Commercial investment for development : analysis of output and employment effects of IFC's investment(s) in India using an input-output model

(1)

2018

Commercial Investment for

Development

Analysis of Output and Employment effects of IFC’s

investment(s) in India using an Input-Output Model

MANTEJ PARDESI mantej.pardesi@student.uva.nl UVA ID: 11443073

MASTERS THESIS

PROGRAMME: MSc Economics SUPERVISOR: N.J. Leefmans SECOND READER: Dr. N. Oomes

(2)

1

STATEMENT OF ORIGINALITY

This is to certify that this thesis is written solely by the author - Mantej Pardesi, who declares to take full responsibility for the contents of this thesis.

The author declares that the contents of this thesis are authentic and that no other sources other than those mentioned in the text and references have been used in creating this thesis. The complete responsibility for the contents of this thesis is borne by the author and the Faculty of Economics and Business is only responsible for the supervision of this work.

(3)

2

ABSTRACT

This thesis seeks to estimate the impact of an IFC investment in a major private sector hospital chain in India using an Input-Output model and apply this model to compare and analyze sectors within the active portfolio of IFC in India on the basis of their capability to produce the largest output and employment multiplier effects. An Input-Output model produces deterministic estimates of a sector’s industrial capability in the form of multipliers. It contains the Input-Output table which shows the economy as a network of inter-related sectors where purchase and sale of inputs between sectors to produce output (purchase-production-sale chain) is used to calculate the economic impact of an exogenous change in the economy. The results show that for an investment of Rs. 450 crore (US $ 68 million), IFC would create 11483 jobs and generate an additional output worth Rs. 807.95 crore (US $ 122.09 million) which is an effect of 1.795 times the investment. Moreover, analysis of IFC’s portfolio supports investing in services sectors (health, education and banking) to solve the problem of job absorption (“jobless growth”) in the Indian economy.

(4)

3

1. INTRODUCTION

Since the beginning of the 21st_{century India has been among the highest growing countries}

registering an average annual growth rate of 7.28% (World Bank Data). In fact, India has been called the “land of opportunity”1_{due to high growth, large endowment of land and a growing}

working age population. Prudent macroeconomic policy and structural policy reforms have improved the business environment2_{and eased inflows of Foreign Direct Investment (FDI)}

(Gupta et al., 2018). The positive economic outlook which India has established for itself has led to a rise of FDI inflows from $ 27.397 billion in 2010 to $ 44.458 billion in 2016 (OECD STAT)3_{. Like most of the developing countries, foreign investments in India have been for}

both commercial as well as development purposes. Till 2003, development flows (ODA and OOF)4_{had the dominant share of the net foreign flows but an emerging private sector attracting}

commercial investments has changed the composition of foreign flows coming to India (Figure 1). These commercial investments in India can be utilized to meet the development goals of

1_{This reference about India is by the Prime Minister of India, Mr. Narendra Modi, in his address to Dutch CEOs}

during his foreign visit to The Netherlands in June 2017. LINK: https://www.ndtv.com/india-news/india-a-land-of-opportunities-pm-narendra-modi-tells-dutch-firms-1717668

2_{This can be seen from the improvement in India’s ranking on the Ease of Doing Business index}2_{from 142}nd_in

2015 to 100th_{in 2018. The Ease of Doing Business index is compiled by the World Bank and rates the}

investment climate of 190 countries on the basis of the regulatory environment and perceptions of market participants.

3_{Data compiled from OECD stat. LINK:}_{https://stats.oecd.org/}

4_{ODA stands for Official Development Assistance and OOF stands for Other Official Flows.}

SOURCE: OECD STAT

-10,000.00 -5,000.00 0.00 5,000.00 10,000.00 15,000.00 20,000.00 25,000.00 30,000.00 35,000.00 199619971998199920002001200220032004200520062007200820092010201120122013201420152016

Figure 1: Net Foreign Receipts of India (in US$,

million)

(6)

5

the country as India performs very poorly on global development indicators such as the Human Development Index5_{. Traditionally, development funding in India comes from a multitude of}

sources including advanced countries, multilateral institutions and non-governmental organizations, however, it is the World Bank Group which contributes the most to India in terms of funds for development projects (Figure 2). India’s attraction as an investment destination in the private sector has created a greater role for the private sector arm of The World Bank Group, International Finance Corporation (IFC).

The International Finance Corporation is a multilateral private sector development institution which provides loans and attains equity stake, asset management and advisory services to private firms in its partner countries. With the objective to promote inclusive growth, improve environmental and sustainability standards, and encourage regional and global integration6_,

IFC investments are spread across various sectors in India. According to IFC, India was the largest recipient of its commitments in South Asia and the third largest in the world next to Brazil and Turkey with cumulative gross commitments of $1.576 trillion (as of June 30, 2017)7_.

India has largely benefitted from IFC’s involvement in providing credible sources of finance

5_{HDI or the Human Development Index is the benchmark index compiled by the United Nations Development}

Programme where 190 countries are ranked on the basis of economic growth, life expectancy rate, literacy rate and infant mortality rates. It shows how developed a country’s standard and quality of living is.

6_{IFC has 6 “focus areas” in India viz.}(1) _{developing infrastructure and energy projects,}(2)_{strengthening capital}

markets, (3)_{promoting financial inclusion,}(4) _{improve domestic competitiveness,}(5)_{creating jobs and}

opportunities, (6)_{expanding the availability of telecommunications.}

7_{This data is compiled from IFC’s Annual Report for the FY 2016-17, p.18. LINK:}

https://www.ifc.org/wps/wcm/connect/f7f19c11-859b-45bd-8cd9-cedfb0ad0e0b/IFC-AR17-Vol-2-Financials.pdf?MOD=AJPERES -4,000.00 -3,000.00 -2,000.00 -1,000.00 0.00 1,000.00 2,000.00 3,000.00 4,000.00 5,000.00 6,000.00 19 81 19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13

Figure 2: Net flows from Multilaterals to India (in

US$, millions)

Net flows from Multilaterals Net flows from WBG Net flows from IFC

(7)

6

and increasing financial inclusion in an erstwhile weak and exploitative micro-finance sector (IFC, 2016). IFC also has a significant effect on ‘job creation’ as their investments are shown to have spill-over effects in the form of positive demonstration effects and other production externalities (IFC, 2013). This contributes to providing jobs for the burgeoning working age population in India (Euromonitor International expects working age population to grow by 17.4% from 2010-2020 in India)8_{which is projected to be the highest in the Asia-Pacific region}

by 2050 (UNDP, 2016). To realize the impending demographic dividend job absorption must be high which is unsatisfactory as statistics show that out of an increase of 300 million people belonging to the working age population from 1991 till 2013 only 140 million got a job (UNDP, 2016). The sad state of job absorption attributes for inter alia low labor force participation rate (53.7% in 2018 and is expected to drop to 52.2% in 2030) (ILOSTAT) and unemployment rate (projected to be constant at 3.5%) (ILOSTAT)9_{. Therefore, there is a need for IFC’s}

investments to have a positive effect on employment creation in India.

