RGGI is INCREASING net CO₂ emissions in PJM — the central failure of the program
The basic carbon-accounting test of RGGI: does it reduce actual CO₂ in the
atmosphere? Twelve months of empirical dispatch data (Apr 2025 – Mar 2026)
against actual PJM zonal LMPs and EPA CAMD operations says NO — it increases
atmospheric CO₂ by ~2.9 million tons per year. The mechanism is straightforward:
RGGI's carbon adder makes clean RGGI-state gas units uneconomic relative to dirtier,
higher-HR units in PA, OH, KY, WV. The displaced demand is met by the dirtier units.
RGGI's books look better; the atmosphere is worse.
All figures below are for the trailing 12 months (April 2025 – March 2026), based on twelve independent monthly dispatch runs. Not a forward-looking annual rate — this is what already happened.
12.9 M MWh shifted out of RGGI states (Apr 2025 – Mar 2026)
12.90 M MWh
Generation that would have run at clean RGGI-state gas plants but didn't, because the RGGI carbon adder made them uneconomic vs PJM zonal LMPs. Same demand, different plants.
RGGI plants squeezed out: emission rate
0.54 t / MWh
Load-weighted CO₂ rate of the displaced RGGI fleet (NJ/MD/DE gas — mostly modern CCGTs at HR 7.0–7.5; some HR > 8 peakers). 12.9 M MWh × 0.54 = 6.94 M t avoided inside RGGI (booked in-region reductions that may be offset by leakage outside RGGI)
Replacement plants: emission rate
0.76 t / MWh
Load-weighted CO₂ rate of the non-RGGI PJM peers that picked up the demand (PA/OH/KY/WV — older steam, coal-to-gas conversions, peakers at HR 9.0–11.5). 12.9 M MWh × 0.76 = 9.80 M t emitted outside RGGI (real atmospheric emissions)
Per-MWh emissions wedge
+0.22 t / MWh
Each MWh that RGGI shifts out → peer plants emit ~41% more CO₂ than the RGGI plant would have. The "compliance gain" inside RGGI is more than offset by the "compliance loss" outside.
NET ATMOSPHERIC CO₂ CHANGE (trailing 12 months)
+2.86 M tons more CO₂ in the atmosphere
12.9 M MWh × (0.76 − 0.54) = +2.86 M short tons. Our analysis estimates RGGI caused a NET INCREASE in actual CO₂ emissions over the last 12 months — counterproductive on a PJM-wide emissions basis. Equivalent to ~565,000 extra passenger cars on the road for a year.
The arithmetic in plain English:
RGGI's carbon adder made 12.9 million MWh of clean RGGI-state gas generation uneconomic over Apr 2025 – Mar 2026. That demand didn't disappear — it shifted to non-RGGI PJM plants. The plants squeezed out are modern combined-cycle gas (heat rate 7.0–7.5 MMBtu/MWh, emitting ~0.54 short tons CO₂ per MWh). The plants that picked up the demand are older / less-efficient (heat rate 9.0–11.5 MMBtu/MWh, including coal-to-gas conversions and peakers, emitting ~0.76 t/MWh).
Same MWh, but ~41% more CO₂ per MWh.
The 6.94 M t "avoided inside RGGI" is real on RGGI's compliance books, but the 9.80 M t "added outside RGGI" is real in the actual atmosphere — and it's bigger.
Net: +2.86 M t more CO₂ in the atmosphere over Apr 2025 – Mar 2026 because of RGGI's dispatch distortion.
This is leakage in its purest form: the program achieves accounting compliance by exporting (and amplifying) the actual emissions. The forward-looking annual rate at $40+ RGGI prices is likely larger — but is not what these figures represent.
Consumer cost framework — at today's $40.30/ton RGGI carbon price
The anchor: RGGI's 2026 cap is ~75 M short tons. At $40/ton, auction
proceeds = $3.0 B/yr. This is the unambiguous floor of consumer cost —
every cent flows from consumers through generators' offers (carbon adder embedded
in dispatch costs) into the auction. Above this floor are 4 additional cost layers,
each a real wealth transfer mechanism documented in peer-reviewed literature
(Murray & Maniloff 2015; Rose et al 2023; PJM IMM 2024 SOM Table 3-74).
