Analyzing solar auctions in India: Identifying key determinants

Energy for Sustainable Development 45 (2018) 66–78

Contents lists available at ScienceDirect

Energy for Sustainable Development

Analyzing solar auctions in India: Identifying key determinants Sapan Thapar a,⁎, Seema Sharma a, Ashu Verma b a b

Department of Management Studies, Indian Institute of Technology, Delhi, India Center for Energy Studies, Indian Institute of Technology, Delhi, India

a r t i c l e

i n f o

Article history: Received 9 March 2018 Revised 10 May 2018 Accepted 10 May 2018 Available online xxxx Keywords: India Solar Auctions Effectiveness Regression

a b s t r a c t Solar technology has been identified as a key tool to fight climate change. The sector, promoted by several policy enablers, has seen a rapid growth in terms of deployment, with the global capacity reaching 390 GW at the end of 2017. In recent years, an increasing number of countries are adopting auctions to award solar contracts, resulting in steep tariff reductions. Researchers, while analyzing solar auctions, focused on ground level deployment, without capturing other factors influencing the investors' decisions. India, with its ambitious solar plan, has seen numerous contracts being awarded under auction schemes run by its federal and state agencies. We regressed eleven variables across thirty-two solar tenders issued in India between 2014 and 2017. Analysis of these auctions brought out a different set of determinants for federal and state programmes. On an overall basis, factors like solar targets, utilities' credentials and the level of subscription came out as strong determinants. Additionally, cost of funds and module price figured as drivers in the federal bids. Possible recommendations include spatial and temporal spacing of bids, sale to multiple off-takers and provision of risk guarantee funds. These factors may be taken into consideration by Indian Policy makers while designing solar tenders. © 2018 International Energy Initiative. Published by Elsevier Inc. All rights reserved.

Introduction The solar photovoltaic sector has witnessed rapid growth during the last decade, with the total global installations reaching 390 GW at the end of year 2017. The top five solar markets were China, the United States, Japan, India and the United Kingdom, accounting for 85% of all increments. Rapid declines in module costs (as indicated in Fig. 1), use of auctions and increased efficiency levels have led to economic competitiveness of solar photovoltaic (PV), with average levelized tariff falling to US 10 ¢ per kWh. Fig. 2 provides the average solar power generation cost for the period 2010 to 2017. India is among the top five countries in terms of greenhouse emissions, even with a low level of per capita emissions.1 Over 70% of India's emissions are attributed to the energy sector2 due to the preponderance of fossil-fuel generators in its energy sector. Classified as a tropical country, India is endowed with enormous solar potential, estimated at 5000 trillion kilowatt hours, or kWh; refer Fig. 3. The

⁎ Corresponding author. E-mail addresses: [email protected], (S. Thapar), [email protected], (S. Sharma), [email protected] (A. Verma). 1 GHG Emission Report, World Resource Institute, 2012. 2 India Biennial Update Report to UNFCCC, Government of India, December 2015.

total solar potential has been estimated at 750 G watts, or GW, spread evenly across the country.3 Theoretically, a small fraction of the total incident solar energy can meet India's power requirements. Solar technology also enables energy access to rural India under decentralized and distributed modes. Looking into the above perspectives, the Indian government launched the National Solar Mission4 in the year 2010 under its National Action Plan on Climate Change to promote ecologically sustainable growth, while addressing India's energy security challenge. The Mission proved to be a milestone for the growth of solar in India. It coincided with a global focus on solar technologies, abetted by a rapid scale-up of photovoltaic module manufacturing facilities across the globe. In line with India's commitments towards climate change mitigation, the solar capacity targets were revised upwards by the Ministry of New and Renewable Energy, MNRE, in the year 2015 from 20 GW to 100 GW, to be achieved by the year 2022. Doing away with the prevalent feed-in-tariff model, or FIT, the Indian government used auctions as an instrument to award solar capacities. Due to interplay of market forces and growing interest of

3

Indian Solar Atlas, Government of India, November 2014. Mission Document can be seen on link http://www.mnre.gov.in/file-manager/ UserFiles/mission_document_JNNSM.pdf.

https://doi.org/10.1016/j.esd.2018.05.003 0973-0826/© 2018 International Energy Initiative. Published by Elsevier Inc. All rights reserved.

4

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

67

Fig. 1. Trends of solar PV module cost (PV Exchange, 2014a). aSolar Module Rates, https://www.solarserver.com/service/pvx-spot-market-price-index-solar-pv-modules.html accessed on March 05, 2018.

the investor groups, significant reduction in solar tariff over the applicable FIT was observed in about fifty auctions conducted since the launch of the Solar Mission. When compared with established policy

instruments like FIT, renewable purchase obligations, or RPO, and tax credits, auctions have been a recent phenomenon and is yet to be researched upon in detail.

Fig. 2. Average solar tariff (IRENA, 2017a, 2017b).

68

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Fig. 3. Indian solar resource map (The World Bank, 2017).

We regressed cost effectiveness5 of over thirty solar tenders issued by federal6 and state7 agencies in India during the period 2014 to 2017 for eleven exploratory variables and found significance of factors like resource assessment, targets utility's performance and cost of funds. Discussion follows in the sections below. Indian power sector Overview In India, both the federal government and the provincial governments are eligible to frame electricity related legislations. Under the Electricity Act, the power generation and distribution sectors are delicensed, facilitating open access through power markets. The

5 Cost effectiveness is defined as the reduction in tariff offered in the bids over the base FIT prices. 6 Both Federal and Central government are used in the same context. 7 Both Provincial and State government are used in the same context.

