RT3, the hidden cost of integrating renewables in Spain's grid: ×10 since 2020

Updated 2026-05-21 · Thanks to Sergio B. for the expert feedback. We initially labeled RT3 as “tertiary reserve” — RT3 is actually the consumer-side cost of technical constraints in the day-ahead market (the mechanism by which REE redispatches the OMIE clearing to ensure system security by forcing synchronous units online). All data, ratios, and the renewable correlation stand; the mechanism is now correct, and the post-blackout ×2.5 step is explained by REE’s “Operación Reforzada” introduced after April 28, 2025 — as documented in PwC’s February 2026 report.

While the public conversation around the Spanish electricity market remains anchored on OMIE — the wholesale price that makes the headlines every day — a component of the bill that almost nobody mentions has changed scale. Technical constraints (RT3), the cost REE incurs to override the OMIE market clearing and keep synchronous generation online for system security, has gone from 1.97 €/MWh in 2020 to 19.53 €/MWh so far in 2026. Nearly ten times more in five years.

The aggregate balancing-services (SSAA) chart doesn’t tell a clean story: they fell between 2012 and 2019, exploded in 2022 with the gas crisis, and have been growing again since then. But when you break it down by component, one line item concentrates the structural rise — and it correlates with the penetration of the two technologies that mechanically force more redispatch: solar PV and wind.

What balancing services are (and why nobody looks at them)

The price residential consumers pay is not OMIE. It’s OMIE plus tolls, charges and balancing services — this last bucket adds €/MWh that retailers pass through to the bill. Here we unpack that last piece.

REE publishes SSAA as 22 technical components with opaque names: BS3, RT3, RT6, AJOM, DSV, BALX… Consumers only see the aggregate in the toll, with no breakdown. Each one pays for a different service: holding a power band available to ramp up or down (secondary band BS3, balance reserves BALX/DSV), or covering the technical constraints REE imposes when the market clearing doesn’t produce a secure dispatch.

Of the 22, three structurally dominate by weight: RT3 (consumer-side cost of technical constraints in the day-ahead market — REE’s redispatch to ensure synchronous generation is online), RT6 (the real-time variant of the same mechanism) and BS3 (secondary band — instantaneous fine-grained frequency adjustment). Separately there’s AJOM, which despite its name is not a balancing service: it’s the consumer-side settlement of the production-cost adjustment mechanism under RD-L 10/2022, popularly the “gas cap” or “Iberian exception”. It only appeared between 2022 and 2023 while the mechanism was in effect; by 2024 it’s back to zero.

The physical mechanism behind RT3 is concrete. The OMIE day-ahead market clears purely on price. The cheapest offers win — in Spain that’s increasingly PV and wind. But the resulting dispatch isn’t always technically secure: the grid needs inertia (from spinning synchronous machines), reactive power, fault response, transmission capacity in the right places. When OMIE’s pure-price clearing leaves the system short of these stability services, REE intervenes via technical constraints — forcing some renewable output off and bringing synchronous units (CCGT, large hydro, nuclear) back on, regardless of price. Those redispatched units get paid pay-as-bid, the difference vs OMIE is charged to consumers via RT3.

The connection to renewables isn’t with all of them: it’s specifically with the non-dispatchable ones — solar PV and wind. Large hydro, combined-cycle gas, and pumped storage are dispatchable: they already provide the synchronous stability services REE needs, so they don’t trigger constraints. Variable renewables don’t: when they win the OMIE clearing they displace synchronous capacity, and REE has to redispatch to put it back. That’s why throughout this analysis we use % variable (PV + wind) over total generation, not the aggregate renewable metric that mixes heterogeneous things.

Decomposition: RT3 explains the entire structural rise; AJOM was a separate phenomenon

Summing the 22 components hides two very different stories. Separating them:

The 2022-2023 spike was AJOM — and AJOM is not what its name suggests. It’s not a balancing component but the production-cost adjustment mechanism under RD-L 10/2022, popularly known as the “gas cap” or “Iberian exception”. When it took effect in June 2022, gas for power generation was capped below market price; the difference is passed through to the consumer as a separate charge accounted for under AJOM. The line item was at zero every month before June 2022, jumped to 22.95 €/MWh on average in 2022, and by 2024 had returned to zero as the mechanism expired. Its correlation with variable generation penetration is r ≈ 0 — it’s not a physical phenomenon, it’s a one-off regulatory charge.

The structural growth is RT3. Comparing 2020 against 2026 YTD, RT3 added +17.49 €/MWh to the total. Other adjustment components (RT6 +1.62, BS3 +0.83) came along. PC3 and RAD1 actually fell. In other words: a single line item — technical constraints in the day-ahead market — explains more than 100% of the structural rise in SSAA over five years.

Technical constraints vs % variable generation (PV + wind): r = +0.76

The correlation between the cost of constraints + reserve bands (RT3 + RT6 + BS3) and monthly % variable generation isn’t marginal: it’s strong and robust.

