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.

Solar and batteries have beaten coal pricing in coal's game: Frequently asked questions

by Nikit Abhyankar, Amol Phadke and Ajay Shah.

We recently published three articles, triggered by the remarkable SECI tender for thermal-mimicking power, where solar power and batteries achieved a lower price than new coal:

1. Shah, Ajay. 2026. "Consequences of cheap electricity." Business Standard, October 12, 2026.
   This article examines how the rapidly falling costs of solar power and battery storage are transforming India's industrial landscape. The author highlights that India is emerging as a favorable location for energy costs compared to the US, creating new strategic advantages for energy-intensive sectors like green steel and aluminum.
2. Shah, Ajay, Nikit Abhyankar, and Amol Phadke. 2026. "A tipping point in Indian energy." Business Standard, September 27, 2026.
   This article examines how the rapidly falling costs of solar power and battery storage are driving a non-linear shift towards renewables in India. This transition, increasingly driven by economics rather than decree, has profound implications for national energy procurement, renewable energy investments, and the future viability of existing coal assets.
3. Phadke, Amol, Nikit Abhyankar, and Umed Paliwal. 2026. "Renewable Energy with Storage Can Match Coal's Reliability and Operating Profile at a Lower, Fixed Price: Evidence from India's Market Test." The Leap Blog, September 23, 2026.
   Despite dramatic declines in clean-energy costs, over $1 trillion in conventional plant investments is globally under consideration because fossil fuels are assumed essential for round-the-clock power. This article demonstrates how firm renewable power bundled with battery storage in India can now match coal's reliability at a lower, inflation-proof fixed price.

These articles generated substantial debate. This FAQ distils those discussions.

Q: Does the SECI Thermal Mimic tender lack an annual Demand Fulfilment Requirement (DFR), allowing excessive permissible non-compliance? The tender requires 15-minute block compliance. Generators must supply at least 90% of contracted capacity during six buyer-nominated peak hours, 70% during other non-solar hours, and 50–60% during solar hours. Shortfalls face a penalty of 1.5 times the contract tariff. The implicit 70% annual DFR deliberately mimics coal plants, which are backed down during solar hours to avoid variable costs. Increasing the DFR to 90% during solar and non-peak hours actually lowers the per-unit cost, as generators receive full payment during inexpensive solar hours (for example, see Paliwal et al. 2026, which assesses cost of achieving 90% annual DFR).

Q: Coal plant maintenance is schedulable based on system requirements. Are renewable energy (RE) shortfalls dictated by unpredictable weather, making them fundamentally less reliable? It is incorrect to assume coal outages are entirely planned. Central Electricity Authority (CEA) daily generation data over 20 years shows coal plants suffer 5–7% forced (unscheduled) outages, frequently during peak demand seasons. While solar and storage face monsoon challenges, their availability during peak non-monsoon seasons exceeds 90%. Moreover, fossil plant reliability is declining globally; capacity credits for gas plants in US markets have been downgraded below 70% due to correlated weather failures.

As an example, consider the CEA report for 12 October 2026. On this day, 12.33% of the 225.19 GW of monitored coal generation (i.e. 27.76 GW) went down. Of this, 5.58 GW was planned and 22.05 GW was not. That day, NLDC reported a 4.27 GW evening shortage at 19:00 and 15.45 GWh of energy shortage over the day. Words like coal and base load evoke something in our minds that is not quite a description of the reality.

Q: Are weather-related failures in renewable generation highly correlated, making fleet-level reliability impossible without massive overbuilds? Assessments of 100 Indian sites over 10 weather years demonstrate that geographic diversity significantly mitigates correlation (Paliwal et al. 2026, similar findings for the US, see Caldeira et al. 2018). A monsoon week does not uniformly eliminate generation across the entire subcontinent. Uncorrelated forced outages in RE plus storage can be managed using a standard reserve margin of 10–15%, identical to the reserve margins required to backstop discrete unit trips in coal fleets.

