Financial Instruments to Address Renewable Energy Project Risks in India

: This paper provides a summary of ﬁnancial instruments to address two biggest risks to renewable projects in India. These risks include the following: ﬁrst, off-taker (or counterparty) risk, which relates to payment delays by public-sector distribution companies to independent power producers, which then impact project level cash ﬂows in the domestic currency; second, currency (or foreign exchange) risk related to currency ﬂuctuations, which impact foreign investor level cash ﬂows in foreign currencies. This paper then describes multiple solutions for each of these risks, using public funding mechanisms. For payment delays, the category of solutions is termed Payment Security Mechanisms; whereas, for currency ﬂuctuations, the category of solutions is termed Foreign Exchange Hedging Facilities. The coverage in this paper shows the evolution of the solutions from theory to practice over time. These solutions are likely to be applicable to other developing countries.


India Has Ambitious Renewable Energy Targets
Electricity is a key driver of socio-economic development. To meet increasing electricity demand, the Government of India has been working to increase electricity capacity addition targets in successive National Electricity Plans. According to the National Electricity Plan, 2018 [1], the net energy requirement in India (accounting for savings on account of demand side management, energy efficiency and conservation measures) is likely to vary at a compounded annual growth rate (CAGR) of 6.18% between 2016-2017 and 2021-2022, and a CAGR of 5.51% between 2021-2022 and 2026-2027. Similarly, the peak demand is projected to grow at a CAGR of 6.88% between 2016-2017 and 2021-2022, and at 5.77% between 2021-2022 and 2026-2027. In contrast, globally, electricity demand is forecasted to grow at a CAGR of 1.9% per annum between 2015 and 2040 [2].
In addition to these energy supply goals, India also has renewable energy growth targets stemming from environmental goals. Under India's Intended Nationally Determined Contribution towards meeting the global climate change goals of limiting global warming to within 2 degrees over 2005 levels, India has targeted 33-35% of emissions intensity reduction of its GDP by 2030 over 2005 levels [3]. India will focus on achieving these targets primarily by increasing the proportion of renewable energy sources in its electricity generation mix to up to 40% by 2030. To this end, it has targeted 175 GW of renewable energy capacity addition by 2022, which does not include large hydroelectric projects.
Considering both growing electricity demand and renewable energy targets, India has adopted a generation planning approach, which includes the following: achieving sustainable development, power generation capacity to meet demand, fulfillment of desired operational characteristics such as reliability and flexibility, most efficient use of resources, fuel availability, and the integration of renewable energy sources.
The National Electricity Plan of 2018 estimates that the total investment required for generation capacity addition is INR 11.5 trillion for the period 2017-2022, which includes the funds required for renewable energy sources' capacity addition, as well as the advance action (i.e., expenditure incurred in advance of actual capacity deployment) on projects in the period 2022-2027. The total investment required for the period 2022-2027 is estimated to be INR 9.5 trillion but does not include advance action for projects coming up during the period 2027-2032 [1]. Here, INR stands for the Indian currency, also referred to as the Indian Rupee, and we typically use 1 USD = 70 INR unless specified otherwise.

Key Project Risks Are Counterparty and Currency
However, given the current policy and institutional environment in India, based on our primary research, investors-both domestic and foreign-may find it hard to meet the requirements due to multiple risks to utility scale renewable energy projects, including the following: foreign exchange, off-taker credit (or counterparty), regulatory/policy, etc. These risks are sometimes also referred to as barriers to deployment of renewable energy projects in practice [4]; however, we have used the word "risk" to indicate investor specificity. Addressing these risks would go a long way towards ensuring India's renewable energy goals receive the required level of investments. These risks are further explained in Table 1, which provides a comprehensive list of the barriers that projects face in practice, beyond high-level policy issues related to target setting as well as market design. Table 1. Risks faced by investors in renewable energy projects in India [5].

Risk Brief Description
Financing Foreign exchange risk Currency risk due to uncertain currency movements and high cost involved with market-based currency hedging solutions. Currency risk is a major barrier to foreign investments in developing countries. Currency crises, defined as a quick decline of a local currency, have triggered regional economic crises [6]. While all projects with foreign investments face currency risk, infrastructure projects are often exposed to greater risk because of longer terms and difficulty in redeployment of assets, making exit difficult for investors.
Off-taker credit (or counterparty) risk The risk that the buyer/off-taker will not fulfill its contractual obligations. It is a key contributor to the overall credit risk of a power project.

Quality of renewable energy Projects
The credit rating of the operational renewable energy assets may be low overall, leading to operational assets not meeting investment criteria.

Lack of instruments for investment
Lack of financial instruments (or pathways)-illiquid or liquid-to invest in renewable energy. Low returns compared to expectations Renewable energy projects not being able to meet the risk-return expectations of investors. Limited availability of debt capital Limited availability of debt capital due to capital market conditions, either domestically or internationally.

Completion
Construction risk Risks related to increase in overall financing cost due to construction related issues-esp. due to delays in construction due to permits.
Land acquisition issues Issues faced in land acquisition, esp. if there is no single window clearance in place, or if the time taken to obtain clearances is high.

Transmission evacuation
The lack of availability of transmission evacuation infrastructure, and time taken to obtains the clearances and permits.
Operational Curtailment issues Wind developers may face this issue during high wind seasons when higher than expected generation creates oversupply situations as well as congestion.

Contract enforceability risk
Drastic reduction in cost of renewable (e.g., solar) power generation may result in poor contract enforceability in the long-term, given that the buyers may want to cancel older, more expensive contracts, and move to newer, less expensive ones.

