Embedded Files
An examination of block-wise aggregated demand–supply data of a major Power Exchange for 16 December 2025, corroborated with Grid-India (NLDC) Power Supply Position (PSP) data, reveals systematic distortions in bid behaviour.
Using AI Analytical tools, one finds that these distortions are inconsistent with competitive, marginal-cost-based price discovery and point to structural MCP inflation. The analysis focuses on three representative periods—00:00–01:00, 12:00–13:00, and 19:00–20:00—covering off-peak, solar-rich, and evening peak conditions. The same patterns recur across all periods, indicating a systemic issue. (link to the reference files is placed below)
1. Demand is perversely inelastic across large price jumps
Look at the buy side across price buckets within the same block:
Example (00:00–00:15):
₹0–1000 → 11,176 MW Option-style, non-marginal bid
₹1001–2000 → 9,128 MW Weak elasticity
₹2001–3000 → 8,723 MW Demand persists despite high price
₹3001–4000 → 5,231 MW Abrupt cliff, not gradual response
What’s unusual
A ₹3,000/MWh increase wipes out only ~6 GW of demand.
That implies extremely low-price elasticity, which is implausible for OA / industrial buyers.
Interpretation.
This is not real willingness to pay. It is classic option bidding:
“Schedule me only if prices crash; otherwise, I don’t really need power.”
These bids exist to shape MCP and benefit in derivative markets, not to consume power.
2. Demand drops sharply at mid-prices, not gradually
Look at Demand Change with Price:
−2,047 MW
−405 MW
−3,493 MW (sudden collapse)
Why this is abnormal
In real demand curves: Elasticity changes smoothly
One does not see cliff-edge demand exits within ₹1,000 bands
What these signals?
Algorithmic or portfolio bidding, Demand stacked mechanically at price steps, not economically
This makes the aggregated demand curve artificial.
3. Supply behaves in the exact opposite way of merit order
Same block, supply side:
₹0–1000 → 3,921 MW Thin supply, inconsistent with surplus
₹1001–2000 → 6,638 MW Below expected coal variable cost
₹2001–3000 → 8,844 MW Still suppressed supply
₹3001–4000 → 16,939 MW Supply wall / volume dumping
What’s unusual?
Supply more than doubles suddenly at higher prices. Very thin supply where coal variable cost should normally sit.
What this means?
Withholding at low prices. Volume dumping at high prices. This is strategic supply placement, not cost discovery.
4. Supply jumps are far larger than demand drops
Compare slopes:
Demand change: −400 to −3,500 MW
Supply change: +2,200 to +8,100 MW
Why these matters
MCP is being driven by supply cliffs, not demand pressure. Clearing price is where a supply wall appears, not where marginal demand meets marginal cost. This is textbook price-setting by offer concentration.
5. Buy–sell curves never “hug” each other near clearing
In healthy markets, Near MCP, demand and supply curves become steep and close
Here, even at higher prices, gap remains huge, Clearing happens because exchange truncates volume, not because curves converge.What is actually happening
Even when the price is pushed higher, buyers and sellers remain far apart
Buyers still want much more power than sellers are offering (or vice versa)
The curves never come close to each other
Yet a price is still declared
Why?
Because the exchange forces the match by cutting volumes
This confirms that MCP is an administrative intersection, not an economic one.
6. Repeated across blocks → not a one-off
This pattern repeats block after block, Heavy low-price demand, Thin low-price supply, Sudden mid-price demand exits, Massive high-price supply entry
That rules out: Weather, Ramp constraints, One-time outages
This is systemic bidding behaviour.
7. What is especially damning from a Grid-India perspective
Grid-India’s actual system demand curve: When the aggregated demand curve published by the power exchange is compared with the physical system demand reflected in Grid-India’s Power Supply Position, the divergence is stark. The exchange demand curve exhibits a stepwise structure with pronounced cliffs and artificial, discontinuous elasticity, with large quantities appearing and disappearing at discrete price points and repeating mechanically across blocks. In contrast, the Grid-India PSP demand profile is smooth, continuous, and load-driven, with gradual changes reflecting real consumption patterns, weather, and operational conditions. Most importantly, the exchange demand curve shows weak physical linkage, as a significant portion of bids does not translate into scheduled drawal, whereas the PSP represents actual system demand that must be met in real time.
This divergence cannot be explained by thin participation alone. A genuinely thin market may show lower volumes or some noisiness, but it does not produce persistent low-price demand walls, abrupt mid-price cliffs, and repeatable patterns across off-peak, solar, and peak hours. Thin participation would reduce depth; it would not create large quantities of non-credible, option-style demand or strategic supply withholding that systematically shapes the clearing price. The repeatability and structure seen in the exchange curves point to bid behaviour enabled by market design rather than a mere lack of participants, indicating that the issue lies in bid credibility and incentives, not market size.