Independent multi-reviewer audit (Apr 2026) verified each component of the cost framework against published methodology and peer-reviewed literature.
Annual gross WHOLESALE cost
~$9–10 B
Sum of 5 cost layers, at $40.30/ton RGGI, on 942 TWh of RGGI region + PJM spillover load
RGGI auction proceeds (cap × price)
$3.0 B
Lower-bound floor — flows from consumers through emitters' offers to RGGI state programs
NET cost to consumers after rebate
~$7 B
Wholesale cost less ~$2.1 B/yr returning to consumers via state energy-efficiency / bill-credit / clean-energy programs
Inframarginal RENT to non-emitters
~$3.5 B
Pure profit captured by non-emitting generation: nuclear ~$1.6B, hydro ~$0.8B, wind/solar ~$0.4B, low-HR gas ~$0.6B
Cost to non-RGGI PJM states (no rebate)
~$2.0 B
PA $0.53B, OH $0.42B, VA $0.34B, IL $0.34B — pay LMP uplift via cross-zonal coupling, receive zero auction rebate
And the climate-impact gap
HIGHER RGGI PRICES APPEAR TO INCREASE NET CO₂ RATHER THAN DECREASE IT
Our analysis estimates consumers and businesses pay $7B/yr (net) for a program that has appeared to raise actual atmospheric CO₂ by 2.86 M tons over the trailing 12 months — see the emissions wedge above. The cost is real; the in-region "reduction" is partially or fully offset by leakage.
Why the price spike is structural — and what to fix
The Feb–Apr 2026 rally from $23 to $40 is not driven by fundamentals tightening. It's driven by a
term-structure mismatch between (a) generators' multi-year compliance hedging needs and
(b) RGGI's quarterly single-vintage auction structure. Roughly $5–10 / ton of the current
forward price is attributable to this mismatch alone — a $400–800 M / yr implicit transfer from
ratepayers to the secondary market.
The problem: term mismatch
Demand side: Generators signing multi-year gas PPAs with data-center offtakers carry compliance obligations of 5, 10, 15, even 25 years. They need to hedge that forward.
Supply side: RGGI auctions sell only the current calendar-year vintage, four times per year. ICE Vintage futures technically list 10 years out, but open interest collapses beyond Dec-2026. There are no forward vintage auctions.
Result: Multi-year compliance demand gets compressed into current-vintage buying. The thin forward curve becomes a short-squeeze chamber — exactly what we observed Feb–Apr 2026: a ~61% rally from the Auction 71 clear of $24.99 (Mar 11) to $40.30 (Apr 27), or ~75% from the Feb low of $23.05 — on no fundamental news, declining emissions data, and a softening PJM load forecast.
Supply side: RGGI auctions sell only the current calendar-year vintage, four times per year. ICE Vintage futures technically list 10 years out, but open interest collapses beyond Dec-2026. There are no forward vintage auctions.
Result: Multi-year compliance demand gets compressed into current-vintage buying. The thin forward curve becomes a short-squeeze chamber — exactly what we observed Feb–Apr 2026: a ~61% rally from the Auction 71 clear of $24.99 (Mar 11) to $40.30 (Apr 27), or ~75% from the Feb low of $23.05 — on no fundamental news, declining emissions data, and a softening PJM load forecast.
Embedded premium at $40+: ~$5–10 / ton
Annual transfer to secondary market: ~$400–800 M / yr
Annual transfer to secondary market: ~$400–800 M / yr
→
The fix: restore advance vintage auctions
This was RGGI's original design. Auctions 3–14 (Mar 2009 – Dec 2011) included advance-vintage tranches of ~1.9–2.2 M allowances per quarterly auction. The mechanism was retired in the 2013 Model Rule because the market was oversupplied at the time ($1.86–$3.51 prints, with no forward hedging demand at the time to absorb the forward tranches).
