditure is not limited to school education. For instance, Vincent-Lancrin (2008) presents a comprehensive analysis on the impact of demographic changes on education expenditure and provisioning of educational services through student–teacher ratios and student enrolment for higher education in the OECD countries.

Demographic Transition and Economic Growth

Studies by Bloom, Canning, Fink, and Finlay (2007), Bloom et al. (2010), Choudhry and Elhorst (2010) show the linkage between demographic transition and economic growth and include India as one of the sample countries. Bloom et al. (2007) forecasted economic growth, using the cross-country regression estimations of 67 countries (including India) over the periods 1960–1980, 1980–2000 and 1960–1980. Log of the initial ratio of working age to total population measured the age structure transition. Two estimation models were distinguished based on the familiar neoclassical conditional convergence growth model: absolute growth model explained why and how fast each country grew, and relative growth model on how fast the country grew relative to the world average. The forecasted growth rate of India over the period 2000–2200 was equal to 2.1 per cent in the absolute growth model and 1.6 per cent in the relative growth model. Out of 90 countries, India ranked 34th by the forecasted value of the absolute growth model (the highest was China at 5.9%). Bloom et al. (2010) compared the sources of growth (or annual average growth rate of GDP per capita) between India and China over the periods 1970–1980 and 1980–2000. In both the periods and countries, the largest source of growth was evident for growth of labour productivity (or growth rate of real GDP per worker) as compared to age structure transition (or growth rate of ratio of population aged 15–64 to total population). Further, using the 5-year panel of countries over the period 1960–2000, determinants of economic growth were estimated by including the log share of initial working-age population and growth of share of working-age population. The results showed positive impact of these demographic variables on economic growth. Choudhry and Elhorst (2010) adopted the augmented Solow–Swan model of economic growth as a framework for estimation of demographic and non-demographic variables on economic growth in India, China and Pakistan. Growth differential between working-age population and total population measured the age structure transition. The coefficient of this demographic variable was found to be positive and significant in all cross-section and panel regression estimations based on the data from 70 countries over the period 1961–2003. Further, the study found that the impact of age structure transition on economic growth was higher in future (about 11% over the period 2005–2050) than in the past (about 5% over the period 1963–2003).

Two important India-specific studies on demographic transition and economic growth are Aiyar and Mody (2011) and Das and Kar (2015).⁶ Aiyar and Mody (2011) adopted the framework of the neoclassical conditional convergence growth model as in Bloom et al. (2007) and estimated the growth effects of demographic transition by using data on 22 states over the period 1961–2001. Log of initial working-age ratio and growth rate of working age are the two important demographic variables that capture age structure transition on economic growth. Positive and significant coefficients of these variables offered evidence for the growth effects of demographic transition. Das and Kar (2015) developed the time-series (OLS and VAR [vector autoregression] models) for the estimation of the growth effects of demographic patterns (or change in dependency ratio) and other macroeconomic variables (i.e., financial depth, interest rate and capital inflows). The OLS estimates over the period 1980–1981 to 2011–2012 showed significant and negative effect of dependency ratio on economic growth. The VAR results showed the predicted negative time trend on dependency ratio variable. Further, the results of the structural model showed the nature of contemporaneous relations between the dependency ratio and other variables. That is, positive relationship with the interest rate and financial depth and moving together in contemporaneous time with capital flows. However, the presence of large unorganized manufacturing sector and financially non-inclusive rural sector are shown to be the limiting factors for reaping the demographic dividend for India.

Demographic Transition, Human Capital and Economic Growth

Most recent literature links demographic transition and economic growth through human capital and public education.⁷ This literature includes Das and Kar (2016) and Mason, Lee, and Jiang (2016). Das and Kar (2016) developed a model of endogenous growth with human capital by health and education (measured by public expenditure on education and health) and infrastructure. This backward-looking model brings in demography into growth model only through changes in effective labour force and is useful to explain and predict the changes in human capital on growth of labour force (or working-age population as a proportion of total population) and economic growth (or growth rate of per capita income). Using the time-series data on India from 1970 to 2010 and the VAR model, empirical effects of human capital on growth of labour force and economic growth are separated by health, education and infrastructure. The evidence shows, among others, the positive (or negative) impact of changes in education expenditure on economic growth (or growth of labour force). However, both the theory and empirics in Das and Kar (2016) do not capture the impact of demographic changes on human capital expenditure and their resultant impact on long run economic growth. In contrast, Mason et al. (2016) developed a forward-looking growth model with explicit demographic impact on human capital and its resultant economic growth. Their approach has three contributions. First, the level of human capital investment is distinguished by public and private education and that of health expenditure by using the Synthetic Cohort Measure (SCM) of human capital spending and is expressed in years of average labour income. Second, income elasticity of human capital spending, among others, is estimated by public and private education and health expenditure. Third, the impact of human capital on economic growth (or growth rate of per capita consumption) is estimated through the Second Demographic Dividend (SDD) and over the period 2010–2100 under varying assumptions on fertility rate. Three results of these analyses for public education expenditure are interesting and insightful. (a) India’s SCM for public education (0.82) is remarkably lower than for the high income countries (3.02) and middle and low income countries (1.37) and lower than all the countries except for Ethiopia (0.27), Cambodia (0.63), Nigeria (0.17), Senegal (0.54) and Vietnam (0.49). (b) The estimated point income elasticity of public education expenditure is 1.2, which is higher than for private education expenditure (0.817). (c) Human capital accounts for the largest source of economic growth beyond 2040 and public education spending composing the higher share within human capital beyond 2070. These results strongly support for the use of public education spending as a policy instrument for the promotion of economic growth through human capital formation.

This article adds to the aforementioned literature by linking between (a) age structure transition and public education expenditure, (b) reallocation of public education expenditure by level of education and human capital investments and (c) human capital investments and economic growth in India. These linkages integrate for the economic relationships between age structure transition, human capital through public education expenditure and economic growth in the framework of NTA’s FDD. Further, the linkages are contextualized by India’s institutional structures of public provisioning and financing of education and aim at answering all the research questions in this article.

