I saw this paper of Nick and really wanted to cover it. We often think that the barrier to adopting green energy is the cost of the technology. And while that is to a large extent true, there is often an overlooked institutional component. Namely, can you trust the energy regulator to actually honor the contract and pay the ex-ante agreed upon payments.
Because if not, the solar energy companies will want a much higher rate of return, and that could price solar out of the market. Which is what happened in India on a grand scale. And this is phenomenon is key to understand what could prevent the adoption of solar in the emerging economies, the countries which will have the biggest impact on how fast we can combat climate change.
With that said, enjoy the episode!
Transcript
Arvid Viaene:
Hi, welcome to another episode.
Today I want to talk about the fact that if we are going to fight climate change, we are going to need a decrease in brown energy and a large expansion of green energy. Within OECD countries, brown energy already looks to have peaked. Outside the OECD, however, brown energy is still growing substantially. And this is despite the fact that solar energy has become remarkably cheaper.
So in this episode, we are going to discuss what might actually be slowing down this green energy revolution, through a really good paper called Holding Up Green Energy: Counterparty Risk in the Indian Solar Power Market by Nick Ryan. (link)
I think this is an incredible paper, and it was recently published in Econometrica, one of the top economics journals. So I’m delighted to have Nick with me today to talk about it.
Nick Ryan is an Assistant Professor of Economics at Yale University. His research focuses on energy markets and environmental regulation in developing countries. His work explores how electricity markets, energy efficiency, pollution regulation, and renewable energy can support economic growth while reducing environmental costs. Nick received his PhD in economics from MIT in 2012 and a BA in economics from the University of Pennsylvania.
Nick, welcome to the podcast.
Nicholas Ryan:
Thanks very much, Arvid. Glad to be here.
Arvid Viaene:
To kick us off, could you talk about the central question you are trying to answer in this paper?
Nicholas Ryan:
As you said, the main strategy the world has been using against climate change, but also to support energy access and economic growth, has been a shift toward renewable energy. People often describe this in a capsule form as “electrify everything and green the grid.” Move more energy demand onto the electricity grid, and then build wind and solar power to supply it.
But when you look across the world, particularly in the early years of renewable energy deployment, the pattern was very uneven. You had real oddities, like more solar power being deployed in Germany than in the entire continent of Africa. That naturally raises the question: why?
The solar resources are very good in many parts of the world, including many developing countries. So the uneven pace of development is not simply because the sun does not shine there. The paper asks what institutional factors might cause renewable energy development to be slower in poorer countries.
The Risk of Building a Power Plant and Not Getting Paid
Arvid Viaene:
What setting do you study this in?
Nicholas Ryan:
There are many reasons why renewable energy projects might be harder to develop in poorer countries. The cost of capital might be higher. It may be logistically harder to build a new power plant. Equipment may need to be imported at higher transportation cost. The expertise needed to construct transmission lines may be less available domestically. Access to capital markets may also be more limited.
So there are a lot of possible explanations. The hard part in an empirical study is isolating one factor. The factor I focus on is the risk that investors in power plants will not get paid.
Why might that matter so much? Think about what a solar power plant does. You put up a large amount of money upfront. You build the plant at substantial capital cost, and then for the next 20 or 30 years you receive electricity at very low marginal cost. Once it is built, it just keeps generating.
That is one of the great attractions of renewable energy. Almost all of the cost is upfront. But imagine you are the investor who puts up the capital. Who is going to pay you back? You need to be confident that over the lifetime of the project, perhaps 20 or 30 years, you are actually going to receive payment for the electricity you produce.
In many countries, the buyers are utilities or state governments. These are natural monopolies that purchase electricity from large generators and distribute it to millions of customers. In India, many of these buyers have a poor historical track record of paying generators on time. There have been cases where investors built power plants in good faith and then were not paid as expected.
So the paper asks: if that is the investment environment, what effect does it have on investors’ willingness to build power plants in the first place? Could weak contract enforcement be one reason investment in green energy is lower?
The Hold-Up Problem
Arvid Viaene:
Could you explain this risk a bit more?
When I think about Europe or the US, there is also rooftop solar, where households may be involved directly. But the setting you study is more about large utility-scale solar projects, where the firm is dependent on a state utility or government buyer. And the risk of not being paid is much more important there.