That condition no longer applies: the market is clearly not oversupplied today (current $40+), and long-dated term demand is now a structural feature of the market — multi-year data-center PPAs and other 5- to 25-year compliance obligations are here to stay.
Rollout — a catch-up term auction in June or September 2026: RGGI Inc. and the participating states can announce an out-of-cycle advance auction at Auction 72 (June 2026) or Auction 73 (September 2026) that releases a one-time tranche of 2027 / 2028 / 2029 vintages — sized to absorb the multi-year hedging demand currently compressed into the current-vintage forward. This is the fastest practical path to relieve the squeeze and re-anchor the curve before Virginia re-entry.
The proven model already exists: California WCI and Washington CCA-WA hold an advance auction alongside each current-vintage auction, selling ~8–10% of vintage three years forward. Those advance tranches consistently sell out, trade at a predictable narrow discount, and provide a transparent forward-price anchor. Holding limits prevent hoarding.
That condition no longer applies: the market is clearly not oversupplied today (current $40+), and long-dated term demand is now a structural feature of the market — multi-year data-center PPAs and other 5- to 25-year compliance obligations are here to stay.
Rollout — a catch-up term auction in June or September 2026: RGGI Inc. and the participating states can announce an out-of-cycle advance auction at Auction 72 (June 2026) or Auction 73 (September 2026) that releases a one-time tranche of 2027 / 2028 / 2029 vintages — sized to absorb the multi-year hedging demand currently compressed into the current-vintage forward. This is the fastest practical path to relieve the squeeze and re-anchor the curve before Virginia re-entry.
The proven model already exists: California WCI and Washington CCA-WA hold an advance auction alongside each current-vintage auction, selling ~8–10% of vintage three years forward. Those advance tranches consistently sell out, trade at a predictable narrow discount, and provide a transparent forward-price anchor. Holding limits prevent hoarding.
Estimated price compression at $40+: $10–15 / ton in Dec-2026 (the over-fundamental gap has widened materially since the cliff paper; compression scales accordingly)
Cap integrity preserved — only the timing of monetization shifts
Cap integrity preserved — only the timing of monetization shifts
Proposed mechanic (CARB-style restoration): hold an advance auction alongside each quarterly current-vintage auction, selling year +3 vintage in volumes of approximately 8–12% of that vintage per quarter. Reserve price tracks the same MRP escalator used for current-vintage auctions. Holding limits sized to compliance need plus bounded surplus to address hoarding. Total cumulative supply is unchanged; only the timing of monetization shifts.
Authority: RGGI Inc., the Independent Market Monitor, and the participating states all have model-rule amendment authority. The 2014 cap-cut precedent (165→91 M tons) established that price management through administrative adjustment is a legitimate policy tool — and that precedent cuts in both directions: if supply can be cut to defend a price floor, it can be expanded to address a price spike.
Authority: RGGI Inc., the Independent Market Monitor, and the participating states all have model-rule amendment authority. The 2014 cap-cut precedent (165→91 M tons) established that price management through administrative adjustment is a legitimate policy tool — and that precedent cuts in both directions: if supply can be cut to defend a price floor, it can be expanded to address a price spike.
Layer stack: where the $9-10 B/yr comes from
Each layer is a distinct economic mechanism. Layers are NOT additive in retail
terms — per audit panel correction, do not apply retail markup to the
inframarginal-rent layer (it is generator profit collected directly via LMP, not
a utility cost being marked up). Wholesale-anchored figures.