Age Structure Transition

Figure 1 shows India’s age structure transition from 1961 through 2100. Data from 1961 to 2011 are from the Indian census reports and 2021 to 2100 from the UN Population Projections under medium fertility variant (United Nations, 2013b). The transition is remarkable: a decline in the share of young (0–14 years) and youth population (15–24 years), an increase in the share of elderly population (60 years and above); and the highest share of working-age population (25–60 years). For instance, before 1991, the share of young population (0–14 years) was higher than the working-age population (25–60 years). Since 1991, young population shows a continuous and rapid decline in contrast with a rising working-age population. Further, youth population (15–24 years) shows a gradual increase from about 17 per cent in 1961 to about 19 per cent in 2011 and a decline from about 17 per cent in 2021 to about 13 per cent in 2050 and 10 per cent in 2100. On the other hand, share of elderly population shows a gradual increase from about 6 per cent in 1961 to about 7 per cent in 2001 and a rapid increase from about 8 per cent in 2011 to about 20 per cent in 2051 and 36 per cent in 2100. This transition is essentially due to a long-term decline in fertility and complemented by a decline in mortality including at later ages.

Demographic Window of Opportunity (DWO) is a transition period during which the proportion of young population (age 0–14 years) permanently falls below 30 per cent and that of elderly population (age 60 years and above) is less than 15 per cent.⁸ Using the age structure transition in Figure 1, India’s DWO is identifiable from 2012 to 2037 with a falling (or rising) share of young population (or elderly population) from 29.66 (or 7.96) per cent to 21.52 (or 14.83) per cent.

A long-term decline in the share of young and youth population implies a corresponding decline in the share of school and college-going population because they include elementary school-going population (6–13 years), secondary-school-going population (14–17 years) and higher education-receiving population (18–24 years). The impact of these demographic changes on education system may be summarized by the EDR. This is measured by the ratio of school-age or college-age population to total population in active/working ages (25–60 years) or potential workers. Figure 2 shows the EDR by levels of education and all levels of education from 2005 through 2100. These ratios decline throughout including the period of DWO and imply that, other things being equal, the number of students by all levels of education falls for every 100 potential workers. However, the highest decline in EDR is evident for the school education. For instance, the EDR for the elementary (or higher) education falls from about 41 (or 32) per cent in 2005 to 24 (or 22) per cent in 2050 and 22 (or 20) per cent in 2100. During the same period, total EDR falls from about 93 per cent to 57 per cent and 54 per cent. The decline in EDR (all education) is in contrast with the transition of Total Dependency Ratio (or ratio of population in 0–14 years and 60 years and above to total population in 25–60 years) in Figure 2.

Thus, other things being the same, age structure transition has important implications on current and future public education spending because the beneficiaries of this spending are projected to decline over the period up to 2100. This is called potential savings in the education system as a result of demographic changes which are investable in educational attainments (Marchionni & Alejo, 2015). Long back, Mincer (1981) noted the importance of this demographic transition (i.e., decline in younger population [below age 15 years] who need more resources for consumption and education) on growth through supply of more productive labour or economic contributors.

India has been experiencing an increase in the share of working-age population. This, combined with more workers with secondary and higher education, may have a positive impact on labour productivity and, hence, the size of demographic dividend. Analyses of these impacts may provide with new explanations and predictions on growth effects of age structure transition through public spending on education. These impacts are of considerable policy importance if age structure transition may lead to more investible resources for higher education without higher taxation, without cut in non-education public expenditures and/or without raising new public debt. Thus, introduction of age structure transition has implications to uncover potential sources of India’s current and future human capital investments through the public education expenditure.

Framework for Empirical Analyses

India’s public education expenditure includes pre-secondary education (comprising elementary, lower secondary and higher secondary education), post-secondary or higher education, and literacy and training programmes. Higher education may be characterized by levels and types. Levels include the university education (regular and open/distance), autonomous institutes of higher learning (affiliated or autonomous and deemed to be universities) and collegiate education (affiliated and autonomous). Types include general, technical, medical, agricultural, management and legal education. Institutions in these types and levels may be distinguished by their ownership, management and financing, such as, union and state government, private-aided and private-unaided. Public financing of education may broadly include all those sources of financial and non-financial resources for institutional provisioning of educational services. Own and non-own sources of expenditure include government’s grants in aid, students’ fees, industry contributions including through corporate social responsibility, alumni associations, asset income (including interest income from corpus), affiliation fees, voluntary and tax-exempted philanthropy and donations. These descriptions show the complexity of reality in public spending on education in India. We simplify these complexities by developing the following aggregate framework by linking age structure transition, public education spending and economic growth in two steps. First, link is established between the age structure transition and public education spending through expenditure forecast model. Second, age structure transition and public education spending forecast are integrated in a model of national economic growth through human capital investments and demographic dividend.

Age Structure Transition and Public Education Spending

This framework is aggregate at national level and all public spending on education are distinguished by age. Age profile of public spending on education is calculated by using the methodology of NTA. To start with, NTA-Flow Account Identity (suffix ‘f’ stands for private sector, ‘g’ for public sector and ‘i’ refers to individual or age group) is formulated as follows:

Y_{L, i} + Y_{A, i} + (T_{f, i^{+}} + T_{g, i^{+}}), = (C_{f, i} + C_{g, i}) + S_{i} + (T_{f, i^{-}} + T_{g, i^{-}}),

(1)

where Y_L,i is labour income, Y_A,i is non-labour or asset income, T_f,i⁺ and T_f,i^– are private transfer inflows and outflows, respectively; C_f,i is private consumption expenditure, C_g,i is public (government) consumption expenditure, S_i is savings, and T_g,i⁺ and T_g,i^– are public transfer inflows and outflows, respectively. The left hand side of Equation (1) shows total inflows and the right hand side shows total outflows. Net exports are indirectly introduced in (1) to take care of Rest-of-World (ROW) by including net compensation of employees from ROW in Y_L,i and net entrepreneurial income from ROW in Y_A,i. This implies that Equation (1) is consistent with an open macro-economy. Further, individual is the fundamental entity in the NTA and all flows are disaggregated at individual level by age.⁹

Public education consumption is included in the public sector’s consumption inflows in Equation (1). The public education consumption profile measures the age-specific public education consumption. The profile is distinguished by aggregate and per capita consumption at each age and by primary, secondary and higher education and other education. The public education consumption profile is converted into public education expenditure profile for expenditure forecasting purposes. The methodology for this forecasting is elaborated further.

