India’s Primary Energy Evolution: Past Trends and Future Prospects Kaushik Deb Group Economics, BP plc

Paul Appleby Group Economics, BP plc

India Policy Forum July 14–15, 2015

NCAER | National Council of Applied Economic Research 11 IP Estate, New Delhi 110002 Tel: +91-11-23379861–63, www.ncaer.org NCAER | Quality . Relevance . Impact

India’s Primary Energy Evolution: Past Trends and Future Prospects* Kaushik Deb Group Economics, BP plc Paul Appleby Group Economics, BP plc

India Policy Forum July 14–15, 2015 Abstract The large increases in energy consumption over the past 30 years that have accompanied rising population and economic growth in India have been shaped by a variety of environmental factors and policy choices. The shifts in energy mix, away from coal and towards oil until 2000s, and the subsequent recovery in coal’s share, followed by a period of competition between coal and gas, do illustrate the potential for change. This change is most affected by changes in domestic production, especially of gas and renewables. However, India nonetheless faces a stubborn energy mix that is unlikely to change very much over the next two decades, with growth in population and GDP driving up energy consumption. This growth in demand calls for increased consumption of fossil fuels. Going forward, India’s primary energy consumption is expected to grow at a rate outpacing most large developing countries. Coal will continue to dominate the energy mix, though it will lose some market share to gas and renewables. Other implications of this slow moving energy mix are observed in overall energy intensity and in carbon emissions. While India’s energy and emissions intensities have declined over time, these gains are mostly due to improving energy efficiency. However, with an inflexible energy mix, the environmental gains from improving the share of more energy and carbon efficient fuels would remain limited. More significantly for India, domestic production has been sluggish in responding to energy demand growth, and imports are likely to continue rising. This increased share of energy imports as a percentage of GDP would place a significant burden on the economy. JEL classification: Q47

Keywords: Energy demand, India, Energy intensity *Preliminary draft. Please do not circulate beyond the NCAER’s preparations for the India Policy Forum 2015, for which this paper has been prepared.

[email protected]; [email protected]

Kaushik Deb and Paul Appleby 1 The large increases in energy consumption that have accompanied rising population and economic growth in India have been shaped by a variety of environmental factors and policy choices. India faces a stubborn energy mix that is unlikely to change very much over the next two decades, with growth in population and GDP driving up energy consumption. This growth in demand calls for increased consumption of fossil fuels. In addition, despite growing domestic production of both fossil and non-fossil fuels, imports would continue to rise. This is likely despite an expected rapid ramp up in renewable energy and nuclear power generation, apart from gains in energy intensity. This paper presents a forecast of India’s energy demand based on the results of the BP Energy Outlook 2035 (BP, 2015). The following section briefly describes the growth in India’s primary energy demand and supply since 1980. Next is a brief review of the prevalent energy demand modelling techniques and the existing forecasts of energy demand in India and the world. The approach and methodology used in the BP Energy Outlook 2035 is then presented, followed by the results for India in the base case, and two alternative scenarios. The two alternative scenarios are constructed to assess the outcomes of a higher GDP growth world; and that of a greater penetration of renewable energy and higher energy efficiency. The impact of the three scenarios is described in terms of energy demand, carbon emissions, and the import dependency.

1. Energy Demand Growth in India From a previous three decade average annual growth rate of 3.7%, the 1980s saw India’s GDP growth rate rise much faster (DeLong, 2003). During this period, the increase in the industrial activity, contributing to the manufacturing sector GDP, was higher than the increase in the GDP (Figure 1). Figure 1: GDP growth rate and share of manufacturing in GDP in the 1980s 9%

15.5%

6%

15.0%

3%

GDP growth rate

14.5%

Share of Industry 0%

14.0% 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990

Source: (Oxford Economics, 2015)

2 India Policy Forum 2015 This resulted in primary energy consumption rising by 5.9% p.a. The growth was led by coal and oil, meeting the 49.5% and 33.6% of the growth in demand respectively (Table 1). While natural gas consumption also increased during this period, it did so from a very small base. Combined with a more than 150% increase in coal based electricity generation capacity, the increase in coal consumption by 5.4% p.a. allowed coal based electricity generation to increase by 12% p.a. during the decade (Central Electricity Authority, 2015). Table 1: Increase in Energy Consumption during 1980s 1980-90 Primary energy (mtoe) 78.22 out of which Coal 49.5% Oil 33.6% Gas 12.5% Hydro 3.2% Nuclear 1.1% Source: (BP, 2015)

In oil and gas, the increase in consumption was supported by the first significant forays into E&P (exploration and production). Domestic oil production, following the significant offshore find in Bombay High in 1974, picked up during this period along with new oil discoveries in the Godavari basin and Assam, and gas in Western India. Both oil and gas proved reserves1 saw some of the fastest growth in energy reserves in India’s history, with proved oil reserves growing by 7.3% and gas by 7.4% p.a. Domestic oil production rose by an average 6.3% p.a. during this period, while gas production increased by more than 9 times, from a negligible base. Overall, oil and gas contributed more than a quarter of the increase in total energy production during the decade (26.6%). The vast majority of the increase in energy production though came from coal (70.8% of the increase in total energy production during 1980-90). The contribution of non-fossil fuels, including renewables, was marginal, adding up to just 3%. Interestingly, oil imports grew very slowly during this period. The increase in oil production kept up with the growth in consumption, leading to only a very gradual increase in oil imports by just 1% p.a. (Figure 2) 2. This is particularly remarkable given that the international price of oil fell sharply during this period, and at its lowest was less than 40% of the decade’s high. Thus, even when faced with an extremely competitive alternative, domestic production was able to meet most of the domestic demand. The significant E&P activities of the previous decade clearly helped boost energy consumption in the 1980s.