8_LINK:_{https://blog.euromonitor.com/2010/11/indias-working-age-population-growing-faster-than-chinas.html} 9_{ILO modelled estimates. LINK:}

http://www.ilo.org/ilostat/faces/wcnav_defaultSelection?_afrLoop=596450441293026&_afrWindowMode=0&_ afrWindowId=null Agribusiness & Forestry 2% Financial Institutions 52% Funds 2% Health and Education

7% Infrastructure

19% Manufacturing

3% Oil, Gas & Mining

1% Others

11% Telecommunications, Media and Technology

1%

Tourism, Retail and Property

2%

FIGURE 3: IFC ACTIVE PORTFOLIO IN INDIA

(IN USD, MILLIONS)

(8)

7

The active portfolio of IFC in India (as on June, 2018) is more than $5 billion (IFC) (World Bank Group Finances)10_{. In figure 3, we see that IFC’s investment portfolio is skewed towards}

the financial markets wherein IFC stimulates banks, insurance companies and other financial institutions to expand credit and perform financial services. Investment in sectors such as the construction sector (infrastructural development – 19% of investments), manufacturing and social services (predominantly health and education – 7% of investments) are direct ways of creating jobs and supplement the productive capacities of the economy as compared to the financial sectors. Since social sectors are empirically shown to have other qualitative effects (improved life expectancy and literacy, etc.), measuring the economic effects of investment in these sectors assumes importance to meet the Sustainable Development Goals. Hence, this thesis seeks to evaluate the output and employment effects of one such investment by IFC in the health sector in India. Due to the heterogeneity in IFC’s portfolio and lack of research on evaluating IFC’s investment decision making, this thesis will also analyze IFC’s portfolio by classifying it into 9 specific sectors viz. (1)_{Construction;}(2) _Banking;(3) _Insurance;(4) _Medical

and Health; (5) _{Communication;} (6) _{Education & Research;} (7) _{Electricity, Gas & Air}

Conditioning Supply sector; (8) _{Oil, Petroleum & Natural Gas;}(9) _{Hotels and Restaurants. These}

sectors are chosen on the basis of their concordance with the broad categories in IFC’s portfolio and the methodology used in this thesis – Input-Output Model (I-O). These sectors will be compared on the basis of their ability to generate a multiplier effect in the economy in terms of additional output and employment using an I-O model.

An I-O model (unlike a regression equation) allows us to obtain deterministic estimates of the economic effect an investment has on the economy. It shows the economy as a network of inter-connected sectors wherein each sector produces its output by using labor and inputs from other sectors11_{. Modern day economies are a complex web of sectoral inter-relations wherein}

each sector is both a supplier to and an end user of other sectors’ produce. When there is an exogenous change in any component of this web, the entire system is affected, i.e., an investment in any sector has an economy-wide multiplying effect. The aggregate effect can then be decomposed into Direct and Indirect effects. Direct effects refer to change in economic variables in the sector where the change takes place (in this case, investee client due to IFC investment) whereas the Indirect effect refers to the effect on the rest of the economy. Although

10_{Data on IFC’s portfolio in India has been compiled from World Bank Group Finances. LINK:}

http://financesapp.worldbank.org/en/summaries/ifc/#ifc-isp/countries=IN/

11_{Complete description of the model and the methodology used to calculate the results is mentioned in Section}

(9)

8

the Monitoring & Evaluation system of the WBG (Development Outcome Tracking System or DOTS for IFC) can measure the direct effects of their projects, the indirect effects are usually either unreported or inaccurately estimated (IFC, 2013).

The objective of this thesis is to measure the output and employment effects of a particular IFC investment in the health sector using an I-O model and applying this model to analyze IFC’s active portfolio in India. The particular investment which will be studied in this thesis is IFC’s investment in Apollo Health and Lifestyle Limited (AHLL) – a major private hospital chain in India (see Box 1). The primary motivation behind choosing this topic and this methodology is that the health sector is a social sector wherein investments focus to improve health service delivery such as availability, accessibility and affordability but not many projects focus on the economic effects of a Hospital. A hospital, like any other enterprise/firm, employs people and provides services for a fee (since AHLL is a private hospital chain it charges fee for its service which is not the case for a public hospital) and hence generates value for its services which is considered to be its output. Each hospital has a set of suppliers and distributors from where inputs are bought for the production of health services like hospital beds, drugs and medicines, sanitary equipment, etc. Since an investment in a hospital is followed by an increase in its output and job creation, the inter-sectoral linkages of the health sector will lead to an increase in demand for inputs and hence an economy-wide purchase-production-sale chain is created12_.

This chain leads to the creation of additional output and employment which is more than the investment amount. Since an I-O model quantitatively shows all the inter-sectoral relations existing in the entire economy, its use in economic impact analysis provides a comprehensive picture of the production induced effects in the economy (Bess & Ambargis, 2011) (Miller & Blair, 2009) and hence an I-O model for the economy of India is used in this thesis.

12_{Purchase of inputs by sector 𝑖 from sector 𝑗 to produce output in sector 𝑖 and then selling the output to sector}

𝑗. The purchase-production-sale chain is understood as the consecutive rounds of exchange of output in the economy due to the initial stimulus/change in the final demand in the economy. Every sector has a chain of suppliers and distributors. For instance, when output in Construction increases by $100, inputs from other sectors such as Manufacturing and Mining will be required as well as inputs from the Construction industry itself. And now to produce the output of Manufacturing and Mining, inputs from agriculture and services are required. This creates a chain of consecutive rounds of exchange between different sectors. The demand for the extra $100 worth of Construction output is regarded as having caused the production of these outputs.

(10)

9

Therefore, this thesis seeks to provide answers to the two following questions: (1)_{What are the}

output and employment effects of the IFC investment in Apollo Health and Lifestyle Limited? and (2)_{On the basis of output and employment creation, how can investing in some sectors have}

bigger multiplier effects on the economy than IFC’s current portfolio in India? The answer to the first question will be in the form of descriptive estimates obtained from I-O analysis. The primary impact analysis will estimate the absolute effect of the investment which would be compiled by multipliers calculated through the I-O model. The second question will be answered by computing multipliers for the above mentioned 9 sectors of the Indian economy and comparing them with the distribution of investment in these sectors within IFC’s portfolio. The comparison of the 9 sectors will be done by making hypothetical investments of equal magnitude as the one in AHLL in each of the 9 sectors to establish a counter-factual viz. “What would have been the output and employment effect if the money was spent in another sector?”. The intuition here is that the more industrial linkages a sector has, the higher will be its economic multipliers and the larger will be the economic impact of investment in that sector as compared to a sector which does not have substantial linkages with other sectors (Sonis et al., 1995) (He & Zhu, 2016). Hence, if a sector has a higher economic multiplier then it warrants more investment to maximize the marginal value of each additional dollar spent in the economy. It should be noted here that this thesis assumes ceteris paribus with respect to

