AEC Market Education Module #7

Renewal Timing Windows

01 / The Forward Curve

The window opens 60 months out. A handful of REPs go to 120.

ERCOT power forwards are quoted across a range that depends on who is doing the quoting. Most retail energy providers will price terms out to 60 months. A handful will go to 120 months from prompt. The number that actually matters is which months out are liquid: where bid-ask spreads are tight enough that a real bid will hold up to a real award, and whether the buyer's credit is strong enough to underwrite the term.

Front-month and prompt-quarter forwards trade with the tightest spreads, since most volume settles there. The next two to three years are where commercial buyers spend most of their time. Past 36 months the participant pool shrinks, but liquidity remains workable through 60 months for credit-tier buyers, and matrix products from the major REPs typically extend to 48 months on that basis.

Past 60 months the picture changes. A few REPs will quote 90 or 120 months, and the curve still trades, but the gating factor stops being curve depth and starts being customer credit. A supplier underwriting a fixed price out to ten years needs confidence that the buyer survives the full term, and that lift gets harder the further out you go. For most loads in the 1,000 to 15,000 MWh range, the practical procurement window runs 12 to 48 months, with 24 to 36 months being where the cleanest pricing tends to live.

Forward curve liquidity
By months out from prompt
Matrix products typically cap here
0 12 36 60 90 120
Prompt Core Deep Thin Edge

Read this as: a 36-month term locked at month 0 lives entirely in the core/deep zone. A 60-month term still lives in liquid territory. Reaching past 60, which a handful of REPs will do, pushes into thinner liquidity (the hatched section), where customer credit rather than curve depth becomes the constraint.

02 / The Pattern

Two soft windows a year, and they're not random.

ERCOT power forwards track natural gas at roughly 85 to 90 percent correlation. Gas has a pronounced annual seasonality that's been visible for two decades. The chart below maps that pattern as a 20-year average, with the front-month gas contract indexed against itself.

SOFT SOFT FEB TROUGH NOV PEAK 15% 10% 5% 0% -5% -10% JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC ☀ AVG 1ST 90°F · APR 19 ☀ AVG 1ST 100°F · JUL 1 ❄ AVG 1ST FREEZE · NOV 22 FRONT-MONTH NATGAS · 20-YR AVG · % VS. SEPT BASELINE
SEASONALITY SHAPE MIRRORS THE 20-YEAR AVERAGE PUBLISHED BY EQUITYCLOCK.COM · WEATHER ANCHORS FROM NWS FORT WORTH/DALLAS AND EXTREMEWEATHERWATCH.COM, PERIOD OF RECORD 1899-2024
WINDOW A · MID-FEB TO MID-MARCH

Cold-weather risk is fading, prompt gas pulls back as winter draws end and storage refill demand sets up. That weakness leaks into the forward curve. Power forwards historically soften in sympathy.

WINDOW B · LATE SEPT TO MID-NOV

Summer cooling load is behind, storage looks at where it ended the injection season, and front-month power eases before the winter gas premium fully prices in. Forward power often dips before climbing into year-end.

The mechanism behind the seasonality is straightforward: gas demand surges during the cold months when residential heating runs flat-out, and ERCOT power demand surges during 100° heat when air conditioning runs flat-out. The two soft windows fall in the shoulder seasons between those peaks. Three real-world DFW dates make the calendar concrete. The average first 90°F day is April 19, the average first 100°F day is July 1, and the average first freeze is November 22. The shape of an average flattens out the actual ranges, which are wide. First 90° has occurred as early as January 31 (1911) and as late as June 12 (1970). First 100° has occurred as early as March 9 (1911) and as late as August 23 (1989). First freeze has come as early as October 22 (1898) and as late as January 4 (1972). North Texas also sees one or two measurable ice storms in a typical winter, and any one of them can spike forwards on a week's notice. Forward curves price expected variance, not just averages.

WHY DFW WEATHER MATTERS BEYOND DFW

The trader-proximity effect

An outsized share of natural gas traders live and work in Houston or DFW. When the heat is brutal outside their windows, or when an ice storm shuts down their commute, that sensation enters the decision loop. Forward curves carry some Texas-weather bias because of who is sitting in front of the screens during a price move. Treat the seasonality and the weather anchors as the structural reasons for the soft windows; treat the trader-proximity effect as the reason a single weather event in Texas can move the curve more than the same event happening anywhere else in the country.

03 / Calendar Geometry

June to June, with a flex band on either side.

June is our default target, with late May through early July as an acceptable flex band. The reason shows up clearly when you walk through the calendar: pick an expiration month below and the timeline rebuilds with soft windows shaded green, the casino zone (the last six months before expiration) shaded red, and DFW weather anchors (avg first 100° and avg first freeze) plotted where they fall in the lookback runway. The exception worth flagging up front is lease alignment: better to expire alongside a lease end-date than to buy too much contract.

Contract expires in:
Alden default Flex band Workable, not optimal
☀ Avg first 90°F: April 19 ☀ Avg first 100°F: July 1 ❄ Avg first freeze: November 22 Sources: NWS Fort Worth/Dallas (100° + freeze), Extreme Weather Watch (90°). Period of record 1899-2024. North Texas also sees one to two ice storms in a typical winter; measurable snow has fallen in 127 consecutive winters.
Verdict

Last soft window before expiration
Best lock target (months ahead)
04 / The Frame

Calendar-driven and market-driven are not the same buyer.