Following Miller (2006), aggregate public expenditure on education is forecast by using a fixed age profile of public education consumption [E(x,t₀)], which shifts upward over time at the growth of nominal labour productivity (ρ), combined with a forecast of population by age (x), P(x, t). The nominal labour productivity (ρ) is equal to sum of growth rate of real labour productivity and inflation. Thus, aggregate public expenditure on education in time-t, E(t), is equal to:

E (t) = Σ ρ E (x, t_{0}) P (x, t)

(2)

where E(x,t₀) = [EX₀{EC(x,t₀)/∑EC(x,t₀)}]/P(x,t₀) is the age profile of per capita public education expenditure in the benchmark year (t₀), where EX₀ is total public expenditure on education, EC(x,t₀) is the total education consumption at age-x and ∑EC(x,t₀) is the sum of public education consumption over all ages in the benchmark year. Thus, E(t) in Equation (2) is the age-specific and population-adjusted forecast of total public education expenditure.

Aggregate labour income is calculated by using a fixed age shape of per capita labour income, L(x, t₀), which shifts upward over time at the growth rate of nominal labour productivity (ρ) combined with a forecast of population by age P(x,t). The fixed shape means that the age profile of labour income is constant throughout the forecasting period or labour income profile varies over years by levels, but shape of the profile remains fixed at the benchmark level.

Y (t) = Σ ρ L (x, t_{0}) P (x, t)

(3)

Next, GDP is forecast by assuming a fixed ratio of GDP to aggregate labour income in the benchmark year. That is,

G D P (t) = [G D P (t_{0}) / Y (t_{0})] . Y (t)

(4)

The forecasting of public education expenditure in Equation (2) is determined by (a) age profile of public education consumption expenditure, (b) aggregate public education expenditure and (c) age structure of population. Given the age profile and aggregate control expenditure, and age structure determines the forecast values of the public expenditure. The important advantage of using the NTA-based age profile of education expenditure in Equation (2) is its macroeconomic basis in Equation (1).

Growth Effects of Public Education Spending

Let Y(t) be the national income in year t, L(t) be the total number of effective producers or workers and N(t) be the total number of effective consumers. Effective number of producers and consumers are measured, respectively, by

L (t) = Σ γ (x) P (x, t)

(5)

N (t) = Σ φ (x) P (x, t)

(6)

where γ(x) is productivity at age-x or productivity age profile, φ(x) is consumption needs at age-x or consumption age profile, P(x,t) is population at age-x and time-t and the summation (∑) is over all ages. ‘Effective’ number of producers (or consumers) refers to weighted sum of producers (or consumers) in total population and per capita productivity (or consumption) at age-x is the weight. Consumption profile is calculated by combining the age profiles of public and private education, health and other expenditures. These age profiles [γ(x) and φ(x)] are calculated in the framework of NTA and, hence, consistent with the country’s macroeconomic equilibrium. They also provide the essential link between the NTA and economic growth as given further.

Using Equations (5) and (6), national income per effective consumer [Y(t)/N(t)] can be expressed as a product of income per effective producer [Y(t)/L(t)] or labour productivity and ratio of effective number of producers or workers to effective number of consumers of goods and services [L(t)/N(t)].

Y (t) / N (t) = {Y (t) / L (t)} {L (t) / N (t)}

(7)

Taking natural log on both sides of Equation (7) and differentiating with respect to time, growth rate (g) of income per effective consumer or economic growth is equal to the sum of growth rate of labour productivity and growth rate of ratio of effective number of producers to effective number of consumers.

g [Y (t) / N (t)] = g [Y (t) / L (t)] + g [L (t) / N (t)]

(8)

In technical terms, L(t)/N(t) is called the ESR. Age structure transition leads to large shifts in the ESR and interacts with the labour productivity to determine the economic growth. The period during which the growth of ESR leads to increase in the economic growth (or growth of national income per effective consumer) is called FDD. In other words, given the growth rate of labour productivity (i.e., g[Y(t)/L(t)] = g[Y(t₀)/L(t₀)], for all t), the FDD is rate of growth of ESR, which rises or falls, subject to the age compositional transformation in the process of demographic transition.

Public education expenditure affects the economic growth in Equation (8) by two ways. First, as a consumption expenditure, it affects through φ(a) or N(t). Other things being equal including the growth rate of labour productivity, an increase in public education consumption may result in reduction in economic growth rate through reduced growth rate of ESR or/and shortened duration of FDD. Second, as a non-consumption (or productive) expenditure or human capital investment, it affects through γ(a) or L(t) and g[Y(t)/L(t)]. The analysis here aims at capturing these two production effects of public education expenditure on economic growth in Equation (8).¹⁰

In fact, India’s public education consumption expenditure is officially measured in three different ways. First, Indian Public Finance Statistics (IPFS) by the Union Ministry of Finance gives the combined revenue expenditure of the Union and State governments on education. Second, National Accounts Statistics (NAS) by the Central Statistical Office in the Union Ministry of Statistics and Plan Implementation measures the consumption expenditure on education services under the Government Final Consumption Expenditure or current expenditure as per the Economic and Purpose Classification of Expenditure (EPCE) of Administrative Departments. Third, Analysis of Budgeted Expenditure on Education (ABEE) by the Union Ministry of Human Resource Development gives the combined revenue expenditure of the Union and State governments on education by Education and Other Departments. In general, the combined revenue expenditure on education in the IPFS is higher than the Government Final Consumption Expenditure on education services in NAS but less than the current expenditure in the EPCE and total revenue expenditure of the Education and Other Departments on education in the ABEE is approximately equal to the current expenditure in the EPCE. The current expenditure on education in EPCE includes consumption expenditure (same as Government Final Consumption Expenditure), subsidies, current transfers to local bodies and other current transfers. Government Final Consumption Expenditure on education is the total expenditure on compensation of employees and net purchase of commodities and services. Thus, the major difference between the combined revenue expenditure on education in ABEE and Government Final Consumption Expenditure on education in NAS is attributable to current transfers.

Union Budget 2011–2012 (Government of India, 2012) introduced the Effective Revenue Deficit as a new concept of budget deficit. It is measured by subtracting those revenue expenditures for creation of capital assets or investment from the conventional revenue deficit (i.e., total revenue receipts minus total revenue expenditure). Using this approach, the public education expenditure (on revenue account) is separable between consumption and non-consumption expenditure. Public education consumption expenditure is equal to Government Final Consumption Expenditure on education services in India’s NAS and is the same aggregate expenditure in NTA in Equation (1). Public non-consumption education expenditure (on revenue account) is equal to total revenue expenditure less public education consumption expenditure.¹¹ In the absence of this separation, however, the impact of revenue (or government consumption) expenditure on education may be underestimated on economic growth in Equation (8).

