⚠ Load shedding is not random. The same communities lose power, decade after decade.
⚠ The operator’s choice of which feeder goes dark is a moral decision dressed up as a technical one.
India has a measure for income inequality. It does not have one for outage inequality. Not yet.
Equity in outages — who stays in the dark?
| When India’s grid is stressed and load has to be shed, the operator decides which feeders go dark. The decision is not random. By every available metric, the burden falls on the same parts of the country, decade after decade. We propose calling that pattern the Grid Gini Coefficient — and we propose making DISCOMs compute it. |
|---|
01 · THE PRIORITISATION HIERARCHY
Feeder cuts follow an internal priority list — and the bottom looks predictable.
Load shedding is not a single event. It is a sequence of operator decisions made under time pressure: when grid frequency starts to drift below 49.9 Hz and the system risks cascading failure, the protection protocol kicks in and feeders are shed in a pre-determined order.
Every DISCOM operates an internal prioritisation hierarchy that ranks feeders by importance. At the top — protected from interruption — are large industrial consumers, defence installations, hospitals, and critical infrastructure. At the bottom are residential rural feeders and subsidised agricultural connections.
The hierarchy is rarely published. When asked about it in regulatory hearings, the standard answer is that “essential services” are protected and “non-essential” feeders absorb the cuts. The definition of “non-essential” is political, not technical.
02 · THE REVENUE-VS-NEED DILEMMA
Commercial logic and social mandate pull in opposite directions.
When power has to be cut, the choice is not between feeders that are equally important. It is between feeders that generate different revenue. The tariff reality:
▸ Industrial feeders: ₹7–9/kWh
▸ Domestic urban high-tier feeders: ₹6–8/kWh
▸ Domestic lifeline-tariff feeders: ₹2–3/kWh, with state subsidy
▸ Agricultural feeders: heavily subsidised or free in many states
The DISCOM’s commercial logic says: cut the feeders that generate the least revenue first. The DISCOM’s social mandate says: cut the feeders whose consumers can least afford the consequences last. These two instructions are structurally in conflict, and revenue logic wins.
A luxury mall in an urban centre and a rural primary health centre may both lose power when the grid is stressed. But the choice of which loses power first is made by an operator following a protocol that was never publicly debated and is not subject to equity review.
03 · THE ENERGY-POVERTY NEXUS
Energy insecurity does not just reflect poverty. It deepens it.
Households in rural and low-income urban areas experience more outages, longer outages, and less predictable outages than higher-income consumers. The downstream effects compound:
▸ Spoiled food and lost cold-chain pharmaceuticals.
▸ Disrupted schooling for children dependent on evening study time.
▸ Missed work hours for casual labour where evening hours matter.
▸ Increased exposure to extreme summer heat without fans or cooling.
Research from the Indian context (Biswas et al., 2022) documents this asymmetric harm. Work from the University of Pretoria (2026) extends the finding: unreliable supply in developing-country grids has measurable impacts on business formation, educational attainment, and infant mortality rates.
Energy insecurity therefore does not just reflect existing poverty. It deepens it. A household that loses power more often has more economic damage to absorb and more recovery cost to pay, with fewer reserves to do it. The compounding effect is the ethical indictment.
04 · THE GRID GINI COEFFICIENT
A standard measure of inequality, applied to load shedding.
The energy-poverty nexus has been described for years. It has never been measured. There is no Indian DISCOM that publishes a quantitative measure of inequality in its load-shedding distribution. This series proposes a specific measure: the Grid Gini Coefficient.
Operational definition
The Grid Gini Coefficient (GGC) for a DISCOM service area, computed over a defined window (typically a financial year), is the Gini coefficient of cumulative supply-hours-lost per capita across all served segments, weighted by income proxy. A GGC of 0 means outage burden is distributed equally across income levels. A GGC of 1 means all outages fall entirely on the lowest-income consumers.
Three computable proxies DISCOMs could publish today
While the full GGC requires consumer-level income data, three practical proxies can be calculated from data DISCOMs already collect:
▸ SAIDI / SAIFI disaggregated by tariff tier. System Average Interruption Duration Index and Frequency Index, broken out by lifeline / domestic / commercial / industrial tariff class. This is the most direct proxy for distributive load-shedding inequality.