‘National Tariff Policy’ suggested tariff determination through an open regulatory process, which included bidding as one of the instruments. The electricity regulatory commissions have been created to regulate the sector. Table 1 Source-wise power generation capacity in India (CEA, 2017). Energy source

Installed capacity (MW)

Coal Gas Diesel Nuclear Hydro Renewable Wind Small hydro Biomass Waste to energy Solar TOTAL

1,92,971 25,150 838 6780 44,963 32,715 4399 8182 114 15,575 3,31,688

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

69

Fig. 4. Growth of Indian renewable sector (MNRE, 2017).

The total installed generation capacity at the end of year 2017 was 330 GW, with over 60% share from coal based thermal sources,8 followed by hydro and renewable sectors9 as shown in Table 1. Even with a high quantum of power generation (about 1200 billion units), the per capita consumption of India is a third of the global average10 at about 1100 kWh. The Indian government has been promoting renewable energy in the country by way of several policy enablers. Under the Electricity Act, the State Electricity Regulatory Commissions, or SERCs, are required to promote renewable technologies by providing suitable measures for grid connectivity and determine remunerative tariffs. Under the National Tariff Policy, SERCs are required to fix a certain minimum percent of electricity to be sourced from renewable resources; the compliance can be in terms of actual purchase of green power, or by way of tradable green certificates, referred to as renewable energy certificates. There are separate procurement mandates for solar and non-solar technologies. Investor companies are eligible to claim depreciation on the asset value of their total capital employed in a renewable project, leading to reduction in their tax obligations. These policy instruments led to a rapid growth in the Indian renewable energy sector, at over 20% during the period 2010–17. At the end of the year 2017, the total installed renewable capacity was over 60 GW, with wind having the maximum share (32 GW), followed by a rapidly growing solar sector (20 GW); refer Fig. 4 for details. The share of renewables in terms of total installed capacity is 18% and in terms of grid penetration, it is about 7%. Indian solar sector To harness the large solar potential, the Indian government came out with solar specific policy instruments. When the Solar Mission was initiated, solar FITs were substantially higher than FITs offered to

8 Monthly Power Sector Report, Central Electricity Authority, CEA, Government of India, December 2017. 9 In India, hydro power projects upto 25 MW capacity are considered as renewable. 10 World Energy Outlook, International Energy Agency, 2017.

Table 2 Cost of bundled power. (Adapted from Thapar, Sharma, & Verma, 2016).

Procurement cost (US Cents per kWh) Number of kWh Total cost Weighted cost (US Cents per kWh)

Solar

Coal

22 1 22 =(22 + 20)/6 = 7

4 5 20

other renewable technologies; it was three times of wind FIT during the year 2011.11 To make solar power palatable to utilities, Solar Energy Corporation of India, or SECI,12 employed auctions to award solar contracts based on the minimum quoted bid, leading to 30% decrease in tariffs (Altenburg & Engelmeier, 2013; Dawn, Tiwari, & Goswami, 2017). This was in line with a global shift to allocate solar capacity under auctions rather than FIT. Further to bridge the price differential between solar tariff discovered under auctions and the average cost of power procurement of the utilities, two mechanisms were used. In the first case, solar power was bundled with cheaper coal power, reducing the average purchase cost for the utilities. Table 2 explains concept of bundled power. In the second case, solar tariff was capped at a value similar to utility's procurement price, with developers being offered funds to meet their expectations on returns. This is locally referred to as viability gap funding support, or VGF. To meet the targets set under the Solar Mission, federal agencies have been issuing solar tenders. Similarly, various state governments have been inviting bids under their respective solar polices.13 In case of the former, solar power is procured by a central agency, which acts as an intermediate entity and arranges power procurers from the host states as well as other geographies, while in latter, solar power is procured by the host utilities themselves. 11 Review of Tariff Orders issued by Central Electricity Regulatory Commission for the year 2011–12. 12 SECI is an Indian government owned agency, www.seci.co.in. 13 Compendium of State Solar Policies, IREDA, 2015.

70

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Fig. 5. Solar capacity growth in India (MNRE, 2017).

Fig. 6. Year-wise solar tariff trends in India (MNRE, 2017).

An example of the first case is the procurement of power from a Madhya Pradesh based solar park by Delhi Metro Rail Company.14 This inter-state mode of power sale is being promoted by exempting solar projects from payment of transmission charges for wheeling power beyond a state's boundary.15 It was felt that smaller scale solar projects lead to inefficient use of capital, requiring creation of infrastructure facilities separately for each project, besides soliciting approvals. To obviate these issues, the Indian Government conceived ‘Solar Park’, as a concentrated zone of solar power generation, with availability of infrastructure facilities and approvals, facilitating projects under ‘Plug and Play’ mode. The federal government has planned to set up 50 solar parks across the country, each with a capacity of 500 MW and above, with a cumulative capacity target of 40 GW.16 14

Rewa Ultra Mega Solar Project; http://rumsl.com/. Government of India has exempted solar projects from payment of inter-sale transmission charges. 16 Solar Park Scheme, Ministry of New and Renewable Energy, Government of India, 2017.

Auctions have resulted in accelerated capacity addition of solar in India (refer Fig. 5) at competitive tariffs, substantially lower than the prevailing FIT, in some cases, as low as US¢ 5 per kWh17; refer Fig. 6. However, on closer examination of tariffs provided in Table 3, it can be observed that the outcome under different auctions has been unsymmetrical. Sections below analyze these bids and suggest probable factors for the variability. Literature review and need for study In recent years, an increasing number of countries have been adopting auctions to simulate their solar energy market because of the associated benefits of tariff reductions (Pablodel & Pedro, 2014). Research has established that auctions can lead to significant savings of public spending per unit of power in the range of 20% and 41% (Mayr, Schmidt, & Schmid, 2014). As per a report, procurement of

15

17

INR is Indian Rupees, Average exchange rate for the year 2017 was 1 US$ = 65.2 INR.