BP Statistical Review of World Energy defines proved reserves as ‘Generally taken to be those quantities that geological and engineering information indicates with reasonable certainty can be recovered in the future from known reservoirs under existing conditions.’ 2 Net imports here are defined as the difference between consumption and production, ignoring stock-building and losses in transit. 1

Kaushik Deb and Paul Appleby 3 Figure 2: Growth in Oil Market in the 1980s

Source: (BP, 2015)

After the sharp increase in GDP in 1980s, growth started to fall from a high of 8% in 1988 to just 2% in 1991. This slowdown in GDP growth and in industrial production was also mirrored in primary energy consumption (Figure 3). The slowdown was more pronounced in oil and gas, reflecting the sharper decline in industrial activity while electricity generation (mostly coal based) ramped up. Annual growth in oil consumption slowed down to 5.4% during 1990-95 from 6.2% during the 1980s while gas consumption grew by 9.3% annually compared to over 26% in the 1980s. In comparison, coal consumption grew slightly faster at 5.5% p.a. during 1990-95 compared to 5.4% p.a. during the earlier decade. Overall, coal met 53.1% of the increased energy consumption during this period while the share of oil and gas combined fell to 42.1% from 46.1%. Non fossil fuels also pushed up their share to just less than 5% of the energy mix on the back of new capacity addition for hydroelectric generation. Figure 3: Change in IIP and Fossil Consumption (1990-2000)

Source: (BP, 2015) (Oxford Economics, 2015)

4 India Policy Forum 2015 As GDP growth recovered in the latter part of 1990s, energy consumption growth continued to slow down. This slowdown mirrored a gradual maturity of the economy with a larger share of services and less energy intensive industry, as evidenced in a slower growth in manufacturing industry and IIP (Figure 3). In particular, coal consumption growth, the primary driver in the early 1990s, stagnated in this period growing by only 2.9% p.a. during 1995-2000. In addition, with a slowdown in coal and gas consumption growth, electricity generation growth during this period also declined to 6.2% p.a. The only positive in electricity generation during this period was a ramp up in nuclear power generation, rising by 15.6% p.a. in these five years. On the other hand, a higher GDP growth and the fastest per capita growth in GDP over the previous two decades, pushed oil consumption growth to a high of 7.1% p.a. As a result, oil overtook coal as the lead contributor to energy demand, meeting nearly 52% of the growth in consumption (Table 2). Amongst non-fossil fuels, nuclear overtook hydroelectricity as the principal fuel in terms of increments. Table 2: Growth in energy consumption in the 1990s (million tonnes of oil equivalent) Oil Gas Coal Nuclear Hydro Renewables Total

1990-95 17.3 6.1 29.5 0.3 2.2 0.2 55.5

1995-2000 30.9 6.8 19.3 1.8 0.2 0.5 59.6

Source: (BP, 2015)

Production collapsed during the 1990s though. A decade of underinvestment as well as no incremental E&P in this decade led to oil production stagnating, even as gas production continued to grow very slowly. Proved reserves also stagnated with proved oil reserves declining by almost 5% during the decade while gas reserves increased by only 8.5% in 10 years. As a result, the marginal increase in oil production in the first half of the 1990s (1.4% p.a. during 1990-95) was completely reversed by declines in the second half (-1.4% p.a. during 1995-2000). The rate of growth of gas production also slowed down from a high of 26.2% p.a. in the 1980s to just 8.1% p.a. in the 1990s. Coal production growth also slowed down to 4.1% p.a. Overall, energy production increased by only 3.3% p.a. over the entire decade compared to 5% p.a. growth in energy consumption. The more rapid growth in energy consumption compared to production, particularly for oil, was supported by a weak international oil market that allowed oil imports to rise without being crowded out by domestic production as happened in the 1980s (Figure 4). Oil imports tripled during this decade from less than 500 Kb/d in 1990 to 1.54 Mb/d in 2000, while Brent oil prices rose by just 20%. Coal imports also tripled during that decade though remained at modest levels; the share of domestic coal consumption met by imports increased from under 4% to 8.3%.

Kaushik Deb and Paul Appleby 5 Figure 4: Oil imports and Brent Prices 1600

$/bbl 30

Kb/d

1200

23

800

15

400

Oil imports

Brent price

0

8 0

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Source: (BP, 2015)

Pursuant to some policy reforms in the oil and gas sector, proved reserves for oil and gas started to increase again after a decade of inactivity and underinvestment. Oil reserves rose by just under 12% during this five year period while gas reserves rose by a remarkable 45%. However, production growth continued to languish, still suffering from a long lull in investment and exploration activity. Oil production grew by a marginal 0.5% p.a. while gas production growth slowed down to just 2.4% p.a. Coal production growth also slowed down to 4.2% p.a. Still, the more modest slowdown in coal production compared to all other fuels implied that it contributed nearly three quarters of the growth in energy production. Overall, energy consumption growth remained muted even as GDP growth recovered slowly. Consumption growth of all fuels slowed down except for hydroelectricity growing by 4.9% p.a. during 2000-05 (hydroelectricity had stagnated during the previous two decades growing by only 1.7% p.a.). The recovery in GDP growth rates during the first decade of the 21st century, were harbingers of a coming growth spurt. India’s energy consumption increased by 6.8% p.a. during 2005-10, and India accounted for 4.2% of the world’s energy consumption by 2010. This period was characterized by particularly strong growth in gas consumption, and gas met an even larger share of growth in consumption. Gas consumption grew by 11.9% p.a. meeting 17% of the increase in consumption, up from under 12% earlier in the century. This was at the expense of coal, which despite stronger growth than the previous 5 years at 6.1%, lost with its share in incremental consumption falling to 61.3%. Oil consumption growth also picked up to 5% p.a. buoyed by rising per capita GDP. Hydroelectricity, which grew rapidly in the early years of the century on the back of new capacity addition, also grew more slowly bringing down the contribution of non-fossil fuels to 6.7% of the increment in demand from 9.3% during the previous five years. Coal (51% of total consumption) remained the dominant fuel in India. Oil (30.5%) was the second largest fuel, with natural gas (11.1%) and non-fossil fuels (7.4%) far behind (Figure 5).