A change in non-consumption revenue expenditure on education (or, in brief, non-consumption public education expenditure) is a form of human capital investment within the government consumption expenditure and may contribute to production through changes in L(t) and [(Y(t)/L(t)]. These production effects are measured by modifying Equation (8) as follows.

g [Y (t) / N (t)] ​ * = g {Y (t) / L (t)} ​ * + g [{L (t) + ω (t)} / N (t)]

(9)

where ω(t) = [∑E(x,t₀)P(a,t) – ∑c(x,t₀)P(x,t)] and g{Y(t)/L(t)}* = g[Y(t₀)/L(t₀)] + h’(t); where ω(t) is total number of effective producers from a marginal increase in non-consumption public education expenditure, E(x, t₀) is the per capita age profile of public education expenditure and c(x) is per capita age profile of public education consumption; g[Y(t₀)/L(t₀)] is the growth rate of labour productivity in the benchmark year and h’(t) is changes in human capital. h’(t) is measured by summing over the age in the cross-sectional per capita profile of ω(t) from age 6 to 24 and expressed by the years of average labour income (YoYLs).¹² Average labour income is calculated by summing over age in the cross sectional profile of per capita labour income from age 30 to 49 years and dividing the resultant value by 20. For the benchmark year 2004–2005, years of average labour income per adult for those from age 30 to 49 years is equal to ₹30,721 and the human capital investment in raising a child is ₹44,122 or 1.44 YoLYs. Thus, the growth rate of national income per effective consumer in Equation (9) is decomposed into modified growth rate of productivity and ESR.¹³

Equation (9) captures the growth effects of public education expenditure from the production side through productivity growth rate and growth rate of effective number of producers. Further, it captures the consumption and non-consumption (or a form of human capital) expenditure effects of public education on economic growth. If ω(t) ≠ 0, growth effects are different between Equations (8) and (9). This difference is attributable for growth effect of marginal increase in public education spending through growth rate of productivity and effective number of producers. It should be emphasized that a higher and positive value of ω(t) raises the effective number of producers and growth rate of ESR. At the same time, given the growth rate of labour productivity at the benchmark year, a higher h’(t) leads to a higher growth of modified labour productivity. However, growth rate of modified labour productivity increases as h’(t) rises but does not fall below g(Y(t₀)/L(t₀). This ensures that productivity gains of non-consumption education expenditure (or a form of human capital investment) are positive.

Equation (9) integrates the age structure transition, public education spending through human capital formation and economic growth in the framework of the FDD. Further, the framework integrates the public education spending forecast by using Equations (2–4) in Equation (9).¹⁴

Fanelli (2015b) shows an interesting relationship between the FDD and lifecycle deficit (LD). This relationship can be explained as follows. First, rearranging terms in Equation (1), (C_f,i⁺ C_g,i) – Y_L,i = (Y_A,i – S_i ) + (T_f,i⁺ – T_f,i^–) + (T_g,i⁺ – T_g,i^–). The left-hand side of this equation is defined as LD (or surplus) if positive (or negative). In economic terms, LD is a measure of total demand for age reallocations or a measure of total value of goods and services consumed by members of an age group less the value of goods and services produced by an age group. Let C_t = ∑_i (C_f,i,t + C_g,i,t), be the aggregate consumption in year t. In the same way, let LD_t, Y_Lt and ESR_t be, respectively, the aggregate LD, labour income and ESR in year t. Accordingly, the direct relationship between the evolution of the ESR, propensity to consume relative to labour income and the aggregate LD is defined by (LD_t/Y_Lt) = [C_t/Y_Lt(1–ESR_t)] and the benefits of FDD on economic growth fade when consumption and, therefore the LD, is higher relative to labour income. This FDD approach to the determinants of growth effects is different from the approach in Equation (9), although both the approaches have a common premise of Equation (8).

Variables, Measurement and Data Sources

Throughout, public expenditure on education refers to the budgetary expenditure of the Education and other departments of the Union and State governments for the provisioning of educational services of elementary, secondary, higher education, training and others on the revenue account. The analysis excludes the private (or out of pocket) education expenditure incurred by students and/or parents for access and utilization of public educational services because such private expenditure cannot be included in public education spending within the government budget.

Aggregate controls and age profiles of labour income [L(x,t₀)] and public and private consumption by education, health and others are calculated for the benchmark year 2004–2005. Aggregate controls or macro economic variables/controls are drawn from the NAS with adjustments for net indirect taxes. Aggregate controls are essential to scale age profiles and to ensure consistency of Flow Account Identify in Equation (1). Age profile refers to the age shape or pattern after it is adjusted to be consistent with the macro control. Age pattern or shape refers to per capita values of a macro control by age before the adjustment to match therelevant macro control (United Nations, 2013a).

Aggregate consumption [∑(C_f,i + C_g,i) = ∑φ(x)P(a,t)] combines the public and private consumption. Age profile of aggregate consumption is essential for calculations of the FDD from Equation (5) through Equation (9). Age profiles of aggregate and per capita labour income and public education consumption are essential for forecasting from Equation (2) through Equation (4). Aggregate controls, age allocation rules and data sources for measurement of variables for calculation of the age profiles are given in Table 1. Throughout, the UN projected population under the medium fertility variant (United Nations, 2013b) is the basis for all calculations and forecasting.

Table 1.

Aggregate Controls, Age Allocation Rules and Data Sources for Calculation of Age Profiles, India: 2004–2005