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

71

Table 3 Tariff discovered under State Solar Schemes.a Sl no.

Year

Month

State

Tender issuing agency

Capacity (MW)

SERC tariff (INR/kWhb)

Minimum tariff (INR/kWh)

Reduction in tariff (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

2014 2014 2014 2014 2015 2015 2015 2015 2015 2015 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2017 2017 2017 2017 2017

April August September October July August September October December December January January January February March March March March April June June June July July August August September February April May May September

Chhattisgarh Karnataka Telangana Andhra Pradesh Madhya Pradesh Telangana Punjab Uttrakhand Andhra Pradesh Haryana Rajasthan Uttar Pradesh Maharashtra Karnataka Andhra Pradesh Andhra Pradesh Karnataka Jharkhand Karnataka Maharashtra Chhattisgarh Telangana Gujarat Rajasthan Odissa Uttar Pradesh Maharashtra Madhya Pradesh Andhra Pradesh Rajasthan Rajasthan Tamil Nadu

State State State State State State State State Federal State Federal Federal Federal Federal Federal Federal State State Federal Federal Federal State Federal Federal Federal Federal Federal Federal Federal Federal Federal State

100 500 500 500 300 2000 500 170 150 150 420 100 380 50 500 350 860 1200 500 450 100 350 225 130 270 100 450 250 250 400 500 1500

7.74 8.4 8.75 8.75 10.44 7.72 7.04 7.72 7.72 7.32 6.74 7.06 7.08 6.51 7.04 7.04 6.51 5.68 6.51 6.04 7.03 5.68 6.33 5.4 5.68 7.06 6.04 5.45 5.68 3.93 3.93 4.5

6.44 6.71 6.45 5.25 5.05 5.17 5.09 5.57 5.13 5.08 4.34 4.78 4.81 4.43 4.63 4.63 4.69 5.08 4.78 4.41 4.88 4.67 4.96 4.35 4.81 4.81 4.58 3.3 3.15 2.62 2.45 3.47

17 20 26 40 52 33 28 28 34 31 36 32 32 32 34 34 28 11 27 27 31 18 22 19 15 32 24 39 45 33 38 23

a b

Data extracted after perusal of solar tenders issued by both federal & state government agencies of India. INR is Indian Rupees, average exchange rate for the year 2016 was 1 US$ = INR 70.

renewable energy through auctions is rapidly driving down costs, by as much as 69% in case of solar PV sector (IRENA, 2017a, 2017b18). Fig. 7 gives trends of global solar tariffs for the period 2010–17. Santana (2016) suggested a mix of technology-neutral and technology-specific policies using auctions to promote renewable energy. Some scholars opined that auctions can play an important role in the implementation of renewable capacity around the world (Atalay, Agni, & Pattberg, 2017; Pablodel & Pedro, 2014). Rego and Parente (2013) compared the outcomes from the old and new energy auctions held in Brazil from 2004 to 2010 and found statistically similar results. Nurcan (2016), while evaluating renewable energy policies across the European Union and the United States over the period 1990–2008, found tenders to be an effective mechanism for stimulating renewable capacity deployment. Buckman, Sibley, and Bourne (2014) found the outcome of the first Australian auction to be highly competitive and identified land availability as a key risk. Experts like Pablodel (2017) and Shrimali, Konda, and Farooquee (2016), while examining the effectiveness of auctions, suggested improvement in their design by ensuring competition, volume disclosure, provision of payment guarantees and imposing penalties for delays. Kreiss, Ehrhart, and Haufe (2017) recommended financial and physical pre-qualifications to achieve a sufficiently high realization rate. Cassetta, Monarca, RubinaNava, and Meleo (2017) assessed Italian wind auctions from 2012 to 2016, in terms of cost efficiency and policy effectiveness. They explored key determinants affecting tariff, including firm-specific factors and auction design elements. Winkler, Magosch, and Ragwitz (2018) found mixed results while evaluating

18 IRENA (International Renewable Energy Agency) is an intergovernmental organization, www.irena.org.

the effectiveness of auctions undertaken in Brazil, France, Italy, the Netherlands and South Africa. Shrimali and Rohra (2012) did a qualitative analysis of the barriers to solar deployment in India, focusing on power sector reforms. IRENA, in its report, noted that countries like Brazil and South Africa have not been effective in terms of getting the desired results from the auctions (IRENA, 2017a, 2017b). Research by Climate Policy Initiative (2015) found certain factors to impact the effectiveness of auctions and suggested appropriate policy designs to mitigate the same. Most of the studies focused on the early period when auctions were evolving as a policy instrument. Moreover, they assessed the effectiveness in terms of ground level deployment, without analyzing the micro elements influencing the bids. However, upon closer scrutiny, it can be observed that the steep reduction in tariff experienced during the period 2010–13 got marginalized in subsequent years (IRENA, 2017a, 2017b). The glut experienced by the solar manufacturing facilities during these years might have led to greater discounts in module costs. Bayer (2018) analyzed Brazilian wind auctions for the period 2009 to 2015 on the discovered tariffs, completion rates and market concentration. He attributed decline in prices to the increased level of competition and noted the impact of factors like regulatory changes and currency movement. Our research covers a much larger number of schemes using a broader set of variables. Another shortcoming in the existing literature is limited research on the Indian solar auctions, whose assessment is vital for many reasons. One, India has set ambitious solar capacity targets (100 GW), with a sizable share coming from large size solar parks, which is also being planned in several other countries. Two, India as a federation of states, is similar to the European Union and the United States in terms of electricity regulations. Analysis of solar tenders issued by Indian states, with varying levels of economic profile, energy markets, solar potential and

72

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Fig. 7. Solar PV price trends (IRENA, 2017a, 2017b).

policy structure may provide useful inputs. Three, more than fifty solar tenders have been issued in India since the launch of the Solar Mission, providing a rich source of data for an in-depth causal analysis. Assessment of solar auctions (refer Fig. 8), would help identify key determinants with respect to tender design so as to help accelerate solar deployment in India.

