Estimating the multiplier effect of social sector expenditure in Karnataka

Choices. Job creation through infrastructure investment in the Middle East and North Africa. Input-output multipliers, general purpose technologies, and economic development. Un[r]

Estimating the multiplier effect of social sector expenditure in Karnataka

Authors: Jyotsna Jha, Archana Purohit, Apurva K.H., Achala S Yareseeme, Centre for Budget and

Policy Studies, Bengaluru, June 2019

Abstract:

Social and economic change can be envisioned through public expenditure On one hand, the national and international commitment such as SDGs, CRC and the entire Rights approach calls for an increased and well-directed domestic public expenditure on social sector including health, early childhood care, education and empowerment, a major focus on fiscal management generally tends to view such expenditures as „consumption‟ and therefore not as desirable as „investments‟ on infrastructure or as crucial as defence While acknowledging the importance of prudent fiscal management, we argue that social sector expenditure is as important for both growth and human development and well-being This calls for an integration of social and economic policies to have a lasting and equitable impact on economic growth Looking at these recurrent expenditures (on education, health, early childhood expenditure etc.,) as investment rather than as mere consumption goods made us use the lens of „multiplier effects‟ to analyse the extent of income generation I-O method has largely been used to study industry specific multipliers, but have hardly been employed in studying the impact of social sector expenditure Hence, the objective of our study is to estimate the multiplier effect of social sector expenditure (mainly education & health) using Input-Output method

as a test case for the state of Karnataka

Introduction:

The element of „time‟ as a categorical variable in an economic model for theoretical analysis is quintessential as it represents concrete reality This means time not only matters in economic analysis but also has to recognise the irreversibility character of time (Historical Time) A logical set of relations that links variables in a unique direction implying causation and reaching equilibrium corresponds to no real position in time (Termini 1981) The marginal revolution of 1870‟s led by Jevons, Walrus and Alfred Marshal transformed the discipline of economics by analysing the economic processes of production, exchange, distribution and accumulation (growth) in linear time devoid of any chronology using the supply and demand framework But the Great Depression of 1929 brought a significant change in the discourse of doing economics Keynes spearheaded this revolution

in economic thinking through his concept of effective demand that eventually led to the formation of Macroeconomics as a separate discipline which primarily dealt with explaining, „fallacy of composition‟ Supply creates its own demand was the hallmark statement of marginal economists where the emphasis is on supply side variables But the role of effective demand in stimulating the

economy especially recessionary economy was explained through applying multiplier, a demand-based spur to growth i.e., Quantity Adjustment (Bhaduri 1986) Multiplier debunks the fallacy put forth in the form of „Pigou Effect‟ Myrdal (1957) lucidly puts it, "The notion of stable equilibrium is normally a false analogy to choose when constructing a theory to explain the changes in a social system What is wrong with the stable equilibrium assumption as applied to social reality is the very idea that a social process follows a direction – though it might move towards it in a circuitous way – towards a position which in some sense or other can be described as a state of equilibrium between forces Behind this idea is another and still more basic assumption, namely that a change will regularly call forth a reaction in the system in the form of changes which on the whole go in the opposite direction to the first change On the contrary, in the normal case there is no such a tendency towards automatic self-stabilisation in the social system The system is by itself not moving towards any sort of balance between forces, but is constantly on the move away from such a situation”

The concept of multiplier is based on the belief that expenditure creates incomes The underlying logic is that economy is an integrated system and subsequently multiplier works as a convergent process over time through rounds of expenditure and income Multipliers have been categorised into various types One categorisation classifies multipliers into type I and type II categories Type I multipliers sum together direct (change in final demand) and indirect effects (supply chain effects to meet that demand) of a change in economic activity Whereas type II multipliers also include the

induced effect (the effects of wages earned in the direct and the indirect supply chains that are used to

buy goods and services in the economy) Additionally, multipliers are also categorised into output, employment, income and value-added multipliers In this study, we propose to estimate the multiplier effect of government‟s expenditure in social sector on the economic growth in the context of Karnataka This study aims to (i) estimate and understand the multiplier effect of public spending in social sector, and (ii) growth of tertiary sector and its implications for the sustained economic growth

in Karnataka Subsequent part of the paper looks at the literature review that has explored the relationship between social sector expenditure and economic growth, irrespective of the methodology they have adopted Latter part of the review looks specifically at the studies estimating multiplier effects and approaches taken in these studies to estimate multipliers