sector-BOX 1: Details of IFC investment in Apollo Health and Lifestyle Limited

IFC’s investment was in the form of an equity stake of Rs. 450 crore ($ 68 million) in AHLL on December 2016. AHLL aims to provide community based, multi-specialty chain of clinics that provides the neighborhood with easy access to specialist doctors, primary care diagnostics services, treatment facilities (treatment room, physiotherapy, dentistry) and preventive healthcare (health checks and vaccinations). These clinics are clustered around Apollo Hospitals operational in major cities of the country. As of FY 2017, AHLL employs around 3000 doctors and 3000 other health workers across 153 small-healthcare centers, 12 maternity facilities, 23 pathology labs and more than 250 collection centers in the country. From this investment, AHLL aims to open 151 small-healthcare centers, 9 maternity facilities, 64 pathology labs and 765 collection centers1_{. This investment will lead to an increase in the}

operational capacity of AHLL and subsequently lead to more employment of doctors, nurses and other health workers. As with all other projects, IFC aims to stimulate other market participants in this relatively unexplored market segment (small scale clinics) to generate sector-wide effects. Since the clinics under AHLL have a large geographical reach (as they are located in smaller localities thereby increasing accessibility for the patients and higher demand for their services) and employ substantial amount of health workers, the output and employment effect generated by these clinics would act as a very useful foundation for future policy or investment decision by other market players. (IFC, 2017)

(11)

10

specific economic changes, labor market developments and other externalities. The main purpose of this thesis is to alienate sector-specific characteristics and assess a sector’s ability to have an economy-wide effect solely on the basis of its industrial linkages. The results would then contribute to the discussion about India’s capacity to provide employment to its growing working age population and the role of IFC and other foreign investors in facilitating employment opportunities for the young in India.

The layout of this thesis is the following: Section II provides a review of the existing literature on (1) _{Input-Output models – as a methodology for impact assessment,}(2)_{the economic impact}

of investment in healthcare through Input-Output Models and, (3) _{the role of I-O models in}

studying the economic effects of different sectors in India. Section III describes the Input-output Model for this paper and the methodology for deriving the Input-output and employment multipliers. Section IV deals with the description of the various sources from where data for this thesis has been procured. Section V provides the results and findings from the I-O analysis and provides an objective appraisal of the IFC project. Section VI contains the discussion on IFC’s investment portfolio in India and how they can improve this portfolio to maximize economic gains for the country. This section shows how this thesis attains relevance in impact assessment; and Section VII concludes the thesis.

(12)

11

2. LITERATURE REVIEW

2.1. Input-Output model: a brief review

The literature on economic impact assessment is intensively dominated by the use of Input-Output models as the main source of methodology (Sargento, 2009). Developed by Wassily Leontief (1936)13_{, the I-O model is designed to study the economic interdependencies among}

the different sectors in an economy. The structure of the model allows us to compute deterministic estimates, in terms of economic multipliers, for any change in the economy (Miller & Blair, 2009). The I-O model consists of the I-O table which is a matrix representation of all the industrial sectors in the economy producing and providing output to the consumers, government, trade and international transactions, and domestic intermediate consumption. The I-O table is followed by the I-O analysis (more on I-O analysis in Section 3). Lately, the use of I-O models in policy planning and computable general equilibrium models has been increasing due to developments within the I-O model. Construction of regional, inter-regional and international (between countries) I-O models, Social Accounting Matrix, Multiplier Product Matrix, etc. have all contributed to these developments resulting in more sophisticated and accurate results obtained from different iterations of the basic I-O framework (Sargento, 2009) (Plosjaz et al., 2015). The applications of a simple I-O Model include estimating production targets, structural analysis of the economy (to what degree certain sectors are linked to each other) and economic impact analyses (Eleish, 1963). For instance, impact analysis of foreign investment for commercial & development related activities, endogenous and exogenous policy changes, industrial ecology, changes in structural composition of the economy, shutdown of a firm or an industry due to foreign competition and any other kind of quantifiable microeconomic or macroeconomic impulse to the economy that will affect the wider economy as a whole, the impact of which can be investigated using I-O analysis (Miller & Blair, 2009). Due to its objective depiction of the inter-industry links and each sector’s

13_{Wassily Leontief compiled the first input-output table in 1932 of the American Economy for the years, 1919}

and 1929. In 1936, he published his first input-output paper demonstrating the importance of input-output economic analysis and in 1941 published his first book demonstrating his general equilibrium theory, The Structure of American Economy, 1919-1929: An Empirical Application of Equilibrium Analysis. He won the Nobel prize in Economic Sciences “for development of the input-output method and for its application to important economic problems”.

(13)

12

production capabilities, many researchers, business professionals and policy makers also use it as an ex-ante and ex-post project analysis tool.

However, the I-O methodology is not a water-tight tool for estimating economic results. Since it is not a regression equation, the results obtained do not establish a cause and effect relationship. In his seminal article, Christ (1955) argues against the I-O model as lacking any optimizing condition and hence not a general equilibrium model. He also mentions technical problems such as collection, concordance, and processing of large quantity of data. He refers to these problems as “errors due to rounding”, “errors due to approximation” and “errors due to inaccurate data”. However, he also states that such errors are “not likely to be dangerous”. These data related problems are compounded by limitations in the assumptions of the I-O framework. Standard economic theory says that inputs in a production process are used relative to their prices, however, I-O analysis assumes inputs are fixed technologically (Christ, 1955, p. 24). Moreover, a common production function for the entire sector might not depict the heterogeneity in the sector (Christ 1955, p. 5) and might very well be approximated with a considerable bias of the researcher (Wilner & Birnberg, 1986). The views of Christ (1955) were further supported and supplemented by Jensen (1990) who provided a systematic review of the problems and developments in the construction of I-O tables. His approach largely focused on analysing the two methods of collecting data (survey and non-survey). Moreover, criticism about the static nature of the analysis (Sonis et al., 1995) and the inability to include the flow of capital goods in the I-O model (Gemmell et al., 2006) are also major drawbacks of this methodology. Hence, there is a general consensus that caution should be expressed while making decisive judgements from the results of an I-O analysis.

The following section shows the studies estimating the economic impact of the health sector using an I-O model and how these results differ on the basis of regional and economic differences. The third section reflects upon studies which have used an I-O analysis to show how the strength of industrial linkages between sectors have affected the structural composition of India.

(14)

13

2.2. Economic impact assessment of the health sector using I-O analysis

From a macroeconomic perspective, an efficient health sector leads to improved health and productivity of an economy’s human capital. This improves the growth prospective of a country (Jamison et al., 2013) (Bloom, Canning and Sevilla, 2001). This efficiency is inter alia determined by the amount of funding (both public and private) the hospitals and other medical facilities get (Mackintosh et al., 2016). The effect of financing or investing in the health sector is not only through the direct channel of labour productivity/ human capital enhancement but also through increasing the productive capacities of the medical infrastructure (hospitals, clinics, etc.). The sectoral interaction of a productive health sector enables the health service providers to achieve economies of scale (Stoney, 2004). If the sectoral interaction/linkages of the health sector are high, then it has a “central” role in the industrial structure of the economy and thus a greater role in output and employment generation (Acemoglu et al., 2012). This is the indirect channel where the health sector affects economic growth.