Data and methodology We regressed thirty-two solar tenders issued by federal agencies as well as provincial governments for eleven exploratory variables to check for their impact on the bids. These eleven parameters have been categorized into six heads – Geographic, Policy, Economic, Energy, Commercial and Scheme and the same is provided in Fig. 9. Ideally, panel data regression would have been a suitable tool for such a study. However, due to paucity of data (available for a 3-year

period) and issuance of multiple tenders in some states, we decided to use simple regression. The eleven determinants included economic profile, energy deficit, solar potential, capacity targets and purchase obligations, cost of funds, module price, grid share, counterparty profile, besides contract size and bid subscription. The states included Andhra Pradesh, Chhattisgarh, Haryana, Jharkhand, Karnataka, Madhya Pradesh, Maharashtra, Punjab, Rajasthan, Telangana, Uttrakhand and Uttar Pradesh. Table 3 provides the details of tenders which have been analyzed. Variables Discussions on dependent variables The dependent variable is the cost effectiveness of solar tenders, computed as the percentage reduction of minimum quoted tariff over the prevailing solar feed-in-tariff.

Fig. 8. Solar tariff trends in India (IRENA, 2017).

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

73

Fig. 9. Exploratory variables [Authors Depiction].

Discussions on exploratory variables i. Solar Resource Potential The pre-requisite for setting up a solar project is the availability of radiation. The Indian government, through its agency, National Institute of Solar Energy, has estimated the solar potential for the entire country at about 750 GW as highlighted in Fig. 10. Investors prefer states with higher resource potential. ii. Solar Procurement and Capacity Targets The state electricity regulatory commissions set solar procurement targets for their respective utilities, signaling a medium to long-term market for the investment community. On similar lines, the federal government has set up solar capacity targets for each of the states as part of the 100 GW capacity plans, envisaged to be achieved by the year 2022. These two variables have been considered as independent factors.

v. Module Price With modules constituting a substantial portion in the cost of a solar project, investors look at module price to evaluate the viability of their bids. Due to economic reasons, a number of solar project developers in India are using imported solar modules and their price has been considered as a determinant. vi. Cost of Funds Capital intensive solar projects, with low generation levels would require low-cost and long-term funds to make them commercially viable in India (Dawn et al., 2017; Shrimali, Nelson, Goel, Konda, & Kumar, 2013). However, it is difficult to estimate the lending cost of projects due to confidential clauses. Banks have been active in lending to solar projects in India and typically use their term deposits so as to avoid asset-liability mismatch. Therefore, we have used the value of term deposit rates as offered by State Bank of India, SBI, as a proxy.

iii. Economic Variables vii. Off-takers Profile A large number of experts have correlated energy consumption with economic growth (Apergis & Payne, 2010; Kraft & Kraft, 1978). A higher earning capability also reflects willingness to pay premium for procuring clean energy as reflected in the Kuznets curve (Bruyn, Bergh, & Opschoor, 1998; Stigka, Paravantis, & Mihalakakoua, 2014). As such, the per capita income has been taken as an exploratory variable. Energy deficit is another factor which provides signals with respect to market demand and has been considered as an independent variable.

Utilities have an important role in terms of solar power procurement, RPO compliance, tariff payment and grid management. The most suitable variable is the technical and commercial performance of a utility, measured in terms of aggregate technical and commercial (ATC) losses and the same has been considered as an independent variable. ATC loss is being annually computed for the Indian states.19 viii. Bid Parameters

iv. Grid Penetration of Renewables Solar generation varies seasonally as well as diurnally, which may dissuade a utility from procuring it beyond a certain threshold. Some experts have highlighted the challenges of integrating infirm renewable power in the grid (Ayodele, Jimoh, Munda, & Tehile, 2012; Hirth, 2013; Debra, 2010). As such, a state's share of renewable power in its grid has been considered as an exploratory variable.

The total contract capacity under an auction scheme and the participation level would have a bearing on the bid outcome and have been considered as exploratory variables.

19 Annual Rating of Utilities, Power Finance Corporation, Government of India.; Years2014, 2015 and 2016.

74

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Fig. 10. State wise solar potential, India (MNRE, 2014).

Data sources

Statistical equation

Details of dependent and independent variables, their nomenclature and data sources are provided in Table 4. Datasheet has been provided in Table 5.

Yst = α + Gx + Px + Ecx + Enx + Cx + Bx + εit, where, Yst = Percent Reduction in tariff over solar FIT for each tender Gx = Geographic Variables

Table 4 Dependent and independent variables and data sources. Type/category

Variable

Nomenclature

Data Sources

Cost effectiveness of the solar tenders (in percent)

Disctariff

Solar Tenders issued by government agencies

Log value of Solar Resource Potential (GW) Log value of state-wise Solar Targets Solar Procurement mandate, or, RPO Log value of Per capita income (INR)

Res Target RPO Percapitainc

Cost of Funds (lending rate in percent) Energy deficit (in percent) Renewable share in grid (in percent) ATC loss of utilities (in per cent) Module rates (in US$ per Watt peak) Contract Capacity (MW) Subscription (as a multiplier of contract capacity)