Others
Lack of trusted intermediaries Lack of trusted financial intermediaries, which allow indirect investments in projects via financial vehicles (e.g., green bonds), may result in new and/or smaller investors staying away from the sector.
Limited understanding of sector Many investors are not aware of renewable energy sector and, therefore, prefer to make investments in mainstream asset classes.
Regulatory/policy risk The risks related to uncertainty in availability of incentive schemes, poor implementation of policies and nonuniform policies across states.
In particular, [5] finds that currency and off-taker (or counterparty) risks are the biggest risks based on discussions with foreign investors, where these investors are asked to assign scores out of 10 regarding risk. The investors included: Bank of America, Blackrock, Generation Investment Management, EIG Partners, Goldman Sachs, Morgan Stanley, Silverlake Kraftwerk, TIAA CREF, UC Regents, etc. Some of this information was used to support the analysis in [5]. Table 2 indicates that currency and off-take risks are at least twice as highly rated as other risks. The research by [5] also finds in a similar discussion with domestic investors that off-taker risk is the top-rated risk for them. Therefore, given that these risks are the highest rated risks for all investors, and by a wide margin, the focus is mainly on these two risks in this paper. Table 2. Scoring of risks faced by investors in renewable energy projects in India [5].

Risk/Barrier
Score (Out of 10) Furthermore, [7] reports that follow-up discussions with investors verified these findings. These investors included: Bank of America, Barclays Finance, Blackrock, Citibank, Goldman Sachs, GE Capital, etc. These follow-up discussions further confirmed the earlier findings related to the high importance of currency and off-taker (or counterparty) risks. These findings have been verified by [8].

These Risks Can Be Partially Mitigated Using Publicly Funded Facilities
Given the focus on increasing renewable energy capacity, the Indian government needs to recognize the role that counterparty and currency risk mitigation mechanisms could play in expanding renewable energy capacity, and the rationale for creating such mechanisms using public funding sources. This goes beyond recognizing the role that policies can play in reducing investor risk perceptions [9]. These innovative financial structures can ensure that renewable energy deployment happens at scale [10], via publicprivate partnership [11].
For counterparty risk, there is an argument that the government should bear this risk because it is due to the primary off-takers, the distribution companies (DISCOMs), that are completely in the public domain. Further, there is an argument that government should bear currency risk in some strategic situations (e.g., renewable energy in the context of the Indian power sector); a key reason being that macroeconomic conditions are key drivers of currency movements and related foreign exchange rates, and government policy, in turn, influences macroeconomic conditions.
Thus, in what follows, designs of public funded counterparty and currency risk mitigation mechanisms are explored, in Sections 1.4 and 1.5, respectively. In Section 2, we provide the methods for calculating the sizing of facilities, whereas in Section 3 we provide the results. In each of these sections, we cover two solutions each for the counterparty and currency hedging risks. Finally, in Section 4, we summarize results, draw implications, and suggest future work. Our working hypothesis is that such risk-mitigation mechanisms can exist in practice, appropriate sizing is theoretically possible, and different methods may result in providing a suitable range of public funding required.
Before proceeding further, we note that the paper is quite conceptual in nature and builds on a significant body of work by the author [12][13][14][15], of which no direct precedent exists in the literature, whether academic or in practice. The key contribution of this paper is to bring these disparate pieces-some in academic literature and some in grey literaturetogether in a cohesive manner. Furthermore, while some of the data may appear a little dated, the issues and corresponding solutions are as relevant today as they were during the original development of these financial instruments during the period 2015-2018 when the original research was conducted. Therefore, the reader is encouraged to focus on the conceptual aspects as much as possible, given that they are applicable to not only India but also other developing countries.