Q: What is the marginal cost of extra reliability from solar and storage? Does covering the worst hours result in an expensive leftover? Forcing a single solar and storage site to meet 100% of demand across all 8,760 hours of the year is an artificial constraint that inflates costs to roughly ₹9/kWh. No power grid demands 100% guaranteed output in every hour from a single plant. Using a portfolio approach with a 10% reserve margin, 24/7 clean power with 99.9% reliability costs under ₹6/kWh. The SECI auction discovered a firm renewable power price of ₹5.25–5.26/kWh, fixed for 25 years. With inflation, this fixed nominal tariff equates to roughly ₹3.79/kWh in real 2026 rupees.

Q: The ₹5.25/kWh product is not perfectly firm. Does it genuinely compare to coal? The SECI contract explicitly tests whether a resource can meet the dispatchability requirements the power grid actually needs, not a hypothetical flat load. Roughly 80% of existing Indian coal plants would fail to consistently meet the 15-minute availability profile enforced by the SECI tender. Recent competitively procured coal PPAs opened at ₹5.38–6.30/kWh. Because coal tariffs include escalating variable fuel costs, average new coal PPA prices could reach ₹7.4/kWh by 2050. The thermal-mimic tariff remains fixed at ₹5.25/kWh, avoiding fuel-price and freight cost inflation. The thermal-mimic price is 10 per cent cheaper than new coal.

Q: Do least-cost capacity expansion models indicate a need for new coal? No. Evaluating hundreds of thousands of resource-mix combinations across multiple weather years from 2030 to 2050 demonstrates that new coal is not cost-effective beyond the roughly 35 GW already under construction. Existing coal and wind assets will optimally backstop solar and storage.

Q: Should the electricity system be centrally planned to guarantee 100% supply-side reliability? Attempting to build hardware that solves for every eventual peak through central planning suffers from the socialist calculation problem. It is societally inefficient to build supply assets assuming perfectly inelastic demand. The SECI tender serves as a clarifying intellectual device demonstrating that solar and batteries can beat coal on its own terms. However, the future of electricity involves a price system with dynamically fluctuating intra-day prices. Price signals drive supply, storage and demand responses. The most important solution to evening shortages in a solar-heavy system is higher prices, not more batteries. High evening prices will prompt large-scale decentralised adjustments, such as modifying air-conditioning loads, shifting agricultural pumping, or discharging electric vehicles. Once the price system operates correctly, greater variation of supply within the day is permissible. Under these conditions, the case for coal base load vanishes, and the economic superiority of solar+batteries is accentuated.

We are grateful to Apoorva Bahadur, Rahul Tongia, Rohit Vijay, Ashwin Gambhir and Shantanu Dixit for valuable conversations.

Cross-Border Frictions in India and Impact on Growth - A practitioner's View

by G. Padmanabhan.

Abstract

Since 1991, India has sought to integrate more closely with the world economy. The scale of this engagement is now substantial. In 2025–26, exports of goods and services were over USD 863 billion, including USD 421 billion of services exports. Remittances are estimated at USD 135 billion. Yet these flows have not translated into commensurate investment. India is not among the ten largest recipients of FDI, and net fixed FDI has declined in nominal terms since 2020. This is striking at a time when global firms are reorganising supply chains and looking for alternatives to China. This paper argues that the lag in India’s growth, and the gap in its deep engagement with the world, is a problem of friction in the flow of data, capital and trade. Some of the problem lies in implementation: rules that are reasonable on paper are not administered efficiently. Some lies in the rules themselves: the regulatory design creates costs and uncertainty more so when implemented retrospectively. These costs matter because a firm does not experience each regulatory obstacle in isolation. A delay in an approval, an uncertain tax treatment, a restriction on data flows or difficulty in moving capital across the border all enter into the future decision to invest. The relevant cost is therefore not just the cost of complying with a particular rule. It is the investment that does not take place because the cumulative uncertainty makes India less attractive. This paper suggests a different approach to achieving efficiency. Cross-border friction should not be treated as a collection of unrelated problems to be fixed one rule at a time by incremental tweaking of rules or policies in isolation. The problems are connected because they affect the same decision: whether a firm will choose to engage with India. The first requirement is therefore to recognise cross-border friction as an overall architecture and a policy problem in its own right. The second is to examine the rules and institutions that create it and chalk out how they are to be addressed on an enduring basis.