Variable	Aggregate Control	Age Allocation Rules and Data Sources
1. Labour income	Compensation of employees + (2/3) of mixed income + net compensation of employees from rest-of-world	Age profile is calculated based on individual income from wages and salaries and household income from self-employment (i.e., farm income and non-farm business income) in the India Human Development Survey 2005 (Desai & Vanneman, 2017) or, in brief, IHDS 2005. Age profile of self-employment income at household level is calculated by allocating self-employment income of household to individuals in a household who reported as self-employed, using the age profile of mean earnings of employees. Accordingly, self-employment income accruing to ith individual in household j[YLS_ij(ϰ)] is equal to YLS_j.ν(ϰ) and ν(ϰ) = w(ϰ).SE_j(ϰ)]/Σ w(a).SE_j(a), where x is the age of ith household; SE_j(a) is the number of people in household j who are self-employed or unpaid workers of age a; w(a) is average earnings of employees. Thus, ν(ϰ) is the share of total household self-employment labour income allocated to each self-employed who is at age x. Summing across all households, total self-employment labour income is computed at age ϰ.
2. Public consumption	Government Final Consumption Expenditure (GFCE)
2.1. Public education consumption	GFCE on education services	Age profile is calculated by public formal and informal education. Public formal education age profile is based on computed per student public education consumption by levels of education. This computation is based on the following enrolment rates and public expenditure by level of education. First, using estimated attendance data from the 61st Round of National Sample Survey Organization (July 2004 to June 2005) on Status of Education and Vocational Training in India 2004–2005, share of attendance in public institutions by levels of education is obtained. This share is applied for total enrolment data in the Government of India’s Education Statistics 2004–2005 to obtain attendance in public institutions (i.e., government, aided and local body institutions). Second, using the Indian Public Finance Statistics 2006–2007, combined revenue expenditure on education by all levels of governments (including non-education departments) is obtained. Public education consumption is presumed to be proportional to revenue expenditure by levels of education. Per student public education consumption is obtained by using the computed enrolment data in public institutions. Public informal education and training consumption is equal to expenditure on adult education, language development and training. This is allocated on per capita basis for age group 25–59 years. Per capita public expenditure by levels of education is calculated by up scaling per student expenditure (on enrolment basis) to aggregate control for public education consumption and dividing it by total population in relevant age groups of education. Accordingly, three per capita values by levels of education are calculated: (a) Public education expenditure per student; (b) Per capita public education consumption and (c) Per capita public education expenditure.
2.2. Public health consumption	GFCE on health and other services	Age profile is calculated by using the individual level data on utilization of public health facilities in the 60th Round of National Sample Survey on Health Care, Morbidity and Conditions of Aged in India in 2004. Public health facilities refer to health services provided by public hospitals and dispensaries (including Primary Health Centres, Sub-centres and Community Health Centres). Utilization is proxied by expenditure incurred on treatment for hospitalized or in-patient (during 365 days prior to the survey), non-hospitalized or out-patient (during 15 days prior to the survey) and other expenditure (e.g., transport expenses to and from the hospital visits).
2.3. Public consumption other	GFCE on non-education and non-health services under GFCE	Public consumption other includes general public services; defence; housing and other community amenities; cultural, recreational and religious services; and economic services (e.g., agriculture, mining, transport and communication). This variable is allocated on per capita basis.
3. Private consumption	Private Final Consumption Expenditure (PFCE)
3.1. Private education consumption	PFCE on education net of indirect taxes.	Age profile is calculated by using the individual level data on private education expenditure in the IHDS 2005. Private education expenditure refers to expenditure incurred by currently enrolled students in elementary, secondary and tertiary levels of education on school/college fees, books uniform and other materials, transportation and private tuition.
3.2. Private health consumption	PFCE on medical care and health services net of indirect taxes.	Age profile is calculated by using the individual level data on private health expenditure in the IHDS 2005. Private health expenditure refers to sum of expenditure incurred for in-patient as well as out-patient treatment services for short-term morbidity during last one month and major morbidity during 12 months. Treatment expenses included hospital surgery, medicine and tests and others (e.g., tips, bus/train/taxi fares or lodging while getting treatment.
3.3. Private consumption other	PFCE on non-education and non-medical care and health services.	Private consumption other includes food and beverages, clothing and footwear; fuel and power; furniture, furnishing, appliances; transport and communication; and recreation and cultural services. Age profile is calculated by the Equivalence Scale and using household data on private consumption other in the IHDS 2005. The scale is equal to 1 for adults aged twenty or older, declines linearly from age 20 to 0.4 at age 4, and is constant at 0.4 for those age 4 or younger. That is, λ(a) = (1 – 0.6), (a ≤ 4); λ(a) = 1 – [0.6.(20 – a)/16], (4<a<20); and λ(a) = 1, (a ≥20). Given these scales, intra-household allocation of private consumption other is equal to: CFX_ij = [CFX_j.λ(x)/Σλ(a).M(a)], where ϰ is the age of ith household member and M(a) is the number of household members at age-a.

Source: Author’s own.

Notes: (a) All aggregate controls are derived and measured by the data in Government of India (2015). (b) Except for public education and public health, age allocation rule for all other aggregate controls follows the NTA general methodology (United Nations, 2013b). (c) Indirect taxes on private consumption are assumed equal to their share in PFCE. (d) In general, eight steps are followed in the calculation of each age profile: (i) Obtain the aggregate controls or macroeconomic variables on labour income and consumption from the NAS. (ii) Select nationally representative surveys or national administrative records for age allocation of labour income, and components of public and private consumption by education, health and others. (iii) Assign labour income and consumption values to each individual by applying the NTA allocation rules. (iv) Tabulate the data using survey weights to obtain per capita age profiles of income and components of consumption. (v) Smoothen per capita profiles to eliminate noise. (vi) Multiply per capita profiles by final population estimates to obtain aggregate age profiles for income and consumption. (vii) Adjust the aggregate age profiles and unsmoothed and smoothed per capita profiles proportionately to match the aggregate control. (viii) Determine each component of public and private consumption, and then aggregate to obtain total consumption, by age.

GDP is measured at current market prices and sourced from the National Account Statistics 2015 (Government of India, 2015) for the benchmark year 2004–2005. Inflation rate is equal to official inflation rate at 5 per cent. Real labour productivity growth rate is measured by per cent change in gross value added per worker over the period 1999–1900 to 2004–2005. Basic data for this measurement is taken from Government of India (2008).

Age profile of per capita public education expenditure is shown in Figure 3. The per capita expenditure varies from ₹501 for elementary education to ₹1,397 for secondary education, ₹4,932 for higher education and ₹11 for informal education. These per capita figures are different from per student public expenditure on enrolment basis (or per capita education consumption in India’s NIPA: ₹2,792 (or ₹261) for elementary education, ₹7,785 (or ₹727) for secondary education, ₹27,492 (or ₹ 2,566) for higher education and ₹60 (or ₹6) for informal education.¹⁵ This implies that changes in public spending on education can be sensitive to measurement of per capita public education spending. Throughout, per capita public education consumption and public education expenditure are used for their consistency with the NTA in Equation (1) and the forecasting model in Equation (2), respectively.

Empirical Results

The results are presented by public education expenditure forecast, intra-sectoral allocation of public education expenditure and growth effects of age structure transition through public education expenditure and human capital investments.

