Saul Kriger Wilson: "The Origins of Land Finance: Local Resistance and Top-Down Imposition": A review

by Anirudh Burman.

Indian public bodies use several distinct land-related instruments for generating revenues from land-use. The National Land Monetization Corporation supports the disposal or lease of non-core public assets; development authorities auction or lease land; planning authorities and municipal bodies collect development or floor-space charges. These activities raise questions about the designs of the precise mechanisms that must be used, and whether the correct mechanisms are in place.

Saul Kriger Wilson's article asks a different, more specific question about China: why did local governments fail to collect the public revenue available from land transfers, and what changed that practice? In his telling, local governments engaged in transfers of land-use that severely undercut public revenue interests until the central government made concerted efforts to ensure transparent, auction-based transfers of land-use rights.

Wilson argues that fiscal opportunity alone did not ensure public receipts. Central political pressure, inspections and consequences for local officials helped make market pricing and payment collection effective requirements. In this review of his article I analyse his account alongside scholarship on fiscal incentives, land as development and credit infrastructure, and political control.

Fiscal opportunity and collected revenue

Wilson's argument relies on a distinction between the revenue a public authority might potentially earn and the actual collection. According to him, land receipts in China remained limited until the early 2000s, although local governments had opportunities to raise more. Negotiated transfers could yield less than competitive transactions, and agreed payments were not always collected. Wilson traces how local resistance and central intervention shaped that gap. His evidence includes historical examples, local cases and accounts of inspections and disciplinary action.(Wilson, pp. 2-4 .)

The conventional explanation links the expansion of land finance to the 1994 tax-sharing reform, which increased local spending responsibilities while reducing some local revenues. Wilson does not deny the reform's fiscal consequences. However, he argues that it cannot by itself explain when market-priced transactions and reliable collection took hold. Guangzhou sought bids for development rights in 1986, and Shenzhen sold term-limited land-use rights in 1987. These precedents show that land transactions predated the reform, while later resistance suggests that fiscal pressure did not automatically align officials' conduct with the public owner's interests. (Wilson, pp. 5-7.)

Wilson's account also qualifies any simple claim that valuable land or a need for revenue would lead officials to maximize public receipts. During the 1992-1993 property boom, officials in Beihai and Haikou continued to make discounted deals. Wilson uses contemporary accounts and disciplinary investigations to describe how officials and their networks could benefit from those arrangements. This is evidence about the cases he examines, not proof that local officials everywhere preferred private gains to public revenue. It does show why fiscal incentives alone cannot stand in for an account of who controls transactions and how rules are enforced.

Changes in land acquisition, housing and bank finance during the late 1990s made profitable transfers more feasible. Yet Beijing's 2002 restrictions on negotiated sales retained broad exceptions, and local self-examination did not reliably discipline municipal leaders. Wilson identifies a shift in 2003, when local leaders were assessed on implementation and failed inspections could threaten construction-land approvals and quotas. He describes Guangxi's response to repeated inspections and the later pressure on Beijing to reject pending negotiated projects and reorganise district land bureaux. ( Wilson, pp. 11-15.)

Wilson reports that Shenzhen collected only 25 per cent of contracted land revenue in 1995. During the later reform campaign, authorities identified 25.2 billion yuan in late, reduced or illegally discounted payments and recovered 17.6 billion yuan. Subsequent audits reported collection rates of 90 per cent or more of fees due. He also reports a 977 per cent increase in land-sale revenue between 2000 and 2004.( Wilson, pp. 2, 4, 6, 15-16 .)

Wilson states that corruption continued nonetheless. Officials could manipulate auctions or accept bribes even as developers paid substantial sums to government. His narrower claim is that sustained central intervention shifted the balance toward public collection and made market pricing more routine. The article's contribution is to explain the political and administrative work behind that change, rather than to treat higher fiscal need as a sufficient cause. ( Wilson, pp. 16-17.)

Three debates around Wilson's account

Fiscal arrangements and incentives:

Fiscal accounts ask how intergovernmental revenue and spending responsibilities shape local incentives. The 1994 reform matters to that question, but Wilson's chronology complicates this account, in which fiscal deprivation alone produced market-oriented land transfers. His outcome is not simply the availability of land or the growth of local borrowing. He focuses on whether public authorities priced transfers accurately and collected payments. The evidence he presents suggests that fiscal need and valuable land created incentives, but did not ensure that officials acted on them. Wilson therefore qualifies fiscal explanations of the timing of collection.

Land as development and credit infrastructure:

Meg Rithmire treats land institutions as part of China's system for managing investment and economic expansion. She writes that municipal governments could convey urban land and plan development, while higher authorities retained controls over development permissions. This helps explain why land transactions mattered to central and local development strategies, but it does not by itself explain whether officials charged the intended prices or collected the amounts due. Wilson supplies that account of implementation. The two arguments address linked outcomes and complete each other's accounts. (Rithmire, working paper, pp. 7-20, 29-32; Wilson, pp. 3-4, 16 .)

Liu, Oi and Zhang examine how local governments financed development after fiscal recentralisation. Their "grand bargain" account describes how the central government offered the promise and creation of locally controlled banks, local government financing vehicles and land collateral as parts of a borrowing arrangement that supported investment. This complements Wilson's narrative by adding to it. Local governments negotiated for autonomy in running credit institutions, financing mechanisms, in addition to, and despite the fact that the central government was pushing them to maximise land revenue. The land-finance story is not merely about selling land, but also about creating mechanisms that helped create a financial market for land finance at the local level.(Liu, Oi and Zhang, pp. 12-22 ; Lin and Yi, pp. 50-79.)

Political control and implementation:

An important aspect of Wilson's contribution is that it shifts attention from formal authority to the officials who apply it. Auction requirements had limited force when local bureaux could not discipline the municipal leaders who appointed and paid them. Inspections, performance assessments, disciplinary pressure and control over construction-land approvals changed the cost of resistance. The mechanism is not merely the existence of a central rule, but the centre's ability to monitor implementation and attach consequences to non-compliance. Wilson's account qualifies institutional explanations that treat formal market rules as self-executing.

These debates are complementary. Fiscal arrangements help explain why local governments sought resources; land institutions and credit explain how land supported investment; political control helps explain why potential transfer proceeds became more reliably collected. Wilson's contribution lies chiefly in the third question and its link to the first.

Thinking about land-value capture using Wilson's account

Wilson's account suggests questions for Indian research. Any comparison of public land auctions, town-planning schemes, development charges or other instruments would need to establish who has authority and ownership; who can acquire land and approve plans; how transactions are priced and receipts collected; which government controls implementation; which body receives and may spend the proceeds; what liabilities accompany the arrangement; and who bears its distributional costs. These dimensions differ across instruments and jurisdictions.

References

Saul Kriger Wilson. 2026. "The Origins of Land Finance: Local Resistance and Top-Down Imposition." The China Quarterly, First View, pp. 1-20.

Meg Elizabeth Rithmire. 2017. "Land Institutions and Chinese Political Economy: Institutional Complementarities and Macroeconomic Management."Politics & Society 45(1), pp. 123-153.