India faces a throughput problem

Over the decades since liberalisation, India has put in place most of the components needed for deeper engagement with the world economy: exports, services trade, remittances and digital infrastructure. The components are no longer the main constraint. The problem is the efficiency with which they translate into investment and growth.

A simple comparison with other Asian economies makes the point. Table 1 scores the intensity of friction across five types of cross-border flow for four economies at different points on the openness spectrum.

Table 1: Relative intensity of cross-border friction, by flow type.

Flow type	India	China	South Korea	Taiwan
Trade	Moderate	Low–moderate	Low	Low
Capital	High	High	Low	Low
Technology	Moderate–high	High	Low	Low
Data	High	Very high	Low	Low
Financial	High	High	Low	Low

Source: Practitioner’s Research

India and China are both high-friction economies. China has been able to carry this friction given the scale it offers: the volume of its trade and capital flows, and the depth of its integration into global supply chains, allow it to absorb costs that are more consequential and costly for India. South Korea and Taiwan sit at the other end, with relatively low friction across each category. The point is not that borders should be frictionless. Every country regulates cross-border flows. The question is the price paid for the manner in which those regulations are designed and administered. India pays a high price.

Address frictions as an architecture

The usual instinct is to treat each of these frictions as a small, self-contained technical problem, owned by the agency that created it. This misses how firms experience regulation. The frictions interact because they affect the beliefs firms form about the policy environment. A firm that receives a trade-reporting notice for a transaction settled years ago, or encounters a new card-data requirement with little warning, does not necessarily treat the episode as an isolated nuisance. It updates its assessment of how India is governed. When it next considers bringing capital into India, that assessment enters the return it requires comparing with other markets. The cost of an individual friction is therefore not limited to the delay or dispute it creates. It is the increase in the risk premium attached to India, and the investment that may be lost as a result. This is why these domains cannot be addressed independently and every regulatory intervention needs to be weighed against higher stakes – the cost to Indian growth.

There are several reasons why India continues to produce such frictions. Some reflect limitations in state capacity. Some reflect isomorphic mimicry: the adoption of the form of an effective institution without its underlying function. And some reflect weaknesses in regulatory design. A state under pressure to appear modern can reproduce the institutional machinery of a successful system without reproducing the conditions that make it work. India provides several examples. Inflation targeting, for instance, exists formally through a statutory mandate and a monetary policy committee, while the exchange rate continues to be actively managed through intervention in the currency market.

The rest of this paper draws on the author’s experience of this machinery, at the Reserve Bank of India and in subsequent roles, to examine how Indian policy and regulation create cross-border friction. The cases that follow are grouped according to their underlying cause. The first are implementation failures: cases where the policy design was sound but its administration reintroduced the friction the reform was intended to remove. The second are design failures: cases where the problem lies in the rule itself, irrespective of how well it is administered. The closing section then turns to what a better approach would require. The cases draw on the author’s direct knowledge of events rather than on publicly available documentation; these are identified as such.

Observed Case Studies: Cross-Border Frictions

Implementation failures

Trade documentation: The Reserve Bank of India (RBI) introduced the Export Data Processing and Monitoring System (EDPMS) and its import counterpart, the Import Data Processing and Monitoring System (IDPMS), in 2014 to bring trade data held separately by the RBI, Customs and the Directorate General of Foreign Trade (DGFT) into a common system, tracking transactions from initiation to final realisation of proceeds. The objective was sound, but the implementation was not. Historical transactions that had never been reconciled were carried into a system designed to track future flows. Reporting gaps at the source compounded the problem. Firms now receive notices for transactions settled years ago and must spend time establishing what should already be clear: the matter is closed. Such notices are not merely an administrative nuisance. They disrupt a firm’s international business and undermine confidence in India’s systems for managing cross-border transactions. That experience shapes how firms assess the risks of doing business in India—and where they choose to invest next.