Nicholas Ryan:
Exactly. When I talk about solar projects here, I mean large utility-scale solar farms covering hundreds of acres and producing hundreds of megawatts — enough electricity for thousands of homes. That is a very different market from rooftop solar.
The theoretical idea I draw on is what economists call the hold-up problem. That is where the title of the paper comes from.
Imagine that I have to make an investment that is particularly valuable to you, but I cannot write a perfectly enforceable contract guaranteeing that you will pay me afterward. If I know I may not get paid after making the investment, I become reluctant to invest in the first place.
Investment can then dry up entirely. Or, alternatively — and this is what I investigate in the paper — investors demand a risk premium. They say: “I think Arvid will probably pay me, but there is some chance he will not. So I need a higher return upfront to compensate for that risk.”
Arvid Viaene:
Exactly. It is like borrowing from a bank. The bank looks at your risk profile, and if you are a riskier borrower, it charges you a higher interest rate.
Nicholas Ryan:
Exactly.
Why Indian States Offer a Useful Research Setting
Arvid Viaene:
One of the graphs in your paper really illustrates how much variation there is across Indian states in the risk of default. Could you talk about that?
Nicholas Ryan:
Yes. From a research perspective, one very useful feature of India is that it is a federal democracy. States have substantial autonomy and largely run their own electricity sectors.
Different states therefore have very different reputations and historical records when it comes to paying electricity generators. India liberalized its electricity sector and allowed private firms to invest in generation around 25 years ago. Since then, states have differed substantially in whether they pay their bills to private generators on time.
One thing I do in the paper is use those payment histories and credit ratings and compare them with the prices states pay to new power plants. There is a very steep relationship.
The states that have poor payment histories tend to pay higher prices. The states that pay their bills reliably get a substantial discount.
That is persuasive, but it does not yet prove the effect is caused by risk. You can imagine many other reasons why a state that does not pay its bills on time might also have high electricity prices.
For example, Bihar might have poorer infrastructure. That could make it more expensive to build and operate a power plant there. So if firms bid a higher price in Bihar, perhaps they are simply covering higher construction costs rather than charging a risk premium.
That is the empirical problem the paper has to solve.
An Unusually Detailed Dataset of Solar Auctions
Arvid Viaene:
That brings us to the data, because I think you have an incredible dataset. Could you describe it?
Nicholas Ryan:
The data are critical, but what really makes the paper work is the institutional structure of how states purchase electricity.
The dataset covers bids from power producers offering to supply renewable energy, particularly solar power, to state distribution companies across India over roughly a decade. There are thousands of bids from hundreds of firms, including bids that were not ultimately accepted.
These are procurement auctions. Several companies might say: “This is the price at which I am willing to build a new solar plant and supply electricity in Bihar, Gujarat, Maharashtra, or another state.” Generally, the lower bids win.
But the really useful institutional feature is that the Indian central government also began running auctions on behalf of the states.
How the Central Government Acts Like a Co-Signer
Nicholas Ryan:
The analogy I like to use is someone graduating from college and trying to rent their first apartment.
You go to the landlord and say: “I just got a job. I’m going to earn enough. I promise I’ll pay the rent.” The landlord might respond: “You have no credit history. You have not even received your first paycheck yet. I’m not sure I can trust you.”
In that situation, a parent might co-sign the lease. That gives the landlord confidence because if the child does not pay, the parent is responsible.
Something similar happened in India. The central government observed that some states were paying much higher prices than others and began running solar auctions on behalf of those states through the Solar Energy Corporation of India, or SECI.
Suppose Andhra Pradesh wants to buy solar power. It can run an auction itself, with its state electricity distribution company directly responsible for paying the generator. Or it can procure power through the central government. SECI runs the auction and acts as an intermediary between the generator and the state. Nothing physical changes. The solar plant is still built in the same place and the electricity still goes to the same state. The difference is financial.
The central government is highly creditworthy. If I am a generator and win a SECI auction, but the state later fails to pay, I have the central government standing behind that payment.
It is very similar to the landlord knowing there is a co-signer on the lease. That mechanism allows me to compare auctions run directly by the states with auctions intermediated by the central government.
Arvid Viaene:
Before we get to the results, there is also a tension between the federal government setting renewable energy targets and the states actually having to buy the power. Could you explain how that works?