George C. S. Lin and Fangxin Yi. 2011. "Urbanization of Capital or Capitalization on Urban Land? Land Development and Local Public Finance in Urbanizing China."Urban Geography 32(1), pp. 50-79.

Adam Y. Liu, Jean C. Oi and Yi Zhang. 2022. "China's Local Government Debt: The Grand Bargain."The China Journal 87, pp. 40-71. Published online 6 December 2021.

Madhu Bharti and Shagun Mehrotra. 2020. "Metropolitan Ahmedabad: Scaling Up with Contiguous Replication of Town Planning Schemes." In Greater Than Parts: A Metropolitan Opportunity , Volume II, edited by Shagun Mehrotra, Lincoln L. Lewis, Mariana Orloff and Beth Olberding. Washington, DC: World Bank. Financing discussion: printed p. 23, PDF p. 30.

OECD. 2022. Global Compendium of Land Value Capture Policies: India.

How courts choose arbitrators

by Animesh Bordoloi and Prashant Narang.

A court hearing an application under section 11 of the Arbitration and Conciliation Act 1996 must answer two questions. Should an arbitrator be appointed? If so, who should it be? Indian arbitration law has paid far more attention to the first than to the second.

Our new working paper, How courts choose arbitrators: Evidence from the Bombay and Madras High Courts, examines the second question. It codes 372 reported and retrievable section 11 orders from 2016 to 2024, naming 387 appointments among 201 individuals, and draws on 36 semi-structured interviews.

A question the Act leaves open

The Act regulates the person once in view, covering agreed qualifications, independence and impartiality, disclosure, and fees. It says little about how candidates enter consideration, whether alternatives are weighed, or why one eligible person is preferred over another. Beyond qualifications agreed by the parties, the Act prescribes no positive qualification or accreditation criteria. Those norms are left to regulations by the Arbitration Council of India, which the Government told Parliament in February 2026 had still not been constituted.

This creates an observability problem. Judges may draw on professional knowledge, party suggestions or informal inquiries that never appear in the order. The record shows who was appointed and which steps were recorded. It cannot necessarily reveal how the choice was made.

What the orders record

The process is largely court-driven. In 315 of 372 orders (84.7%), the court named the arbitrator without recording both parties’ agreement to that person; such agreement appeared in 50 (13.4%). Consent to arbitration, or to “a retired judge”, is not agreement on who should hear the dispute.

Retired judges received 63.3% of named appointments, and advocates and senior advocates 33.9%. Technically qualified persons received six appointments across the two courts over nine years; former bureaucrats or senior officials, five.

Orders rarely explained the choice. Only 18 (4.8%) recorded any reason bearing on it, mostly a category of person such as a retired judge or legal expert. Six (1.6%) gave a ground tied to the individual appointed, and none compared candidates.

Disclosure practice differed sharply. Bombay directed a section 12 disclosure in 115 of 123 orders; Madras in four of 249. Section 11(8) places disclosure before appointment, yet 111 of the 119 directions came after the arbitrator was named. Some appointments may have remained conditional on disclosure, and an order’s silence does not prove none was obtained.

The orders recorded fee frameworks rather than the amounts eventually charged. High Court rules appeared in 51.1% of orders; 22.8% left the fee to the arbitrator, and 15.9% recorded no basis. Interviewees described the difficulty of negotiating remuneration with a tribunal already chosen to decide one’s dispute.

Repetition was substantial, with 186 of 387 named appointments going to someone the same court had already appointed. The ten most frequently appointed individuals received 31.6% of Bombay’s appointments and 34.3% of Madras’s. In Madras, most repeats came from the same judge appointing the same person, however they are counted.

What the record cannot show

These patterns do not establish favouritism. Repetition may reflect expertise, availability, reputation or a thin alternative pool. But professional knowledge and professional familiarity look alike from outside, and the orders rarely contain enough to tell them apart. The limit cuts both ways, since an arbitrator appointed twenty times on merit has no published record capable of showing it.

The study covers two High Courts and reported orders only. It cannot say how other courts appoint, or whether institutions do better. Its argument is about information. Identifying candidates, checking conflicts, confirming availability and settling fees need records that persist across cases. An institution with a continuing caseload can keep them; a rotating section 11 roster does not itself provide that continuity. That favours parties choosing an institution, appointment method and fee framework when they contract, with institutional performance still open to scrutiny.

Where courts continue to appoint, a narrower step is available. High Courts could publish periodic aggregate figures on appointment numbers, professional backgrounds, unique appointees, repeat appointments and concentration, all compiled from the orders themselves. That would constrain no individual appointment. It would make a public function observable, and let competing explanations be tested rather than asserted.

References

Animesh Bordoloi and Prashant Narang, How courts choose arbitrators: Evidence from the Bombay and Madras High Courts, TrustBridge Rule of Law Foundation Working Paper No. 25, 13 October 2026.

Prashant Narang and Vishnu Suresh, The Blind Spot in Indian Arbitration: Fees, Power, and Structural Oversights, The Leap Blog, 27 February 2025.

Animesh Bordoloi is Senior International Case Counsel at the Asian International Arbitration Centre, Malaysia. Prashant Narang is a researcher at the TrustBridge Rule of Law Foundation, New Delhi. We thank Renuka Sane for comments on earlier drafts.

A District-Level Dataset of Building Volume and Nighttime Lights

by Ayush Patnaik and Kapilan Mahalingam.

How fast is Pune growing, compared with Nagpur? Is Coimbatore catching up with Chennai? In India, these simple questions are surprisingly hard to answer.

The official statistics that describe the Indian economy are, for the most part, produced at state-level, once a year, often with a significant lag. Policy questions do not respect this. Infrastructure, industrial concentration and urbanisation play out at the district level, and the district is where the Indian statistical system is weakest. There is no official district-level measure of monthly economic activity. There is no official measure whatsoever of the stock of physical capital in a district, or of how fast it changes. For much of what we would like to know about a district, the source is still a census from fifteen years ago.

We have built a new dataset to address this very problem, using satellite data. For each of 640 Indian districts, it gives two measures. The first is nighttime lights, monthly, from 2014 onwards: a flow measure of economic activity. The second is building volume, annually, from 2016 to 2023: a stock measure of physical capital. Both are derived from satellite imagery, both are released as CSVs, and the code that produces them is public.

Satellites were not designed as instruments of economic measurement. But every few days they record the surface of the earth, and economic activity leaves traces on that surface. Buildings appear. Towns spread outwards. Settlements brighten or dim after sunset as activity ebbs and flows. Henderson, Storeygard and Weil (2012) showed that the growth of nighttime lights tracks economic growth in places where the national accounts are weak, and Chen and Nordhaus (2011) worked out the conditions under which luminosity adds information to conventional statistics rather than restating them.

The difficulty is that raw satellite data is not usable as-is. It arrives in unfamiliar formats. It must be corrected for a long list of biases, and processing it at any scale requires serious computational heft. A researcher who wants a district-level series faces months of engineering before they can run their first regression. The purpose of this dataset is to do that engineering once, publicly, reproducibly, so that nobody has to do it again.