Fdrate Deficit Reshare Atcloss Module Capacity Bid

National Institute of Solar Energy, Government of India Ministry of New and Renewable Energy, Government of India Ministry of New and Renewable Energy, Government of India Central Statistics Office, Ministry of Statistics and Programme Implementation, Government of India State Bank of India Central Electricity Authority, Government of India Central Electricity Authority, Government of India Annual Reports of Power Finance Corporation, Government of India PV exchange www.solarserver.com Solar Tenders issued by government agencies --do --

Dependent Independent Geographic Policy Economic

Energy Commercial Scheme

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

75

Results and discussions

Table 5 Variables dataset. Slno

Disctariff

Atcloss

Capacity

Subscription

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

16.8 20.1 26.3 40.0 51.6 33.0 27.7 27.8 33.5 30.6 35.6 32.3 32.1 32.0 34.2 34.2 28.0 10.6 26.6 27.0 30.6 17.8 21.6 19.4 15.3 31.9 24.2 39.4 44.5 33.3 37.7 22.9

24 17.9 13 13 24.5 13 16.66 20.2 13 27.5 26 38 20 17.5 13 13 18 39 18 19 22 12 22 26 37 35 22 25 12 22 22 19

100 500 500 500 300 2000 500 170 150 150 420 100 380 50 500 350 860 1200 500 450 100 350 225 130 270 100 450 250 250 400 500 1500

2.5 2.8 3.7 2.6 2.3 2.4 3.3 2.8 5.7 4.3 9.8 3.5 3.55 1.4 10.8 6 2.9 2.1 4.5 0.5 1.7

Module

8.125 11 142.1052632

35 33 33 30.5 30.5 30.5 30 30 30 31 31 31 31 31 31 31 31 28.5 28.5 28.5 28.5 26.5 26.5 26.5 26.5 24 22 22 22 22 24

2.3 1.26 2.6 30

Slno

RPO

Fdrate

Deficit

Percapitainc

Pot

Target

Reshare

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

0.5 0.3 0.3 0.3 100 0.3 0.19 0.08 0.3 0.25 2 1 0.5 0.25 0.3 0.3 0.25 1 0.25 0.5 100 0.3 1.5 2 0.3 100 0.5 1.25 0.25 2.5 2.5 5

8.5 8.5 8.5 8.5 7.75 7.25 7.25 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 6.5 6.25 6.25 6.25 6

0.4 5.6 14 7.3 0 1.4 0 1.9 0.2 0.2 0.4 12.7 0.4 5.3 0.2 0.2 5.2 2.3 3.5 0.4 0.4 0 0 0.1 0 2.3 0.1 15 0.1 0.9 0 0

4.81 5.04 5.01 4.89 4.67 5.05 4.99 5.10 4.94 5.13 4.81 4.58 5.05 5.06 4.94 4.94 5.06 4.73 5.06 4.71 4.85 5.09 5.07 4.99 4.79 4.58 5.05 4.71 4.98 4.81 4.81 5.09

4.26 4.39 4.31 4.58 4.79 4.31 3.45 4.23 4.58 3.66 5.15 4.36 4.81 4.39 4.58 4.58 4.39 4.26 4.39 4.81 4.26 4.31 4.55 5.15 4.41 4.36 4.81 4.79 4.58 5.15 5.15 4.25

3.25 3.76 3.69 3.69 3.75 3.69 3.68 2.95 3.69 3.62 3.76 4.03 4.08 3.76 3.69 3.69 3.76 3.30 3.76 4.08 3.25 3.69 3.90 3.76 3.38 4.03 4.08 3.75 3.69 3.76 3.76 3.95

5 16.2 2 4.8 2.7 2.1 2.4 5.5 6.2 2.8 9.8 3.4 7.6 16.5 6.2 6.2 16.5 0.3 16.5 7.6 4.7 2.1 7.7 9.8 1.6 3.5 7.6 8 10.1 11.8 11.8 14.5

Px = Policy Variables Ecx = Economic Variables Enx = Energy Variables Cx = Commercial Variables Bx = Bid Scheme Variables εit = independent random variables

We regressed eleven exploratory variables by excluding insignificant variables in successive iterations. Finally, four variables came out as significant determinants. To maintain equivalence in figure size, we used log values for three factors, namely, per capita income, solar potential and solar targets. In order to minimize the impact of duplicate values (as some states appear in multiple times in our dataset), we suppressed the constant terms. Descriptive statistics are provided in Table 6. Correlation matrix, placed at Table 7, suggests that multicollinearity is not an issue with our data, except for one case discussed below. Module cost and bank deposit rates displayed an unexpected high level of correlation. Several factors influence global module prices including manufacturing size of facilities, market demand, climate pledges, solar policies and customs duties. India imports more than 90% of its solar module requirements due to price differential. On the other hand, fixed deposit rates offered by an Indian bank largely reflect the domestic money markets including strength of the economy, type of deposits and the debt requirements. These two being totally unrelated factors, we have ignored this correlation. Regression results are provided in Table 8. Detailed analysis I. Geographic Variables India is endowed with solar potential across most of its landmass, with most places receiving radiation in the range of 4–7 kWh per square meter. Ramachandra, Jain, and Krishnadas (2011) found 60% of India's land area amenable for setting up solar projects. Further, satellite based solar data is available for most of the Indian sites. Project developers do not have to undertake micro-siting to identify suitable locations for putting up modules as is the case with wind projects. As such, this factor got reflected as an unimportant variable. II. Policy Variables Solar targets came out as a significant factor. The federal government came out with state-wise targets to meet its international commitments, with a large capacity being developed under the ‘Solar Park Scheme’ administered by the federal agencies. These state-wise targets provide a long-term direction to the investor community. Similar to solar targets, Solar RPO also got reflected as a significant factor. Though there have been challenges in terms of RPO compliance, they give an indication to the utilities to prepare for a certain quantum of solar uptake in their grid. Utilities are also allowed to pass through the cost of solar power procurement to the rate-payers.