6 India Policy Forum 2015 Figure 5: Share of Energy Consumption (2000-2010) 55% 44% 33% 22%

Coal

Gas

Oil

Non fossil

11% 0% 2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Source: (BP, 2015)

This growth in energy consumption found support in a rapid increase in production with a growth rate of 6.3% p.a. compared to just 3.6% during 2000-05. Boosted by the deregulated E&P regime and investment in exploration activity, gas production growth jumped by to 11.9% p.a. during this period, and contributed 21.1% of the increased energy production in the country. This was also mirrored by oil production to a lesser extent with production growth recovering to 3.4% p.a. Coal production also started to recover, rising to 6.1% p.a., supporting a faster growth in electricity generation at 6% p.a. The contribution of non fossil fuels also rose to 10.6%. This fed the growth in energy consumption that in turn fuelled faster GDP growth. After having emerged mostly unscathed from the 2008-9 financial crisis, a surprising policy paralysis affected the economy after 2010, and. GDP growth slowed down sharply and industrial activity fell. This policy paralysis manifested itself in the energy sector in two ways. First, it stalled new E&P activity with a variety of unresolved issues related to licensing, price regulation, and other regulatory and administrative issues. India’s energy production fell to just 1.6% p.a. compared to 5% p.a. during the previous decade. This slowdown is attributable to gas and oil, which fell to 19.3% of total energy production in 2014 from 25.4% in 2010 (Figure 6). Oil production growth fell to just 0.4% p.a. during this period while gas production started to decline sharply by 11.2% p.a. Coal production, kept growing by 2.9% p.a., even though it was lower than earlier. In fact, coal production in India fell twice in these five years, the first volumetric declines since 1999. Even so, coal regained its share in energy production having been slower to moderate growth than other fuels. Over the last few years, oil production has essentially stalled while gas production is declining with mature fields not being supplanted by new discoveries and production.

Kaushik Deb and Paul Appleby 7 Figure 6: Change in Energy production (2005-14) 100 Mtoe

Renewables Nuclear Gas

80 60

Hydro Coal Oil

40 20

0 -20 -40 2005-10

2010-14

This decline in energy production growth has had significant effects. Gas consumption declined (-5.2% p.a.), though more slowly than production, with coal consumption rising sharply (8.5% p.a.) to fill this space (Figure 7). However, the increase in coal production was significantly lower than the increase in coal consumption. As a result, coal imports increased sharply to nearly twice of the 2010 levels. This has allowed electricity generation to continue rising at 7% p.a. even with coal and gas production not keeping up. There have been increases in gas imports as well by 12.4% p.a. with India emerging in the top 5 LNG importers in the world. In general, coal consumption and imports have responded quickly to changes in gas availability, thus balancing the energy markets. Oil consumption growth also slowed as the commodity super cycle intensified. This reversal took India back to the 1980s in terms of fuel shares. Figure 7: Change in Energy Consumption (2005-14) 160 140 120 100 80 60 40 20 0 -20

Mtoe

2005-10 Renewables

2010-14 Hydro

Nuclear

Coal

Gas

Oil

8 India Policy Forum 2015

1.1 Implications Over the last few years, even with economic growth slowing down in India, energy consumption has been robust. GDP growth slowed down from 7.4% p.a. during 2000-10 to 6.1% p.a. in 2010-14, while energy consumption growth is up from 5.6% p.a. to 5.8% p.a. As a result, improvements in energy intensity of GDP also slowed down. The fuel mix had implications for CO2 emissions from energy use as well. More significantly, the sharper slowdown in domestic production compared to consumption implied that the share of India’s energy consumption met by domestic sources fell to 57% by 2014, the lowest on record.

1.2 Unchanging Energy Mix The energy sector is generally slow moving and changes in consumption and production profiles are a result of lumpy investment decisions and also only gradual improvements in efficiency. As a result, India’s energy sector profile appears largely unchanged since 1980, with coal and oil dominating the energy mix. This broad average, however, masks a significant shift away from coal and towards oil until 2000, and the subsequent recovery in coal’s share in the early part of this century. The competition between coal and oil in the last century has now been played out between coal and gas over a much shorter period. A rapid rise in gas consumption during 2005–10, followed by a decline from 2010–14, was offset by equivalent changes in coal consumption. However, an increase followed by a decline in gas production during the same intervals was not matched by an equivalent trend in coal production.

1.3 Rising Import Dependency India’s net energy imports increased by 5.5% p.a. during 2010-14, compared to 4.8% p.a. in the first decade of the century. While coal led the trend with imports rising by 28.6% p.a. during this period, gas and oil imports rose by 12.4% p.a. and 4.9% p.a. respectively. Underlying this rapid increase in coal imports is the trend in domestic gas production during this period described above and the tight Asian LNG market. As domestic gas production collapsed in 2010, energy demand shifted to imports. LNG, on the other hand, entered a three year supply growth lull in 2011 along with the Fukushima nuclear disaster pushing Asian demand (and prices), to record highs and making gas imports much more expensive than coal imports. The result was this dramatic increase in coal imports during 2010-14.

1.4 Higher Energy Intensity and Emissions from Energy Use Energy markets are sluggish in their response to economic drivers. As a result, while GDP growth slowed, energy consumption growth remained stable, thus slowing improvements in India’s energy intensity as well. From a decline of 1.6% p.a. during 2000-10, energy intensity fell by only 0.3% p.a. during 2010-14. These gains in the first period were when GDP growth was much faster than the increase in energy consumption; GDP rose by 7.4% p.a. during 2000-10 while energy consumption increased by 5.6% p.a. Therefore, improvements in energy intensity were 1.6% p.a. In the following four years (2010-14) GDP growth came down to 6.1% p.a. while energy

Kaushik Deb and Paul Appleby 9 consumption growth increased to 5.8% p.a. This led to energy intensity decreasing by 0.3% p.a. The growth in CO2 emissions from energy consumption accelerated in India – from 5.6% p.a. during 2000-10 to 6.2% p.a. during 2010-14. This implies that the carbon intensity of energy consumption was broadly unchanged during 2000-10, but then increased over the 2010-14 period as coal has gained shares from gas rapidly.

2. Energy Modelling Approaches Modelling techniques in the energy sector are mostly applied to modelling energy demand, while energy supply models are usually simple aggregation of individual supply sources. Each supply source in turn would be projected based on the resource availability and likely utilization rates. The likely decline rate of fossil fuels supply source also brings another factor to the forecasting supply model The literature on energy demand modelling is rich and varied. Econometric top down modelling and bottom up end-use modelling are the two prevalent approaches for modelling the energy sector. The end-use or engineering models take a ‘bottom-up’ approach and estimate demand based on equipment saturations, efficiencies, and usage. This section provides a very brief description of the more widely applied energy demand modelling techniques, followed by a review of the recent energy sector modelling exercises in India, the latter using the survey reported in (Navroz K Dubash, 2015).