Nighttime lights: a measure of economic activity

Nighttime lights are the older of the two measures, and economists have been using them for more than a decade. The data comes from the Visible Infrared Imaging Radiometer Suite (VIIRS) (Elvidge et al., 2021). We use the monthly series from 2014 onwards, at a resolution of 500 metres. For each pixel and month, VIIRS reports the average radiance and the number of cloud-free observations that went into it. Each district-month row in our dataset gives the sum of radiance across the district, the mean radiance per pixel, and the number of pixels.

Raw VIIRS data is messy. In earlier work at XKDR Forum (Patnaik, Shah and Thomas, 2022) we catalogued some of the more notable problems. Four, in particular. First, bloom: bright pixels spill light into their darker neighbours. Second, background noise: pixels with no lights at all do not report zero. Third, outliers: gas flares and fires produce readings that have nothing to do with economic activity. Fourth, and most important for India, seasonal attenuation: there is a marked and poorly understood correlation between the number of cloud-free observations and the measured radiance (Patnaik et al., 2021). Consequently, in the monsoon months, there are fewer cloud-free observations, and the recorded radiance is biased downwards. A new user of the data will not know to correct for this, and the seasonality of clouds shows up as spurious seasonalities of the economy.

Because lights have been studied for so long, solutions exist. We take the raw data and run it through the PSTT2021 pipeline of Patnaik et al. (2021), as implemented in NighttimeLights.jl, which corrects the bulk of these problems, and then aggregate to the district. The map below shows the result for 2021.

Nighttime lights by district, dark is dim

Lights are a flow. They respond quickly: a neighbourhood brightens as it becomes more active, and dims when activity slows. This is what makes them useful at a monthly frequency. Beyer, Chhabra, Galdo and Rama (2018) worked with cleaned VIIRS data at the district level and traced the effect of demonetisation on Indian districts month by month, an episode that annual, state-level data would have largely smoothed away.

A word of caution. As household and street lighting shifts to LEDs, measured radiance can fall even where activity is rising, because LEDs emit less in the band that VIIRS is sensitive to (Kyba et al., 2017). A dimmer district is not necessarily a slower one. Gibson, Olivia and Boe-Gibson (2020) survey this and the other traps in the economic use of nighttime lights, and are worth reading before treating radiance as a direct measure of output.

Building volume: a measure of the capital stock

The second measure is new. Google's Open Buildings 2.5D Temporal dataset (Sirko et al., 2023) is built from Sentinel-2 imagery at an effective resolution of 4 metres. It estimates not just where buildings stand but how tall they are, for every year from 2016 to 2023. Where the lights are high frequency, the buildings are high resolution. We resample the raw data to 100 metres, which is far quicker to process and is plenty for district-level work, and chunk the requests to stay within Google Earth Engine's limits. Each district-year row gives the total building footprint in the district, the mean building height, and the total building volume, which is the product of the two.

Building volume is a stock. It changes slowly, through the accumulation of investment in housing, industry and commerce. Buildings are rarely demolished on a scale that would show up at the district level, so in the ordinary course the series should rise, or stay flat, from one year to the next. The map below shows building volume per square metre of footprint in 2021, which is a rough measure of how tall a district builds rather than how much of it is built on. The dense corridors are visible without any further processing.

Building volume per square metre by district

These are the same districts in the same year, seen in concrete rather than in radiance. The densely built regions broadly match the bright ones, but not exactly. The Gangetic plain and the southern coasts are brighter than their building stock alone would suggest, and parts of the interior are dimmer. That gap is the reason for carrying both measures.

Here the contrast with nighttime lights is instructive. A decade of work has gone into understanding what is wrong with VIIRS data and how to fix it. For building volume, that literature does not yet exist. There are no papers cataloguing its biases and no established cleaning pipeline. We therefore release the building volume data as published by Google, without a cleaning pass of our own, and note its most significant problems in our appendix.

There is one respect in which building volume has more potential than lights, and it is the reason for optimism despite these problems. The height of a building is an objective, verifiable measure. Municipal building registers, property tax rolls and structures of known height give a ground truth against which the satellite estimate can be checked and corrected. Seth, Singh and Uday (2026) have begun to do this in Karnataka: working with building volume inside urban local body boundaries, they find that the summed volume of a body explains much of the variation in its property tax demand. There is no comparable ground truth for the radiance of a district. We are working on a bias correction for building heights that uses known buildings as anchors for the distribution, and we expect that the cleaning methodology for this data will develop in the coming years, as it did for lights.

Putting the two together

Each measure answers a straightforward question on its own. Together they answer a more useful one. Building volume is the stock of capital in a district; lights are the flow of activity running through it.

Across districts in a single year the two agree closely. In 2023 the correlation between log building volume and log radiance is 0.83, and the fitted slope is about 1.5: a district with twice the built volume of another is, on average, close to three times as bright. Some of that is size, since a larger district holds both more buildings and more lit pixels. What is useful is the spread around the line. The six districts furthest from it all lie on the same side, dim for what has been built, by factors of twenty to eighty, and five of the six are in the Himalayas or the Northeast.

Every district in 2023: building volume against summed radiance, both on log scales, with the fitted line. The named districts are the six furthest from it, all of them dimmer than their building stock would suggest

The difference in tempo is easiest to see when the two are drawn on the same axis. Building volume gives eight annual points across the period; lights give twelve times as many.

Mumbai Suburban, Pune and Surat: monthly radiance against annual building volume, both indexed to 2016

Sorting districts by whether the stock and the flow are rising gives four kinds of district. Where both are climbing, a district is likely in the midst of genuine, sustained growth. Where construction is rising but lights are flat, capital is being put in place ahead of use. This is what one sees in a place that is being built out, such as a new industrial area. It is also what one sees in a real estate bubble, where the buildings are made and nobody comes. Where lights are rising with little new construction, existing capital is being used more intensively. That is consistent with constrained land, with limited confidence in long-term growth, or simply with a choice to extract more from what is already there. Where both are falling, the district is in decline.

Every district, 2016 to 2023: the trend in building volume against the trend in nighttime lights. Most sit in the upper right; the named districts are the starkest case of each pattern, plotted in the next chart

Most districts sit in the upper right, where both measures are rising. The interesting cases are where the two disagree. Splitting each measure three ways, into rising, flat and falling, gives nine combinations, of which eight are occupied. The empty one is rising volume with dimming lights. The chart below shows the starkest district in each of the eight: Wayanad, where building volume and lights climbed together; Pithoragarh, where both fell; Central Delhi, where both are flat, as one expects of a district that is built out; Lucknow, where construction ran ahead of measured activity; and so on through the remaining four.

Building volume and nighttime lights, indexed to 2016, for the starkest district in each of the eight occupied combinations

Obtaining the data

This dataset is for public consumption. It is released with the generating code, so that every number in it can be reproduced. The data is available as CSVs (nighttime lights, building volume, and district boundaries). The processing pipeline runs end to end in two Colab notebooks, one for nighttime lights and one for building volume. One note: the nighttime lights CSV was computed at native resolution (500 metres), while the notebook downsamples in order to stay within Colab's usage limits. Point the notebooks at a different set of boundaries, whether another country or another administrative level, and the pipelines will run unchanged.