The effect of the health sector through the indirect channel has been studied by many researchers using an I-O model. The results are reported in terms of multipliers which are interpreted as the level of economic activity (in this thesis, focus is on output and employment) generated by the health sector over and above its own. Schwärzler & Legler (2017) analyze the impact of Germany’s health system on employment to provide evidence for targeted investments in the health sector. Using an I-O table14_{derived from a health-specific Supply}

and Use table they find that the health sector had an employment multiplier of 0.62 which showed that for every job created in the health sector, 0.62 jobs will be created in the rest of the economy. Their study shows that the health sector assumes a substantial status in Germany with 11 million jobs (16% of employment) supported by the health sector which is significant when compared to other social sectors such as education and research, employing 6% (ILOSTAT) of the workforce. A similar study was conducted by Perobelli et al. (2015) for Brazil for two years, i.e., 2000 and 2005 who disaggregated the entire health sector into 8 sub-sectors by harmonizing the National Health Accounts with the System of National Accounts. This disaggregation allowed them to perform a structural analysis of the Brazilian health sector. Their results showed how sub-sectors such as manufacturing of pharmaceutical products had high output multipliers but weak employment multipliers showing productivity gains overtime

14_{They use the Input-Output table for the entire German economy for the year 2015, hence the results show the}

(15)

14

(similar to industrial sectors) whereas sub-sectors such as health-services trade and provision of private social services have low output effects and high employment effects mainly due to low linkages with other sectors as their sales are destined for final demand (private consumption). The health sector in Brazil, on average, had a peripheral role in the economy with below-average multipliers.

The discussion about the health sector moves forward from the sub-sectors to more disaggregated units. In the case of the United States, this has been enabled due to the creation of an IMPLAN (Impact Planning) software, the core of which is an I-O table of 536 sectors for the country at county (district) level. This software enables research to be focused on spatial units such as medical institutions and health workers on a geographically bound area such as a town or a state or a community within the United States. In a systematic review of investments in the health sector in United States, Doeksen, Johnson & Willoughby (1997) find that the employment multipliers for most of the studies using IMPLAN to be more than 2 signifying the doubling in number of jobs by the health sector (See Table 1). However, regional disparities existed. For instance, Doeksen & Schott (2003) who conducted an analysis of the regional health sector in Aloka county in south-eastern Oklahoma using IMPLAN, found the employment multiplier for a sub section of the health sector, i.e., Hospitals, to be 1.70 which means that for every job created in the hospitals of Aloka county, 0.70 jobs were created in the county. Therefore, in a large country the size of the United States with a dominant private health system, the economic effects of various components of the health system widely differ in size.

Out of all the studies using I-O analysis to assess health sector’s economic impact (Table 1), studies on the Japanese healthcare system were found to have the largest size of output multiplier effects. Yamada & Imanaka (2015) evaluate the health system in Japan using an I-O table for the year 2000 and data from Statement of Profit & Loss of all medical institutions in the country. Their study was unique as they provided an estimation range over their results from a probabilistic sensitivity analysis using Monte Carlo simulation15_{. This was made}

possible due to the availability of data from individual institutions which allowed them to create sampling distributions for key results16_{. Their results produced an output multiplier of 2.78 for}

the health sector with aggregate output generation of $ 661.5 billion (with a 95 % CI: $651.5

15_{Monte- Carlo Simulation is a simulation technique used to comprehend the impact of risk and uncertainty in}

prediction and forecasting models.

(16)

15

billion - $690.8 billion). This means for an investment of $100 in the health sector $278 worth of output will be created in the entire economy. Their results show an above average impact of the health sector as typically the output multiplier for this sector in any country ranges from 1.5 to 2.5. Tsukahara (1996) in a similar study found the economic impact of the health sector to be slightly lower with an output multiplier of 2.33. This shows that the health sector has evolved in its economic role in Japan mostly due to the growing age burden and increasing number of old-age dependents and thereby rise in demand for medical service (Yamada & Imanaka, 2015).

Overall, there is no clear conclusion as to the place of the health sector in the economy according to its industrial linkages. Much of this depends upon the network structure of the health sector, the type of health system (private or public), whether it is purchases its inputs domestically or imports them and other epidemiological (demand-inducing) factors, and the epidemiological characteristics of the country. In the Indian context, there have been no major studies conducted on the economic impact of the health sector. Analysis of investments in a hospital chain which operates on a country wide scale have been very few. Therefore, the present research contributes to the existing body of literature on I-O models and creates new scope for evaluations of organizations which operate on a national level.

(17)

16

Table 1: Economic multipliers of investments in the health sector

Country Year Health sector component Multiplier

Employment Output

Germany 2018 Broader health system 1.62 1.82

Japan 2015 Broader health system - 2.78

Japan 1996 Broader health system - 2.33

Brazil 2000/2005 Health system (average) 59/38* 1.88/1.92

Manufacture of pharmaceutical products

26/27* 1.75/1.79

Manufacture of apparatus 34/17* 1.36/1.40

Health services trade 89/54* 1.51/1.53

Complementary care 30/22* 1.83/1.80

Hospital care activities 44/27* 1.87/1.89

Other health-related activities 54/43* 1.54/1.65

Private social services 133/103* 1.78/1.70

Public health 58/36* 1.57/1.62

USA

Montana 1987 Hospitals chain 1.30 1.60

New York 1974 Specific Urban Hospital - 2.63

Pittsburgh 1986 Urban Hospitals - 2.69

Rural Oklahoma 1991 Health Worker - Physicians 1.78 1.52

Aloka, Oklahoma 2003 Hospitals 1.70 -

Humboldt, Nevada 2006 Regional Health system 1.39 -

*Perobelli et al. define the employment multiplier as the number of jobs created in the economy per 1 million Brazilian Reals in the final demand for a particular sector.

2.3. Review of inter-sectoral linkages in India

The use of I-O analysis was initiated in India since the 1950s. After independence in 1947, India embarked on a path of heavy industrialization after being predominantly an agrarian economy. Economic planning policy therefore required the computation of I-O tables for better understanding of the economy (Kuwamori & Sato, 2009). The compilation of Input-Output tables in India dates back to the work by Chowdhury (1954) where the I-O table was based on

(18)