Counterparty Risk Can Be Addressed Using Payment Security Mechanism (PSM)
The counterparty risk is related to the risk of non (or delayed) payment by the power purchaser (also known as the off-taker) to the power producer. From a lender's perspective, this may result in the power producer missing the debt payments on the power project, thus diminishing the credit quality of the project. The typical power purchasers in India are the public-sector, state-level distribution companies, also known as DISCOMs. Figure 1 shows the main components of power purchase, with arrows depicting the flow of energy and money flowing in the reverse direction. In this structure, the main problem lies with the DISCOMs who, due to their poor financial health, regularly delay payments. While more recent data is available (Power Finance Corporation, 2019), the overall situation has not changed much, and it is instructive to look at the DISCOM financial situation from about 5 years ago (i.e., in 2014-2015), and its implications for solar auction prices in 2015-2016 [16]. During 2014-2015, the DISCOMs had booked cumulative losses on the order of INR 633 billion [17], for two reasons. The first is economic-the DISCOMs did not even recover costs due to power tariffs being kept artificially low because of political pressures: in the same year, the average cost of purchase of power for the DISCOMs was INR 5.20/kWh whereas the average consumer tariff was INR 4.62/kWh. The second is operational; in the same year, the aggregate transmission and commercial (AT&C) losses stood at 24.62% [17].
This poor performance resulted in a combined negative net worth of DISCOMs at INR 1164 billion as of 31 March 2015, with loans outstanding at INR 6730 billion, receivables outstanding against banks at 92 days, and receivables outstanding against independent power producers at 121 days. The receivables outstanding gap clearly indicates that power producers are exposed to the risk of the poor financial health of the DISCOMs and the consequent risk of delays and/or defaults in payment.
In fact, based on author conversations with industry stakeholders, state DISCOMs have a history of delaying payments to independent power producers (IPPs) by up to as much as 24 months. This poses a direct risk to the ability of IPPs to meet their credit obligations and exposes debt investors to default scenarios. Based on our primary research, this causes banks and other debt providers to limit their investment to the renewable energy sector or otherwise raise the cost of debt provided due to the perception of higher risk.
The higher cost of capital available to the IPPs may have ultimately resulted in higher power tariffs ( Figure 2). This was evidenced in solar auctions in year 2015-2016, wherein the PPA prices were lower when the well-rated National Thermal Power Corporation (NTPC) was the off-taker. For example, a state auction held in Karnataka resulted in an average price of INR 5.07 per kWh while a NTPC auction-also held in Karnataka-achieved a price of INR 4.78 per kWh, equivalent to a saving of INR 0.29 per kWh [18]. The long-term solution to off-taker risk lies in proper management-operational and financial-of the DISCOMs, where states assume full responsibility for running the utilities on sound commercial principles [19]. A comprehensive set of measures is required to do so, including financial restructuring, tariff setting, revenue realization, subsidies' management, metering, and audit and monitoring. Recognizing this, in the past, the central government has introduced many schemes for financial restructuring of DISCOMs, but none of them have produced the intended outcomes [20].
The most recent example with some operational history is a financial restructuring scheme called UDAY [21]. The scheme involved four initiatives: improving operational efficiencies of DISCOMs; reducing the cost of power; reducing the interest cost of DISCOMs; and enforcing financial discipline on DISCOMs through alignment with state finances. The scheme aimed to achieve a reduction of average transmission and commercial (AT&C) loss to 15% by 2018-2019 as well as a reduction in gap between average cost of supply and average revenue realized to zero by 2018-2019. (Average transmission and commercial (AT&C) losses refer to not only electrical losses due to transmission and distribution but also commercial losses due to theft and nonpayment.) However, it is not clear if UDAY has brought about the expected changes [22], and another scheme called ADITYA has been announced as a follow on [23].
While a financial overhaul of DISCOMs is the necessary long-term solution to mitigate off-taker risk, and efforts should continue, there are also short-term solutions that can help drive renewable energy investments. Depending on the creditworthiness of the off-taker, a liquidity facility and/or a sovereign guarantee could support the off-takers' obligations [24]. This paper examines such a short-term solution, called a payment security mechanism (PSM). A PSM is a standalone fund that is a form of guarantee that covers the risk of payment delay in a power purchase agreement. The basic idea in a PSM is that this fund pays the power producer in the event of a payment delay and recovers this payment from the off-taker.
Recognizing the need for a PSM, the Indian government has proposed various PSM schemes [25]. However, the designs of these PSM schemes have been opaque, and one of the goals of this paper is to transparently investigate the theory behind PSMs. In this context, multiple approaches to PSMs for solar projects may be possible in the near term, as covered in Sections 2.1 and 2.2. The former is based on [14], whereas the latter is based on [22]. The basic difference between these two schemes is that the former uses an expected loss-based sizing approach whereas the latter uses a more direct approach for required credit enhancement. Both show that the expected PSM size would be 10-20% of the initial capital deployment in solar projects. This corresponds to a leverage, defined as the ratio of private capital deployed to the public capital provided to support the publicly funded facility, of 5-10. Note that 10% is equivalent to a leverage of 100%/10% = 10, and 20% is equivalent to a leverage of 5.