The PX aggregated demand curve is decoupled from physical load
A large part of the order book is non-physical, non-marginal
MCP is being discovered on financial-style bids, not on real power demand.
Bottom line
This file shows three red flags simultaneously:
a) Non-credible demand at low prices
b) Strategic supply cliffs at higher prices
c) Artificial elasticity patterns inconsistent with system load
Together, they fully support MCP inflation, “Cheap power illusion” for OA consumers & Structural bias of ~₹0.5–0.8/kWh estimated
I’ll be blunt and forward-looking, because the data now points to behaviour, not accident.
8. Who is most likely doing this
A. Large generator portfolios (especially thermal + RE hybrids)
Why they fit the pattern
Ability to withhold at low prices and dump volume at high price bands
Portfolio optimisation, not unit-cost bidding
Can afford to lose a few blocks to move MCP across many
Typical strategy
Keep low-price supply thin → force clearing near high-price wall → earn MCP uplift on entire cleared volume (including long-term contracted surplus)
B. Trader aggregator entities acting for multiple generators
Why?
They see the entire book, they place stepwise bids for multiple clients, they optimise price, not dispatch
Signature in the data
Vertical supply cliffs, Identical price steps across blocks
C. Large OA consumer aggregators / power managers
This may surprises many, but the buy side is complicit.
Why
They place deep low-price bids as free option, No penalty if not scheduled
They hope to get lucky in surplus blocks Or push price signals downward for future blocks
Net effect
They pollute demand curve while claiming victimhood later
D. Exchange-facilitated behaviour (design, not intent)
The exchange is not “placing bids”, but:
a) Unconstrained MCV publication
b) Aggregation without credibility filters
c) No distinction between firm and optional bids
This enables gaming at scale.
9. How this gaming actually works (mechanism)
This is important for regulators.
Large buyers dump 10–12 GW at absurdly low prices
Generators keep supply thin up to ₹3,000+
MCP jumps to first big supply wall
Only a fraction of the book clears
Everyone points to “market forces”
This is coordinated but not collusive — hence hard to prosecute under current rules.
10. What the exchange can do (purely technical fixes)
a) Credibility-weighted bids
Require bidders to declare:
Firm vs optional quantity
Penalty-linked commitment
Weight bids by historical execution ratio
Result
Placeholder bids lose influence on MCP
b) Two-sided minimum participation rules
Large portfolios must bid across multiple price bands
Ban single-price vertical walls above a threshold MW
Used in EU gas and power markets.
c) Publish constrained demand, not unconstrained MCV
MCP is discovered on constrained schedules
Publishing unconstrained curves misleads consumers and policymakers
This single change would expose the illusion.
d) Slope and cliff detection (real-time surveillance)
Algorithmically flag:
ΔMW/Δ₹ spikes beyond thresholds
Repeated block-wise identical ladders
This is trivial tech.
11. What the regulator (CERC) can do — today, without Act amendment
a) Define “economic bids”
Mandate:
Bids must be cost-justified within a band
Extreme outliers require explanation
This already exists in spirit in DSM and ancillary rules.
b) Bid-linked penalties
If bid deviates wildly from cleared price and is not scheduled repeatedly → penalty or bid discount
Similar to phantom capacity rules in capacity markets
c) Separate price discovery from volume discovery
First discover price on credible tranche
Then expand volume
EUPHEMIA effectively does this through welfare maximisation.
12. So why is none of this being done?
a) Everyone benefits except the final consumer
Generators → higher MCP
Traders → higher spreads
Exchanges → higher notional turnover
DISCOMs → pass-through, no pain
Only OA consumers lose — fragmented, weak voice.
b) Regulatory capture by technical complexity
The system is defended with:
“Algorithm is neutral”
“Market is competitive”
“Volume proves liquidity”
Few regulators interrogate curve quality, not just MCP.
c) Fear of destabilising prices
Tightening rules will:
i) Reduce MCP
ii) Expose overcapacity
iii) Create political backlash from generators
So, inertia is safer.
d) Market coupling distraction
Coupling is sold as the silver bullet.
But coupling a distorted book just spreads the distortion.
13. Hard truth -This is not a market failure.
It is a market design choice that tolerates gaming because it raises prices without visible misconduct. This data now gives empirical backing to say this — not as opinion, but as forensic evidence.
14. What would force change- Historically, meaningful correction of such market design issues has occurred only when three conditions come together. First, consumer harm must be quantified and made explicit, whether in terms of ₹/kWh distortion or aggregate ₹ crore impact, so that the issue is no longer abstract but measurable and attributable. Second, the outcomes must enter the public domain, through reasoned analysis, comparative evidence, or informed debate, creating visibility beyond technical forums and limiting the comfort of quiet inaction. Third, reform is often triggered by regulatory embarrassment arising from transparency gaps, when discrepancies between stated objectives and observable outcomes become difficult to defend, compelling corrective action to restore credibility.
Page updated
Google Sites
Report abuse