17

the year 1948-49 and had 23 sectors. In the beginning, the majority of the work was done by individual researchers, however, since 1950s the Indian Statistical Institute commenced publishing I-O tables every 4 years. Soon after, the statistical branch of the Government of India – Central Statistical Organization officially started compiling economy wide-data to produce I-O tables in every 4 years (Kuwamori & Sato, 2009). The latest I-O table for India was published in 2016 by the National Council for Applied Economic Research. This table is based on the financial year of 2013-14. This is the I-O table whose aggregated version will be used in this thesis. The I-O tables published so far in India allow for a dynamic structural analysis of the Indian economy as multipliers for each sector can be compared overtime. The studies which analyze the structural change of the Indian economy divide India’s economic history into two parts: “inward-looking” phase or the period before the 1990s and the “outward-looking” phase or the period post 1990s liberalization (Dholakia et al., 2009). Economic liberalization in India started in 1991 when India overcame its Balance of Payments problem through IMF’s intervention whose conditionality led to a new wave of economic policies focused on free markets. Post-1991, the Indian economy witnessed a rise in imports, investments, and hence final demand but not in the agriculture sector which witnessed a decline in gross capital formation till the 2000s (Bhattacharya & Rao, 1986) (Jha, 2010). This downward trend goes in line with Kuznets’ (1966) perception of modern economic growth which denoted economic growth to accompany a falling share of agriculture in value added and workforce. However, Sastry et al. (2003) conclude that though the share of investments in agriculture is falling, its role in economic growth is still important due to the linkages it has with rest of the sectors in the economy. This view is also acknowledged by Vyas (2004). Traditionally, it is seen that the sectors which have the higher industrial linkages in the economy are more efficient in increasing the economic gains from investments in those sectors as compared to sectors with low industrial linkages (Das, 2013). The sectors with the highest linkages are likely to stimulate rapid growth of production, income, and employment (Hirschman, 1958)17_{(Rasmusen, 1956). On an aggregate level, Saikia (2011) analyzed the}

industrial linkages between the agriculture, industry and services sector from 1968-69 till 2003-04. He found that sectoral linkages from agriculture to industry have grown, while from industry to agriculture it has decreased. This means that agriculture’s production function contains more of industrial machines and equipment but industry does not require much of

17_{Hirschman’s theory of ‘unbalanced growth’ describes a sector’s relationship with the rest of the economy}

(19)

18

agriculture produce. Regarding services, agriculture lags behind industry sector which has strong linkages with the services sector. The weak relation between industry and agriculture is unexpected as since 2001, there has been considerable investment in the rural primary sector which has led to increased income, productivity and agriculture’s terms of trade. This should have led to more demand for industrial inputs and hence stronger inter-dependence between the two. In his study on finding the sources of growth in the Indian economy through a dynamic Input-Output system, Das (2013) estimated industrial linkages for all the sectors from the I-O tables available from 1993-94 till 2007-08. His results aligned with Kaldor’s growth theory, i.e., the faster the growth rate of manufacturing sector, the higher the growth rate of GDP of the country (Kaldor, 1966). Within the manufacturing sector, he found that the iron and steel industry had the highest production linkages in the 1990s but had a much smaller impact than electricity and electrical appliances sector in 2000s. In the services sector, hospitality, trade, banking and other financial activities were the leading source of growth during the post-1990 period but not as much as the manufacturing sector. Hence, he concluded that expansion of the manufacturing sector would lead to an overall expansion of the economy and stimulate the growth of the services sector (not vice-versa).

Another important aspect of an inter-sectoral structural analysis is the difference in employment generation capacity between sectors. Although a rise in output and industrial linkages may seem to cause a rise in employment generation due to the increased demand for factors of production, however, increased output might lead to factor substitution as well. This is because the most productive sectors are dependent on capital intensive intermediate products so higher output will lead to less than proportional employment generation effects18_(Acemoglu

and Guerrieri, 2008) (Baumol, 1967, p.415). This concept is important for the Indian context as critics argue that although India has had a high annual growth rate, it is mired with a phenomenon called “jobless growth”19_{(Mehrotra et al., 2013). Research by Sarma & Ram}

(1989) on employment linkages identified agro-based industries to have greater potential to generate income and employment in 1979-1980. A more recent work on employment linkages was done by Bhattacharya & Rajeev (2014) using I-O tables of the year 2003-04 and 2007-08.

18_{As per Acemoglu and Guerrieri (2008), sectoral differences in the responsiveness of factor proportions to}

changes in output leads to structural change. This is because when capital intensiveness is high in a sector, increase in output or Total Factor Productivity will lead to further capital accumulation and labor substitution.

19_{Tejani (2015) provides statistical evidence for the jobless growth in India by analyzing output growth and}

employment growth under Kaldor’s growth theory (Kaldor, 1966). Her results argue that India has “leapfrogged into a high-productivity regime without the broad-based expansion of labour-intensive production…”. Hence, “jobless” growth has been a characteristic feature of India’s growth “miracle”.

(20)

19

On an aggregate level, the agriculture sector assumed a lion share of total employment (60%) with employment in services growing from 19.92% in 2004 to 21% in 2008. “Jobless” growth was shown with a decreasing employment share and employment linkages of the manufacturing sector. After accounting for specific sectors and computing output linkages, they concluded that sectors like mining, FMCG products and textiles had both a high employment and output generation effect. Agriculture, “Wood and wood products” and “textiles” were the best in employment generation and “petroleum products”, “chemicals” and “metals” were the best in creating output. Therefore, to ensure inclusive growth investments should be directed to sectors which have high output and employment effects.

(21)

20

3. RESEARCH METHODOLOGY

3.1. Objective

The main objective of the Input-Output Analysis is to show how output of a sector is produced through consumption/utilisation of production factors which are supplied by other sectors in an economy20_{. This industrial production process of an economy is visualized through an}

Input-Output Table. An I-O table records the “flows of products from each industrial sector considered as a producer to each of the sectors considered as consumers” (Miller and Blair, 1985, p. 2). It shows the economy as a matrix of sectoral inter-relations quantifying the

value of supply chain for each sector in the economy. Each row of such a matrix represents how the supply of a sector’s output is used by other sectors in the economy whereas each column represents how each sector uses inputs (provided by other sectors) in its production process (the industry’s production function). For example, the health sector in the economy

20_{Some sectors might use production factors from outside the economy. Factors which are imported will be}

shown as Imports for the sector which produces that specific factor of production. This is a simplifying assumption widely used and accepted in I-O analysis.

(22)

21

buys inputs from the construction sector for new infrastructure (hospitals and clinics), drugs & medicines, insurance and communication services for essential medical support services, hotels and restaurants sector for hospitality services, etc. These relations are quantified in an I-O Table. From the I-O table, we first derive the Direct Requirements Matrix which, as the name suggests, shows the input-requirements for each sector. Each cell in this matrix is obtained by dividing the corresponding cell in the I-O table with its column total (output produced by the sector). Thereafter, we use matrix algebra to derive the Leontief Inverse Matrix or Total Requirements Matrix which shows sectoral estimates (output multipliers) of the total output required to be produced in the entire economy to produce 1 monetary unit worth of output in the respective sector.

3.2. Assumptions of the I-O model

As mentioned above, an I-O table acts as the foundation of I-O analysis. This table is built on a basic set of assumptions/axioms (Jansen & ten Raa, 1990). These assumptions act as the starting point of our I-O model.

1. Material Balance (supply = use). The total output of a commodity/service produced by each sector is distributed so as to meet input requirements of the entire economy. This means that the total quantity of supply by a sector should be equal to the sum of intermediate uses by all the other sectors and final demand.

2. Scale Invariance Functional fixation of production processes is applied throughout an I-O table. There is a fixed input structure for each sector such that the technical coefficients of the production function are invariant to changes in output of the sector. The production function thus assumed is called a Leontief Production Function. 3. Due to the fixed input structure, there are no economies of scale in production activities.

In the I-O analysis, each industry has constant returns to scale. This means that if final demand for output of a sector increases by 1%, all the inputs will increase by 1%. This is also called the Principle of proportionality.