Nicholas Ryan:
Yes. This is a broader tension in the Indian electricity sector. Both the central government and the states have roles, but ultimately the states are the ones buying electricity and distributing it to customers.
If the central government runs an auction, there still has to be a state willing to buy that power. The central government cannot simply keep procuring unlimited electricity without an eventual buyer.
So you need both sides. You need demand from the states — they must actually want renewable energy — and you need the intermediation that reduces risk for the private generators supplying it.
Risky States Pay About 17 Percent More
Arvid Viaene:
So you have a really useful setting. On one side, states differ substantially in creditworthiness. On the other, you have central-government intermediation that can largely remove doubts about repayment. What do you find?
Nicholas Ryan:
The key idea is to compare how much states pay when they procure power themselves with how much they pay when the central government intermediates the transaction.
If the higher prices in risky states were simply due to higher construction costs, those costs would remain regardless of who runs the auction. If Bihar is simply an expensive place to build a solar plant, then a centrally run auction for a plant in Bihar should still result in a high price.
But that is not what we observe. When a risky state procures solar through the central government, it gets very low prices — roughly the same prices as the less risky states. The large difference appears only when the risky state runs the auction itself.
For risky states, the gap is around 17 percent between the price they pay on their own and the price they pay with central-government backing.
What convinces me that this is really about risk is what happens in low-risk states. Take Gujarat, which has a strong reputation as a place for businesses to invest. Gujarat pays essentially the same price whether it runs the auction itself or whether the central government runs it.
The risk premium appears only where the risk of hold-up is actually salient.
A 17 Percent Risk Premium Is Like Losing Two Years of Solar Progress
Arvid Viaene:
When you say the difference is 17 percent, how large is that economically?
Nicholas Ryan:
One possible reaction is: solar prices have fallen so dramatically, so who cares about another 17 percent?
It is true that prices have fallen enormously. During my sample period, solar prices were declining at around 8 percent per year, largely because solar PV modules and panels were becoming cheaper globally, driven in large part by enormous investment in China.
That decline in cost is one of the remarkable stories of the energy transition. But in that context, a 17 percent risk premium is roughly equivalent to two years of technological progress.
So you could say that counterparty risk puts the riskier Indian states around two years behind in the renewable energy transition. Equivalently, central-government intermediation accelerates their transition by roughly two years of solar cost declines. It is not as large as the entire technological revolution in solar. But it is still a meaningful drag against that backdrop.
Arvid Viaene:
And it is also comparable to the normal profit margin on some of these projects.
Nicholas Ryan:
Yes. Depending on where you look in the distribution, a typical project might earn a markup of around 10 or 12 percent. So a 17 percent increase is certainly large enough to wipe out the profit margin of a project.
Firms With More Leverage Charge Smaller Risk Premia
Arvid Viaene:
You also do some interesting checks to make sure that what you are observing really is a risk premium. One I particularly liked is that you look at solar companies that also operate in other parts of the energy sector. Could you explain that result?
Nicholas Ryan:
The idea is somewhat more speculative, but it is interesting.
Suppose I am worried that a state might not pay me. That concern could be smaller if I have some leverage over the state.
Some companies supplying solar also have large existing thermal power businesses. States depend on them for electricity from coal-fired power plants, which remain a major source of electricity in India.
If I am one of those companies, I may worry less about not being paid for solar because I know that if the state refuses to pay me, I could become less willing to continue supplying electricity from my other power plants.
So the hypothesis is that hold-up risk should be smaller for firms that have more leverage over the state. That is basically what I find.
Firms with broader generation portfolios — companies that operate both thermal and solar plants — do not charge the same risk premium in states where they already have substantial generating capacity.
The interpretation is that they may feel more confident they will be paid, or that they will be first in line for payment. A company that supplies only solar power has much less leverage. Once the solar plant has been built, it simply continues generating.
How Risk Changes the Amount of Solar That Gets Built
Arvid Viaene:
So you establish that there is a clear differential in the risk premium. The next question is how that translates into the actual supply of solar. A risk premium matters for prices, but what does it do to how much solar gets built?
Nicholas Ryan:
Solar, like any source of electricity, is never the only option. States have solar, wind, existing coal plants, some natural gas, nuclear, and hydro. They are assembling a portfolio of energy sources to supply power reliably while also trying to keep costs low.
If the risk premium disproportionately raises the cost of a technology with large upfront capital costs, like solar, then states may become more inclined to buy other sources of electricity instead.