Review of Literature

Social sector expenditure in the New Normal Era characterised by austerity is an anathema Fiscal consolidation is the rule rather than an exception and this lens generally tends to view social sector expenditure as consumption and therefore not as desirable as investments on infrastructure or as crucial as defence We argue that social sector expenditure is as important for both growth and human well-being Majority of the academic literature have looked at the welfare enhancing aspect of social sector expenditure by focussing on changes in social outcomes such as educational attainment and health status While we do not deny the importance of these welfare enhancing aspects of expenditure,

we emphasise investment in human capital also can have growth enhancing effects These studies have been conducted at different levels including cross-country, national and sub-national / local levels

The impact of increased public investment in education on economic growth has been widely studied Many of these studies found a positive impact on economic growth (Jung and Thorbecke (2003); Annabi et al (2011)) Some looked at the differential impact on economic growth depending on the way in which the increased public investment was financed They found that the growth and welfare enhancing effects are most pronounced when they are financed through a re-composition of public expenditure rather than when they are financed through increased taxation Similarly, many studies exploring the linkages between increased public investment in healthcare and subsequent economic growth found a positive relationship between them The positive impact found in these studies ranged from 4% increase in output due to one-year improvement in populations life expectancy (Bloom et al (2004)) to a 0.06 – 0.10% increase in per capita GDP growth rate due to a 1% increase in total health expenditure (Beraldo et al (2009))

Various studies also looked at the relationship between gender inequality and public expenditure that aimed at reducing these gender inequalities and subsequent economic growth Hakura et al (2016) found that gender inequality impeded growth in the initial stages of development resulting in large growth losses Ahang (2014) also found a negative impact of gender inequality on economic growth Cross-country panel data regressions investigating the impact of gender inequality in education and employment on long-term economic growth show a direct negative impact by lowering the average level of human capital as well as an indirect impact through investment and population growth channels (Klasen (2002), Klasen and Lamanna (2009)) Kim et al (2016) and Agénor & Canuto (2013) found that improving gender equality can contribute significantly to economic growth by changing females‟ time allocation and promoting accumulation of human capital and showed that gender policies are likely to produce tangible economic results only in the long term Amiri & Gerdtham (2013) undertook an econometric study to examine whether there are relationships between maternal and child health outcomes and economic growth in different countries at different income levels They suggest that the effect of marginal health investments on health outcomes is higher at low levels of GDP, i.e in countries where the level of health investments is generally lower

There have been studies which have looked at the impact of increased public investment in social sectors in the context of India Empirical analysis using panel data was undertaken to find that the increase in education, health and development expenditures helps reduce poverty They examine how the share of Government spending on public goods such as health, education and basic infrastructure affect per capita GDP and poverty and found a large, positive and significant impact on per capita GDP growth as well as significant reduction in poverty Interestingly, they found the reallocation of expenditures to increase public good spending could on an average increase per capita GDP growth rate by 2.7% points and reduce poverty headcount index by up to 6.6 percentage points ((Hong & Ahmed (2009) Jha at al (2001)) Bhakta (2014) studied the linkages between health status of children

and their educational achievements and showed how public expenditure on Supplementary Nutritional Program has indirect positive impact on education through the improvements in health status of children whereas additional expenditure on elementary education has positive impact on enrolment rates, but at a diminishing rate However, Ganesh Kumar et al (2017) studied the macro-economic impact of different types of public expenditure in India and found that focusing government expenditure in social sectors does not affect GDP growth Largely the studies pertaining to international and national spheres have shown that there exists either positive or negligible impact of social sector expenditure on economic growth

Impact of Public expenditure growth through the lens of Multiplier

Scholarly literature concerning the effects of public investment on varied economic variables give mixed results The approaches, methodologies, techniques and tools in these studies towards explaining the impact of investment spending is diverse Further, the research studies use multiple theoretical frameworks to derive myriad conclusions The review of few studies is discussed below Fiscal policy as a stabilisation tool are primarily analysed in the context of advanced economies and low-income countries Aggregate multiplier for the year 1997 was estimated for OECD countries using Input Output database show the domestic aggregate multiplier to be 1.61, with imports, the domestic multiplier reduces to 1.03 and it is 1.65 for intermediate sector multiplier (Jones, C 2007) Similarly, another study (ADB, A A., Furceri, D., & IMF, P T 2016) for advanced economies shows the evidence concerning multiplier effect of public investment using a dynamic stochastic general equilibrium model This study found that an unanticipated increase in government investment spending by 1% point increases the level of output by 0.4% (short term investment multiplier) in the same year whereas four years later, the medium-term fiscal multiplier turns out to be 1.5% Further, fiscal multiplier was calculated for 25 EU countries for the 1995 and 2010 using Vector Auto Regressive Method to estimate cross-national fixed effects The multiplier for total government spending was found to be 1.61 (Reeves, A et al 2013)