Currency Risk Can Be Addressed Using a Currency Hedging Facility (FXHF)
It is found that, to achieve India's renewable energy targets cost-effectively, more debt is required at attractive terms-i.e., with reduced costs and extended tenors [26]. From debt perspective, high costs (more than 12%), short tenors (less than 10 years), and variable rates (as opposed to fixed), end up increasing the cost of renewable energy in India by 24-32% compared to renewable energy projects elsewhere [26]. The difference in cost/terms of capital has not changed much over time and, therefore, this conclusion is unlikely to change.
Foreign loans (e.g., in USD) appear attractive for Indian policymakers, given that seemingly cheaper cost of capital (e.g., 5-7% USD), longer-term (12 years or more), and fixed-rate foreign loans have the potential to not only reduce the cost of renewable energy significantly but also reduce the cost of government support by making renewable energy more competitive with fossil based-electricity [12,27]. This raises the question as to why this issue cannot simply be fixed by developers borrowing directly in USD. The answer is that when developers borrow in foreign currencies, they must then worry about foreign exchange rate risk, as described below.
The reason that foreign exchange risk is an issue is that renewable projects earn revenues in local currency (e.g., in INR), when financing a renewable energy project by a foreign loan (e.g., in USD), the mismatch in the currency of debt obligations (i.e., USD) and currency of revenue (i.e., INR) exposes the project to the risk of devaluation in INR over time. This can result in reduced investments in the country due to currency risk, necessitating the use of a currency hedge (or currency swap) with a third-party provider to protect against these devaluations. A currency swap is an agreement to make a currency exchange between two parties. The agreement consists of swapping principal and interest payments on a loan made in one currency for principal and interest payments of a loan of equal value in another currency. Borrowers can lock in currency swaps with a third-party provider that takes the currency risk and charges a currency swap fee. These third-party providers are typically from the private sector (e.g., commercial banks).
In India, market-based currency hedging solutions are not only limited in availability (e.g., beyond 5-year tenors) but also have significant costs, increasing the final cost of debt, and almost eliminating the benefit of seemingly cheaper foreign loans. Further, most of the currency hedging in India happens in short-term (1 year or less) contracts, which makes the market for longer-term currency hedging less liquid. This does not mean it is not available at all. In fact, currency hedging contracts of up to 10 years may be available in the Indian market. For example, the typical cost of currency hedging in India is around 7% per year [28]; making completely hedged foreign loans as expensive as domestic loans-i.e., at 12-13% [26]. The difference in cost/terms of capital has not changed much over time and, therefore, this conclusion is unlikely to change.
Further, depending on the credit risk of the borrower, based on primary research, an additional credit-risk premium may increase the cost of currency hedging by another 100 basis points (bps), where 100 bps is equal to 1 percentage (%) point. Recall that the price of a market-based currency hedge reflects three components: cost of managing currency risk itself, cost of managing the credit risk of the counterparty, and margin for the currency hedge provider. Credit risk is the risk that a party (e.g., a project developer) to the swap agreement will default on its obligations. Currency swaps have high exposure to credit risk as they involve the exchange of money (e.g., USD and INR) over an extended period. Since a premium is charged for default risk, currency swaps lead to a double counting of credit risk as the borrower (e.g., the project developer) already pays a credit risk premium for the underlying debt to the creditor (e.g., a bank). Given that the debt provider and currency hedge provider can be different parties, credit risk becomes priced into not only the debt rate but also the price of the currency hedge.
Governments need to recognize the role that cheaper (and longer term) currency hedging mechanisms could play in expanding renewable energy capacity; essentially, these would enable the increased availability of cheaper, longer-term, fixed-rate debt to projects, reducing their cost of capital, thereby reducing the delivered cost of renewable energy, and making renewable energy more competitive.
The standard principle of risk allocation is based on allocating the risk to the party that may be able to best manage it. For power projects, the parties that can bear the currency risk are-the project developers, the government, or the customers. Project developers often bear currency risk. Sometimes, currency risk is passed on to the consumers. However, the government may be in a better position to bear currency risk as it can influence this risk. Further, given that government policies can influence macroeconomic conditions, which in turn are primary drivers of currency rates [29], there is an argument that governments should bear currency risk in some strategic situations.
The Indian government has in the past offered currency protection. However, this protection only applied to investments in roads and-most importantly-only in an event of default. This still does not cover the much more likely situation that the project does not default but the local currency depreciates significantly [30]. Given that governments may be in the best position to bear (and respond to) currency risk, they can choose to bear this risk in certain strategic situations, such as the deployment of renewable energy. In the case of India, another strong argument for government-sponsored currency hedging solutions is that bearing the currency risk for renewable energy today offsets the currency risk the economy would have borne in future on purchasing imported fossil fuels that the renewable energy would displace. This is particularly relevant for imported coal, which is the marginal fossil fuel that additional renewable energy is likely to replace [31].
Given that currency depreciation is a direct consequence of macroeconomic conditions, such as inflation, the long-term solution to control currency risk is to reduce inflation via sound macroeconomic policy that, for example, targets disciplined government spending and borrowing. However, controlling inflation may not always be possible in a fast-growing economy such as India and, therefore, short-term fixes may be required.
Multiple solutions may be possible in the near term. One potential solution is to use a structure where public money is used to provide a buffer against the risk of unexpected currency movements (Section 2.3), beyond the risk of expected movements that are already absorbed by the project developer. Another approach is to use market-based instruments (e.g., swaps) as much as possible to provide a tail risk guarantee (Section 2.4), where the project developer again absorbs the risk up to a certain threshold of currency devaluations. These approaches have been previously covered in detail in [12,13], respectively.

Payment Security Mechanism: Approach 1 (Z-Scores)
Payment risk is similar to credit risk. Both are legal obligations; where credit risk is directly related to the default risk in debt payments; payment risk is related to the default risk in accounts payable. For this discussion, a simplifying assumption is made, that defaulting on any legal obligation is equivalent, and hence the defaulting on debt payments is the same as defaulting on accounts payable. This assumption can be further defended on the grounds that both reflect an issue with the financial health of the off-taker. This allows us to use existing techniques for creating contingent facilities for credit risk management [32,33].
The framework for calculating the size of this contingent facility (i.e., PSM) uses elements of credit guarantees, specifically the following: (1) the probability of default (i.e., the likelihood that default would occur); (2) exposure at default (i.e., the amount not paid due to default); and (3) recovery after default (i.e., the percentage of exposure at default that is eventually recovered) [32,33]. The next paragraph discusses how each of these is calculated.
The probability of default is estimated using a modified version of the popular Z-score methodology [34,35], which uses key financial characteristics of the firm in a linear function, where the coefficients are derived from regression techniques on past data on default. Based on typical delays and power purchase agreement legalities in India, the exposure at default is estimated as the payment for one-year's-worth of electricity produced at the contracted per unit price. Finally, given that payments are always made eventually, the recovery after default as 100% of the guaranteed payment after delay. Since the DISCOMs are public sector entities, though they delay payments, they do not default due to regular bailouts by the central government.
The focus is on an existing PSM for a central solar power aggregator that buys power from multiple generators and sells power to multiple off-takers under the Jawaharlal Nehru Solar Mission. The Jawaharlal Nehru National Solar Mission initially set a target of 20 GW of solar power by 2022. This target was later revised to 100 GW of solar power by 2022 under the National Solar Mission. Recognizing that attracting investment for this target would necessitate a PSM, the government of India allocated some funds; however, we show that this amount was not enough. This existing PSM, from the Solar Energy Corporation of India, for 750 MW of solar capacity, was capitalized at INR 1700 million by the government of India. The expected size of this PSM-equal to the product of the probability of default and the exposure at default-is retroactively estimated. For this purpose, a sample of DISCOMs is selected representing the credit spread of all the DISCOMs [36].