2.1 Econometric Modelling Top down econometric models are aggregate models of the entire economy based on past trends to predict the relationship between the sectors of the economy (IPCC 2001). Technological change is usually incorporated exogenously in the model rather than explaining it within the model. For example, Autonomous Energy Efficiency Improvement describes the rate at which sectors require less energy over time to produce a given level of output (MIT 1997). Econometric models would then apply differing rates of change among sectors and regions based on historic data or projected assumptions. Thus, in general, top-down models are useful for forecasting in cases where historical development patterns and relationships among key underlying variables hold constant for the projection period (Hourcade, J C et al. 1996). Top down models are often used when there is little available information on equipment or appliance stocks (Zarnikau, J 2003). The econometric methods require a consistent set of information over a reasonably long duration and are hence easier to execute for individual sectors. There is a large literature on develops specific functional forms for econometric analysis of electricity demand in the short-run (Batancourt, R R, 1981). Other papers have examined the demand for electricity in India by developing a logarithmic linear econometric model, wherein a relation between electricity consumption and variables like income, price of electricity, price of diesel, etc. was developed to estimate the short-run and long run elasticities (Bose, R K et al. 1999). Systems based approaches include residential, industrial, and total electricity demand in the United States by a partial-adjustment approach and by a simultaneous equation approach (Kamerschen, D R et al. 2004). In estimating residential energy demand, energy prices, disposable income and other attributes of the consumers are usually

10 India Policy Forum 2015 incorporated. Crucial here, is having a long enough dataset that allows for sufficiently large number of variables and rich functional forms to be used.

2.2 Bottom-up Modelling Bottom-up modelling focuses on counting equipment and stocks, and adding up the energy consumption by analysing the efficiency and frequency of use of the equipment. These models allow for more comprehensive analysis by aggregating demand across sectors, regions, and fuels. Total energy demand is then a product of activity levels and, energy intensity (energy demand per unit economic output) or process efficiency (energy demand per physical output). These models incorporate development of new technology and processes that improve the efficiency of energy use equipment and usually forecast demand based on engineering costs of a wide range of technologies (IEA 1997; IPCC 2001). Bottom up models often make use of the stock vintage concept to project demand or supply into the future. In other words, they model future energy use based on costs, timing and market shares of technologies or equipment/stocks. These models often use simulation and backcasting to project into the future and are most suited in cases where there are new technologies penetrating into the market, or new policies changing preferences and behavioural patterns. However, such models are weak in incorporating feedbacks between structural evolution of a particular sector and the overall economic development pattern, such as the influence of non-energy consumer behaviour and changes in the size and spread of various sectors of the economy.

2.3 Energy Modelling in India Energy modelling and analysis in India in general, has been bottom up optimization (Sengupta, R 1993; Shukla, P R et al. 1997a; Pandey, R 1998). Several bottom-up and input output studies though aggregate in nature, have modelled energy at a sector level ((Shukla, P R et al. 1997a), (Pandey, R 1998), (Tiwari, P 2000) (Pachauri, S 2002)). The AIM-end-use model (Shukla, P R 1996) has been set up for forty years horizon from 1995 by minimizing the discounted energy system costs at end-use subsector levels. The MARKAL model, originally developed for Canada (Berger, C et al. 1992), has also been adopted for India (Shukla, P R et al. 1997b; Pandey, R 1998). Garg integrates the top-down AIM-End-use model with the bottom-up MARKAL model to provide insights into the implications of mitigation commitments on energy and technology mix, energy costs, mitigation costs and competitiveness of Indian industries (Garg, A et al. 2001). A stochastic Indian MARKAL model (Loulou, R et al. 1997) reflects long-term uncertainties in technology and fuel substitution within the Indian context. Most studies focus only on technical and economic variables influencing energy use, without including other social, demographic and structural transitional dynamics at the household level that impact energy use and development, particularly in a developing country (Pandey, R 2002).

Kaushik Deb and Paul Appleby 11 Table 3: Recent energy modelling studies in India Study Expert group on Low Carbon Strategies

Approach Activity Analysis Model

Type Top down

The Energy Report - India

MARKAL location Model Integrated Assessment Model Integrated Energy Model

Bottom up

Energy-Emissions Trends and Policy Landscape A Sustainable Development Framework for India's Climate Policy Energy Intensive Sectors of the Indian Economy India Energy Security Scenarios

Hybrid Bottom up

World Energy Model Bottom up Excel based simulation

Bottom up

(Planning Commission, 2014) (TERI, 2013) (P R Shukla, 2015) CSTEP, 2015) (The World Bank, 2011) (Niti Aayog, 2015)

The review quoted above, and most of the studies were motivated by climate change and environmental considerations. A number of these studies also highlight issues around energy security, both nationally and as well as in the form of household access to energy. For instance, the objectives of the modelling aspects of studies such as (The World Bank, 2011) and (TERI, 2013) are to quantify the impact of sector specific mitigation activities on GHG emissions. Energy security and the rising share of imports in meeting domestic energy demand is a consideration in (Planning Commission, 2014).

2.4 Contribution of this Study This forecast differs from these studies in three ways. First, being prescriptive in nature, most of these studies develop baseline and alternative scenarios as different cases, with the baseline being ‘more of the same’ type projection of the past, while alternative scenarios build in more aggressive policy reforms. One outcome of such a distinction is that each policy reform can be evaluated in terms of its impact. The baseline forecast in this paper, on the other hand, takes into account a likely policy reform process and builds in autonomous technology development, recognizing that policies do change and evolve over time, and natural progression of technology development works through the system. Thus, the targets for wind, solar, and coal production in India are recognized while making policy assumption for the future. Second, the current paper contextualizes India’s energy demand, especially energy imports, with the overall global demand and supply of energy. This allows for adjusting demand on a yearly basis depending on the excess demand volumes for oil, gas, and coal, and the availability of these resources for import in the regional markets. For the oil market, this assessment is based on the global oil balance, while for coal and gas, the regional supply scenario provides a boundary for import volumes. Although energy prices are not explicitly taken into account in the country level forecasts, this balance provides an indication of the tightness of the market.