Table 6 Descriptive statistics. Variable

Obs

Mean

Std Dev

Min

Max

Atcloss Capacity Disctariff Bid Module Rpo Fdrate Deficit Percapitainc Pot Target Reshare

32 32 32 28 31 32 32 32 32 32 32 32

21.25813 443.9063 29.33125 4.865536 28.59677 10.15531 7.109375 2.515625 4.918438 4.501563 3.70875 7.296875

7.64474 420.3124 8.717518 5.670063 3.543584 29.37703 0.6284181 4.211207 0.1598812 0.3810521 0.2554155 4.854761

12 50 10.6 0.5 22 0.08 6 0 4.58 3.45 2.95 0.3

39 2000 51.6 30 35 100 8.5 15 5.13 5.15 4.08 16.5

76

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Table 7 Correlation matrix.

Disctariff Atcloss Capacity Module Target Deficit Percapitainc Reshare Pot Rpo Bid Fdrate

Disctariff

Atcloss

Capacity

Module

Target

Deficit

Percap~c

Reshare

Pot

Rpo

Bid

Fdrate

1 −0.2878 −0.2436 −0.074 0.2428 0.0794 −0.248 −0.0219 0.3967 0.3489 0.3731 0.0356

1 −0.1895 −0.2506 −0.0674 0.0571 −0.6184 −0.3228 −0.0312 0.2585 −0.0116 −0.2616

1 0.016 0.025 −0.116 0.2455 0.0301 −0.0999 −0.2414 −0.1383 −0.0735

1 −0.115 0.1269 0.2701 −0.1671 −0.3478 −0.0842 −0.4654 0.777

1 0.1215 −0.1777 0.2957 0.2913 −0.0405 0.0565 −0.0768

1 −0.1838 0.0151 −0.0175 −0.1713 0.3931 0.3478

1 0.3634 −0.3851 −0.4499 −0.2306 0.1031

1 0.3034 −0.2518 0.1033 −0.2329

1 0.0023 0.376 −0.2034

1 −0.1666 0.0687

1 −0.3085

1

Table 8 Regression results - all bids. Source

SS

df

MS

Model Residual

25,785.0666 842.003243

4 24

6446.26664 35.0834685

Total

26,627.0698

28 9

950.96677

Number of obs

28

F(4, 24) Prob N F R-squared Adj R-squared Root MSE

183.74 0 0.9684 0.9631 5.9231

Disctariff

Coef.

Std. Err.

t

PNt

[95% Conf.

Interval]

Atcloss Target Rpo Subscription

−0.4159445 9.214206 0.1397521 0.6378675

0.147365 0.9294171 0.038208 0.2043602

−2.82 9.91 3.66 3.12

0.009 0 0.001 0.005

−0.7200908 7.295984 0.0608946 0.2160888

−0.1118 11.13243 0.21861 1.059646

III. Economic Variables Both the economic variables, the per capita income and cost of funds, came out as insignificant factors. There have been contrasting findings on per capita income and renewable deployment. While Menegaki, Hanley, and Tsagarakis (2007) found no causal relationship, Apergis and Payne (2010) found a long-run equilibrium relationship. Baek and Gweisah (2013) observed higher sensitivity to environmental matters with better income level. However, low levels of annual per capita income20 in India may lead to indifference of citizens to endorse it. With regard to funds, a project developer arranges them from the cheapest possible source in order to maximize the returns. Banks and financial institutions undertake credit assessment of a solar power project and accordingly determine risk premiums based on their risk appetite. Corporates raise money through green bonds and multilateral agencies. The lending rates vary from case to case and are not available on public platforms. As such, it is difficult to relate bank deposit rates with the cost of lending to a solar power project, which duly gets reflected in its low significance. IV. Energy Variables Like the economic factors, both the energy variables came out as insignificant factors. Utilities harness energy resources in order to bridge the demandsupply gap. However, with many states reducing their energy deficit, this factor is not relevant any further. Energy deficit in India reduced from 4.7% in the year 2014 to 1.6% in the year 2017.21 Further, the share of solar in the grid over the period of our analysis was insignificant for the utilities to take them as an important energy resource. As such, energy deficit came out as a low impact factor. 20

Average per capita income of US$ 1700, World Bank dataset, 2016. Power Sector Monthly Report, Central Electricity Authority, Government of India, December 2017. 21

A major part of our analysis covered the period when the deployment of solar was at a nascent stage in India, with an inconsequential share in the grid. As such, the quantum of renewable power in the grid did not figure out as a key driver. Even with a high share of renewable energy in its grid, the state of Tamil Nadu could not influence the results as it had issued only one solar tender in these years as per our dataset. However, with an accelerated pace of solar capacity addition expected in the coming years, this factor may influence the bids in future. V. Commercial Factors ATC (aggregate technical and commercial) loss came out as a strong negative determinant in our analysis. It reflects upon the technical and commercial competence of a utility, with a lower value indicating a better credit profile, ensuring timely payments to project developers and capability to absorb intermittent solar power in the grid. To our surprise, the prices of solar photovoltaic modules did not reflect as a significant factor. As opined above, the steep reduction in module prices happened during the initial years of Solar Mission, from 2010 to 13 and the module price curve showed a flat profile thereafter; refer Fig. 2. However, our assessment focused on the bids issued in the period April 2014 to September 2017, leading to its moderate influence. VI. Bid Parameters Due to low entry barriers and an assured market, a number of domestic and international entities have been participating in the Indian solar tenders. As per our dataset, the average bid subscription level came as five, indicating the high level of participation. A lower capacity contract issued after a long hiatus may see aggressive bidding, while the parameters may be starkly different in case of a large capacity ‘Solar Park’ funded by development banks. These would have made it reflect as a strong determinant.