Age Structure Transition Effects of Public Education Expenditure

Other things being equal, public expenditure on education declines over the period 2005–2050 as a consequence of age structure transition. Figure 4 shows this decline as a percentage of GDP from 2.98 per cent in 2005 to 2.27 per cent in 2025 and 1.76 per cent in 2050. This decline is the result of three changes: (a) age-specific level of public expenditure on education, (b) age structure transition and (c) interaction between (a) and (b). Figure 5 shows the decomposition of change in aggregate education expenditure as a percentage of GDP by these sources. The age structure transition effect [(c)–(a)] is negative and biggest as it shows the largest decline in public spending on education as a percentage of GDP. Although changes in age-specific spending levels have a positive effect due to annual rate of inflation and productivity growth rate, the interaction effect is negative due to a stronger age transition effect than age-specific spending level effect. In the absence of age structure transition, however, the share of aggregate public education expenditure in GDP is constant because the education profile is scaled up by the nominal growth rate of productivity, but so too is labour income and, by assumption, GDP. Consequently, if age structure does not change, aggregate education expenditure as a share of GDP remains a constant throughout. These results offer new evidence for the negative effects of age structure transition on India’s public education expenditure over the period up to 2050.

The negative impact of declining school enrolment has already begun by number of schools facing closure and merger.¹⁶ For instance, as per the District Information System for Education (DISE), number of government schools with no students (or with less than 20 students) in Karnataka State was 675 (or 8,903) in 2013–2014 and 535 (or 9,503) in 2014–2015. In the same way, number of private (i.e., aided and unaided) schools with no students (or with less than 20 students) in the state was 146 (or 680) in 2013–2014 and 153 (or 861) in 2014–2015. More recent data shows that the number of government, private aided and private unaided schools with no students in Karnataka, respectively, increased to 684, 40 and 474 in 2015–2016.¹⁷ Teachers from government and aided schools with zero students have been deputed to other schools to teach. However, there is no reporting of newer utilization of physical infrastructure of closed schools. Most recently, the Government of Odisha state has taken a decision to close down as many as 165 primary schools with less than or just five students by October 2015.¹⁸

Intra-sectoral Reallocation of Public Education Expenditure

The benchmark results in Figure 4 assumed that intra-sectoral allocation of public expenditure on education remained the same as in year 2004–2005 and grow at the same rate by all levels of education. Sensitivity of these results to the assumptions is examined by considering a counter-factual case in which public expenditure on elementary education would grow at the rate of inflation and that of secondary and higher education would grow at the rate of nominal labour productivity. Interestingly, this pattern of resource allocation within the education sector results in a larger expenditure share to higher education than to the elementary and secondary levels for the following reasons. First, given the inflation rate and fixed expenditure profile, public expenditure on the elementary education (6–13) declines over time due to age structure transition effect. Second, public expenditure on secondary education (14–17) declines over time but remains higher than elementary education mainly due to higher per student expenditure than the elementary education and higher growth of expenditure linked to growth rate of nominal labour productivity. Third, public expenditure on higher education (18–24) increases over time due to higher cost and growth rate of nominal labour productivity. Consequently, as shown in Figure 6, the ratio of expenditure on the elementary and secondary education to higher education within the total public education expenditure declines over the entire period.

Further, the earlier pattern of intra-sectoral resource allocation may result in remarkable availability of extra budgetary resources or potential savings. This result is shown in Figure 7. The magnitude of the potential savings in total public education expenditure (relative to 2005 level) is 0.83 percentage points in 2020, 4.92 percentage points in 2030 and 6.79 percentage points in 2050. Thus, more resources are expected to be available within the education sector for quality improvement and higher human capital investments (including for skill development and training).¹⁹ In addition, resources may be reallocated to early child education, nutritional programmes of school children, access to technology for teaching learning, and in lagging states with higher fertility to ensure greater quality of education development there.

The Human Capital Report 2015 (World Economic Forum, 2015) has distinguished the HCI ranking of countries by age groups of population. India’s overall ranking in the HCI for all age groups is 100 out of 124 countries. This ranking is lower as compared to the HCI ranking for under 15 for age group (67th rank) and 15–24 age group (98th rank) but higher as compared to 25–54 age group (109th rank), 55–64 age group (115th rank) and 65 and over age group (114th rank). This underlines a need for improvement in overall human capital in general and for working and early elderly age groups in particular. In this context, the results of this article in Figure 6 imply such improvements in human capital investment by age groups. For instance, skill formation and training are particularly relevant for improvement of human capital for working age and early elderly age groups.²⁰

Growth Effects

The growth effects are distinguished by the baseline and three alternative scenarios based on Equation (9).

Baseline scenario: Uses the data in the benchmark year 2004–2005 without inclusion of non-consumption public education expenditure and calculates growth effects by using Equation (8).

Scenario 1: Extends the Baseline scenario with inclusion of public non-consumption education expenditure, and growth effects are calculated by using Equation (9).

Scenario 2: Extends the Baseline scenario with inclusion of public non-consumption education expenditure; assumes that public education expenditure grows at nominal growth rate of productivity by using the total education expenditure forecast from Equation (2), and calculates growth effects by using Equation (9).

Scenario 3: Extends the Baseline scenario with inclusion of public non-consumption education expenditure; assumes that public education expenditure for secondary and higher education grows at nominal growth rate of productivity and that of elementary education grows at annual rate of inflation by using the total education expenditure forecast from Equation (2), and calculates growth effects by using Equation (9).

The results are shown by the growth rates of ESR and economic growth in Figure 8 and Figure 9, respectively. Figure 8 shows no remarkable difference in growth rate of ESR between the Baseline and Scenario 1. This difference is about 0.5 per cent or less up to 2040 and declines thereafter due to population ageing. In contrast, the growth rate of ESR in Scenario 2 and Scenario 3 is higher and rises from 1.5 per cent in 2006 to about 5.5 per cent in 2050. Interestingly, the growth rate of ESR in Scenario 2 scenario is higher than in Scenario 3. This result offers a new evidence for a long run positive impact of higher spending on secondary and tertiary education to enhance the growth rate of ESR or to offset the negative growth effects of population ageing and extending the duration of demographic dividend throughout the period up to 2050 for India.

The growth rate of ESR does result in similar pattern of economic growth in all the scenarios due to positive growth rate of labour productivity over time. This is evident in Figure 9 by the growth effects in the Baseline and Scenario 1. For instance, the growth effect in the Scenario 1 is higher than the baseline throughout. This implies that the current public education expenditure pattern on human capital investment is conducive for achieving a higher economic growth. On the other hand, growth effects are highest in the Scenario 4 but its trends are comparable with the Scenario 3. Growth effects in both Scenario 3 and Scenario 4 are higher than in the Baseline or Scenario 1. The main explanation for this positive, higher and longer growth effects (or demographic dividend) is higher public education spending on secondary and higher education which results in higher growth rate of labour productivity and ESR. Or, negative population ageing effects on growth can be delayed if more public education expenditure is reallocated on the secondary and higher education to attain a higher economic growth.