Leontief’s production function observes inputs in fixed proportions to produce 1 unit of output for each sector. The fixed nature of input-requirements leads to constant elasticity of substitution in a sector’s input-structure. In an I-O table, differences in production functions at

(23)

22

firm-level is averaged thus resulting in a sectoral production function which appropriates production activities for the concerned sector.

Thus, an I-O analysis is a static analysis that assumes constant returns to scale and particularly applicable when assessing relatively small changes in the economy. Moreover, I-O analyses are non-causal and assume an equilibrium state of the economy (Koopmans et al., 2011). Changes in the structural composition of the economy can only be captured by comparing different rounds of I-O tables (Munjal, 2007).

3.3. Input-Output model

In this section, we show the methodological components of the I-O model; the I-O Table, system of equations and final derivation of our desired economic effects of an investment in the economy.

3.3.1. Input-Output Transactions Table

The first step is to show the I-O table. A simplified format of an I-O table and the one which will be used in this analysis is given below:

Figure 4. A graph depicting Leontief’s production function with fixed input proportions.

(24)

23 Table 2: Simplified I-O table

SUPPLY FROM 𝑖𝑡ℎ_INDUSTRY INTERMEDIATE USES (USES IN 𝑗𝑡ℎ

INDUSTRY) FINAL USES _TOTAL

OUTPUT

Ind-1 Ind-2 Ind-n PFCE Other

Final Use Final Demand

Ind-1 F11 F12 F1n C1 G1 Y1 X1

Ind-2 F21 F22 F2n C2 G2 Y2 X2

Ind-n Fn1 Fn2 Fnn C_n G_n Yn Xn

Value Added V1 V2 Vn

Value of Output X1 X2 Xn

 Net Final demand = PFCE + GFCE + GFCF+ CIS + Net Exports;  Other Final Use = GFCE + GFCF + CIS + Net Exports

 GFCE: Government Final Consumption Expenditure; PFCE: Private Final Consumption; CIS: Change in stock; GFCF: Gross Fixed Capital Formation

 Intermediate Uses + Final Demand = Total Output

 In the table to be used for this thesis, n = 52. Hence a 52x52 matrix is developed for the entire economy.

In the given table, the columns (𝑗 ) are shown as the consuming sectors and each industry on the rows (𝑖 ) is shown as the supplying sector. In other words, industries represented on the left most column show the producing sectors who supply output to the economy which is used as inputs by different sectors in the economy and used as goods and services to meet the final demand. Industries mentioned under the heading “Intermediate Uses (Uses in the 𝑗 industry)” are the consuming sectors who use the inputs to produce output and generate value addition to the sum of all inputs.

The output of a sector is also used to satisfy final demand and its components which is shown as a separate column titled “Final Uses”. Output not used in the industrial production process goes to the households, to the government, outside the economy as exports, etc. Entries under the heading “Final Uses” show the components of Final Demand which includes PFCE (Private Final Consumption Expenditure), GFCE (Government Final Consumption Expenditure), GFCF (Gross Fixed Capital Formation), CIS (Change in Stock) and Net Exports. These

(25)

24

columns show the final demand components for each sector in the economy. All entries are made in monetary value terms (in Rs. 100,000)21_{. Algebraically, this is represented as,}

𝐶 + 𝐺 = 𝑌 , ∀ 𝑖 ∈ {1,2,3, … , 52} Where, 𝐶 represents Private Final Consumption Expenditure

𝐺 represents Other components of Final Demand such as GFCE, GFCF, CIS and Net Exports

For example, the cell 𝐹 shows the amount of sector 1’s output used as an input by sector 2; 𝑉 shows the value added over and above the sum of inputs by Sector 2; 𝑌 is the supply of output of sector 2 used to fulfil the final demand in the economy. Each cell in the row for ‘Value Added’ includes wages, salaries and supplements and gross operating surplus generated to produce output in that sector.

The next step is to show the system of equations which guide our analysis. The end result of these equations is what is called a Leontief Inverse Matrix which acts as the multiplying factor over any change in demand in one part of the economy. This is used to estimate the output and employment effect on the economy as a whole.

STEP 1: EQUATION FOR FINAL DEMAND/OUTPUT IN EACH SECTOR

If we assume that the economy is categorized into 𝑛 sectors, then the assumption of material balance can be represented by the following equation,

𝑋 = 𝐹 + 𝑌 (1)

Where, 𝑋 is the total output (supply) of sector 𝑖

𝐹 is the Output of 𝑖 sector used by the 𝑗 sector (Inputs for 𝑗 sector) 𝑌 is the Final Demand (In the given table 𝑌 = 𝐶 + 𝐺 )

In an I-O table, the row total is equal to the column total for each sector as can be seen in Table 1 where 𝑋1 is the total value of output of Sector 1 as the column total and the total output as the row total.

21_{According to the Indian system of money denomination Rs.100,000 is called Rs. 1 lakh and Rs. 10,000,000 is}

(26)

25

𝑋 = 𝐹 + 𝑉

∴ 𝑋 = 𝑋 → 𝐹 + 𝑌 = 𝐹 + 𝑉

However, this must not be confused with equating the residual terms, i.e., the final demand and value added, because

𝐹 ≠ 𝐹

This means that the total output supplied for industrial production process by a sector should not necessarily be equal to the sum of inputs used by that sector in its production process. Hence, final supply must be equal to final use, intermediate supply must not necessarily be equal to intermediate use.

STEP 2: INPUT-OUTPUT (TECHNICAL) COEFFICIENTS

Equation (1) forms the starting point of the I-O analysis. In this equation, 𝐹 is shown as the measure for inter-industry sales (intermediate usage) between sectors. Each 𝐹 can also be written as

𝐹 = a 𝑋 (2)

where 𝑎 is called the technical coefficient or input coefficients in the production function. It shows the proportion of sector 𝑗 𝑠 output used by sector 𝑖 in its production process. For example, in our I-O table the technical coefficient of pharmaceuticals and medicinal products sector in the production function of medical and health sector is 0.15. This number is derived by dividing the input from “Pharmaceuticals and Medicinal Products” sector by the total output of “Medical and Health” sector. In fact, the total input composition (intermediate use of output from other sectors) for each unit of output of the health sector is 0.35 which shows that the health sector uses less inputs than what its value addition is to produce one unit of output.