In the context of low-income countries, the study (Ianchovichina, E., et al 2012) analysed the multiplier effect of infrastructure investment in the Middle East and North Africa using I-O table Type II multiplier estimated varied between 1.09 and 1.82 Because the data requirements are huge in the national context to calculate multiplier effect of public investment, scholars have limited the study

to understand the impact multiplier in the local economy IMPLAN (Impact Analysis of Planning), REMI (Regional Economic Model), are tools used to understand the local multiplier The effect of opening a manufacturing facility in SEZ of Mielec shows that the overall economic multiplier was 1.55 while the local multiplier rooted in SEZ was 1.35 Supply side effects in the local economy is found to be greater than income side effects implying the need for demand driven economy to have larger multiplier effects (Domański, B., & Gwosdz, K 2010) The results are corroborated also by (Micek, G 2011) who say that indirect multiplier (Supply side effects) is greater than direct (Income Effect)

Multipliers have also been calculated at the disaggregated level of sectors in various studies Furceri

& Zdzienicka (2012) conducted an analysis for nine different social policy areas for a panel of OECD countries from 1980 to 2005 and found that social spending devoted to health and unemployment benefits are those that have greatest effects Further, the same results were found in a Maltese economy using I-O analysis where social work, education, health sector had seen positive and had large Type II multiplier (Cassar (2015)) It is interesting to see that social sectors like health had a larger multiplier effect of 4.3 over -7.58 in defence (Reeves et al., 2013) in a study covering 25 EU countries from 1995 to 2010, both before and during the recession that began in 2008 The difference across different types of spending was explained by varying degrees of absorption of government spending into the domestic economy

However, there is scholarly literature which views that the effect of multiplier is either negative or negligible The results for the EU countries (Pereira & Andraz (2015)) show that social spending increases the unemployment rate in all countries Similarly, Kraay (2012) conducted a study on 29 aid-dependant low income countries using two-stage least squares (2SLS) method to calculate contemporaneous spending multiplier and the estimates show that impact multiplier is mere 0.48 The effect of government spending in Kenya using Structural VAR for the period 1991-2012 is found to

be weak and this is linked to high debt ratio levels and high marginal propensity to import (Mahrous, 2016) The studies that have showed that there is negative or negligible effect of multiplier also hold a viewpoint that higher government spending would lead to higher debt levels thereby reducing the multiplier effect These results have a theoretical underpinning of Ricardian Equivalence However, these results are put to disbelief by other studies that hold a Keynesian theoretical framework The following section describes in detail the method and methodology followed to construct I-O table for the state of Karnataka

Methodology followed for the construction of Input-Output table for Karnataka 2013-14:

Input-output multipliers are principally underpinned by the interrelations between production sectors

A significant advantage of utilizing input-output methodology is that the resulting multipliers incorporate not only the direct effects, but also the indirect and the induced effects on the economy as

a result of an exogenous shock to one of the components of final demand Tracing the macroeconomic implications through the lens of multiplier, scholarly studies have largely employed econometric time series techniques like Vector Auto Regressive methods (Reeves et al., (2013), Jones (2007)) which reflect Walrasian general equilibrium models or use simple regressions which are Marshallian partial equilibrium in essence Our study aims to estimate the multiplier effect using the Input-Output (I-O) Method which reflects the essence of the concept of Keynesian multiplier that captures the demand creation due to public expenditure and takes into account the concern of demand deficiency in a capitalist economy I-O provides a detailed disaggregated quantitative description of the structural characteristics of all component parts of a given economic system illustrating vertical integration

The Input-Output table gives a comprehensive view of the linkages across the different sectors of the economy It represents the total output produced in a country in a particular year Hence, it becomes important to capture all the components of the gross output in the matrix format, accounting for the value added for each sector-sector combination Given below is a sample Input-Output table:

Outpu

t (GVO )

Sector / Sector Secto

r 1

Secto

r 2

Secto

r 3

Secto

r 4

Tot

Interm ediate use

PFC

E

GFC

E

GF

CF

Expo rts

Impo rts

Total Input

Gross Value

Added (GVA)

Net Taxes

Total Output

(GVO)

Total Output (GVO) = Intermediate Consumption + Gross Value Added (GVA)