With 148 months of data (2014-01 to 2026-04), the Pearson correlation is r = +0.76. The immediate question — “isn’t it just that when wholesale prices rise, everything rises?” — has a quantitative answer: the partial correlation controlling for OMIE is also +0.76, virtually identical. The effect comes from the structural displacement of synchronous generation by variable renewables in OMIE’s price-only clearing, not from the wholesale price itself.

The slope of the linear fit over the 148 months is +0.54 €/MWh per additional point of % variable. Restricting to the post-2020 era (where the slope is steeper), the coefficient rises to +0.69 €/MWh per point. For context: % variable went from 27.6% in 2020 to 43.7% in 2026 YTD — those 16 extra points roughly explain the observed rise in constraints cost.

A note on the denominator: we use % variable over total P48 peninsular generation (wind + PV + nuclear + hydro + combined cycle + cogen + biomass + thermal solar + pumped storage + coal). If you calculate it over “total renewable” (including large hydro), the correlation is similar (r ≈ +0.78) but the mechanism gets muddy: regulated hydro is itself synchronous and dispatchable — it doesn’t trigger constraints. The most honest metric for this thesis is variable.

The coupling has tightened in recent months — and Operación Reforzada explains the jump

The numerical correlation has its temporal counterpart, with a nuance:

Between 2020 and 2024 the % variable (PV + wind) rose steadily (from 28% toward 41%), but RT3 stayed relatively flat below 5 €/MWh. The connection existed — the monthly correlation captures it — but the constraints cost wasn’t responding with the same intensity yet. Neither the COVID lockdown in March 2020, nor the gas shock following the Russian invasion of Ukraine in February 2022, nor the entry and exit of the “gas cap” (RD-L 10/2022, June 2022 to January 2024) moved RT3 in a sustained way. All of that passed “above” — through wholesale price and through AJOM — without touching the constraints line item.

The regime changes abruptly in April 2025. Right in the month of the Iberian blackout on April 28, RT3 makes its first real spike (from ~5 to >15 €/MWh in a few weeks), drops temporarily, and from late 2025 accelerates without brakes past 20 €/MWh. The mechanism is now documented: REE introduced Operación Reforzada — substantially stricter security criteria requiring more synchronous capacity online at all times to prevent another collapse. Per PwC’s February 2026 analysis: CCGT volume used to resolve technical constraints rose +39% (May–Dec 2025 vs same months 2024), and the monthly cost of constraints rose +55%. December 2025 alone hit 11.50 €/MWh (15% of the wholesale price that month). The ×2.5 step we observe in RT3 is the consumer-side print of this operational change.

The physical mechanism is direct: the more non-dispatchable generation in the mix (sun + wind), the more synchronous capacity gets displaced in OMIE’s clearing, and the more REE has to redispatch to bring it back for security. What these last 12 months suggest is that the system has entered a zone where adding variability is no longer absorbed for free — every additional point triggers more redispatch, and the redispatch is now priced under a much stricter security regime.

What the bill doesn’t tell you

Three quantitative and defensible conclusions:

1. The cost of redispatching synchronous capacity to cover PV and wind displacement has multiplied by 10 since 2020 (RT3: 1.97 → 19.53 €/MWh). This item sits in the electricity bill — aggregated under “tolls and charges” — but doesn’t appear as a separate line.

2. Each additional point of variable generation (PV + wind) adds ~0.54 €/MWh to constraints cost over the full period, and ~0.69 €/MWh in the post-2020 era (r = +0.76, n=148 months; partial r controlling OMIE = +0.76). The effect is independent of wholesale price and specific to non-dispatchable renewables — not to regulated hydro, pumped storage, or thermal generation.

3. The 2022-2023 shock (total SSAA jumped to 32 €/MWh) was AJOM, not constraints — the consumer-side settlement of the Iberian gas exception (RD-L 10/2022), a one-off regulatory charge now extinct, not a physical phenomenon linked to the renewable transition. Separating it from the analysis is methodology, not cherry-picking.

This isn’t an argument against the renewable transition. It’s honest accounting of the cost of the integration mechanism — REE has to redispatch synchronous generation to compensate for what OMIE doesn’t price. While public debate focuses on OMIE, a line item of similar magnitude — already above 20 €/MWh in many months, accelerating under Operación Reforzada — grows without scrutiny.

Anyone analyzing the future cost of the Spanish power system needs to model RT3 explicitly. The data also suggests something uncomfortable: the variable→constraints relationship doesn’t look linear, it looks convex. The system absorbed the rise from 28% to 41% between 2020 and 2024 without major cost, but the 41% → 44% stretch in 2025-2026 has quadrupled RT3 — partly because the underlying renewable share is now hitting threshold levels where synchronous displacement triggers more redispatch, partly because Operación Reforzada amplified the per-MWh constraint cost. If that convexity continues, the next penetration points will cost more, not less.

---

This data is open in ESIOS Data — ask the Spanish electricity market anything you want.