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78 Table 9 Regression results - federal bids. Source

SS

df

MS

Model Residual

17,908.75 453.7732

6 14

2984.79113 32.4123692

Total

18,362.52

20

918.125997

Number of obs

20

F(6, 14) Prob N F R-squared Adj R-squared Root MSE

92.09 0 0.9753 0.9647 5.6932

Disctariff

Coef.

Std. Err.

t

PNt

[95% Conf.

Interval]

Atcloss Module Deficit Pot Rpo Fdr

−0.527627 0.918313 1.151696 14.87028 0.134705 −8.321149

0.200377 0.6161548 0.3446683 4.282724 0.0506373 4.722192

−2.63 1.49 3.34 3.47 2.66 −1.76

0.02 0.158 0.005 0.004 0.019 0.1

−0.9573926 −0.4032082 0.4124563 5.684752 0.0260988 −18.44924

−0.0978609 2.239833 1.890936 24.05581 0.2433112 1.806946

The size of solar contracts did not figure out as an important factor. The average capacity of the size of the contracts in our dataset was only 400 MW, making it inconsequential compared with the huge solar potential in India. Assessment of federal bids Assessment of twenty tenders issued by federal agencies, mostly under the ‘Solar Park Scheme’ gave a different set of significant variables. These included solar potential and RPO, energy deficit, utility's credentials and lending rate. Module cost reflected as a partially significant driver. Regression analysis is presented in Table 9. Electricity, being a concurrent subject under the Indian constitution, federal government facilitated fulfilment of India's climate commitments by way of national RPO mandates and state-specific solar capacity goals. As such, factors like solar potential and RPO reflected as strong determinants in these auctions. With an established resource assessment, pre-developed infrastructure and assured power offtake offered under the ‘Solar Park’ auctions, companies tapping low-cost funds and procuring equipment at cheaper prices stand a high chance of winning. Therefore, cost of funds and module price got reflected as marginally important factors, the former being a negative variable as a high interest rate impacts project returns. Utilities' performance also come out as key factor. Even the higher credit rating enjoyed by federal agencies, who act as intermediaries between project developers and the host utilities, did not provide a strong incentive to alleviate the concerns of investors. Though these agencies have started offering risk guarantee instruments to mitigate payment delays, this is yet to be tried on a large scale. Qualitative factors Qualitative issues like profile of the bidder and the periodicity of tenders might have influenced the bid outcome. India has been categorized among the top five global destinations for renewable investments by the EY Group in its Renewable Energy Country Attractiveness Index, or RECAI.22 This has led to a large number of domestic and international investors participating in the solar tenders. These include private equity investors, pension and insurance funds, energy companies, utilities, start-ups and entrepreneurs. Objectives vary from business expansion for an established company, derisking strategy for a utility, to operational synergies for project management companies. The other parameter is the spatial and temporal spacing of bids. A long lean period would make the investors and equipment suppliers

22 The last RECAI Index was released in May 2018, in which India was ranked fourth globally.

77

jittery, who may seek the next available opportunity, leading to aggressive bidding. On the other hand, a large number of tenders offered in a particular geography may reduce the bid discounts. As per our dataset, bids issued after a gap of four months or more saw a sizable reduction in the discovered tariffs. Conclusion and ideas for further research Accelerated penetration of solar projects in India would depend upon early grid parity between solar and conventional power. Auctioning of solar capacity has been effective towards reducing solar tariff in India, though there have been variations in the outcome. This study focuses on identifying key factors impacting the bids, while suggesting areas for design improvements. Analysis of solar auctions in India bring out a different set of determinants for federally administered schemes and state-specific contracts. On an overall basis, factors like solar targets, utilities' credentials and the level of bid subscription came out as strong determinants. On the other hand, key drivers in federal bids included solar potential, cost of funds and module price. Possible mechanisms which can improve the efficacy of bids are discussed. Tenders can be issued with spatial and temporal considerations besides empowering regulators to ensure solar RPO compliance. Tying with multiple set of consumers and provision of risk guarantee funds can preclude off-taker risk. Instruments like green bonds and yield-cos can help reduce the cost of funds. Similar to oil and gas markets, an open platform providing list of viable solar proposals, can be developed for the investors. Future work can include analyzing outcome of solar rooftop and wind tenders along with in-depth analysis of qualitative factors. An empirical study on the efficacy of renewable policy instruments including FIT, RPO and auctions may further help the policymakers choose the most appropriate instrument. Acknowledgements We are grateful to the editor as well as the reviewers for providing very valuable comments which helped in improving the quality of this paper. We are thankful to our colleagues Mr. Amit Kumar and Dr. Gopal Sarangi who shared useful inputs for this research. References Altenburg, Tilman, & Engelmeier, Tobias (August 2013). Boosting solar investment with limited subsidies: Rent management and policy learning in India. Energy Policy, 59. Apergis, Nicholas, & Payne, James (2010). Renewable energy consumption and economic growth: Evidence from a panel of OECD countries. Energy Policy, 38(1). Atalay, Yasemin, Agni, Kalfagianni, & Pattberg, Philipp (May 2017). Renewable energy support mechanisms in the Gulf Cooperation Council states: Analyzing the feasibility of feed-in tariffs and auction mechanisms. Renewable and Sustainable Energy Reviews, 72. Ayodele, Temitope Raphael, Jimoh, Adiasa, Munda, Josial, & Tehile, Agee (2012). Challenges of grid integration of wind power on power system grid integrity: A review. Renewable energy research. Baek, Jungho, & Gweisah, Guankerwon (November 2013). Does income inequality harm the environment?: Empirical evidence from the United States. Energy Policy, 62. Bayer, Benjamin (January 2018). Experience with auctions for wind power in Brazil. Renewable and Sustainable Energy Reviews, 81. Bruyn, S. M., Bergh, J. C., & Opschoor, J. B. (May 1998). Economic growth and emissions: reconsidering the empirical basis of environmental Kuznets curves. Ecological Economics, 25(2). Buckman, Greg, Sibley, Jon, & Bourne, Richard (September 2014). The large-scale solar feed-in tariff reverse auction in the Australian Capital Territory, Australia. Energy Policy, 72. Cassetta, Ernesto, Monarca, Umberto, RubinaNava, Consuelo, & Meleo, Linda (November 2017). Is the answer blowin' in the wind (auctions)? An assessment of the Italian support scheme. Energy Policy, 110. IREDA (Ed.). (2015). Compendium of state-wise solar policies of India. Dawn, Subhojit, Tiwari, Prashant Kumar, & Goswami, Arup Kumar (August 2017). An approach for efficient assessment of the performance of double auction competitive power market under variable imbalance cost due to high uncertain wind penetration. Renewable Energy, 108.