The values X1 to X16 represents the intermediate consumption in each destination sector of the input coming from the originating sector For example, X10 represents the value of intermediate consumption in sector 2 of the inputs coming from sector 3 T1 to T4 gives the total intermediate use

of all inputs coming from the particular sector In order to calculate the GVA, we need to capture each

of the final demand components that include PFCE, GFCE, GFCF, Exports and Imports

Let us now discuss in detail how each of the items of the Input-Output table is calculated including the data sources used The construction of an Input-Output table is an extensive exercise and requires

us to refer to multiple data sources It usually takes 2-3 years to construct an I-O table and this exercise is carried out in India once every 5 years by the Central Statistical Organization (CSO) The table was last constructed by CSO for the year 2008 To undertake the construction of an I-O table for the state of Karnataka, we have used to I-O table constructed for India in the year 2013-14 published

by NCAER (Saluja & Singh 2016) as the base table, which is also the latest available I-O table for

India, and used different methods to fine tune the table to arrive at the I-O table for Karnataka for 2013-14 This needs to be done due to lack of data availability at state level in India

Background of I-O table:

The I-O table constructed for India in 2013-14 is a detailed table and contains 130*130 sectors (commodity*commodity table) For the purpose of our study that aims at estimating the multiplier effect of social sector expenditure in Karnataka, we have tried to construct the I-O table for Karnataka for the year 2013-14 taking 23 sectors into consideration that includes education & research and medical & health as separate sectors Since the India I-O table contains 130*130 commodity*commodity matrix, we have aggregated them to form 23*23 commodity*commodity matrix to correspond with 23 sectors considered for the construction of Karnataka I-O table These 23 sectors (Table 1: Concordance Matrix) are primarily based on the data available in the State Domestic Product Report (SDP) of 2016-17 for these same sectors Corresponding to our objective of the study,

we have disaggregated the category „Other Services‟ to identify education & research and medical & health as separate sectors to obtain the respective multipliers

Estimation of individual components of the I-O table:

Gross Value of Output (GVO):

The gross value of output data for the state of Karnataka is not made available in any of the data sources However, the State Domestic Product Report (SDP) computed at the State level gives us the value of the gross value added at the state level across each of the 23 sectors that we have discussed The gross value of output data is available only for four sectors i.e crops, livestock, forestry & Logging, fishing as these are state subjects and procurement of data for these sectors becomes easier Therefore, to estimate the data of GVO for Karnataka, we have used the value of GVA available in SDP Report for each of the 23 sectors and multiplied this with the ratio of GVA to GVO obtained from the India I-O table and arrived at the GVO for Karnataka state

Intermediate consumption:

The intermediate consumption table is a rigorous exercise as we have to capture the inputs across each sector-sector combination This exercise has been done across 130 sectors for India However, for our purpose, these 130 sectors have been aggregated to 23 sectors for which no data is available In order

to obtain the inter industry linkages, we calculated the coefficient matrix from India I-O table The underlying assumption is that the inter sectoral linkages for Karnataka reflects that of India which is also a services sector growth driven economy Once these coefficients were obtained, we used these coefficients and multiplied them with the GVO to obtain the intermediate flow matrix for Karnataka

Gross Value Added table and its components:

Before getting into the details of the methodology for the computation of the components of the GVA, let us understand the data sources that we have referred to

State Budget: The state government budget gives us the detail of the total budget for the financial

year for a particular state It captures the detailed capital and consumption expenditure across different departments

Local Budgets: Although the state budgets contains data on the transfers to the rural local bodies and

urban local bodies, it does not capture the own sources of revenue for these bodies Hence, a detailed

analysis of the local budgets is required to understand the different categories of expenditure incurred

Public Sector Corporations: The limitation of the state budget is that they don‟t capture data from

the corporations as they function autonomously Hence, in order to capture the detailed data from the corporations like the gross value added, GFCE and GFCF, extensive field work was carried out We visited most corporations (80 Public Sector Corporations) personally and approached some of them

through Right to Information Act (RTI), 2005 and these accounts will be examined thoroughly

Using the above data sources, each of the components of the gross value added is calculated based on the methodology given below

Private Final Consumption Expenditure (PFCE):

It is the total „out of pocket‟ consumption expenditure of the entire population residing in a particular country during the year In India, the state wise PFCE data is unavailable In order to capture this data for the state of Karnataka, the following steps were followed:

We took the ratio of the PFCE of India to the GVA of India and multiplied this with GVA of Karnataka to get the overall PFCE for Karnataka In order to distribute this across the different sectors, what we need is the consumption expenditure across the different sectors of the economy This data is not directly available in any of the data sources available in India Hence, to calculate this,

we investigated the NSSO 67th Round on Consumption Expenditure for the survey year 2011-12 This

is a sample survey which captures the consumption pattern of the people in the country for the survey years across all item groups We captured the consumption expenditure solely for Karnataka state and based on the item groups, we classified it into that particular sector/industry and hence the sector wise PFCE for Karnataka was calculated

Government Final Consumption Expenditure (GFCE):

The government final consumption expenditure of administrative departments is equivalent to the current expenditure on compensation of employees, purchase of non-durable goods and services net

of sales and the Consumption of Fixed Capital Karnataka publishes the Economic cum Purpose Classification of Karnataka State Budget, 2011-12 to 2018-19 (ECP) This report publishes the GFCE

data for the state of Karnataka of the general government However, this has some limitations This data does not include the data for the public sector corporations Though it captures the transfers to the local bodies, it does not classify them as consumption expenditure or capital expenditure Hence, detailed analyses of the three data sources including general government budget, public sector corporations and local budgets will be explored in detail to obtain the GFCE for Karnataka

Gross Fixed Capital Formation (GFCF):

The GFCF is essentially the net investment in a particular year in the entire economy This data is not available at the state level for Karnataka In order to estimate this we do the following:

The Government Capital Formation data is available in the ECP report for certain services sectors which indicates the capital formation in the government sector alone In addition to this, we will also explore the government budgets, local budgets and the public sector corporation data to calculate the capital formation for the public sector For the private sector, no data is available at the state level Hence, we will refer to the Annual Survey of Industries (ASI) data for the organized sector and the NSSO round for the unorganized sector to calculate the total capital formation for the private sector

Exports and Imports:

We plan to refer to the DGCIS data to get the state level data for exports and imports We are also trying to explore data available with the Commercial Taxes department for the Government of Karnataka to get sector wise exports and imports data

Indirect Taxes:

The total indirect taxes data is available as aggregate for the entire state as a whole in the RBI State Finances 2013-14 report To estimate the Indirect Taxes for each of the 23 sectors, we took the national level proportions of indirect taxes for each sector from the India I-O table and this has been multiplied with total indirect taxes to obtain the indirect taxes for each sector of the state

Once all the data is collected from the relevant sources and the values of the I-O table computed, we

go ahead with the calculation of the Leontief inverse matrix to obtain the multipliers These multipliers are further analysed to understand the inter industry linkages

Conclusion:

The relevance of Input-Output model as a tool helps assess the effect on an economy of changes in elements that are exogenous to the model of that economy 'Input-output analysis is a practical extension of the classical theory of general interdependence which views the whole economy of a region, a country and even of the entire world as a single system and sets out to describe and to interpret its operation in terms of directly observable basic structural relationships' (Kurz and Salvadori, 2000) The above sentence reflects the viewpoint that economy is a circular flow and

concerns the foremost issues of interdependence in production, distribution and relative prices as against Walrus view of production that views production as a one-way avenue that leads from the services of the original sources of production, land, labour and capital to final goods The tool of Leontief Input-Output model largely derives their significance from the fact that output multipliers measuring the combined effects of the direct and indirect repercussions of a change in final demand can be calculated India has had a history of constructing I-O table at the national level in the planning period that began in the early 1950‟s and the first official national I-O table was prepared for 1968-69 jointly by Central Statistics Office and the Planning Commission, comprising 60 sectors The last official I-O table constructed for the country was in 2007-08 This attempt to construct the I-O table for the state of Karnataka is to explain the linkages among sectors as it also helps ascertain the importance of balanced growth of the economy as against policies being sector oriented that leads to lop-sided development

Appendix:

Table 1: Concordance Matrix

Sectors in Karnataka I-O table Sectors in India I-O Table

Crops

Paddy, Wheat, Jowar, Bajra, Maize, Gram, Pulses, Sugarcane, Groundnut, Coconut, Other oilseeds, Jute, Cotton, Tea, Coffee, Rubber, Tobacco, Fruits, Vegetables, Other crops

Livestock Milk and milk products, Animal services (agricultural), Poultry &

Eggs, Other livestock products Forestry & Logging Forestry & Logging

Mining & Quarrying

Coal and lignite, Natural gas, Crude petroleum, Iron ore, Manganese ore, Bauxite, Copper ore, Other metallic minerals, Lime stone, Mica, Other non metallic minerals