We built this for two audiences. The first is researchers working on regional inequality, urbanisation and the geography of growth, who have, until now, had to choose between district level data and time series data. The second is anyone in government who needs to know what is happening in a place where the official numbers arrive late, or not at all.

This is a first step. There is a great deal of work to be done: on the bias correction of building heights, on the LED problem in lights, and on validating both measures against ground truth wherever it exists. Economic development leaves visible traces on the landscape long before it reaches official databases. We hope that researchers will use this data, find its problems, and help us improve it.

References

Measuring districts' monthly economic activity from outer space, Robert Beyer, Esha Chhabra, Virgilio Galdo and Martin Rama, World Bank Policy Research Working Paper No. 8523, July 2018.

Using luminosity data as a proxy for economic statistics, Xi Chen and William D. Nordhaus, Proceedings of the National Academy of Sciences, Vol. 108, No. 21, May 2011.

VIIRS night-time lights, Christopher D. Elvidge, Kimberly Baugh, Mikhail Zhizhin, Feng Chi Hsu and Tilottama Ghosh, in Remote Sensing of Night-time Light, Routledge, 2021.

Night lights in economics: Sources and uses, John Gibson, Susan Olivia and Geua Boe-Gibson, Journal of Economic Surveys, Vol. 34, No. 5, December 2020.

Measuring economic growth from outer space, J. Vernon Henderson, Adam Storeygard and David N. Weil, American Economic Review, Vol. 102, No. 2, April 2012.

Artificially lit surface of Earth at night increasing in radiance and extent, Christopher C. M. Kyba and others, Science Advances, Vol. 3, No. 11, November 2017.

But clouds got in my way: Bias and bias correction of VIIRS nighttime lights data in the presence of clouds, Ayush Patnaik, Ajay Shah, Anshul Tayal and Susan Thomas, XKDR Forum Working Paper No. 7, October 2021.

Foundations for nighttime lights data analysis, Ayush Patnaik, Ajay Shah and Susan Thomas, XKDR Forum Working Paper No. 19, December 2022.

Estimating property tax potential in urban local bodies using satellite imagery, Abhishek Seth, Manish K. Singh and Diya Uday, XKDR Forum Working Paper No. 47, March 2026.

High-resolution building and road detection from Sentinel-2, Wojciech Sirko, Emmanuel Asiedu Brempong, Juliana T. C. Marcos, Abigail Annkah, Abel Korme, Mohammed Alewi Hassen, Krishna Sapkota, Tomer Shekel, Abdoulaye Diack, Sella Nevo and others, arXiv:2310.11622, 2023.

The authors are researchers at XKDR Forum, Mumbai.

Better public procurement for IT in India

by Ashok Pal Singh, Karthik Suresh, and Susan Thomas.

Information technologies (IT) are now ubiquitous in government. The Government of India has more than tripled its expenditure in real terms on large-scale "mission mode projects" over the last fifteen years. Yet, we see persistent evidence of failures (see examples here, here, and here). Failures in IT services therefore affect not only convenience but also people's ability to comply with the law. At the same time, there is a wide gap between contracting volumes and contracting capabilities in the Indian state. This makes government IT systems particularly vulnerable to changes in scope, features, and scale that may happen during the course of a project. These problems pose challenges for contract design through which this flow of change requests is processed by private vendors without exorbitant monopolistic pricing. IT systems are not one-off jobs that is akin to building bridges or even repeat jobs like purchasing medicines. This is a tricky problem to address.

In our new paper, Better public procurement for IT in India, we delve into three cases of large-scale IT projects in India --- Aadhaar, Passport Seva, and MCA-21. We ask, what factors drive the performance of IT procurement by the Indian government? We answer the question by analysing the performance of contracts and the underlying factors driving this for IT projects using a contract life-cycle approach. This involves looking at six aspects of each project: (i) the vision, purpose, and objectives of the project which help the government write the specifications and delegate all the other aspects of the project; (ii) the project's governance structure, (iii) the project's contracting model, (iv) how each project dealt with change requests, (v) how did the project deal with closure and transition into newer versions of the IT system, and (vi) did the project reasonably abide by cost and time estimates.

Our general findings are that the Indian government finds it difficult to take on the IT project after the vendor has completed the contract. Further, the government is unable to communicate, handle, and account for change requests in IT projects efficiently. IT projects carry some level of inherent risk and they require some level of adaptive contracting. At the same time, government projects are conducted using public funds and there are rules against profligate spending. Therefore, IT contracts should be designed in such a way that they confine to the strictures of the rules on public spending, while at the same time they are flexible enough to allow the project managers to adapt to change requests, technological improvements, and other developments. On these lines, we suggest the following prescriptions:

• Restricting the mandate of the project to a single goal/ outcome when multiple goals conflict with each other.
• Sufficiently empowering the project leadership team in terms of skillsets, administrative competences, and spending capabilities.
• An inter-disciplinary team that can complement each other's skills and allow for a space for deliberation, while also guarding against technological solutionism.
• Evidence-based decision making that follows the scientific process of discovery, evidence, and analysis.
• Government should take ownership of the project and use the services of consultants and experts for specific outcomes within the project framework.
• Transition management is as important as project management --- a well-defined transition plan that is written into the project document for the first iteration itself.
• Government's IT system should be available in the open source domain. There are significant positive spillovers when government IT systems are digital public goods i.e., open-source code on published websites which speak open APIs.
• Highly specialized IT projects are susceptible to opportunistic hold-ups and sunk costs. Bidders in India therefore tend to price these costs ex-ante i.e., by overstating the bid. This in turn incentivizes the government to use state power to pay only the marginal production cost to the bidder. As a solution, the literature suggests that the government "owns" the system/ asset.
• Contracts are differentiated on the basis of financial risks, operational risks, and ownership. In our view, the government should procure a working proof-of-concept, after which it can scale up the project.

Government IT projects can be thought of as being implemented in the following three phases: Phase 1 being the development of a system that demonstrates the proof of concept (PoC) ending in the delivery of a pilot implementation; Phase 2 where the pilot is scaled up to the state or national level; and Phase 3 where the project then moves into running an operational system. Successful implementations of public IT systems ideally are a mix of "make" and "buy" rather than just "buy". Phase 1 should done through "make" -- where the government itself builds the PoC and delivers a pilot implementation. Once the PoC is established, the IT system is like a well-defined product. With this in place, scaling up of the system over a state or the entire country can be optimally done through "buy", where private vendors can be contracted to deliver on a clearly defined output.

Ashok Pal Singh is a retired civil servant who served in various positions, including that of Deputy Director General with the Unique Identification Authority of India. Karthik Suresh is Assistant Professor at the National Law School, Bengaluru. Susan Thomas is Co-founder and Senior Research Fellow at XKDR Forum.

Spending as Planned: Efficiency of Drug Procurement in Tamil Nadu

by Pavithra Manivannan, Aakriti Narang, and Susan Thomas.