(27)

26

STEP 3: SYSTEM OF EQUATIONS REPRESENTING ′𝑛′ SECTORS The Material Balance assumption can be expanded into a system of equations:

𝑋 = 𝑎 𝑋 + 𝑎 𝑋 + ⋯ + 𝑎 𝑋 + 𝑌 𝑋 = 𝑎 𝑋 + 𝑎 𝑋 + ⋯ + 𝑎 𝑋 + 𝑌

⋮ ⋮

𝑋 = 𝑎 𝑋 + 𝑎 𝑋 + ⋯ + 𝑎 𝑋 + 𝑌

Next, we shift the ‘𝑋’ terms on the left and group like terms together, we get, (1 − 𝑎 )𝑋 − 𝑎 𝑋 − ⋯ − 𝑎 𝑋 = 𝑌

−𝑎 𝑋 + (1 − 𝑎 )𝑋 − ⋯ − 𝑎 𝑋 = 𝑌

⋮ ⋮ ⋮

−𝑎 𝑋 − 𝑎 𝑋 − ⋯ + (1 − 𝑎 )𝑋 = 𝑌 STEP 4: DIRECT & TOTAL REQUIREMENTS

When written in Matrix form, this system of equations can be expressed as (𝐼 − 𝐴)𝑋 = 𝑌 𝑋 = (𝐼 − 𝐴) 𝑌 (3) Where, 𝑋 = 𝑋 ⋮ 𝑋 ; 𝐴 = ⋯ ⋮ ⋱ ⋮ ⋯ = 𝑎 ⋯ 𝑎 ⋮ ⋱ ⋮ 𝑎 ⋯ 𝑎 ; 𝐼 = 1 ⋯ 0 ⋮ ⋱ ⋮ 0 ⋯ 1 ; 𝑌 = 𝑌 ⋮ 𝑌

(𝐼 − 𝐴) is known as the Leontief Inverse Matrix or the Total Requirements Matrix and 𝐴 is known as the Direct Requirements Matrix. The Leontief Inverse Matrix acts as a catalyst to measure economy wide effects of a change in final demand. That is, if ∆𝑌 measures the change in final demand (for instance, an investment in the health sector), then the corresponding change in the entire economy is equal to

(28)

27

This investment is substituted in the ∆𝑌 matrix as an impulse22_{and when multiplied with the}

Leontief Inverse Matrix, it gives us the change in output in the entire economy (as per different sectors). In this way we can also derive the employment effect by using the Leontief Inverse matrix and calculating the employment effect in other sectors for an investment in the health sector. After deriving the Leontief Inverse Matrix, the next step is to compute the economic multipliers in our analysis.

3.3.2. Multiplier derivation

This is the final section in our methodological derivation of economic effects. Since this thesis talks about the additional (1) output generated and (2) jobs supported in the entire economy due to investments by IFC, it is beneficial to look at the multiplying effect of such an exogenous change. The use of multipliers is very common in empirical I-O analysis (Lapeyre, 2010) (Plosjaz et al., 2015). A multiplier is a summary measure which quantifies the impact of the change in final demand and differentiates the initial impact with the total impact23_{on the}

economy.

To explain this in simple terms we take a hypothetical example. Suppose you receive a lottery of $1000 and choose to spend it on purchasing an automobile (probably a car). To produce this car, the car manufacturer has to hire 5 more workers and buy the requisite parts for the car. This will induce the suppliers and distributors of the car manufacturer to hire 3 additional workers and produce 10 extra units of intermediate inputs for the car (car parts). However, to produce these parts the suppliers need steel/metal which they will buy from their raw material supplier who will in turn hire 2 additional workers for mining and quarrying. And this will go on until the end of the production chain. In this hypothetical example, the initial/direct impact is the $1000 worth of output produced and 5 workers hired by car manufacturer. The 10 units of car parts and additional steel/metal is the output produced and 5 is the number of workers hired by suppliers and distributors. These are categorized as Indirect effects.

In our analysis, we decompose the total impact into direct and indirect effects. These are defined as follows:

22_{This impulse can be visualized as the following.}

If there are 3 sectors (Sector A, B & C) in the economy and there is an investment of Rs. 20 lakhs in Sector C, then the ∆𝑌 matrix looks like: ∆𝑌 =

0 0 20

.

(29)

28

 The direct effect is the change in the economic variable (in our case, output and

employment) of the sector where the investment takes place.

For example, the total number of jobs supported in the Medical and Health sector from a $68 million investment in a particular hospital chain is the direct effect on employment of this investment.

 The indirect effect is the change in the economic variable (in our case, output and employment) of the suppliers and distribution chain partners of the sector where the investment takes place. It is to be noted that, our I-O analysis computes all the production induced effects, i.e., effects which are caused due to the increase in demand for factors of production.

For example, the total number of jobs supported in the pharmaceuticals’ sector, construction sector, etc. due to an investment of $68 million in a particular hospital chain is the indirect effect. These effects can be calculated for a large group of variables such as output, employment, value added, income, imports, etc. However, the focus of this thesis is restricted to output and employment.

3.3.3. Calculating the Input-Output Multipliers

As stated above, calculation of I-O multipliers requires the computation of the Leontief Inverse Matrix. This matrix is important because an I-O analysis is based on production linkages of the economy and quantifies economic impact keeping output as the cause of changes in other economic variables (production-induced).

The rationale behind the importance of the Leontief Inverse Matrix to the I-O analysis is purely an arithmetical one. Using series approximations, we can write it as,

(𝐼 − 𝐴) = 𝐼 + 𝐴 + 𝐴 + 𝐴 + ⋯ = 𝐴

Therefore,

∆𝑋 = (𝐼 − 𝐴) ∆𝑌 = (𝐼 + 𝐴 + 𝐴 + 𝐴 + ⋯ )∆𝑌 Where, 𝐼∆𝑌 are the direct effects

𝐴∆𝑌 are the First-round effects

(30)

29

In our analysis, we look at the output multiplier and the employment multiplier. Before going to the calculation, below is a table of the various effects that will be calculated in our analysis.

S. No. Effect Definition

1. Direct Effect Change in output/employment in sector 𝑗 caused due to investment in sector 𝑗 .

2. Indirect Effect Change in output/employment in sector 𝑖 (𝑖 ≠ 𝑗) caused due to investment in sector 𝑗 .

3. Initial Effect Change in final demand caused due to investment. 4. First Round Effects Change in output/employment in all the sectors to produce the Initial effect. This is the first round of indirect effects. For eg. In the Automobile example, the change in output/employment in the manufacturer of car parts is measured with this variable.

5. Industrial Support Effects Change in output/employment in all the sectors after the First-Round effects. This is the sum of all the subsequent rounds of Indirect effects. For eg. Change in output/employment in the metal/steel industry, etc. 6. Simple Multiplier Sum of Direct and Indirect effects. This is also defined

as the sum of Initial effects and all the subsequent round of effects.

(31)

30

3.3.3.1. Output Multiplier

The output multiplier for sector 𝑗 refers to the total value of output in the entire economy required to meet a ₨. 1 worth of final demand for sector 𝑗’s output (Miller & Blair, 2009, p.245). The I-O analysis proceeds by first calculating the initial effect followed by the direct effect and then the subsequent rounds of effects which are called the production induced effects as these effects are caused by increase in the investee’s demand for inputs to produce its output.  The direct output effect of a change in final demand for sector 𝑗’s production is always more than or equal to 1 (in this case, Rs. 1). This is because an initial impulse of Rs. 1 in a sector will require that sector to produce Rs. 1 worth of output plus any extra output required to be produced due to increase in demand for 𝑗’s output as input by other sector.

 The First-Round effect for sector 𝑗 is calculated as

𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠 = 𝐴 ∗ ∆𝑌

where, ∆𝑌 is the impulse matrix which measures the change in final demand (caused due to investment) in sector 𝑗. This effect can also be measured as the column total for sector 𝑗 in the Direct Requirements Matrix (in our case, 𝐴 matrix).