Payment Security Mechanism: Approach 2 (Credit Enhancement)
As mentioned earlier, a Payment Security Mechanism is a standalone funded capital reserve that provides working capital to its beneficiary projects in the case of payment defaults by the DISCOMs. The aim of this section is to provide a framework for designing a PSM to explicitly achieve the intended credit enhancement-to investment grade credit rating (i.e., AA domestic)-in its beneficiary projects. Investment grade is the minimum rating-AA domestic or BBB-international-at which most investors, especially foreign investors, would invest. Typically, in India, this rating is enjoyed only by one off-taker-NTPC. The basic idea is to make the investors to be as comfortable with the beneficiary project's off-taker (i.e., a DISCOM) default risk as they would be with an investment grade off-taker.
Thus, the primary question is: How to size an efficient PSM for given project(s) exposed to a given DISCOM as off-taker that can help achieve the intended credit enhancement to investment grade? This essentially means sizing the PSM such that the probability of default of the project becomes that same as it would be under an investment-grade off-taker, such as the NTPC. In this process, intermediate questions are explored such as:

•
How does one create a probability distribution for payables defaulted (i.e., delayed) on by a given DISCOM? • How does the probabilistic cash flow distribution, and consequently, the probability of default of a given renewable energy project differ with and without a PSM of a given size?
The payment security mechanism developed in this section builds on existing frameworks for credit guarantees [14,33]. Such guarantees used in project finance are typically for enhancing the credit quality of the underlying project by means of providing protection against the defaults/delays in payment obligations due towards the project.
A stochastic methodology is used in this section, which is calibrated on empirically derived proxies for past payment history. Using these proxies, the optimum PSM size is derived, and the credit enhancement achieved for a project selling electricity to a particular DISCOM. As part of the methodology, two component risk scenarios are examined: • Scenario 1: The default by the project due to all risks outside of the specific counterparty (i.e., the DISCOM) risk. • Scenario 2: The default by the project due to the default/delay in payment by the counterparty (i.e., the DISCOM) risk.
While more complex methods for assessing credit risk are possible, estimation of the probability of default is usually a first step [37]. As a measure of credit risk, the probability of default by the project is examined in two cases: (1) without the presence of payment support (the base case), and (2) in the presence of a given payment security mechanism (postintervention). For each, the probability of default of a project has a one-one mapping with its credit ratings [20]. The underlying assumption is that, for project finance, credit ratings are completely specified by the probability of default [38]. For this paper, Indian domestic ratings are used, unless specified otherwise.
As mentioned above, the base case corresponds to the probability of default associated with the credit rating of the project without the PSM, whereas the second one corresponds to the probability of default associated with the target credit rating intended to be achieved by the intervention (i.e., the PSM). The target credit rating of the project is what could be achieved with an investment-grade off-taker, such as NTPC. The process below describes how these probabilities of default are calculated.
A five-step process is used to arrive at the required size (in number of months of payment obligations) of the PSM for a given DISCOM to achieve a target credit rating, as expected by investors:

•
Calculating the standalone probability of payment default by the DISCOM.

•
In the event of payment default, determining the probability distribution of the number of months of payment outstanding by the DISCOM. This is required for Step #4.

•
Calculating the probability of credit default for a sample renewable energy project with the DISCOM without PSM support.

•
Calculating probability of credit default for the sample renewable energy project with the DISCOM with a PSM support for a fixed number of months of payment.

•
Calculating the size of payment support (i.e., the PSM) to achieve the target reduction in probability of default.

Currency Hedging Facility: Approach 1 (Risk Buffer)
In providing currency hedging solutions for renewable projects, the following questions need to be considered: first, what are the expected costs of providing such hedging solutions? Second, how can the risks related to unexpected and extreme movements in foreign exchange (FX) rates be managed? Third, what is the market risk premium for taking these risks? Insights into these questions can be gathered by examining a governmentsponsored foreign exchange rate hedging facility ("FXHF"), as described below.
Under an FXHF, the government would provide project developers or off-takers a currency hedging solution through a standalone fund that covers debt payments for underlying hard currency (e.g., USD) loans. In this case, the government is not providing a sovereign guarantee, but rather is pre-committing public money for creating a standalone fund that can be used to provide cheaper currency hedging solutions. Typically, governments are averse to providing sovereign guarantees against their own currencies since that amounts to taking positions against their own macroeconomic policies. As seen below, the FXHF provides an indirect way to subsidize currency hedging without providing an explicit (or direct) subsidy to the project developer.
The working of the FXHF for a local currency power purchase agreement (PPA) can be explained as below. Under a local currency power purchase agreement, the project developer borrows in foreign currency (i.e., USD) and therefore, the foreign exchange risk exposure is borne by the project developer. In this case, the FXHF can enter a swap-via a contract for differences-with the project developer.
Under a contract for differences, the two counterparties-FXHF and developerswould sign a contract at a fixed (typically initial) foreign exchange rate and, over time, exchange payments for the differences between the actual and the contracted foreign exchange rates (see Figure 3). The frequency of this payment would be similar to the debt payment obligations of the project developer. For example, assuming that the fixed rate is 1 USD = 63 INR, then, at fixed periods when debt payments are due, if the foreign exchange rate is higher than 1 USD = 63 INR, the FXHF would make a net payment to the project developer. This net payment is equal to the difference of a variable payment (USD debt payments at the actual/current foreign exchange rate on the day of settlement) from the FXHF to the developer and the fixed payment from the developer to the FXHF (USD debt payments at the contracted foreign exchange rate of 1 USD = 63 INR). In the reverse situation, if the foreign exchange rate is lower than 1 USD = 63 INR, the project developer would make a net payment to the FXHF in a similar way.
The final design of the FXHF would depend on the underlying mix of loans. Here, an indicative analysis is provided based on assumptions from primary and secondary research. It is assumed that the underlying USD loan is at 5.5% cost of capital and is for 10 years tenor. It is also assumed that the market cost of providing a 10-year USD to INR currency swap would be 7 percentage points. That is, the eventually landed cost of capital for the 10-year loan would be 12.5% (= 5.5% + 7%), if the currency swap is procured from the market.