Introduction

A robust public financial management (PFM) system ensures that resources allocated in the budget are deployed as planned. One way to test this is the spending gap: the difference between what a state budgets for an item and what it actually spends. This gap offers a useful measure of a state's fiscal marksmanship. When looked further, a large spending gap may indicate that the state has either misestimated demand or struggled to manage suppliers, including ensuring timely delivery and payment. Both are critical to effective public procurement. The spending gap, therefore, offers one measure of a state's capability to procure well.

We apply this test to drug procurement. Tamil Nadu is widely regarded as the gold standard for drug procurement in India, and many researchers attribute this to the institutional design of its procurement agency, the Tamil Nadu Medical Services Corporation (TNMSC). Following TN's lead, many states have since set up similar bodies. In this article, we compare Tamil Nadu's spending gap in drug procurement with Karnataka, a state of comparable budget size, and population, that set up its own corporation in 2019. We ask - has Tamil Nadu's institutional development of its public healthcare system translated into capability in budgeting efficiency, for drug procurement?

Data and Methodology

We use the Detailed Demand for Grants (DDG) released by each state, to get the annual budgeted and actual expenditure numbers. "Drugs" include medicines, medical supplies and medical equipment¹. Hence, to identify expenditure on drugs we extract each of these specific line-items from the Department of Health and Family Welfare's budget of both the states.

This is not a straightforward exercise. State budgets are published only as PDFs running into hundreds of pages, and the data in them is, therefore, not research ready. Without structured data, there is no practical way to trace a budget line item across years and states. We overcome this with the India Public Finance Lab Portal, developed by XKDR Forum, which uses LLM pipelines to convert these budget PDFs into granular, research-ready data. Every line item sits in neatly labelled columns, so the data can be sliced, filtered and manipulated as needed. This lets us isolate and aggregate the different heads for both states and follow them year by year.

The portal covers data of ten years for Tamil Nadu and nine for Karnataka. To keep the comparison robust, we start in 2018-19, the first year with data for both states. Additionally, the actual expenditure for a year appears in the budget documents with a two-year lag, so actuals for 2025-26 and 2026-27 are not yet available. Therefore, we stop our analysis at 2024-25.

For each state and year, we calculate the spending gap from 2018-19 to 2024-25 (Study Period) as:

Spending gap_year = 100 × (Actual Expenditure_year − Budgeted Expenditure_year) / Budgeted Expenditure_year

A positive gap means the state's actual spending is more than its Budgeted Estimates (BE), and a negative gap means it spent less. We then compute the average and the standard deviation of the annual gap for each state. This helps us test our hypothesis; if Tamil Nadu has greater state capability for drug procurement, its average spending gap and variability of spending gap must be lower than Karnataka's.

Finding 1: Tamil Nadu has a higher budget for drugs

In absolute figures, we find that Tamil Nadu budgets a higher amount for drugs than Karnataka in every year of the Study Period, and by more than twice as much in most years. In Figure 1, we report the per capita drug expenditure of both the states. We find that, over the Study Period, Tamil Nadu's per capita drug expenditure averaged to Rs. 114 per person, about 2.2 times that of Karnataka's average of Rs. 53.4 per person.

Figure 1: Per capita drug expenditure for Tamil Nadu and Karnataka

Finding 2: Tamil Nadu's spending stays closer to its budget

Tamil Nadu's actual spending on drugs tracks its budget far more closely than Karnataka's, with a spending gap that stays within a much narrower band around zero. Over 2018-19 to 2024-25, Karnataka's average spending gap is more than three times Tamil Nadu's. The variability differs even more. The standard deviation of the gap for Karnataka is about nine times higher than for Tamil Nadu. Karnataka's gap peaked at over 400% in 2020-21, while Tamil Nadu's highest gap was 48%, in 2018-19.

Figure 2: Spending gap for Tamil Nadu and Karnataka, BE v. Actuals

	Average spending gap	Standard deviation
Karnataka	56.1%	184.7%
Tamil Nadu	17.3%	21.2%

Finding 3: The difference is not just COVID

Plausibly, much of this volatility in spending may just be driven by the unanticipated shock of COVID-19. We therefore recalculate the summary statistics after excluding 2020-21 for both states.

Without the year 2020-21

	Average spending gap	Standard deviation
Karnataka	-7.8%	81.7%
Tamil Nadu	13.4%	20.2%

Without 2020-21, Karnataka underspends on average, and its standard deviation falls by more than half. Although, it still remains about four times Tamil Nadu's. Tamil Nadu does not see as drastic a difference as Karnataka's. Figure 1 also shows that the two states responded differently to COVID. Tamil Nadu roughly doubled its drugs budget in 2020-21, while Karnataka raised its budget only gradually over the following years. This early increase may explain why the COVID year hit Tamil Nadu's spending gap far less: its budget absorbed the surge, whereas Karnataka's did not.

Finding 4: Karnataka corrects itself

We also know that the states publish revised estimates (RE), which update the budget estimate with more realistic figures. After looking at these numbers along with the budget and actuals, we see that Karnataka's revised estimates track its actual spending far more closely than its original budgets do.

Figure 3: Karnataka: BE vs RE vs Actuals

We also repeat the calculation against the RE:

(Revised) Spending gap_year = 100 × (Actual Expenditure_year − Revised Expenditure_year) / Revised Expenditure_year

	Average spending gap (vs RE)	Standard deviation (vs RE)
Karnataka	-18%	24.4%
Tamil Nadu	0.38%	15.9%

Against its revised estimates, Karnataka's gap falls. It spends about 18% less than its revised estimate on average, and its standard deviation falls from 184.7% to 24.4%. However, Tamil Nadu still seems to do better. Its actual spending lands almost exactly on its revised estimates, with a lower deviation as well.

Discussion

The main finding of this article is that Tamil Nadu's average spending gap (17.3%) remains substantially lower than Karnataka's (56.1%). The difference becomes even more pronounced during a crisis, when procurement systems are tested hardest. In 2020-21, Karnataka's spending gap rose above 400%, while Tamil Nadu's remained below 50%. The ability to estimate expenditure requirements accurately and spend broadly in line with budget allocations, even amid unexpected shocks, is an important marker of state capability and procurement efficiency. The pandemic emerged in December 2019, while the budgets for 2020-21 were presented in February 2020. It is plausible that Tamil Nadu anticipated the increased demand for medicines and budgeted accordingly. Although the spending gap alone cannot establish whether this was the case, Tamil Nadu's performance during this period suggests a greater capacity to maintain alignment between budgeted and actual expenditure under uncertain conditions.

We conjecture that this capability rests on two complementary sets of institutional capabilities: the budgeting capability of the TN Department of Finance and the procurement capability developed within TNMSC. The former ensures that resources are budgeted and released in line with requirements, and the latter enables the state to estimate demand accurately and manage the supply of medicines effectively.