 The Output Multiplier or the total effect on the output of the economy due to a change

in final demand in sector 𝑗 is calculated as

(𝐼 − 𝐴) ∆𝑌

 After calculating the output multiplier, we can calculate the Industrial support effects which is he increase in output caused by indirect suppliers & distributors to the investee,

𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑖𝑎𝑙 𝑆𝑢𝑝𝑝𝑜𝑟𝑡 𝑒𝑓𝑓𝑒𝑐𝑡𝑠

= 𝑆𝑖𝑚𝑝𝑙𝑒 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟 − 𝐼𝑛𝑖𝑡𝑖𝑎𝑙 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 − 𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠  The production induced effect is calculated as the total output generation in the entire

economy to meet the input requirements for producing the initial effect. 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝐼𝑛𝑑𝑢𝑐𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠 = 𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 +

(32)

31

3.3.3.2. Employment Multiplier

The employment multiplier of sector 𝑗 refers to the total number of jobs supported in the entire economy required to meet a Rs.1 worth of final demand for sector 𝑗 s output. Here in, the I-O analysis proceeds similar to the output multiplier, albeit with some changes.

 The first step is to make an employment coefficient matrix. This matrix is constructed by dividing employment per sector with the total value of output of the sector. The result is employment coefficients for each sector which quantifies the total number of workers required to produce Rs. 1 additional worth of output. Such calculation will yield a row matrix where the 𝑖 cell (in our case 𝑗 ∈ {1,2,3, … , 52} ) represents the employment coefficients of sector 𝑖. This process is illustrated as follows,

 Let number of employed persons per sector be 𝐸 .

Therefore, the employment coefficients can be represented as 𝑒 = The matrix will then be: 𝐸 = [𝑒 … 𝑒 ]

 The employment coefficient matrix is transposed to give a vector for matrix calculations.

𝐸 = 𝑒

⋮ 𝑒

 The direct employment effect or the number of jobs supported in sector 𝑖 due to Rs. 1 worth of increase in final demand in sector 𝑖 is represented through the employment coefficient24_,

𝐷𝑖𝑟𝑒𝑐𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 = 𝑒

 The First-round employment effect for sector 𝑖 is the increase in employment in the immediate suppliers and distributors of the investee which is given as,

𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 = (𝐴𝐸) ∗ ∆𝑌

 The Employment multiplier or the total number of jobs supported in the entire economy due to a change in final demand in sector 𝑖 is calculated as

[(𝐼 − 𝐴) 𝐸] ∗ ∆𝑌

24_{It is to be noted here that in case of employment multiplier, the direct effect is equal to the initial effect caused}

(33)

32

 The Industrial Support effects is the number of additional jobs supported in the

indirect suppliers and distributors of the investee,

𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑖𝑎𝑙 𝑆𝑢𝑝𝑝𝑜𝑟𝑡 𝐸𝑓𝑓𝑒𝑐𝑡𝑠 = 𝑆𝑖𝑚𝑝𝑙𝑒 𝐸𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟

− 𝐼𝑛𝑖𝑡𝑖𝑎𝑙 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 − 𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡𝑠  The production induced effect is the total number of jobs supported in the entire

economy to produce the additional output (initial effect) of the investee. 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝐼𝑛𝑑𝑢𝑐𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠

= 𝐹𝑖𝑟𝑠𝑡 𝑅𝑜𝑢𝑛𝑑 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑖𝑎𝑙 𝑆𝑢𝑝𝑝𝑜𝑟𝑡 𝐸𝑓𝑓𝑒𝑐𝑡𝑠

The basic undertaking of the I-O model is that it provides the reader with deterministic results and helps us understand the broad structure of the economy. The methodology described above is the fundamental iteration of an I-O model.

(34)

33

4. DATA DESCRIPTION

This section specifies the various sources used to collect data for our I-O analysis along with the various assumptions undertaken to create the I-O table. Section 4.1 provides information about the I-O Table and its derivation. Section 4.2. will highlight the data sources from where the employment statistics are compiled and appropriated.

4.1. I-O Table

This study uses the most recent National Input-Output Table for India for FY 2013-14, published in 2016 25_{. This table shows inter-sectoral linkages between 130 sectors of the Indian}

economy as a whole. All values in this table are in terms of Rs. 100,000. In the results and analysis chapter, a conversion of the results is made in terms of US Dollar ($) using the exchange rate US $1= Rs.66.1467 (average exchange rate for FY 2016-17). The high level of disaggregation of this table (economy divided into 130 sectors) led to problems in finding data for employment per sector for 130 sectors and hence aggregation or grouping up of similar sectors to compute an aggregated sector was done. For instance, in the 130x130 I-O table, there were separate sectors for agricultural crops such as paddy, maize, bajra, wheat, fishery, etc. for which the employment data was not available. However, employment data was available for the agriculture sector as a whole. Hence all the sectors related to the agriculture sector were grouped together. Similar aggregation was done for a number of other sectors. This aggregation was done on the basis of the National Industrial Classification (NIC) Code, 2008 and National Accounts Statistics for comparison and uniformity with international classification procedures. Hence this matching/aggregation provides legitimacy as it based on both the National Accounts Statistics and the official industry classification authority in India.

The NIC Code 2008 is the benchmark statistical standard for classification of all industries/sectors in India. Developed and maintained by the Central Statistical Organization (Government of India), the NIC Code 2008 also helps in international standardization and comparison of all the sectors in India as it is based on International Standard Industrial Classification (ISIC, Rev. 4) which is published by the U.N. Statistical Commission. The aggregation of sectors resulted in a new Input-Output Table, an inter-industry 52x52 matrix

25_{This table is produced by the National Council for Applied Economic Research (NCAER), India and is freely}

(35)

34

(see Appendix 1). This means that the economy of India will be represented through 52 sectors as against the 130 sectors which were provided by the National I-O Table 2013-14.

It is important to note here that although the IFC health investment was done in 2017, we are using the national I-O Table for the year 2013-14 which shows flows between sectors on the basis of the economic condition prevailing in India during FY 2013-14. The rationale for this assumption is based on a comparison between output multipliers for the I-O tables published for the year 2006-07 and 2013-14 (see Appendix 6). It was found that on an average only 7 out of 130 sectors had more than 0.5 as the difference in output multipliers and only 26 out of 130 had less than -0.5 as the difference in output multipliers. This means that although some sectors improved in sectoral-linkages (and some worsened), the difference was minute. Hence the output multipliers remained fairly constant during the period between 2006-07 and 2013-14 and hence we assume that it stays constant for FY 2016-17. Hence, even after 2 years of the creation of the I-O table, it can be used to investigate the impact of an investment. Again, this is a simplifying assumption which helps us to use the I-O table 2013-14 until the next and more updated version is available.

4.2. Data for employment per sector and health workforce composition

To compute the employment multipliers, I-O table requires statistics on absolute employment per sector. This data was obtained from the Report on Fourth Annual

Commercial investment for development : analysis of output and employment effects of IFC's investment(s) in India using an input-output model