In simple terms, a state’s demand for solar slopes downward. If the price goes up, it buys less. The research question is how to discipline that relationship empirically. How much less will a state buy if solar becomes more expensive?
A useful institutional feature is that I can see how many auctions states run, and I can also observe something close to their willingness to pay.
Many states set what is called a ceiling price in their auctions. They might say: “I will buy 500 megawatts of solar, but only if the price is below three rupees per kilowatt-hour,” or some other threshold.
States often looked across the country at the prices other states were getting. A state like Bihar might look at Gujarat and say: “A solar panel is a solar panel. Why should I have to pay more than Gujarat?”
So they set these aggressive ceiling prices. For an economist, that is useful because it directly tells us something about the state’s willingness to pay.
Why Risk Pushes Projects Above the Ceiling Price
Nicholas Ryan:
Once we observe those ceiling prices, we can ask what happens when risk pushes bids above them. Suppose a developer would have bid 2.8 rupees per kilowatt-hour in a low-risk state, which is below the ceiling. If the same project requires an additional risk premium in a riskier state, the bid may move above the ceiling. At that point, the state will not accept it.
Using a model of the auction, I can then ask: if this state had the same risk as a lower-risk state, how much more solar power would it have procured?
What I like about this comparison is that we do not need to make a very abstract assumption about how much the state is willing to pay. The state has already told us in the auction data: “This is our ceiling. We will not go above it.”
Risky States Respond by Using the Central Government More
Arvid Viaene:
Were the riskier states themselves aware that solar developers would require a risk premium? They might know they have a history of not always paying on time.
Nicholas Ryan:
That is a great question.
A state is made up of many people and institutions, and for any particular politician there is always the argument: “My predecessor may have broken a contract, but I will not. Why can’t investors trust me?”
Investors can judge for themselves whether that is credible. But what I think happened over time is that the riskier states began to observe the difference between procuring directly and procuring through the central government. You can see this beginning toward the end of my sample period and probably even more clearly afterward.
The riskier states gradually ran fewer of their own auctions and relied more heavily on central-government intermediation.
So over time, the higher-rated, more creditworthy states increasingly bought solar on their own, while less creditworthy states relied more on intermediated auctions.
Aggressive Price Ceilings Reduced Actual Procurement
Arvid Viaene:
Once you put all of this together — the ceiling prices, risk premiums, and auction bids — what do you find?
Nicholas Ryan:
The main thing is that because states were looking across the country at what solar cost in low-risk places, they often set their ceiling prices quite aggressively.
This was particularly important around 2017, 2018, and 2019, when there was a huge solar boom across India and prices differed substantially across states. Riskier states saw the low prices available elsewhere and set low ceilings themselves. That meant many projects that were close to being viable, and would have been viable without the risk premium, no longer cleared the auction.
The fraction of solar capacity offered in auctions that was actually procured therefore fell dramatically. In 2015 or 2016, a state might try to buy 100 megawatts and end up procuring 70 or 80. By 2019, a state might try to buy 100 megawatts and end up with only 30 or 40.
There was a large volume of developers effectively saying: “I would invest in this state, but not at the price you are offering.” I think the later growth in central-government intermediation helped address some of that. States that could not procure enough solar at acceptable prices began relying more heavily on the intermediated part of the market.
Why the Price Ceiling Is Not Really the Fundamental Problem
Arvid Viaene:
One thing I really liked in the paper is that as an economist, you see a price ceiling and immediately think: “There is the distortion.”
But you make the point that the ceiling itself is not necessarily the underlying problem. Even if you removed it, the fundamental issue would remain.
Nicholas Ryan:
Exactly.
You can criticize the ceiling price as a policy, and I do implicitly. But the ceiling is really just reflecting the fact that states have alternative sources of electricity. In most of the world, green energy ultimately needs to be cost competitive with thermal generation, particularly given characteristics like intermittency and unpredictability.
These states are not operating under something like the EU ETS, where there is a strong carbon price pushing them toward green energy. They were often adopting solar because it had become cost competitive. So even if you removed the explicit ceiling-price rule, the underlying trade-off would still exist.
The states could simply say: “Fine. If you will not let me cap the price, I will not run the auction. I can buy thermal electricity from the market at this price. Come back when solar can beat it.” That fundamental demand response remains.