Who collects the rent? Attribution by generation source at $40/ton
Carbon-adder LMP uplift on infra-marginal generators is ~$3.5 B/yr at
$40 RGGI. These resources do not pay RGGI compliance costs (they don't emit) but
receive the LMP uplift as pure rent. The single-largest beneficiary
class is existing nuclear — plants that have been operating for decades and
whose dispatch decisions are unaffected by the carbon adder. RGGI's revenue largely
transfers to these existing assets rather than driving new low-carbon investment.
| Generation source | Footprint | Share | Annual rent at $40 RGGI |
|---|
Methodology: rent_by_source ≈ (annual MWh from that source in RGGI markets) × (avg LMP uplift from carbon adder when gas marginal, ~$17/MWh at $40 RGGI) × (marginal-share of hours, ~70% gas-on-margin in NYISO / NEPOOL / PJM-RGGI). Source-level attribution (not company-level) — derived from EIA-923 and ISO public generation-by-fuel data.
State allocation — gross / rebate / net at $40.30/ton RGGI
Allocated by state load share within each market (NYISO / NEPOOL / PJM RGGI / PJM
spillover). Rebate column shows the share of $3.0 B RGGI auction proceeds returned
to consumers via state energy-efficiency and bill-credit programs (~70% pass-through
typical). Non-RGGI PJM states get zero rebate.
RGGI states (10 + DC)
| State | Load (TWh/yr) | Gross cost ($B) | Auction rebate ($B) | Net cost ($B) |
|---|
Non-RGGI PJM states (cross-zonal spillover, no rebate)
| State | Load (TWh/yr) | Gross cost ($B) | Rebate | Net cost ($B) |
|---|
How much carbon emissions leakage is RGGI causing in PJM?
In addition to the cost transfer, RGGI's carbon adder also drives physical
emissions leakage — clean RGGI-state gas plants kept out of dispatch
while higher-heat-rate non-RGGI plants (in PA, OH, WV, KY etc.) ramp up to fill
the demand. Headline figures from twelve independent SCED v1.1 dispatch runs
against actual PJM zonal day-ahead LMPs and actual CAMD unit operations. No scaling.
±30% range reflects uncertainty in the marginal-displacement assumption.
March 2026 — leakage
—
MWh of clean RGGI-state generation displaced to non-RGGI PJM states
March 2026 — emissions leaked out
—
CO₂ tons emitted in non-RGGI states instead of RGGI states
March 2026 — extra CO₂
—
Net additional CO₂ tons in PJM caused by RGGI's dispatch distortion
Apr 2025 – Mar 2026 — leakage
—
MWh shifted out of RGGI states over the trailing 12 months
Apr 2025 – Mar 2026 — emissions leaked out
—
CO₂ tons emitted in non-RGGI states instead of RGGI states
Apr 2025 – Mar 2026 — extra CO₂
—
Net additional CO₂ tons in PJM over the trailing 12 months
Per-month dispatch, Apr 2025 – Mar 2026
Twelve independent dispatch runs — one per calendar month — against that
month's actual PJM zonal day-ahead LMPs (~184k hourly prices across 23
zones) and actual EPA CAMD unit operations. No scaling: each bar is the
direct output of dispatching the real RGGI-state PJM fleet against real
market clearing prices for the month, with the RGGI carbon adder priced
at the month's actual ICE close. Calibration (% of model dispatch
decisions matching observed CAMD) shown for transparency.
Multi-year context: RGGI price trajectory (2022 — Apr 27, 2026)
RGGI clearing prices have more than tripled from $13 in early 2022 to $26 by end of 2025.
Auction 71 cleared $24.99 on March 11, 2026; the secondary market has since run to
$40.30 (Apr 27, 2026 close) — a 60%+ premium to the most recent auction clear and the
reference price for this dashboard's cost framework.
Independent verification — what was audited
Each headline figure was verified against a separate, independent data source or established methodology:
- EIA-923 vs EPA CAMD cross-check (Jan 2024 – Feb 2026): 721 plant-months matched. Median net/gross ratio 0.93. Top CCGTs (CPV St. Charles, Keys Energy Center, Red Oak, West Deptford) 0.97–0.99 — within published net/gross literature range. EPA CAMD generation foundation independently corroborated by EIA-923
- Per-plant top-40 comparison (March 2026 RGGI gas plants in NJ/MD/DE): plant-by-plant March 2026 dispatch vs an "as-if-no-RGGI" offer-stack run produced the leakage volumes. Comparison was reviewed for methodological soundness; main acknowledged caveat is LMP endogeneity (we use observed-with-RGGI prices), which makes results conservative.