78

S. Thapar et al. / Energy for Sustainable Development 45 (2018) 66–78

Debra, Lew (2010). How do high levels of wind and solar impact the grid? The western wind and solar integration study. National Renewable Energy Laboratory. Ministry of New and Renewable Energy (Ed.). (2017). Growth of Indian Renewable Sector. Hirth, Lion (2013). The market value of variable renewables: The effect of solar wind power variability on their relative price. Energy Economics, 38, 218–236. https://wri.org/blog/2014/11/6-graphs-explain-world%E2%80%99s-top-10-emitters Ministry of Environment, Forest and Climate Change, Government of India (Ed.). (December 2015). India first biennial update report to UNFCCC. Indian solar resource map. The World Bank (Ed.). (2017). Solar resource data: Solargis. Kraft, John, & Kraft, Arthur (1978). On the relationship between energy and GNP. The Journal of Energy and Development, 3. Kreiss, Jan, Ehrhart, Karl-Martin, & Haufe, Marie-Christin (February 2017). Appropriate design of auctions for renewable energy support – Prequalifications and penalties. Energy Policy, 101. Mayr, Dieter, Schmidt, Johannes, & Schmid, Erwin (June 2014). The potentials of a reverse auction in allocating subsidies for cost-effective roof-top photovoltaic system deployment. Energy Policy, 69. Menegaki, Angeliki, Hanley, Nick, & Tsagarakis, Konstantinos (April 2007). The social acceptability and valuation of recycled water in Crete: A study of consumers' and farmers' attitudes. Ecological Economics, 62(1). NurcanKilinc-Ata (April 2016). The evaluation of renewable energy policies across EU countries and US states: An econometric approach. Energy for Sustainable Development, 31. Pablodel, Río (December 2017). Designing auctions for renewable electricity support. Best practices from around the world. Energy for Sustainable Development, 41. Pablodel, Río, & Pedro, Linares (July 2014). Back to the future? Rethinking auctions for renewable electricity support. Renewable and Sustainable Energy Reviews, 35. Ramachandra, T. V., Jain, Rishabh, & Krishnadas, Gautham (August 2011). Hotspots of solar potential in India. Renewable and Sustainable Energy Reviews, 15(6). Reaching India’s Renewable Energy Targets: Effective Project Allocation Mechanisms', undertaken by Climate Policy Initiative, web link of report is https:// climatepolicyinitiative.org/wp-content/uploads/2015/05/150512_Auctions_FINAL. pdf

Rego, Erik Eduardo, & Parente, Virginia (April 2013). Brazilian experience in electricity auctions: Comparing outcomes from new and old energy auctions as well as the application of the hybrid Anglo-Dutch design. Energy Policy, 55. IRENA (Ed.). (2017). Renewable power generation costs. IRENA (Ed.). (2017). Renewable power generation costs. Santana, Paulo Henriquede Mello (May 2016). Cost-effectiveness as energy policy mechanisms: The paradox of technology-neutral and technology-specific policies in the short and long term. Renewable and Sustainable Energy Reviews, 58. Shrimali, Gireesh, Konda, Charith, & Farooquee, Arsalan Ali (November 2016). Designing renewable energy auctions for India: Managing risks to maximize deployment and cost-effectiveness. Renewable Energy, 97. Shrimali, Gireesh, Nelson, David, Goel, Shobhit, Konda, Charith, & Kumar, Raj (November 2013). Renewable deployment in India: Financing costs and implications for policy. Energy Policy, 62. Shrimali, Gireesh, & Rohra, Sunali (October 2012). India's solar mission: A review. Renewable and Sustainable Energy Reviews, 16(8). Ministry of New and Renewable Energy (Ed.). (2017). Solar energy sector in India. Ministry of New and Renewable Energy (Ed.). (2017). Solar tariff trends in India. National Institute of Solar Energy (Ed.). (2014). State wise Solar Potential in India. Stigka, Eleni, Paravantis, John, & Mihalakakoua, Giouli (2014). Social acceptance of renewable energy sources: a review of contingent valuation applications. Renewable and Sustainable Energy Reviews, 32, 100–106. Thapar, Sapan, Sharma, Seema, & Verma, Ashu (2016). Economic and environmental effectiveness of renewable energy policy instruments: Best practices from India. Renewable and Sustainable Energy Reviews, 66. PV Exchange (Ed.). (2014). Trends of solar PV module cost. Winkler, Jenny, Magosch, Magdalena, & Ragwitz, Mario (April 2018). Effectiveness and efficiency of auctions for supporting renewable electricity – What can we learn from recent experiences? Renewable Energy, 119. International Energy Agency (Ed.). (2017). World energy outlook.