▸ Rural-versus-urban supply-hour gap. Published in CEA quarterly reports as a geographic proxy for income inequality in outage burden. The gap has narrowed nationally but remains significant in many states.
▸ AT&C losses by feeder type. Rural agricultural feeders typically show higher AT&C losses and lower restoration priority; the gap between feeder types serves as a proxy for revenue-driven prioritisation.
Why no DISCOM publishes this
The reason the GGC is not already a standard metric is not technical. The data exists. The reason is political: a published number that quantifies inequality in outage burden becomes a regulatory and legal liability. Measurement creates accountability. That is precisely why it should be mandatory.
05 · FROM TOTAL BLACKOUTS TO EQUITY-AWARE LOAD LIMITING
Smart meters make a different protocol operationally possible.
In the old model, when a feeder had to be cut, the entire feeder went dark — every household, every business, every essential service on that wire lost power simultaneously. Smart meters with remote-control capability make a different approach viable: load limiting rather than total disconnection.
Under load-limiting, a household does not lose power entirely. It loses the ability to draw above a defined threshold — say, 200 watts. That is enough to keep a fan running and one light on, but not enough for a television, a pump, or an air conditioner. The grid sheds load. The consumer retains basic function. The dignity threshold is preserved.
[PILOT] Indian DISCOM load-limiting at scale Tata Power Mumbai is the most documented Indian load-limiting pilot. Its FY22–FY24 demand-response (DR) programme scaled from 2,000 to over 100,000 residential customers. The smart-meter infrastructure permits curtailed supply — limiting a household to a bare-minimum draw of around 500 W (enough for lights and a fan) during grid emergencies rather than a full blackout. |
|---|
The instrument exists. The protocol that mandates it as the default response — rather than total feeder cut-off — does not yet exist in most Indian DISCOM operating procedures. That gap between technology and policy is the ethical frontier.
06 · RECOGNITION JUSTICE IN PRACTICE
Not every consumer can absorb a 200-watt limit.
Even within the load-limiting framework, a further refinement is required. A household with a member on home dialysis needs full power for the dialysis machine. A farmer at a critical irrigation moment in a monsoon-timed crop cycle needs the pump motor. A rural pharmacy storing temperature-sensitive vaccines cannot absorb even a one-hour limit.
The Recognition Justice pillar from Part 02 says these specific needs must be acknowledged in the design of the load-shedding protocol — not as exceptions handled by manual appeal, but as a registered category with automatic protection.
In operational terms, this means DISCOMs need to maintain a registry of vulnerable-load consumers (medical, agricultural critical-hour, essential services in rural areas) and ensure that load-limiting thresholds are calibrated to their documented needs. The technology to do this exists in 2026. The regulatory mandate does not.
07 · THE VERDICT
The protocol can be redesigned. The question is whether the incentive exists to do it.
Load shedding is the single clearest example of how grid operating decisions become ethical decisions. The current protocols allocate burden by revenue logic, which is to say by income logic, which is to say by injustice. This is not a necessary feature of grid management. It is a design choice.
The 2026 toolkit — smart meters, load-limiting, vulnerable-consumer registries, and the Grid Gini Coefficient as a public metric — makes a different protocol operationally possible. The question is whether India’s regulators and DISCOMs will require it.
| The Grid Gini Coefficient is proposed. The three computable proxies are available now. The question for every DISCOM in 2026: what does your number look like — and would you publish it if you computed it? |
|---|
→ · COMING UP IN PART 4
Operational Conservatism and the Hidden Carbon Tax — 23 GW of renewable capacity was curtailed in 2025. Part 4 asks why grid operators default to gas, what that costs the climate, and what the CERC Must-Take Level amendment changes about the ethics of dispatch.
#EthicsofGridStability #RenewableEnergy #IndianRESector #IPP #Governance #Sustainability #DISCOMReform #EnergyJustice #GridGiniCoefficient #LoadShedding