Conclusions and Implications

This article provides an accounting framework for economic linkages between the age structure transition, public education spending and economic growth for India based on the public expenditure forecasting model and NTA. Using these frameworks, age profile of public education expenditure is calculated, public education expenditure by levels of education is forecast up to 2050 and growth effects of public education expenditure through human capital investment is calculated up to 2050. Major conclusions and implications of these analyses are as follows.

Age structure transition impacts public education expenditure because of a long-term decline in school-going population and EDRs which reduces the public education expenditure especially on pre-secondary education. The decomposition analysis shows that age structure transition is a major source for the decline as compared to changes in per capita spending levels and interaction effects between age structure transition and changes in spending levels. Thus, age structure effects are important for design of long-term public education expenditure policies on size and pattern of spending by levels of education and their contributions to the potential savings.

Other things being equal, a reduction in public expenditure on pre- secondary education, as a consequence of age structure transition, may be a new way of financing of secondary and/or higher education through changes in intra-sectoral allocation of resources. However, reallocation of resource for quality improvements in early child education may be crucial to enhance the preparedness of children for secondary and higher education.

This article has separated the government revenue expenditure on education between consumption and non-consumption expenditure. Non-consumption expenditure is considered a form of human capital investment with production implications. Using the NTA framework, growth effects of public education spending operate through the growth rate of ESR and labour productivity. The results offer evidence for unintended positive growth effects of the current public education expenditure through human capital formation. However, growth effects can be positive, higher and longer through the demographic dividend, if more public education spending on human capital formation is allocated to secondary and higher education in India and the growth of those spending is linked to the growth rate of nominal productivity. This approach broadens the scope of determinants of economic growth and useful to identify key policy determinants to promote growth through higher and longer demographic dividend as they are related to public education expenditure policies and programmes. Further, the results offer empirical support for India’s policymakers’ efforts to provide more resource allocation for public education spending in general, and secondary and higher education in particular, with greater share of revenue expenditure for non-consumption items. Thus, both size and patterns of public spending policies on education by levels of education matter for India’s long-term growth due to age structure transition effects. The above approach, results and implications add to the existing knowledge in the FDD studies on India, such as, Narayana (2015).

The conclusions of this article must be qualified due to data limitations in calculation of age profiles and assumptions on expenditure forecast and construction of NTA. Age profiles in this article refer to 2004–2005 and are consistent with the macroeconomic equilibrium under the NTA-Flow Account Identity. They are fixed throughout the long forecasting and growth effect analyses up to 2050. The profiles need to be revised as parts of complete and new construction of NTA Flow Account in future. Analyses based on such revised age profiles and changes in parameters in forecasting (e.g., growth rate of productivity and inflation rate) may offer newer insights into the age structure transition effects on public education spending.

India’s demography is characterized by remarkable inter-state differences in size and growth of school- and college-going population. States in India have considerable autonomy in fiscal policy including spending on education. A comprehensive study is needed in future to analyse the impact of age structure transition on the State level public expenditure on education and its growth effects. Further, subject to the comparability of socio-economic structure and age structure patterns, the methodology and policy analyses of India in this article may have relevance and applicability for other developing countries in South Asia for analyses of current and future public expenditure policies on education as they are related to age structure and growth effects.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

Footnotes

Acknowledgements

Early versions of this article were presented for the (a) UNFPA–EWC–IDRC sponsored Regional Meeting on National Transfer Accounts in Asia (3–4 December 2015, Bangkok, Thailand), (b) Annual International Conference of Comparative Education Society of India at Azim Premji University (14–16 December 2015, Bengaluru, India), (c) Global Conference on Prosperity, Equality and Sustainability: Perspectives and Policies for a Better World (1–3 June 2016, New Delhi, India) and (d) Demographic Dividend and African Development: 11th Global Meeting of the NTA Network (20–24 June 2016, Dakar and Saly, Senegal). Grateful thanks are due to Professors Ronald Lee, Andrew Mason and Sang-Hyop Lee and Dr Tim Miller for technical guidance on the National Transfer Accounts methodology and education forecasting model used in this article; Dr Nicholas McTurk, Professor Saumen Chattopadhyay, Professor Geeta Kingdon, Professor Alok Kumar and other participants in the aforementioned conferences for useful suggestions; Institute for Social and Economic Change (Bengaluru) for research support up to July 2017; Professors N. V. Varghese, Mona Khere and Jinusha Panigrahi at National Institute of Educational Planning and Administration (New Delhi) for helpful discussions; and an anonymous reviewer and the Editors of this SMP for constructive comments. However, the usual disclaimers apply.

Notes

References

Agarwal

(2009). Indian higher education: Envisioning the future. New Delhi: SAGE Publications.

Aiyar

, & Mody

(2011). The demographic dividend: Evidence from Indian states (IMF Working Paper No. WP/11/38). Washington, DC: International Monetary Fund.

Bloom

D. E.

, Canning

, Fink

, & Finlay

J. E.

(2007). Does age structure forecast economic growth? International Journal of Forecasting, 23(4), 569–585.

Bloom

D. E.

, Canning

, Hu

, Liu

, Mahal

, & Yip

(2010). The contribution of population health and demographic change to economic growth in China and India. Journal of Comparative Economics, 38(1), 17–33.

Choudhry

M. T.

, & Elhorst

J. P.

(2010). Demographic transition and economic growth in China, India and Pakistan. Economic Systems, 34(3), 218–236.

Das

P. K.

, & Kar

(2015). A study of demographic and financial changes in India. In Fanelli

J. M.

(Ed.), Asymmetric demography and the global economy: Growth opportunities and macroeconomic challenges in an ageing world (pp. 213–242). New York, NY: Palgrave McMillan.

Das

P. K.

, & Kar

(2016). Public expenditure, demography and growth: Theory and empirical evidence from India (IZA Discussion Paper No. 9721). Bonn: Institute for the Study of Labor.

Desai

, & Vanneman

(2017). India Human Development Survey 2005. New Delhi and Ann Arbor, MI: National Council of Applied Economic Research and Inter-University Consortium for Political and Social Research. Retrieved from https://www.icpsr.umich.edu/icpsrweb/DSDR/studies/22626

Ernst & Young (2013). Higher education in India: Vision 2030. Kolkata: Author.

10.

Fanelli

J. M.

(2015a). Asymmetric demography and the global economy: Growth opportunities and macroeconomic challenges in an ageing world. New York, NY: Palgrave Macmillan.