First, the Department of Finance must have the capability to budget and spend as planned. Even when a procurement agency estimates demand accurately, the treasury must translate those estimates into realistic budget allocations and ensure that funds are available for expenditure as and when they are needed. Tamil Nadu's consistently narrow spending gap suggests that its Finance Department's broader budgeting framework is effective at aligning planned expenditure with actual spending. The inverse is also true. That is, procurement efficiency depends not only on the capabilities of the agency procuring medicines but also on the state's ability to finance and execute its plans. The spending gap is therefore an outcome of both procurement and budgeting capabilities.

Second, TNMSC appears to have developed three institutional capabilities that support more accurate demand estimation and effective supply management:

• First, is a rigorous approach to planning demand. Karnataka bases its demand on the requests raised by public healthcare institutions. TNMSC uses these requests too, but it also draws on a wider set of variables such as past consumption patterns, the changing demographics of the population served, the frequency of outbreaks and the disease profile of the population, and a 10% contingency stock on top of the estimated requirement.
• Second, these practices are not left to the discretion of individual officials. They are written into TNMSC's manuals and policies, so the forecasting method survives changes in personnel. Karnataka, by contrast, does not have a drug procurement manual.
• Third, Tamil Nadu's approach is also supported by technology. TNMSC uses a customised management information system, the Drug Distribution Management System (DDMS), which tracks stock and consumption at each warehouse and facility, and feeds it back directly into its demand estimates. Karnataka uses e-Aushadhi software, a standard platform used mainly for indenting and inventory management rather than for forecasting.

Tamil Nadu's experience suggests that procurement capability cannot be created through agencification alone. Establishing a dedicated procurement corporation modelled on TNMSC is an important first step, but it does not automatically confer the capabilities required to make the model work. These capabilities are built over time through experience, institutionalised procedures, and systems that enable decisions to be informed by reliable data. TNMSC, incorporated in 1994, has had considerably longer to develop these capabilities than Karnataka's Karnataka State Medical Supplies Corporation Limited (KSMSCL), established in 2019. This difference in institutional experience may help explain the gap in their performance, although it does not, by itself, establish causation.

This article compares two states, and the comparison could be extended in several directions. One, the same test could be applied to more states with different procurement models (such as long standing corporations vis-a-vis states that procure through their health departments) and varying information management systems. This would show whether the gap between Tamil Nadu and Karnataka reflects a broader pattern linked to institutional design, procurement planning and technology or experience and expertise. Second, states differ in how their health spending is funded. Where a larger share of drug spending depends on external funding, such as central schemes like the National Health Mission, the timing and size of fund releases are partly outside the state's control. The spending gap in these states may then reflect delays in funding as much as the state's own capability, and separating the two would sharpen the measure. Third, TNMSC procures only drugs, so a further question is whether Tamil Nadu's capability extends beyond them. How does Tamil Nadu perform on public healthcare procurement as a whole, and separately for goods, services and works?

Notes

¹ Section 3(b) of the Drugs and Cosmetics Act, 1940, read with the Drugs and Cosmetics Rules, 1945 and the Medical Devices Rules, 2017. ↩

References

Aggarwal et al (2025), Information Extraction from Fiscal Documents using LLMs, XKDR Forum.

Chitgupi, Gorsi and Thomas (2022), Learning by Doing, The Leap Blog, November 2022.

Hart, T., et al (2025), Rewiring public finance for the health sector, ODI Global Report, London: ODI Global.

Manivannan, Mehta and Thomas (2024), Improving supply chain management of medicines for public healthcare in India, XKDR Forum.

Manivannan (2026), Designing procurement information systems for improved access to public healthcare, Public Finance in the Digital Era: Emerging Themes 2026, ODI Global Report.

Parthasarathi, R. and Sinha, S. (2016), Towards a Better Health Care Delivery System: The Tamil Nadu model, National Library of Medicine.

Pavithra Manivannan, Aakriti Narang, and Susan Thomas are researchers at XKDR Forum. We thank Jay Kulkarni, Amrita Aggarwal, Ajay Shah and all the participants of the Roundtable on {X}ray of State Finances in India at XKDR Forum for their valuable suggestions and comments on this work.

Supply, stress and shortfalls: How India met its high electricity demand in September 2026

by Upasa Borah and Renuka Sane.

India has been touching new electricity demand peaks this year: 256.1 GW on 25th April, 265 GW on 20th May, and an all-time peak of 270.8 GW on 21st May. This demand has also been higher than the levels projected by the Central Electricity Authority (CEA). As demand grows, the ability of the system to provide sufficient supply when it is needed becomes increasingly important.

September 2026 offers a useful window into how the system balances demand and supply under pressure. Despite higher electricity generation relative to the previous year and high utilisation of the coal fleet, electricity shortages were the highest recorded since August 2023. The power exchange was also stressed during the month: average market clearing prices rose by 105% compared with September 2025, while purchase bids increased by 281% and sell bids fell by 45%. Over 100 captive power plants have been ordered to run at full capacity from October and sell surplus power on the exchanges to meet rising demand. Reports of critically low coal stocks at several coal power plants have added to concerns about supply for the coming months.

Against this backdrop, we examine electricity demand, supply and shortages at the transmission level in September 2026, defining shortages as the gap between electricity requirement and available supply. The article is descriptive: it sets out how the system behaved during the month rather than offering causal analysis. Establishing how demand, supply and shortages moved through the month gives a factual baseline against which causes and policy responses can be assessed. We ask three questions:

1. When did shortages emerge, and how did they vary across the days over the month?

   Daily and intra-day variations show how the system responded to high demand, and identify points at which it came under stress. This high-frequency analysis provides a snapshot of the state of the system, offering a useful starting point for understanding the operational challenges that are to become more important as electricity demand and renewable generation grow.

2. How did the power system respond to changes in demand?

   1. What was the composition of supply and its variation on the days of highest and lowest demand?
   2. How intensively were different generation sources utilised?
   3. How much renewable generation was curtailed?

   Installed capacity alone does not tell us how much electricity is available to meet demand at any given time. This depends on how much different sources can generate, how quickly they can respond to changes in demand and whether electricity can be delivered where it is needed. In the absence of storage, solar generation can meet a substantial share of daytime demand but falls as the sun sets. Thermal generation then needs to increase output, often quickly, though it can create technical and cost constraints. Hydro, oil and gas generation and storage can help manage these fluctuations, so their availability and responsiveness also matter. Solar and wind generation can get curtailed at times even when the plants are capable of generating more electricity due to transmission constraints limiting evacuation, or because the system has more power than it can absorb. Understanding these patterns is important for assessing how the power system manages both periods of scarcity and surplus.

3. How was the power exchange used during periods of high and low demand?

   Short-term markets also show how the system responds to changes in demand. Most of India's electricity is procured through long-term power purchase agreements, but buyers can buy additional power on the country's three exchanges. These offer segments across different time horizons, including the Day-Ahead Market (DAM), where electricity is traded a day before delivery, and the Real-Time Market (RTM), where trades take place an hour before. Such near-real-time markets help buyers respond to unexpected shifts in demand and generation, but their usefulness depends on how much electricity is offered and at what price. Both the DAM and RTM have a regulated price ceiling of Rs 10/kWh, which is often hit during high demand, leaving markets to clear at volumes well below what buyers bid for. A well-functioning market matters because it lets prices signal scarcity and guide decisions by generators and buyers.