Currency Hedging Facility: Approach 2 (Tail Risk Guarantee)
When the INR interest rate is higher than the USD interest rate, risk-neutral and rational investors should expect the INR currency to depreciate against the dollar by the difference between the two interest rates. This way, borrowing at home and lending abroad or vice versa produces a zero-excess return. This is known as the uncovered interest rate parity (UIP) condition. There are currency hedging instruments available in the market that are priced using UIP or one of its variants.
For example, based on primary research, a long-term INR-USD cross currency swap costs around 750 basis points, which is priced based on parameters such as expected long-term interest rate differential (UIP component), volatility risk (due to fluctuations on UIP), liquidity risk (due to how shallow or illiquid the market is) and the cost of regulatory risk (due to changing regulation) capital. The cost mentioned is the average cost of principal plus interest commercial currency swaps. Hence, the market cost of currency swap increases the cost of a dollar debt available at a rate of 4-5% to more than 12% (post hedging) or nearly equal to the domestically available loans.
Further, depending on the credit risk of the counterparty (e.g., the project developer), the additional credit risk premium is around 50 basis points which takes the cost of currency hedging to~800 bps. Counterparty credit risk is the risk that a party (typically the buyer, i.e., the project developer) to the swap agreement will default on its obligations. Cross-currency swaps have high exposure to counterparty credit risk as they involve the exchange of notional amounts over an extended period. This risk increases with the length of the contract and can become a major barrier for long-term currency swaps. Swap providers assess the credit quality of the counterparty in determining whether to enter into a swap agreement. Hence, a premium is charged for this default risk. In a way, this leads to a double counting of credit risk as the counterparty already pays a premium for the underlying debt to the creditor.
A commercially available cross currency swap is used as a reference to compare the cost and benefits of the proposed FX Hedging Facility. A cross currency swap has three different components of cost as shown in Table 3 out of which:

•
The currency risk cost component can be directly subsidized, ignoring reduction in cost due to diversification such as in a portfolio of currency.

•
The counterparty credit risk and the liquidity risk cost components can be eliminated, where counterparty credit risk is a measure of risk of the loss of the amount that would not be recovered if a counterparty to a financial contract default in its payments, and liquidity risk is being implicitly subsidized but for simplicity it is assumed as eliminated here. Table 3. Cost components of a commercial swap, in basis points (bps) [13]. Both currency risk and counterparty credit risk components are derived from the cost provided by a commercial bank. The FX Hedging Facility is an alternative currency hedging mechanism to the market swap that incorporates components that aim to reduce the cost of currency hedging and increase leverage of private investment by the user (e.g., the project developer) of the facility, specifically through:

Cost of
• A direct subsidy for FX tail risk by a donor, such as the Indian government, or even other sources of subsidized capital such as family foundations. • Elimination of the counterparty credit risk and liquidity risk.

•
Passing on benefits to the user and/or donor if currency depreciation is lower than expected.
It is assumed that a sample user (e.g., a project developer) operating in the clean energy space and looking to raise foreign capital (e.g., the US dollar) wants a cheaper currency hedging solution and is willing to enter a contract with an FX hedging facility for a fixed FX depreciation rate of 4.5% per annum. The entity will absorb the FX depreciation until 4.5% (p.a.). If the FX depreciation is less than 4.5%, say 3%, then the upside, i.e., 1.5% will be transferred to the hedging facility. At the end of the hedging tenure, the accumulated capital in the facility can be shared between donor and user, as structured at the time of structuring the facility.
For the purposes of this paper, it is assumed here that such an upside will be retained by the donor and, hence, the effective cost of hedging will be fixed for the user. Beyond the FX depreciation of 4.5% per annum, a market-based FX tail risk guarantee will be used to cover the FX risk till a pre-specific limit, which we assume to be at the 3-sigma (i.e., P99.7) level of the FX probability distribution, which is assumed to follow a geometric Brownian motion process as explained in Section 2.3. Here, P99.7 means that there is 99.7% probability that the INR-USD exchange rate would be 87.64 or lower in 2017. The FX Tail Risk Guarantee will cover all losses due to currency risk till P99.7 levels (i.e., 3-sigma). It can be assumed that a rational guarantor will not provide unlimited risk coverage. However, the maximum guarantee coverage can be changed to any other level, for example P99.7, based on the comfort of the guarantor and the investor. The rationale behind taking a 4.5% per annum depreciation rate is discussed subsequently.
There can be several possible designs for the FX Hedging Facility to achieve the mentioned risk coverage. One potential design, which is likely to work in India, is shown in Figure 4. The project developer borrows from the foreign investor in foreign currency (e.g., in USD) and pays the debt back in the same currency. In this transaction structure, the project developer enters two contracts: • One with the FX hedging facility to cover currency risk up until 4.5% annual FX depreciation, and; • The other with an FX tail risk guarantor (e.g., a commercial bank) to cover currency risk beyond 4.5% till the P99.7 level. The FX hedging facility will keep (i) the donor capital to pay the FX tail risk premium upfront to the FX tail risk guarantor and will also hold (ii) the payment from the project developer (which includes the upside, if any) as the risk capital.
As indicated in the structure, the equivalent annualized cost of currency depreciation from 0% to 4.5% per annum will be paid by the project developer to maintain the risk capital for the mentioned risk coverage. When compared to a commercial swap, this annual depreciation rate of 4.5% translates into an annualized maximum cost of~528 basis points [13]. The maximum risk capital would be equivalent to the maximum cost of hedging sourced from the project developer. The maximum cost to the project developer = Total currency risk cost-FX tail risk cost (calculated in Table 4) = 650-134 = 516 bps (~528 bps). This cost translates to a maximum depreciation of~4.5% (p.a.). This is for the cross validation of the pricing. Using standard finance principles, this cost is arrived by calculating the annualized net present value-using the risk-free rate = 6.77% (Source: Bloomberg), tenor = 10 years-of the difference between the debt service payment at 0% and 4.5% per annum of FX depreciation and then dividing it with the notional size of the debt transaction. Beyond the FX depreciation of 4.5% per annum till the P99.7 level of depreciation, the risk coverage will be provided by the FX tail risk guarantee. The FX tail risk guarantee component will have two strike prices (i.e., INR-USD FX rate) in each of the 10 years, as shown in Table 5. Strike prices are the future FX rate (INR-USD) at which the FX tail risk guarantee will be exercised by the entity. The pricing of the guarantee depends upon these strike prices. The first strike price is derived from the minimum depreciation rate of 4.5% per annum and the second-strike price is derived from FX depreciation at the P99.7 level. Note that the FX probability distribution in each year is given by the assumed geometric Brownian motion process.