12 India Policy Forum 2015 Finally, the India centric studies listed above, account for efficiency improvements and technology changes based on the domestic industrial conditions and the best in class experiences from the rest of the world. This forecast is global by design, and approaches efficiency improvements and technology changes based on global trends and their transmission through global trade. For instance, improvements in vehicle efficiencies are reproduced around the world with the automobile industry meeting demand in the OECD countries from manufacturing facilities in the non OECD countries that allows for countries like India to benefit from these efficiency improvements. Similarly, policy developments in the OECD countries such as strengthening EURO emission standards in the European Union (EU) imply non OECD manufacturing facilities need to improve product efficiencies.

2.5 Other Global Energy Models There are a number of other global energy projections that present energy consumption and supply forecasts, differentiated by regions and fuels, using the same approach as this paper. The key amongst these are the International Energy Agency’s World Energy Outlook, and Energy Information Administration’s International Energy Outlook, and the Shell Scenarios. Again, this list is only indicative of the vast literature and products available to researchers and analysts. The Shell Scenarios are described as ‘ask(ing) “what if?” questions to explore alternative views of the future and create plausible stories around them. They consider long-term trends in economics, energy supply and demand, geopolitical shifts and social change, as well as the motivating factors that drive change. (Shell, 2014)’ These present two opposing narratives where the demand and supply of energy is driven by a complex interplay of politics, economics, social development, and technology change. More significant than the demand and supply forecasts that result from the exercise, are impact that these alternative narratives have on growth paths. That helps identify the fault lines in the world of energy that governments and businesses need to be cognizant of. Compared to the Shell scenarios, the forecasts from the IEA and EIA are closer in approach to our paper. The International Energy Outlook of the EIA presented two oil prices and two global GDP growth cases in its last edition to examine a range of potential interactions of supply, demand, and prices in world liquids markets. The model adjusts energy demand and supply growth to balance the market in each scenario (EIA, 2013). The World Energy Outlook of the IEA contains three scenarios, and the IEA emphasises that none of the scenarios is a forecast – they are not designed to predict likely outcomes, but to explore possibilities under different policy assumptions (IEA, 2014). “New Policies Scenario” assumes that announced national policy objectives are fully implemented, while the “Current Policies Scenario” assumes very little change in policy. An aspiration forecast is also presented in the 450 Scenario that sets out an energy pathway consistent with the goal of limiting the global increase in temperature to 2°C by limiting concentration of greenhouse gases in the atmosphere to around 450 parts per million of CO2. IEA identify the New Policies Scenario (NPS) as their central case.

Kaushik Deb and Paul Appleby 13

3. Our Approach and Methodology The forecast presented in this paper does not rely on a single, all-encompassing, general equilibrium model of the global energy economy. Such models do exist, and they can be very useful in highlighting the interdependencies within the energy system, and identifying some of the potential unintended consequences of policy interventions. However their complexity and high maintenance cost (in terms of the time and data required to keep them up to date and calibrated against the real world) limits their usefulness as tools for forecasting a “most likely” outcome. The approach taken here, which is similar to the (IEA, 2014) and (EIA, 2013), is to apply a range of modelling strategies across different sectors and geographies, and then to aggregate the results in an accounting framework that ensures that everything balances. On the demand side, the forecast comes from a hybrid of top down econometric modelling and activity level models applied in conjunction with bottom up aggregation of energy consumption in individual units based on their utilization rates. The final outcome relies heavily on expert judgment, applied to the most up-to-date data that is available, with modelling tools used where possible to inform and support that judgment. Transport sector demand for fuels is modelled in two different ways. First, a technology-rich, “bottom up” model is used to simulate the evolution of the vehicle fleet, based on a range of parameters that enter into vehicle choice decisions, including importantly the constraints imposed on auto manufacturers in terms of the environmental performance of vehicles (e.g. CAFE standards in the US, tailpipe CO2 per km emission standards in the EU). Second, an econometric analysis of the relationship between transport fuel demand and incomes and fuel prices provides the basis for a “top down” projection, giving assumptions about the growth of income and changes in prices. The results from these two different modelling strategies are compared, and expert judgment is applied through an iterative process of discussion to agree to a final set of numbers for the projection of transport fuel demand. Some of the key results from the modelling exercise here are the following:   

The global vehicle fleet (commercial vehicles and passenger cars) more than doubles from around 1.2 billion today to 2.4 billion by 2035. Most of that growth is in the developing world (88%), while some OECD markets are already at saturation levels. Fuel economy gains are likely to accelerate over the Outlook, with vehicle fleet fuel economy forecast to improve by 2.1% p.a. between 2013 and 2035, having improved by about 1.5% p.a. over the past decade (2003-2013). Transport fuel demand will continue to be dominated by oil (89% in 2035), but the share of non-oil alternatives will increase from 5% in 2013 to 11% in 2035, with natural gas the fastest growing transport fuel (6.3% p.a.).

Industrial demand is based on the levels of economic activity represented by GDP, an assessment of the energy intensive sectors within each region, and availability and competition between alternative fuels. For instance, China’s industrial demand is based on the likely trends in the share of industry in GDP in China, competition between coal and gas for market share, and likely trends towards meeting energy intensity and GHG reduction targets in the economy. The fading impact of industrialization is apparent in the split of primary energy consumption by sector. Industry has been the fastest growing sector since 2000, averaging 2.7% p.a., but projected growth slows to 1.4% p.a.

14 India Policy Forum 2015 Power generation is the one sector where all fuels compete and so will play a major role in how the global fuel mix evolves. Demand for primary fuels for power generation is based on generation capacity augmentation in all regions and for all fuels, policy trends and regulatory changes likely to comply with GHG targets, and technology improvements in electricity generation. The latter is especially important for forecasting the generation of electricity from renewable sources where reduction in the cost of generation through achieving economies of scale, learning by doing, and autonomous efficiency improvements result in significant achievements. The outcome by 2035 is a more balanced and diversified portfolio of fuels for power generation. Coal remains the dominant fuel, accounting for more than a third of the inputs to power generation, but that share is down from 44% today and the gap between the shares of coal and of other fuels narrows significantly. Supply forecasts are more bottom up. The projection of nuclear power in India is a good example. This is based on current data on projects under construction and planned, and on announced policy targets, all subject to expert overview on the likelihood of plans being implemented and targets being met. Forecasts for the supply of fossil fuels are based on adding up likely decline rates of existing producing regions balanced with the likely demand for each fuel to mimic the price effect. These are then moderated by the proved resource base and expectations of changes in market structure due to policy reforms that may encourage greater E&P activity. For instance, the confluence of abundant resources with supportive policy and market structure (the world’s largest rig fleet, access to extensive pipeline networks and deep financial markets, etc) that led to the rapid growth in liquids production in the United States is unlikely to be replicated as widely anywhere else in our forecasts. Given the considerable inertia in the energy system, the long lives of assets and the long lead times on new builds, the key ingredients for a good forecast are the most up to date data, to establish the starting point and initial momentum of the system, and people with deep knowledge of how the individual parts of the system work and how they connect with one another.