Data sources

The study uses the following data sources:

1. Grid-India's Power Supply Position Reports: These reports are compiled by Grid India using data reported by State Load Dispatch Centres to Regional Load Dispatch Centres, which are then compiled at the national level. They provide information on energy supplied, energy shortages, peak demand met during solar and non-solar hours, and shortages during those periods. These values are measured at transmission buses and do not capture behind-the-meter generation like rooftop solar, or distribution-level constraints and outages. Our analysis thus focuses on system-level supply adequacy rather than delivery to end consumers or more localised grid constraints.

   The power supply position reports define shortage as the difference between unrestricted requirement and energy available, as per the Central Electricity Authority (Furnishing of Statistics, Returns and Information) Regulations, 2007. Energy availability implies supply at ex-bus, while unrestricted requirement is the sum of available energy, frequency correction, scheduled and unscheduled load shedding. The precise definitions and methodologies of these terms, however, could not be located, and it has been flagged in some Regional Power Committee meetings that there are inconsistencies in the methodologies adopted by states and a lack of transparency in the estimation of shortages. The results presented in this article should thus be interpreted with this caveat and a closer examination of the methodology underlying these estimates remains an important avenue for future work.

2. National Power Portal Daily Generation Reports: These are used to analyse how the generating fleet was operating and include data on daily capacity utilisation of thermal, nuclear and hydro power plants. Capacity utilisation here is defined as the ratio of a day's maximum generation to the available capacity on line, where available capacity is the installed capacity minus the capacity under outage. It is worth noting that these are daily aggregates rather than hourly measures. An hourly or 15-minute interval measure would provide a better indication of how much generation capacity was actually available to meet demand at different points in the day, but such data was not available.

3. Grid India's Daily VRE Report: These reports are used to measure the quantum of solar and wind generation that was curtailed daily. Data for 28, 29 and 30 September were not available.

4. India Energy Exchange's Day Ahead Market (DAM), Real-Time Market (RTM) and High Price Day Ahead Market (HP-DAM): These provide daily 15-minute blockwise data on aggregated purchase and sell bids, market-clearing quantities and market-clearing prices.

Demand and shortages in September 2026

India's electricity requirement in September 2026 stood at 1,62,736 GWh. Of this, about 545 GWh went unmet, implying a shortage of around 0.3% of total demand. These shortages varied across days and hours, but occurred on almost every day of the month, with 25 September being the only exception. Figure 1 shows the daily electricity required and met. Daily demand averaged around 5,424 GWh, with the highest requirement of 5,791 GWh recorded on 9 September; 51 GWh of this went unmet. The highest daily shortage was 54 GWh, recorded on 11 September. As demand declined towards the end of the month, shortages also fell.

Figure 1: Daily energy required and met

Peak demand refers to the highest quantum of electricity drawn from the grid at a point in time. There are two peaks generally observed, one during solar hours of 6 am to 6 pm, and another in the remaining non-solar hours. Figure 2 shows the daily peak demands during solar and non-solar hours, along with the shortage during these peaks. With the exception of a few days towards the end of the month, non-solar peak demand was lower than the solar-hour peak demand. Shortages, however, occurred during the non-solar peaks. On 9 September, for instance, the peak shortage was 7.72 GW at around 7:06 pm.

Figure 2: Solar and non-solar peak demand and shortages

Shortages emerging in non-solar hours, while the system met its highest solar-hour demand, is a defining feature of a grid with a growing share of solar. India's solar generation has increased significantly over the years, allowing the system to meet high levels of daytime demand. But this also leaves the question of how to meet demand as solar generation declines in the evening, particularly when demand remains high or increases.

Composition of electricity supply

Thermal plants generate around 71% of India's electricity, and non-fossil sources (renewables, nuclear and large hydro) the remaining 29%. The mix shifts by time of day and season. In September, coal accounted for an average of 65% of total generation, followed by renewables at 17% and hydro at around 11%.

When it comes to grid balancing, supply matters more in real time rather than in daily or monthly aggregates. We, therefore, study two days in September: 9 September, the day of highest electricity requirement, and 27 September, the day of lowest requirement. Figure 3 plots demand and supply at 15-minute intervals on the two days: On 9 September, solar accounted for 23% of total generation during solar hours (6 am to 6 pm), reaching a peak share of 33% at 1:15 pm when it generated 84.25 GW. On 27 September, solar accounted for 21% of generation during solar hours, reaching a peak of 30% at 1:45 pm when it generated 61.8 GW.

Figure 3 also plots the net load (dark solid line), defined as demand minus generation from solar and wind. The resulting profile shows the duck curve: as solar generation increases during the day, net load declines, reducing the amount of generation required from other sources. As solar generation wanes in the evening, while demand often stays high or increases, net load rises, and other sources must ramp up supply to meet the gap. This pattern is visible on both days, but on 9 September, higher demand meant that the evening ramp-up began earlier, around 2 pm onwards. Gas and hydro generation also increased during the morning and evening periods of lower solar generation.

This flexibility is important for grid operations. Thermal power plants often need time to increase and decrease output, and frequent, fast changes in generation can damage equipment or impose additional costs. Hydro and gas-based generation are better suited to respond quickly to changes in net load, and their greater use on 9 September reflects this role. Storage systems like pumped hydro plants and battery storage can also provide flexible, fast response, and play an important role in shifting generation across time, smoothing the net load associated with increasing renewable generation (Sanyal & Dev 2026). While storage was used on both 9 and 27 September, charging during solar hours and discharging in non-solar hours, its contribution was relatively small.

Figure 3: Source-wise generation at 15-minute intervals, on 9 and 27 September

Capacity utilisation

Figure 4 plots the maximum daily output of thermal, nuclear and hydro plants as a share of the generation capacity that was available to operate (installed capacity minus the capacity under outage). Hydro output remained high throughout the month, exceeding 90% and nearing 95% on several days. The high utilisation may partly reflect lower hydro availability this year due to weaker rainfall. Thermal generation also operated at close to 90% of available capacity. It is worth noting that daily aggregates hide real-time variations within a day. Maximum daily output need not coincide with the national solar or non-solar peak, and available capacity may not be fully dispatched because some generation is retained as operating headroom or reserves. Nonetheless, high utilisation rates suggest that there may be relatively little headroom to respond to unexpected increases in demand or sudden outages elsewhere in the system (Spencer and Vautrin 2026).

Figure 4: Capacity utilisation (maximum power output as a percentage of capacity on line)

Curtailment of renewable energy

On the other side of the supply story is the curtailment of solar and wind generation. Curtailment can occur for different reasons. Transmission constraints, grid-security requirements and other technical constraints may restrict the amount of renewable power being evacuated. Curtailment can also occur when there is surplus generation, particularly when demand is low or when thermal generators cannot reduce their output sufficiently. In such scenarios, the system operator can use the TRAS Down mechanism under the ancillary services framework to ask renewable generators to reduce their output even when they are technically able to generate more. This means available renewable energy is not fully utilised. In the first quarter of 2026, 300 GWh of solar and wind generation was curtailed. Better transmission infrastructure and deployment of storage systems can reduce such curtailment by allowing renewable energy to be evacuated or shifting surplus generation to periods when it is needed.