- Cost framework triangulation: $9-10 B wholesale total cross-checked against four independent benchmarks — (1) RGGI auction proceeds anchor (cap × price = $3.0 B at $40), (2) Murray & Maniloff 2015 (~7% retail premium implies ~$14 B regional retail), (3) Rose et al 2023 ($5–7/MWh sustained retail premium), (4) PJM IMM 2024 SOM Table 3-74 ($1.94/MWh PJM-wide CO₂ cost component, scales to $3.50/MWh at $40 RGGI = $2.7 B PJM-wide alone, a subset of our framework). All four benchmarks reconcile with the $9-10 B wholesale total
- Source-level rent attribution: rent-by-source figures cross-checked against EIA-923 generation-by-fuel data and ISO public capacity-mix reports for NYISO, ISO-NE, and PJM-RGGI zones. Aggregate $3.5 B/yr rent capture cross-checks against the gap between cap-x-price auction proceeds ($3.0 B) and total LMP uplift on RGGI region load (~$6.5 B wholesale).
About this dashboard
RGGI Carbon Leakage Monitor — Cost & Emissions Impact.
Independent empirical analysis of RGGI's effect on (1) PJM electricity dispatch and CO₂ emissions, and (2) the layered cost burden on consumers and businesses across the RGGI footprint and PJM spillover zones, with full attribution of who collects the carbon-adder rent.
Independent empirical analysis of RGGI's effect on (1) PJM electricity dispatch and CO₂ emissions, and (2) the layered cost burden on consumers and businesses across the RGGI footprint and PJM spillover zones, with full attribution of who collects the carbon-adder rent.
Period: 12 monthly dispatch runs Apr 2025 – Mar 2026 + cost framework at current $40.30/ton RGGI. Updated —
Detailed research available to Alpha Inception advisory clients. The full empirical research underlying this dashboard — including the comprehensive cliff-pricing report, detailed dispatch modeling, source-level rent attribution, the 12-month per-plant dataset, and forward-looking sensitivity analysis — is available through Alpha Inception's advisory engagements. Alpha Inception is currently considering new advisory engagements in energy and emissions markets.
For inquiries, contact Andre Templeman, Principal:
Email: andre@alphainception.com
Phone: 801-455-3033
For inquiries, contact Andre Templeman, Principal:
Email: andre@alphainception.com
Phone: 801-455-3033
Methodology summary & caveats
Volume (leakage MWh, emissions): twelve independent monthly dispatch
simulations against that month's actual market data. For every PJM unit-hour we
compute variable cost (fuel × heat-rate + VOM + RGGI adder if in NJ/MD/DE) and
compare to the unit's PJM zonal day-ahead LMP. When a RGGI-state unit is economic
without the carbon adder but uneconomic with it, the model counts it as kept-out.
Heat rates and CO₂ rates per unit from EPA CAMD load-weighted multi-year averages.
Per-unit observed capacity factor and per-zone empirical marginal CO₂ rate drive
the displaced-generation and emissions calculations. RGGI prices use ICE RJ6 +
ICE daily close. Leakage volume EIA-923 cross-checked.
Cost framework: 5-layer stack anchored on cap × price = $3.0 B/yr
auction proceeds floor. Inframarginal rent per owner = (annual non-emitting MWh in
RGGI markets) × (LMP uplift when gas marginal) × (~70% gas-on-margin share).
NYISO + NEPOOL treated as 100%-RGGI exposed (correct: every gas plant subject to
compliance). Cross-checked against Murray & Maniloff 2015 (~7% retail premium →
~$14B), Rose et al 2023 ($5–7/MWh retail), and PJM IMM 2024 SOM Table 3-74
($1.94/MWh PJM-wide CO₂ cost component, scales to $3.50 at $40 RGGI).
The $9–10 B wholesale total is the audit-passed canonical figure.
Caveats (fairness disclosures):