Data localisation: Data is now a cross-border flow in its own right, much like goods or capital. India’s approach to regulating data shapes firms’ ability to participate in global markets. In 2018, the Reserve Bank of India (RBI) mandated that payment data be stored in India through a brief circular that offered little guidance on implementation. Some card networks subsequently withdrew services dependent on cross-border data flows, including emergency card replacement for Indian travellers abroad. Uncertainty over localisation and tokenisation has also led foreign counterparties to abandon Indian use-cases rather than assume an unclear compliance burden. The Digital Personal Data Protection (DPDP) Act has not resolved all these uncertainties; the rules governing consent and cross-border transfers, for instance, still need clarification. For a firm operating across several markets, an activity permitted elsewhere may not be worth offering in India if compliance requirements remain unclear. The resulting cost is not confined to activities that regulation prohibits, as uncertainty alone can deter cross-border business.

Investment screening: The implementation of Press Note 3, introduced in 2020, created similar uncertainty for investors. In the days following its announcement, basic questions remained unanswered: what ownership threshold triggered the restriction, which entities were covered, and whether it applied to investments routed through India’s international financial centre. Subsequent clarifications did not settle every issue, and in some cases enforcement followed after investments had already been made. For an investor, the terms on which capital can be committed are part of the investment decision. A known restriction can be factored into the deal’s structure and expected return. An uncertain one introduces the risk of additional approvals, unforeseen costs or a change in regulatory treatment after the event. Investors may respond by spending more on legal advice and structuring, demanding higher returns, delaying transactions or abandoning them altogether. Retrospective changes, including reversals of earlier rulings, compound the problem. The rules have since been amended and clarified, but the delay itself has imposed costs. Uncertainty is not costless simply because the rules are eventually settled.

Design failures

Foreign Exchange Management: In many cases, India’s cross-border frictions also stem from conception and design. The Foreign Exchange Management Act (FEMA), 1999 replaced the Foreign Exchange Regulation Act (FERA), 1973, marking a shift from conserving foreign exchange through restrictive controls towards facilitating external trade and payments. The change was fundamental: under FERA, transactions were generally prohibited unless specifically permitted. FEMA reversed this presumption for current-account transactions, while retaining a permission-based framework for capital-account transactions.

This shift was partly reversed in 2017 by the insertion of money-laundering-style provisions under Section 13 of FEMA. Cases closed years earlier were reopened, including some relating to periods before the money-laundering provisions underpinning the new enforcement approach came into force. This did not just cause uncertainty for existing investors but also increased future regulatory risk, especially for investors committing long-term capital.

Duplicated KYC: The Central KYC Registry was created to reduce duplication such that a customer verified once could be considered approved across institutions, instead of being re-verified at each juncture. In practice, however, the registry has been bottlenecked by major system updates, rigid data-matching requirements and fragmented implementation across banks. When records cannot be retrieved, accepted or matched, institutions must conduct KYC again. For firms, particularly those operating across borders, this means additional compliance costs and uncertainty about whether existing documentation will be accepted. The problem extends beyond duplication: when firms cannot predict what compliance will require at each stage of a transaction, they face regulatory risk that can make investing in India less attractive.

Aircraft financing: Aircraft financing shows how a narrow regulatory interpretation can affect an entire industry. Indian airlines rely on sale-and-leaseback arrangements, a standard global practice that allows them to finance fleet expansion without buying aircraft outright. Treating these arrangements as merchanting transactions brought them within restrictions on the movement of funds, disrupting their use in India for some time. A question of foreign-exchange classification thus became a constraint on how airlines finance fleet renewal. It also carried an environmental cost, since newer aircraft are generally more fuel-efficient. A classification that appears technical to one regulator can change the terms on which capital is available to an industry, shape international financiers’ perceptions of the Indian market, and add to the frictions that constrain cross-border flows.

From silos to a system: what smarter regulation requires

Frictions like these interact across the system, so reform cannot proceed one rule at a time. Removing one constraint may simply leave a firm facing another further down the transaction. The aim should be to reduce the total friction a firm encounters when moving capital, goods, data or payments across borders. Three principles follow.