Figure 5 shows daily wind and solar curtailment for September, distinguishing between curtailment due to technical constraints and curtailment through TRAS Down. As electricity requirements eased by the end of the month, TRAS Down curtailment increased, reaching 93.58 GWh on 26 and 82.6 GWh on 27 September.

Figure 5: Solar and wind curtailment

Role of power exchanges

When the long-term contracted capacity of distribution companies is insufficient to meet demand, power exchanges provide a platform for short-term procurement. There are three power exchanges in India, but most of the trade happens on the India Energy Exchange (IEX). In our analysis, we focus on three of IEX's market segments that reflect procurement under different system conditions; Table 1 summarises their key features.

Table 1: Comparison of DAM, RTM and HP-DAM

	Day Ahead Market (DAM)	Real-Time Market (RTM)	High Price Day Ahead Market (HP-DAM)
Purpose	Procurement planned one day in advance	Balance demand and supply close to real time	Additional day-ahead procurement during periods of expected scarcity
Bids	One day before delivery	Every 15 minutes for delivery one hour ahead	One day before delivery
Activation	Used routinely for scheduled short-term procurement	Used to manage unexpected changes in demand and generation	Activated when expected prices exceed the DAM price ceiling
Price discovery	Market-clearing price at the point of demand and supply intersection, capped at Rs 10/kWh	Market-clearing price at the point of demand and supply intersection, capped at Rs 10/kWh	Market-clearing price at the point of demand and supply intersection, capped at Rs 20/kWh

In September, both the DAM and RTM were heavily stressed. Purchase bids were high in both markets, while sell bids and, consequently, market clearing volumes, were significantly lower. In the DAM, purchase bids were, on average, four times the volume of sell bids, reaching a maximum of eight times the sell bid on 17 September. In the RTM, purchase bids were, on average, five times sell bids. It was the highest at 13 times the sell bid on 16 September. The gap narrowed towards the end of the month as demand pressures eased (See Figures 6 and 7).

Figure 6: Purchase and sell bids, and market-clearing volumes in the IEX DAM

Figure 7: Purchase and sell bids, and market-clearing volumes in the IEX RTM

The divergence between purchase and sell bids can be understood by looking at both the available supply on the exchange and the market-clearing prices. Since much of India's generation capacity is tied up in long-term contracts, the amount of generation available for sale on the exchange can be limited, particularly during days of high demand. At the same time, the DAM and RTM operate under a CERC-mandated price ceiling of Rs 10/kWh. Figures 6 and 7 show the number of 15-minute blocks in each day during which the prices reached this ceiling. On most days, prices hit the ceiling in more than 60 of the 96 blocks. On 17 September, when the DAM recorded its highest purchase-to-sell bid ratio, the market clearing price reached the ceiling in 75 blocks. Similarly, in the RTM, prices reached the ceiling in 72 blocks on 16 September when the purchase-to-sell bid ratio was at its highest.

Figures 8 and 9 plot the purchase bids, sell bids, market clearing volumes and market clearing prices for each 15-minute block on 9 and 27 September in the DAM and RTM, respectively. The contrast between the two days illustrates how market conditions changed with changes in electricity requirements. On 9 September, the day with the highest electricity requirement, purchase bids were substantially higher than sell bids. Supply increased during the day, bringing down market clearing prices (MCP): in the DAM, the MCP fell to around Rs 2/kWh at 1 pm, while the RTM saw the lowest MCP of Rs 2.6/kWh at the same time block. But as solar generation declined and less supply was available, prices rose sharply and repeatedly hit the Rs 10/kWh ceiling.

On 27 September when electricity requirement was lower, market clearing volumes followed the quantum of purchase bids more closely and prices were considerably lower. The average MCP was Rs 2.76/kWh in the DAM and Rs 4.05/kWh in the RTM.

The difference in available short-term supply is also visible when comparing the two days: sell bids were consistently lower on 9 September than on 27 September for comparable time periods.

Taken together, these patterns suggest that both factors played a role during the period of highest demand. First, the price ceiling likely limited the market's ability to attract additional supply, and second, the small volume of sell bids indicates that spare generation capacity available for short-term trading may itself have been limited.

Figure 8: 15-minute blockwise buy and sell bids, MCV and MCP in the DAM

Figure 9: 15-minute blockwise buy and sell bids, MCV and MCP in the RTM

Another market in the IEX is the high price day ahead market (HP-DAM), which allows trades at prices above the ceiling price. The market was active from 4 till 17 September but traded volumes remained small, averaging around a total of 120 MWh per day. Unlike the DAM and RTM, it had higher sell bids relative to purchase bids, but on days that the market was active, HP DAM also hit its regulatory price ceiling of Rs 20/kWh (Figure 10).

Figure 10: Purchase and sell bids, and market-clearing volumes (MCV) in the IEX HP-DAM

Figure 11 plots market conditions in 15-minute blocks on 9 and 27 September. On 9 September, the HP-DAM was active in just 16 of the 96 blocks. Prices hit the Rs 20/kWh ceiling in all 16 blocks, with an average of only 9.5 MW clearing in each block. On 27 September, the market was not active, with no purchase or sell bids.

Figure 11: 15-minute blockwise buy and sell bids, MCV and MCP in the HP-DAM

Conclusion

India's electricity system is entering a period in which both demand and renewable generation are increasing rapidly. Our analysis of September 2026 shows that the power system was strained in four connected ways:

1. Shortages emerged on most days, and were concentrated during non-solar high-demand periods, when solar generation was low and the system had to rely more heavily on other sources. On 9 September, the peak shortage was 7.72 GW at around 7 pm.
2. Thermal and hydro plants ran near or above 90% of available capacity. This suggests little room to absorb demand spikes or outages.
3. The system faced both scarcity and surplus within the same month, as there was high curtailment of renewables through TRAS Down.
4. Prices on the IEX were lower when solar generation was abundant, and rose sharply as solar output declined and demand stayed high or increased. Prices hit the Rs 10/kWh ceiling in more than 60 of 96 blocks on most days.

As demand peaks rise and the share of renewable generation grows, ensuring reliability will increasingly depend not just on how much capacity is installed, but on when and where electricity is available, how quickly supply can respond, and whether it can be delivered to consumers. September's experience highlights the importance of examining these dimensions together when assessing India's electricity system.

References

Sanyal, Sanjeev and Dev, Satvik. 2026. The Duck and The Camel: Tracing the Net Load on the Indian Power Grid. EAC Working Paper Series EAC-PM/WP/53/2026.

Spencer, Thomas and Vautrin, Anthony. 2026. India's electricity demand grows at night: Managing rising cooling demand. International Energy Agency.

The code used to extract the monthly data and generate the figures in this blog is available on our GitHub repository.

The authors are researchers at TrustBridge Rule of Law Foundation. We thank Aryan Anand and Akshat Shukla for their assistance with the initial data extraction, and two anonymous referees for their helpful comments.