Why De-Risking Can Matter as Much as Climate Policy
Arvid Viaene:
Which is why the central intermediation is so important. It is lowering the effective price of solar.
Nicholas Ryan:
Exactly.
Economists have traditionally placed a lot of emphasis in climate policy on regulations that deliberately raise the cost of fossil fuels because of their externalities. That is important, especially in places such as the EU or California.
But much of the emissions-reduction progress in the world has not come primarily from those policies. It has come from solar and wind becoming dramatically cheaper. Countries adopt them because they are economically attractive. The same may happen with electric vehicles if they continue to become cheaper and eventually outcompete internal combustion engine vehicles.
Could the Indian Model Work in Other Developing Countries?
Arvid Viaene:
One of the broader implications you raise is that not every country has something like the Indian central government available to play this intermediary role. How relevant do you think this hold-up problem is elsewhere?
Nicholas Ryan:
This is always the hardest part of an empirical paper.
You zoom into a particular institutional setting and use differences across places or auctions to measure something carefully. Then the question becomes: how relevant is this mechanism for the rest of the world?
My informal sense, based on talking to people and presenting this paper, is that it is everywhere. Counterparty risk is a primary concern for renewable energy investors in many countries.
One example I like is a World Bank project called Scaling Solar in Africa. The World Bank arranged a fairly elaborate system of financing and guarantees that effectively backed a solar project in Zambia.
That created a relationship broadly similar to what the Indian central government provides to Indian states. The difficult issue is that the World Bank’s capital is scarce. An intermediary cannot simply guarantee everything. It also needs leverage over the frontline payer. If the payer refuses to pay, the intermediary must be able to exert pressure and make sure the obligation is honored rather than simply absorbing the loss itself.
The Indian central government can do that with its states because it has a multidimensional relationship with them. It provides transfers, controls funding, and has other forms of influence.
The World Bank has a more arm’s-length relationship with a country like Zambia. The Scaling Solar project was viewed as a success and procured solar at a very low headline price, but it involved only one project and was not necessarily a scalable model for financing projects across many countries in Sub-Saharan Africa.
So an important question is whether institutions like the World Bank or IMF can back projects in a more capital-light way without creating incentives for countries to default on their obligations.
Could something analogous to SECI be built more generally to reduce the cost of solar in developing countries? That is an important open question.
Why Learning About Contract Risk Takes Time
Arvid Viaene:
I imagine there is also a lot of learning by doing in setting up these institutions.
Nicholas Ryan:
Certainly. There is always learning in the solar industry and in project deployment.
But learning about risk happens slowly and only at intervals. With long-term contracts, you do not really know how good the contract is until something changes. I saw this in another paper I wrote about coal contracts in India. Those contracts initially seemed to procure coal-generated electricity at very low prices, and everyone was happy. But there was a large untested risk: what happens if coal prices rise dramatically?
The contracts had not fully anticipated a two- or three-fold increase in coal costs. When that eventually happened, the risk was revealed.
The same applies to guarantees. You do not really learn how strong the guarantee is until the frontline payer actually defaults. That may happen only infrequently and after long periods.
So a lot of this is fundamentally about reputation and institutions — whether there is a credible authority behind the contract that can ensure the frontline payer honors its obligations.
The Solar Revolution May Still Be Just Beginning
Arvid Viaene:
Is there anything else about the paper that you would still like to highlight?
Nicholas Ryan:
I think we are still in very early days. Everybody knows that renewable energy prices have fallen dramatically. But the implications of that decline for the structure of energy markets have not yet fully played out.
I was recently reading a report from IRENA, the International Renewable Energy Agency. They calculate that the full installed cost of solar declined by around 88 percent globally between 2010 and 2025. Call it roughly a 90 percent decline. That is astonishing.
You could say: “That is great, and solar deployment has already grown enormously.” But the investment response to that price decline is really just beginning.
It takes a long time to replace old, long-lived power plants and to develop new uses for newly cheap renewable energy. It may seem as if the solar boom is already well underway, but in many respects it is only just out of the starting gate.
Closing Thoughts
Arvid Viaene:
I think that is a great message to end on. Nick, thank you for taking the time to discuss the paper, and congratulations on the publication.
Nicholas Ryan:
Of course. It has been a pleasure. Thank you, Arvid.
Arvid Viaene:
Thank you.