3.1 Comparing with other Forecasts TABLE 4 contains a comparison against other publicly available projections. Our outlook is within the range of publicly-available forecasts. We see weaker growth in OECD energy demand than others, while our projections for non-OECD growth are stronger than most. Comparing to the IEA scenarios (which are not forecasts but assessments of potential outcomes based on defined sets of policy assumptions) our outlook lies above the IEA’s “New Policies Scenario”, which assumes that announced national policy objectives are fully implemented, and close to the “Current Policies Scenario”, which assumes no change to existing polices. This probably reflects our differing views on the outlook for rapidly industrialising economies, and the speed with which China in particular can move to a less energy-intensive growth path.

Kaushik Deb and Paul Appleby 15 Table 4: Comparison to other global forecasts3

World Oil Biofuels Gas Coal Nuclear Hydro Renewables

Primary Energy Consumption growth rate BP IEA NP IEA CP EIA 1.5% 18.8% 1.5% 1.6% 0.8% 10.0% 0.9% 0.9% 3.2% 24.6% 1.4% 1.9% 22.6% 1.8% 1.7% 1.1% 10.7% 1.6% 1.6% 1.9% 45.9% 1.9% 2.7% 1.8% 36.3% 1.8% 7.0% 207.4% 1.8% 2.7%

Shares in 2035 BP 100.0% 27.4% 0.7% 26.4% 27.1% 4.9% 7.1% 6.4%

IEA NP 100.0% 28.0% 10.7% 22.7% 26.0% 6.3% 2.8% 3.5%

IEA CP 100.0% 29.1% 10.0% 21.6% 29.2% 5.4% 2.5% 2.3%

EIA 100.0% 28.5% 0.0% 22.8% 27.9% 6.9% 0.0% 14.0%

3.2 Building Blocks Population growth and increases in income per person are the key drivers behind growing demand for energy, so the assumed path for these variable is a critical input to the outlook. The population projections are taken directly from the United Nations Population Division, Revision 2010. By 2035, India’s population exceeds 1.5 billion, which means an additional 250 million people will need energy. The economic growth assumptions are based on projections provided by Oxford Economic Forecasting, and sit well within the range of forecasts for the global economy that are available. The GDP numbers are expressed in real 2011 US dollars and Purchasing Power Parity (PPP) exchange rates. Using PPPs instead of market exchange rates to convert currencies makes it possible to compare the output of economies and the welfare of population in real terms (using the same prices for the same goods in all countries and all years). Over the outlook period, global GDP is expected to more than treble, with India contributing nearly 13% of the total world’s GDP growth. India’s per capita GDP increases by 166% (2013-35), growing at an average of 4.6% p.a. On this measure, by 2035 India’s income per person is just above where China is today, and less than half the current European Union level. Given those assumptions for population and income growth, both the level of energy demand and the fuel mix are heavily influenced by policy. The primary focus of India’s policy throughout the Outlook period remains securing affordable and reliable energy to support economic development. But there is also an increasing emphasis on clean energy, driven both by local environmental concerns and by India’s desire to play its appropriate part in addressing the global climate change issue. Some specific policy assumptions are illustrated in TABLE 5.

3

EIA forecasts are for 2010-35 and biofuels are included in renewables here.

16 India Policy Forum 2015 Table 5: Policy assumptions Current status Gas Administered prices with some incentives for E&P Coal Target of 1 billion tonnes of production by 2020 Solar Target of 100 GW by 2022 Wind Target of 100 GW by 2022 Nuclear Target of 63 GW by 2032

Policy assumption Gradual price deregulation by 2025 Coal production reaches 1 billion tonnes by 2027 80 GW by 2035 110 GW by 2035 27 GW by 2035

Given India’s stage of economic development, energy demand is expected to remain closely linked to economic growth. That linkage weakens gradually as the economy matures, and this weakening is reinforced by policy efforts to improve efficiency (including, for example, the removal of energy subsidies). The effects can be seen in the elasticity of energy consumption with respect to GDP (i.e. the ratio of energy growth to GDP growth), which has averaged 0.85 over the past decade and which declines to 0.72 over the next decade, and to 0.65 in the final decade to 2035. Table 6 provides more detail by fuel type. Table 6: Energy demand elasticity Demand Oil Gas Coal Nuclear Hydro Renewables Total GDP_PPP

Elasticity with respect to GDP 2015-2025 2025-2035 0.67 0.62 0.98 0.44 0.63 0.63 1.48 0.75 0.50 0.68 1.76 1.28 0.72 0.65

Annual growth rate 2015-2025 2025-2035 4% 3% 6% 2% 4% 3% 9% 4% 3% 4% 10% 7% 4% 3% 6% 5%

The potential for India’s own fossil fuel production to meet energy demand growth is constrained by its resource endowment. The forecasts of production are based on existing proved reserves, likely extraction and decline rates, and the prospects for finding and developing new reserves – all conditional on the expected policy reforms (Table 7). Table 7: Energy supply growth Oil (Billion bbl) Gas (Tcm) Coal (Billion tonnes) Total Source: (BP, 2015)

Proved reserves 5.74 1.43 60.60

2015-2025 -2% 1% 4% 5%

2025-2035 0% 5% 3% 3%

Kaushik Deb and Paul Appleby 17

3.3 Reference Case Based on those assumptions, India’s primary energy consumption is expected to grow by 128% between 2013 and 2035, achieving an average growth rate of 3.8% per annum (Figure 8). That is almost double the average rate of growth for non-OECD energy markets; India’s share of global energy demand rises to 8% in 2035, still some way behind China (at 26%), but ahead of Russia (5%), and Brazil (3%). Figure 8: India's Energy Consumption (2013-35)

Source: (BP, 2015)

Despite that rapid growth in total energy consumption, India’s per capita consumption of energy remains relatively low in 2035, less than half the global average in 2035. To put it into perspective, in terms of per capita energy use, India in 2035 is roughly where South Korea was in 1978 or Thailand in 1995. India’s energy intensity (the amount of energy consumed per unit of GDP) also remains low, declining by 1.6% per annum. India’s economic development is expected to be much less energy-intensive than China’s recent experience. India’s energy mix continues to evolve slowly, with fossil fuels accounting for 87% of demand in 2035, down from 92% today (and compared to a global average of 81% in 2035). Coal continues to dominate the energy mix, accounting for 51% of energy consumption in 2035, though it does lose some market share, notably to renewables and nuclear (Figure 9). Coal accounts for nearly half of the growth in India’s energy consumption by 2035.