Most procurement decisions get made for one of two reasons. Either the contract is about to expire and someone needs a new one, or the curve has moved into territory that justifies acting regardless of where the calendar happens to sit. The first is the default. The second takes some setup.

Default mode

Calendar-driven renewal

The phone rings around 60 days before expiration. Someone gets a renewal letter from the incumbent supplier, or accounting flags it. Quotes go out, two or three come back within a week, and a contract gets signed against whatever the curve happened to be doing that morning.

If the curve is soft, this works out fine. If the curve is up against a multi-year high, it works out poorly, and the company is locked into that outcome for 24 to 36 months. The decision was driven by the calendar, not the market.

What triggers action A contract expiration date moving inside 90 days.
Strategic mode

Market-driven renewal

The forward curve gets watched starting two years before expiration. A range gets defined in advance: a price level that triggers locking the next term, a level that triggers waiting, and a hard backstop where exposure becomes unacceptable regardless of the curve. When the trigger hits, the contract gets executed even if the existing one still has 14 months on it.

This requires a procurement structure that allows mid-term locking, supplier relationships that can move quickly, and the willingness to hold a price discipline through periods where the curve is moving against the trigger. Most companies don't have that setup. The ones that do have meaningfully lower long-run pricing.

What triggers action The forward curve crossing a pre-defined price level inside the procurement window.
05 / Inside Six Months

Inside three to six months of expiration, you're at the casino.

Locking is still a hedge. It prices in expectations and removes uncertainty. Hedging at six months out, with seasonal patterns visible and runway to wait for them, is a deliberate act. Hedging at six weeks out, against whatever the curve hands you that morning, is something else.

/01

Hold-Over Rates

When a contract expires without a replacement signed, the serving REP decides what happens next.

Two patterns dominate. Some REPs bill at a flat hold-over rate two to four times the prior fixed price (the $120 to $330 per MWh range is not unusual). Others roll the meter to real-time index pricing with a penalty adder of $15 per MWh or more on every settlement interval. Both are designed to make signing the renewal feel urgent, not to be competitive products.

/02

Bridge Contracts

90- and 180-day bridge contracts exist for buyers who got caught and need to move now.

Pricing reflects that. Suppliers know the buyer has limited optionality, and short-tenor fixed pricing carries a premium even in soft markets. A bridge that runs through summer can cost more than a 36-month lock priced two months earlier.

/03

Variable Rate Fallback

If no contract gets signed and no hold-over is in place, the meter rolls to a variable rate.

Variable means monthly index, which during ERCOT summer scarcity events has produced individual monthly invoices several multiples of normal. The 2021 and 2023 summer events made this a board-level issue.

The deeper issue

The price is the visible cost. Loss of control is the hidden one.

The point most customers miss is not the price; it is the loss of control. A post-term customer no longer holds a contract with negotiated terms. They hold a meter being served by a REP that has unilateral discretion over rate, billing cadence, and how to handle whatever shows up next in the market. Texas customer protections for accounts above 50 kW are already narrower than the protections small-commercial and residential customers receive. Post-term, those protections narrow further. If a scarcity event spikes prices, if a billing dispute arises, or if a regulatory question lands, the customer is generally at whatever position the REP decides to take, with limited room to push back through escalation.

None of these outcomes are inherently bad if exposure to short-dated pricing is part of an actual strategy. The problem is that almost no one arrives there on purpose. The vast majority of companies sitting in hold-over, bridge, or variable rate exposure ended up there because the calendar got away from them, and the strategy was to renew the contract before that happened.

06 / The Runway

What should happen at each milestone before expiration.

A clean renewal cycle has six checkpoints between the moment a contract gets signed and the moment its replacement gets signed. The labels below use T-minus notation, borrowed from countdown timing in aerospace: T-24 reads as "twenty-four months before expiration," and T-0 is expiration itself. Most companies engage at one or two of these checkpoints. The companies that consistently get good pricing engage at all six.

T-24
Begin tracking the curve

Forward prices for a 36-month term starting at expiration are now quotable with reasonable liquidity. Establish a price reference point and a target range.

T-18
First trigger checkpoint

Two full seasonal soft windows still ahead. If the curve is at or below your target, this is the cleanest place to lock. Long runway, full optionality.

T-12
Comfortable execution zone

One full year of soft-window opportunities ahead. Most disciplined renewals get executed somewhere between T-18 and T-9.

T-6
Casino zone begins

One soft window remaining at most, and only if the calendar lines up. For June expirations (Alden default), this is Thanksgiving, which falls five days past the average first freeze. Pencils-down date for budget-sensitive buyers.

T-3
Whatever the market gives

Tradeable terms exist, but the calendar is now setting the price, not the buyer. Bid-ask spreads widen as suppliers price in the urgency.

T-0
Hold-over and bridge territory

Hold-over rates engage automatically for unrenewed meters. Bridge contracts price at a premium. Variable fallback exposes summer scarcity. The worst pricing of the cycle.

Where does your contract sit on this runway?

If you know your expiration month, we can map the soft windows ahead of you and show what a market-driven renewal would look like for your load.

Get Started