Payment Security Mechanism: Approach 1 (Z-Scores)
For the supported capacity (750 MW) of the existing payment security mechanism and based on the realistic assumption that the exposure at default is 12 months, the size of the payment security mechanism is estimated as less than 10% of capital costs of the solar power deployed, but at almost three times the size of the existing payment security mechanism deployed by the government [14]. That is, these results indicate that the existing provision for a payment security mechanism may not have been adequate in covering the risk of delayed payment from DISCOMs and highlight the need to use analytically rigorous approaches.
However, this PSM solution (based on expected loss sizing) does not assess the impact such a PSM would have on the credit ratings of the covered projects, or alternatively, the size of the PSM needed to achieve the desired credit enhancement (e.g., from BBB to AA). Additionally, this solution also misses a crucial piece of analysis comparing the expected benefits of such a facility with the cost of maintaining such a pool of capital. A sizing that considers the differential credit quality of DISCOMs would ensure a fair and efficient allocation of capital. We examine such an approach next in Section 2.2.

Payment Security Mechanism: Approach 2 (Credit Enhancement)
Using this methodology, the indicative number of months of payment support, needed to be provided by a PSM to ensure that a project with a given DISCOM (as the off-taker) would achieve a target credit rating, is calculated. These results are enumerated in Table 6 for eight representative DISCOMs. An "N/P" result indicates that it is not possible for a PSM of any size to achieve the given credit rating, i.e., even by eliminating the counterparty credit risk completely, such a rating may not be achieved. Table 6. DISCOM-wise table of size of Payment Support Needed to achieve target credit ratings [15]. "N/P" indicates PSM is not needed/cannot achieve the target rating. MoP stands for Ministry of Power. Some key insights and possible explanations are below. First, in this sample, the maximum possible credit rating that a PSM can help achieve is BB. It is found that a project with NTPC as the off-taker (i.e., with zero risk of credit default) has a probability of default of 1.51%, which corresponds to a BBB rating. Therefore, given that DISCOMs have higher credit risk than NTPC (the highest credit rated off-taker), projects facing varying levels of DISCOM default risk would naturally have credit ratings of BBB or worse. Since a PSM, however sizeable, will not completely negate the risk of off-taker default, it is rational that the maximum credit enhancement that can be affected using a PSM would be one band below a BBB rating, i.e., BB.
Second, it is found that most DISCOMs require moderately high payment support. Barring the outliers discussed above, the rest of the DISCOMs require 8-17 months payment support for associated projects to achieve a domestic BB rating. On average, it was found that these DISCOMs need about 12 months' worth of payment support, which is consistent with the assumption in Section 2.1. This translates to a PSM fund size equivalent to approximately 10-20% of the total capital expenditure of the project being supported; that is, the PSM size calculated in Section 2.1 is at the lower end of this estimation, and an expected loss-based sizing approach, while simple, may underestimate the size of the PSM required. The size requirement for this fund may be further reduced to 6-18% of the capital expenditure by requiring the DISCOM to furnish a revolving letter of credit of 3 months' payment, thus off-loading part of the cost of payment support to the DISCOMs.