18 India Policy Forum 2015 Figure 9: Shares of Primary Energy India

Source: (BP, 2015)

The consumption of fossil fuels more than doubles over the outlook, with natural gas up 145%, oil up 117%, and coal up 112%. Renewables and nuclear grow even more rapidly, expanding by 564% as and 363% respectively. Large-scale hydroelectricity shows the slowest growth, but still achieves a very respectable 98% increase in output. India’s own energy production meets just over half of the increase in energy consumption, growing by 117% (3.6% p.a.). Amongst fossil fuels, only coal is able to keep up with the growth rate of demand, with production expanding by 119%. Gas production also grows (+78%), but less rapidly than consumption; and oil production declines. Coal remains the dominant fuel produced in India with a 66% market share in 2035. Renewables in power overtakes oil as the second largest, increasing from 3% to 11% in 2035 as oil drops from 12% to 4%. India contributes the third largest increment to renewable energy generation during 2013-35 in the world (Figure 10). Figure 10: Renewable energy growth 2013-35

Source: (BP, 2015)

Kaushik Deb and Paul Appleby 19

3.4 Policy Implications The outlook described above poses two key policy challenges. First, with energy demand growth outpacing the expansion of domestic energy supply, India’s dependence on imported energy increases. India’s energy production as a share of consumption declines from 59% today to 56% by 2035 as imports rise by 143%. Oil imports rise by 161% and account for 61% of the net increase in imports, followed in volumetric terms by increasing imports of coal (+96%) and gas (+270%). This would place a significant burden on the macroeconomy as India currently consumes 4.3% of the total world oil consumption and 7% of the total trade in oil. By 2035, India’s share of the world’s oil consumption would rise to 7% while it would account for 11% of total imports. India’s share of the global LNG trade would increase from 6% today to 8% in 2035. The second key challenge is the growth of carbon emissions. While India’s energy intensity of GDP declines by 1.6% p.a. by 2035, the slow moving energy mix means that the carbon intensity of India’s energy consumption declines only modestly, by 0.3% p.a. The net result is that CO2 emissions from energy use more than doubles, averaging growth of 3.5% p.a. Over the final decade of the projection, from 2025 to 2035, India accounts for more than a third of the growth in global emissions, adding more than twice as much CO2 as China during that decade. However, this still allows India to meet its stated goal of reducing CO2 intensity of GDP by 20% by 2021, a year later than the target date (UNFCCC, 2011).

4. Alternative Cases There are, of course, many uncertainties surrounding any projection of India’s energy future. Just to illustrate this, two alternative cases are described below. One explores the implications of assuming a higher GDP growth path. The other examines the possibility of a “greener” growth path with greater gains in energy efficiency and a stronger push on renewables. Neither of these examples is a full-blown scenario: they represent sensitivities where we adjust a few key parameters, relative to our base case. Nor do we attempt to assign any probabilities to these cases, they are simply designed to illustrate the range of possibilities for India.

4.1 Implications of a Higher GDP Growth Path India aspires to higher GDP growth than the assumed growth rates in the base case described above. What might be the impact of assuming GDP growth of say 7.5% p.a. during 2013-35, while keeping the relationship between economic growth and sectoral energy demand the same as in the base case? Total energy demand growth increases to around 5% p.a. (Table 8 “High” case). Both oil and gas grow slightly faster than 5%, and coal slightly slower. This case results in CO2 emissions growing at 4.7% p.a., which is a concern in a world that is increasingly likely to be carbon-constrained. Moreover, if India’s own production of fossil fuels remains at the base case level, the higher demand in this higher GDP growth case would result in a significant increase in energy imports. Today India imports a little over 40% of its energy consumption. By 2035 that has risen to 44% in the base case, and to 54% in the high GDP growth case. In volume

20 India Policy Forum 2015 terms, net imports of fossil fuels exceed 900 mtoe in 2035 in the high case, compared to around 600 mtoe in the base case. That level of imports is not infeasible – global supply flows could accommodate India’s requirements, albeit with pressure on fuel prices. In this high growth case for India, China and India together would be importing around a quarter of the world’s oil production and more than 40% of the world’s LNG. In terms of oil trade, 16% of global oil exports would find their way to India. The major challenge to this alternative case would, of course, be the question of whether the Indian economy could afford the import bill. Table 8: Energy consumption projections Annual average growth 2013-35 Base High Green Oil 3.6% 5.2% 2.2% Gas 4.1% 5.4% 3.6% Coal 3.5% 4.4% -2.0% Nuclear 7.2% 8.0% 7.6% Hydro 3.1% 3.9% 3.5% Renewables 9.0% 9.8% 16.0% total 3.8% 5.0% 2.4% Source: (BP, 2015) and authors’ calculations

2013 29% 8% 55% 1% 5% 2% 100%

Share of primary energy Base High Green 28% 31% 28% 8% 9% 10% 51% 49% 21% 3% 2% 4% 4% 4% 6% 6% 5% 31% 100% 100% 100%

4.2 Green Growth India would like to play its part in addressing the risk of climate change. Under what conditions is it possible that India’s carbon emissions in 2035 would be no higher than they are today? The arithmetic of this case is relatively simple; finding a feasible policy set that could credibly deliver it is much more challenging. Compared to the base case, this greener alternative requires much a faster decline in energy intensity, and a much more rapid shift from fossil to non-fossil fuels4. The “green” case assumes that energy intensity declines at 3% p.a. This is quite a stretch for India, given that it already starts at a relatively low level, but similar rates of decline sustained for at least ten years have been seen in other rapidly developing Asian economies (e.g. Taiwan, Philippines). The result is total energy demand growing at just 2.4% p.a. (Table 8). This case also assumes that India’s renewables grow at the same rate as in the European Union (EU) over the past 20 years, and that both nuclear and hydroelectricity achieve production growth rates that sit between the base case and high case rates. Finally, among fossil fuels, gas is assumed to gain share of primary energy, while oil maintains the same share as in the base case. That leaves coal being squeezed out, with quite a dramatic decline in its share of energy to just over 20% by 2035.