Treat cross-border friction as one subject

Any improvement in ease of doing business of investors requires an initial intellectual shift on the part of Indian policymakers. Cross-border friction should be treated as a single policy problem spanning trade, capital, data and payments, rather than as a collection of issues belonging to different agencies. This does not require a new super-regulator. It requires a common framework for assessing these frictions and clear responsibility for the points where they interact. India’s domestic digital infrastructure offers a useful precedent: identity, payments and data-sharing were developed as parts of a common architecture rather than as separate systems. Cross-border infrastructure needs the same approach. If the problem is treated as a collection of separate regulatory issues, reform will remain fragmented as well.

Assess regulatory interventions carefully

Viewing cross-border frictions as a single architecture means that each regulatory intervention has to be weighed against the frictions it creates elsewhere and the cost those frictions impose on growth. This raises the stakes of technical regulatory changes and policy innovations that seek to serve several objectives at once. A measure designed to address one problem can alter the cost or certainty of transactions elsewhere in the system. The assessment therefore cannot stop at whether the measure achieves its immediate regulatory objective; it must also consider what it does to the overall cost of engaging with India.

How can this be fixed? Rules are often drafted without sufficient understanding of how firms operate or consultation with those they regulate. Reform must therefore address not just the rules themselves, but how they are made and revised. Draft rules should be published for meaningful consultation, with rule-makers required to identify likely unintended consequences and offer explanations for unaddressed stakeholder concerns. Once implemented, rules should be reviewed after a defined period and modified in light of experience. Where the underlying design is flawed, they should be withdrawn rather than patched incrementally. The final decision should document which concerns were addressed, which were not, and why. This would make rule-making more responsive, transparent and credible.

Further, rules should arrive with the guidance needed to comply with them. Questions of interpretation should be settled in advance rather than through after-the-fact action, and the state should treat predictability as a key outcome in the process. Much of the friction documented here would dissolve through clear and predictable rulemaking.

Separate interpretation from macroprudential judgement

A related problem arises when regulators both make the rules and interpret and enforce them. Many frictions stem from routine interpretive questions: whether a sale-and-leaseback constitutes a merchanting transaction, an investment screen applies to a particular holding, or a deposit qualifies under a remittance scheme. Such decisions can have significant consequences for firms, yet are difficult to challenge when the same institution makes, interprets and enforces the rules.

Routine interpretation is not macroprudential judgement. Regulators need discretion to respond to systemic risks. But routine questions about how existing rules apply to specific transactions do not warrant the same deference. Treating the two alike shields ordinary regulatory interpretations from scrutiny that their consequences demand.

The two functions should therefore be separated. Regulators should retain discretion over macroprudential decisions, while routine interpretations of existing rules, including under FEMA, should be subject to independent appellate review. Firms must be able to challenge an interpretation before a body other than the regulator that issued and enforced the rule. This would make regulatory uncertainty contestable rather than a cost firms must simply absorb, while checking the gradual expansion of restrictions through interpretation.

Conclusion

India has spent three decades building the capacity to engage with the world. The friction that now blunts that capacity is not a scatter of technical irritants to be filed away with the agencies that produced them. It is a connected system of drag, in which each obstacle raises the price at which capital is willing to come, and it has to be understood and reformed as one thing.

The stakes are significant. Growth since the early 2010s has run below the country’s ambitions, and net foreign direct investment has weakened at a moment when supply chains are being reconfigured. Economies that have already stripped out friction will benefit more from that reconfiguration than those unwilling to change. Removing cross-border friction is one of the more direct levers India has for turning its hard-won global engagement into growth. India has shown that it can build complex domestic digital infrastructure by designing different systems to work together. The next task is to apply the same approach to its engagement with the world: to treat the movement of capital, goods, data and payments as a connected system, and reduce the frictions that prevent that system from working well.

G. Padmanabhan is the Former Executive Director, Reserve Bank of India. Several of the cases in Section 2 rest on the author’s direct experience of the events described rather than on published documentation, and are presented as practitioner testimony.