Currency Hedging Facility: Approach 1 (Risk Buffer)
Starting with the first question: What are the expected costs of providing such hedging solutions? Analysis reveals that the expected cost-or the average cost across all potential outcomes represented by a probabilistic model-to provide a 10-year currency hedge via the FXHF is approximately 3.5 percentage points per year [12]; 50% below market rates, bringing the developer cost of capital to 9% (= 5.5% + 3.5%). In the context of a probabilistic model, the expected (or average) cost means a statistic that is higher than 50% of the potential cost outcomes and lower than the other 50%. This (i.e., 3.5%pts) is what the FXHF would charge the developer for providing the currency hedge.
However, governments should be aware of the risk exposure of the FXHF due to unexpected currency movements. That is, they should be aware of what will happen to the FXHF if the Indian currency depreciates more than the expected value in an extreme way. The FXHF would need to manage this risk; a risk that is typically managed by market.
This brings up the second question: How can the risks related to unexpected and extreme movements in foreign exchange rates be managed? One way to protect against this risk, and to ensure that the FXHF does not default, is to use a capital buffer, which is essentially a funded facility managing the risk of currency fluctuations.
Based on the analysis in [12], for the FXHF to achieve India's current sovereign rating of BBB-(i.e., to have the same default probability), the cumulative capital buffer requirement for 10 years would be almost 30% of the underlying loan amount; that is, with a leverage of approximately 3. The basic idea is to enable investors to view this investment as being as good as investing in the government of India's securities. The government of India is rated at BBB-from a global perspective, which is also considered investment grade, which is the minimum rating that a lot of investors would invest in. The basic steps behind this analysis are as follows: The difference time series (i.e., C = B − A) is the expected net payment from the FXHF to the developer. Calculate the net present value (NPV) of "C" at the government's cost of capital, denoted by "D". • Find an annuity on the original loan whose NPV is equivalent to "D". The implied rate on this annuity is the expected cost of the hedging provided by the FXHF-i.e., FXHF expected cost.
The government should also be aware that the expected cost of the FXHF of 3.5 percentage points does not consider the market cost of a capital buffer-i.e., the risk-premium that the market would place on taking the risk of unexpected and extreme movements in foreign exchange rates and maintaining this capital buffer.
This brings up the third question: What is the market risk premium-i.e., the difference between what the market would charge vs. what the FXHF is charging-for taking currency risks? Using foreign exchange option pricing theory [39], the risk-premium can be calculated as 2.76 percentage points, which largely accounts for the difference between the cost of currency hedging in the market and the expected cost of the FXHF [12]. The basic idea is to find 10 pairs of call and put options, one for each year, that hedge the currency risk. The market cost of the currency hedge is, essentially, the net present value of the difference of these option pairs. That is, this result confirms that the government is indirectly subsidizing the FXHF by keeping the capital buffer but not charging for the risk it mitigates.
In closing this subsection, a relevant question is whether other solutions are possible with higher leverages. A potential solution to avoid such large public commitments is to use a structure where public money is used to provide protection against currency devaluation in particular range, via a market-based instruments such as currency options [13], as discussed in Section 2.4. This approach shows that much higher leverages (up to 10) for public money can be achieved.
3.4. Currency Hedging Facility: Approach 2 (Tail Risk Guarantee) Table 4 shows the parameters and cost of FX tail risk guarantee based on the equivalence premium principle [40], and using the market-based option pricing model [39]. In actuarial science, under the equivalence principle, the premium is determined such that the expected size of the future loss is zero. The FX tail risk guarantee component is thus envisaged as a guarantee product which can operate outside the purview of Securities and Exchange Board of India (SEBI) regulation thereby allowing entities other than commercial banks to provide such a guarantee; hence the cost of the guarantee fee was estimated using the equivalence premium principle. However, to validate the pricing, a market-based option pricing method is also used, as discussed below.
To summarise, the developer pays the annual payments equivalent to 528 bps to the hedging facility to maintain the risk capital for FX depreciation from 0-4.5% p.a. This is a reduction of 30% from a commercially available FX swap. From 4.5% to P99.7 of FX depreciation, the FX tail risk guarantor provides the risk coverage and is paid the guarantee premium of 134 bps by the donor capital upfront through donor capital as a direct subsidy [13].
As shown in Table 3, the market cost of currency risk is 650 bps. Based on above calculations, the cost of currency risk in an FX tail risk guarantee transaction is 528 + 134 = 662 bps, which is quite close to the market cost of managing currency risk, i.e., 650bps. Hence, our pricing is not violating the efficient (market) pricing hypothesis.
It can then be shown that this donor capital achieves a leverage of 9 [13], i.e., approximately three times higher than the leverage obtained under the previous scheme that uses an explicit risk buffer (Section 2.3). The increased leverage comes from the observation that the donor capital just covers the price of the currency hedge between 4.5% to P99.7 of FX depreciation, as opposed to supporting the risk buffer itself.

Conclusions
This paper provides a summary of financial instruments to address key risks to renewable projects in India. These risks include the following: first, payment delays by distribution companies to independent power producers, which impact project level cash flows in the domestic currency; second, currency fluctuations, which impact foreign investor level cash flows in foreign currencies.
Multiple solutions are then described for each of these risks, using public funding mechanisms. For payment delays, the category of solutions is termed as Payment Security Mechanisms (PSM); whereas, for currency fluctuations, the category of solutions is termed as Foreign Exchange Hedging Facilities (FXHF). The coverage shows the evolution of the solutions from theory to practice over time. These solutions are likely to be of value to other developing countries, given that they may face similar risks [41].
In case of PSMs, two solutions are explored, one based on Z-scores to cover the expected loss, and another based on direct credit enhancement to have the underlying debt rated at the highest level possible. In the former case, the size of the PSM is found to be less than 10% of capital costs of the solar power deployed; whereas in the latter case, the size of the PSM is found to be 10-20% of capital costs of the solar power deployed. Thus, independent of the method used, the size of the PSM can be significant, from a low of 10% to a high of 20%, with corresponding leverages of 5-10. Note that 10% of public capital requirement with respect to private capital refers to a leverage of 10, whereas 20% refers to a leverage of 5.
In the case of FXHFs, two solutions are again explored that are fundamentally different. The first one is based on creating a contingency buffer to directly absorb tail FX risk, whereas the second one focuses on absorbing certain tranches of FX risk while utilizing market available FX swaps. It is found that the leverage, i.e., the amount of private capital deployed per unit of public capital committed, varies in the range 3-9, with the lower number corresponding to the first scheme and the higher number corresponding to the second scheme.
The implications of our results are manifold. First, they advance applied theory. Second, these results provide concrete recommendations on the sizing of various funded facilities that could address key risks to renewable energy projects. Third, they suggest that policymakers utilize public money to fund these facilities, while appropriately structuring them in consultation with various stakeholders, such as ministries and financial institutions. Finally, we acknowledge that these solutions are applicable to broader international contexts, given that counterparty and current risks are commonly the biggest risks that investors face, in particular in developing countries.
We acknowledge that our proposed solutions are not the final word. The limitations of our work include issues around the availability of trusted data. Thus, we believe that our results can be improved via the application of more reliable data. Furthermore, other methods may be applied to the sizing problems at hand, and corresponding results could be compared with ours. Finally, a lot more work needs to be conducted to implement these schemes in practice.