11.

Fanelli

J. M.

(2015b). Demographics and macroeconomics: Opportunities and risks in dividend-era Argentina. In Gragnolati

, Rofman

, Apella

, & Troiano

(Eds.), As time goes by in Argentina: Economic opportunities and challenges of the demographic transition (pp. 302–338). Washington, DC: The World Bank.

12.

Government of India (2008). Report on definition and statistical issues relating to informal economy. New Delhi: National Commission for Enterprises in the Unorganized Sector.

13.

Government of India (2012). Budget at a glance: A paper of the Union Budget 2011–12. New Delhi: Ministry of Finance.

14.

Government of India (2014). Statistics of school education 2011–12. New Delhi: Bureau of Planning, Monitoring and Statistics, Ministry of Human Resource Development.

15.

Government of India (2015). National accounts statistics 2015. New Delhi: Ministry of Statistics and Programme Implementation.

16.

Government of India (2016). Education statistics at a glance. New Delhi: Bureau of Planning, Monitoring and Statistics, Ministry of Human Resource Development.

17.

Gragnolati

, & Troiano

(2015). Public finance implications of population aging in Argentina: 2010, 2050, 2100. In Gragnolati

, Rofman

, Apella

, & Troiano

(Eds.), As time goes by in Argentina: Economic opportunities and challenges of the demographic transition (pp. 93–112). Washington, DC: The World Bank.

18.

Grob

, & Wolter

S. C.

(2005). Demographic change and public education spending: A conflict between young and old? Education Economics, 15(3), 277–292.

19.

Hanushek

E. A.

(2013). Economic growth in developing countries: The role of human capital. Economics of Education Review, 37, 204–212.

20.

Lee

, & Mason

(2011a). Population aging and the generational economy. Cheltenham: Edward Elgar.

21.

Lee

, & Mason

(2011b). Introducing age into national accounts. In Lee

& Mason

(Eds.), Population aging and the generational economy (pp. 55–78). Cheltenham: Edward Elgar.

22.

Lee

S. H.

, & Mason

(2014). Are current tax and spending regimes sustainable in developing Asia? ADB Economics (Working Paper Series No. 414). Manila: Asian Development Bank.

23.

Mankiw

G. N.

, Romer

, & Weil

D. N.

(1993). A contribution to the empirics of economic growth. Quarterly Journal of Economics, 107(2), 407–437.

24.

Marchionni

, & Alejo

(2015). Funding basic education throughout the demographic transition in Argentina. In Gragnolati

, Rofman

, Apella

, & Troiano

(Eds.), As time goes by in Argentina: Economic opportunities and challenges of the demographic transition (pp. 203–234). Washington, DC: The World Bank.

25.

Mason

, Lee

, Abrigo

, & Lee

S. H.

(2017). Support ratios and demographic dividends: Estimates for the world (Technical Paper No. 2017/1). New York, NY: Population Division, Department of Economic and Social Affairs, United Nations.

26.

Mason

, Lee

, & Jiang

J. X.

(2016). Demographic dividends, human capital, and saving. The Journal of the Economics of Ageing, 7(2), 106–122.

27.

Mehta

A. C.

(2013). Elementary education in India 2012–13: Analytical tables 2012–13. New Delhi: National Institute of Educational Planning and Administration and Department of School Education and Literacy, Ministry of Human Resource Development, Government of India.

28.

Miller

(2006). Demographic models for projections of social sector demand, Latin American and Caribbean. Santiago: Demographic Centre (CELADE), Population Division.

29.

Miller

, & Castanheira

H. C.

(2013). The fiscal impact of population aging in Brazil: 2005–2050. Revista Brasileira de Estudosde População [Brazilian Journal of Population Studies], 30(Supplement), S5–S23.

30.

Mincer

(1981). Human capital and economic growth (NBER Working Paper No. 803). Cambridge, MA: National Bureau of Economic Research.

31.

Narayana

M. R.

(2006). Privitisation policies and post-privitisation control devices in India’s higher education: Evidence from a regional study and implications for developing countries. Journal of Studies in International Education, 10(1), 46–70.

32.

Narayana

M. R.

(2008). Education, human development, and quality of life: Measurement issues and implications for India. Social Indicators Research, 90(2), 279–293.

33.

Narayana

M. R.

(2015). India’s age structure transition, sectoral labor productivity and economic growth: Evidence and implications based on National Transfer Accounts. Population Research and Policy Review, 34(3), 381–415.

34.

NSSO (2015). Key indicators of social consumption in India: Education, NSS 71th Round (January–June 2014). New Delhi: Ministry of Statistics, and Programme Implementation, Government of India.

35.

Pelinescu

(2015). The impact of human capital on economic growth. Procedia Economics and Finance, 22, 184–190.

36.

Poterba

J. M.

(1997). Demographic structure and the political economy of public education. Journal of Policy Analysis and Management, 16(1), 48–66.

37.

Poterba

J. M.

(1998). Demographic change, intergenerational linkages, and public education. The American Economic Review, 88(2), 315–320.

38.

Prakash

(2007). Trends in growth and financing of higher education in India. Economic and Political Weekly, 42(37), 3249–3258.

39.

Rani

(2014). Education loans and financing higher education in India: Addressing equity. Higher Education for the Future, 1(2), 183–210.

40.

Singh

A. K.

(2016). India’s demographic dividend: A skeptical note. Indian Journal of Human Development, 10(1), 10–26.

41.

Tilak

J. B. G.

(1993). Financing higher education in India: principles, practice, and policy issues. Higher Education, 26(1), 43–67.

42.

Tilak

J. B. G.

, & Varghese

N. V.

(1991). Financing higher education in India. Higher Education, 21(1), 83–101.

43.

UNDP (2015). Human development report 2015. New York, NY: Author.

44.

United Nations (2013a). National transfer accounts manual: Measuring and analysing the generational economy (Working Paper No. ESA/P.WP/226). New York, NY: Population Division, Department of Economic and Social Affairs.

45.

United Nations (2013b). World population prospects: The 2012 revision. New York, NY: Population Division, Department of Economic and Social Affairs.

46.

Vincent-Lancrin

(2008). What is the impact of demography on higher education systems? A forward-looking approach for OECD countries. In Higher education to 2030. Vol. I: Demography (pp. 41–103). Paris: OECD.

47.

World Economic Forum (2013). Human Capital Report 2013. Geneva: Author.

48.

World Economic Forum (2015). Human Capital Report 2015. Geneva: Author.