To keep the analysis simple, we have ignored the potential for carbon capture and sequestration in this case. If that technology was available to be deployed at scale, it would allow a larger share for coal to be consistent with the goal of stabilising carbon emissions. 4

Kaushik Deb and Paul Appleby 21 This case delivers zero growth in carbon emissions from energy use. It also sharply reduces the net energy import requirements. The decline in coal demand leaves the level of coal consumption in 2035 below the current level of coal production, thereby eliminating the need for net coal imports. There would still be a requirement to import oil and gas, but overall the net energy import in 2035 will be about half the base case level, and a third of the high case level. So this alternative case has much to commend it in terms of the outcomes. However it is clearly built on some very challenging assumptions. To illustrate the challenge, consider the policy intervention that has been required in the EU to secure the rapid and sustained growth of renewables. The Renewable Energy Directive, 2009, established a binding target of 20% for the share of renewables in EU final energy consumption by 2020. Individual EU countries have then committed to a range of differentiated targets to achieve this EU aggregate goal, and submitted National Renewable Energy Action Plans which lay out sectoral targets and policy measures. A simple measure of the degree of policy support for renewables is the estimated amount of subsidy that has been paid to renewable energy sources. The most recent report for the European Commission (Ecofys, 2014) cites a figure of €40 billion for renewable energy support in 2012.

5. Conclusion Rising GDP and the changing structure of the economy in India has resulted in a significant growth in energy consumption over the past 30 years, even as the energy mix appears to be stubbornly dominated by fossil fuels. However, the significant shifts in the energy mix, away from coal and towards oil until 2000s, and the subsequent recovery in coal’s share, followed by a period of competition between coal and gas, illustrates the potential for change. This change is most affected by changes in domestic production, especially of gas and renewables. Going forward, India’s primary energy consumption is expected to grow at a rate outpacing most large developing countries. Coal will continue to dominate the energy mix, though it will lose some market share to gas and renewables. Other implications of this slow moving energy mix are observed in overall energy intensity and in carbon emissions. While India’s energy and emissions intensity has declined over time, these gains are mostly due to improving energy efficiency. However, with the inflexible energy mix, the gains from improving the share of more energy and carbon efficient fuels would remain limited. More significantly for India, domestic production has been sluggish in responding to energy demand growth, and imports are likely to continue rising. This increased share of energy imports as a percentage of GDP would place a significant burden on the economy.

22 India Policy Forum 2015

References BP. (2015). BP Energy Outlook 2035. London: BP. BP. (2015). Statistical Review of World Energy 2015. London: BP. Central Electricity Authority. (2015). Growth of Electricity Sector in India from 19472015. New Delhi: Central Electricity Authority, Government of India. CSTEP. (2015). A Sustainable Development Framework for India’s Climate Policy: Interim Report. Bangalore: Centre for Study of Science, Technology and Policy. DeLong, J. B. (2003). India Since Independence: An Analytic Growth Narrative. In D. Rodrik, In Search of Prosperity: Analytic Narratives on Economic Growth (pp. 184–204). New Jersey: Princeton University Press. Directorate General of Hydrocarbons. (2014). NELP-X oil & gas blocks : salient features. Petrotech 2014. Noida: Ministry of Petroleum & Natural Gas, Government of India. EIA. (2013). International Energy Outlook 2013. Washington: U.S. Energy Information Administration. Graczyk, S.-J. A. (2012). Understanding energy challenges in India-Policies, Players and Issues. Paris: International Energy Agency. IEA. (2014). World Energy Outlook 2014. Paris: International Energy Agency. Ministry of Industry. (1991). Statement on Industrial Policy. Ministry of Industry, Government of India. New Delhi: Ministry of Industry, Government of India. Navroz K Dubash, R. K. (2015). Informing India’s Energy and Climate Debate: Policy Lessons from Modelling Studies. Centre for Policy Research, Climate Initiative, Research Report. New Delhi: Centre for Policy Research. Niti Aayog. (2015). India Energy Security Scenarios (IESS) 2047. New Delhi: Niti Aayog. Oxford Economics. (2015). Global Economic Databank. Oxford Economics. P R Shukla, A. G. (2015). Energy-emissions trends and policy landscape for India. New Delhi: Allied Publishers. Panagariya, A. (2004). India in the 1980s and 1990s: A Triumph of Reforms. (W. I. Paper, Ed.) Planning Commission. (1999). Hydrocarbon Vision 2025. New Delhi: Planning Commission. Planning Commission. (2014). Low Carbon Strategies for Inclusive Growth. New Delhi: Planning Commission, Government of India. Reuters. (2009, February 29). Two top economists agree 2009 worst financial crisis since great depression; risks increase if right steps are not taken. Retrieved 1 20, 2015, from Reuters: http://www.reuters.com/article/2009/02/27/idUS193520+27-Feb2009+BW20090227 Shell. (2014). Shell Scenarios.

Kaushik Deb and Paul Appleby 23 TERI. (2013). The Energy Report – India: 100% Renewable Energy by 2050. New Delhi: WWF India. The World Bank. (2011). Energy Intensive Sectors of the Indian Economy. Washington, D.C.: Energy, Environment, Water Resources and Climate Change Units Sustainable Development Department, South Asia Region, The World Bank. UNFCCC. (2011). Compilation of information on nationally appropriate mitigation actions to be implemented by Parties not included in Annex I to the Convention. United Nations Framework Convention on Climate Change, United Nations. United Nations. (2015). World Urbanization Prospects: The 2014 Revisions. New York: United Nations, Population Database, Population Division, Department of Economic and Social Affairs.