Why Your Electricity Bill Is So Complicated — And Why It Has to Be

A guide to different electricity markets: spot, day-ahead, energy-only, capacity.

Electricity Isn't Like Any Other Product

When you buy a bottle of water, it sits on a shelf waiting for you. Electricity can't do that. The moment a generator produces a watt, it has to go somewhere instantly. There's no warehouse, no inventory, no "we'll deliver it Tuesday." Battery storage and pumped hydro exist, but cover only a fraction of what's needed moment to moment. Hence, supply and demand on the grid must be perfectly balanced, every second of every day, or the whole system becomes unstable.^[1]

This one physical constraint makes electricity markets uniquely complex and uniquely fascinating.

The Energy Market — Buying the Actual Electrons

The most basic layer of the electricity market is where generators sell the power they actually produce and utilities buy it for their customers. This happens across two timeframes.

The Day-Ahead Market

Every morning, utilities look at their weather forecasts, industrial schedules, and historical patterns, then submit bids for the electricity they expect to need tomorrow, hour by hour. Generators submit offers specifying how much they can produce and at what price. A central grid operator — like PJM (covering the Mid-Atlantic and Midwest), MISO, CAISO (California), or ISO-NE (New England) — matches buyers and sellers and sets a clearing price for each hour.^[2][3]

The day-ahead market handles roughly 80–95% of all electricity transactions. Why? Because utilities prefer certainty. Locking in tomorrow's prices eliminates the anxiety of not knowing what you'll pay in real time.^[4]

The Real-Time Market (Spot Market)

The real-time market runs continuously. It reprices electricity every five minutes. This is where the day-ahead plan meets reality. A cloud rolls over a solar farm. A factory unexpectedly goes offline. A generator trips. The real-time market absorbs all of these surprises, dispatching faster generators to fill gaps and repricing accordingly.^[3]

This is what people usually mean when they say "spot market" — the live, right-now price of electricity. On a calm spring afternoon, spot prices might sit at $25/MWh. During a heat wave with every air conditioner in the region running at full blast, that same price can spike to $500/MWh or beyond in minutes.^[5]

Why Your Price Depends on Your ZIP Code: LMP

Here's something most people don't know: electricity doesn't have one price, it has thousands. Grid operators use a system called Locational Marginal Pricing (LMP), which calculates a separate price for every physical point — called a node — on the transmission network. PJM alone has over 10,000 pricing nodes.^[6]

LMP has three components:^[7]

Energy cost: the base cost of generating that next megawatt, set by whichever generator is cheapest and still available
Congestion cost: if transmission lines between cheap generation and high-demand areas are full, prices diverge sharply
Loss cost: electricity loses some energy as heat while traveling through wires; this accounts for that

The result is that two factories 20 miles apart can pay dramatically different electricity prices at the exact same moment, simply because of where they sit relative to a congested transmission line. This also acts as a price signal telling developers exactly where the grid needs the most amount of investment.^[8]

Why Energy Markets Alone Aren't Enough

So if prices spike during peak demand, shouldn't that be enough to keep generators running and new plants being built? This is the central question of electricity market design and the answer is a reluctant no, for three uncomfortable reasons.

Reason 1: Price Caps Kill the Signal

Regulators hate politically explosive price spikes. When electricity hits $5,000/MWh during a crisis, the headlines are brutal. So most grid regions impose administrative price caps that prevent prices from rising to their true scarcity level. This is rational politics but terrible economics — it amputates the very signal that would tell investors "build more generation here." Economists call the resulting shortfall the "missing money problem": generators that are critical for grid reliability during peak demand emergencies can't earn enough revenue across the year to justify staying open.^[9][10]

Reason 2: Peaker Plants Can't Survive on Spikes Alone

A gas "peaker" plant might only run 50 hours a year during the hottest days and coldest nights. But it has fixed costs every single day: maintenance crews, insurance, property taxes, debt service. Even without price caps, 50 hours of energy revenue rarely covers 8,760 hours of fixed costs. The plant rationally retires. The grid loses exactly the resource it needs most during emergencies.^[11]

Reason 3: New Plants Take Too Long to React

Even if today's price spike is perfectly visible, a new power plant takes 5–10 years to permit and build. A developer can't look at today's prices and decide to build something that won't come online until the crisis has long passed. The market needs a forward signal — a commitment made years in advance — to trigger investment at the right time.^[12]

These three frictions are why energy markets alone — what economists call "energy-only markets" — struggle to guarantee long-run reliability without additional mechanisms.

The Capacity Market — Paying Generators to Exist

This is where capacity markets enter the picture. Instead of paying for megawatt-hours of electricity delivered, a capacity market pays generators simply for being available and for committing to show up when called.^[13]

Think of it like a hospital paying an on-call surgeon a retainer fee. The surgeon might not be needed most nights. But the hospital needs to know that when a trauma case comes in at 3 AM, someone will be there. The retainer covers the surgeon's fixed costs during the quiet periods and guarantees availability during the critical ones.

How a Capacity Auction Works

Grid operators run capacity auctions 3 years in advance, giving the market enough lead time to actually build new resources if supply is tight:^[12]

The grid operator calculates how much total generation capacity the region needs at peak demand, plus a safety buffer
Generators submit sealed bids — the minimum payment they need to stay available
Bids are stacked cheapest to most expensive; the auction "clears" when enough capacity is secured
Every winning bidder receives the same clearing price — the price of the last accepted bid
Those that fail to show up when called face stiff financial penalties

The cost gets spread across every utility in the region, proportional to their share of peak load. This passes through to consumers as the capacity charge on their electricity bill, typically around 25% of a commercial customer's total electricity cost.^[14]

The Free-Rider Problem: Why You Can't Just Let the Utilities Handle It

At this point, a reasonable question is: why does this need to be centralized and mandatory? Can't utilities just sign long-term contracts — PPAs — directly with generators and solve the problem privately?

PPAs (Power Purchase Agreements) are long-term contracts where a utility agrees to buy electricity from a specific generator at a fixed price for 10–20 years. They're everywhere. Every major utility uses them. And they do solve one problem well: price volatility. A utility that has locked in $45/MWh for 15 years doesn't care if spot prices swing between $20 and $500.^[15]

There are two broad problems why utilities buying PPAs can’t replace capacity markets.

First, standard energy-only PPAs are contracts for energy delivered — megawatt-hours that actually flow. They are generally not contracts for physical availability. If a generator under a PPA trips offline during a polar vortex, it simply doesn't deliver that hour's energy. In such settings, there's generally no reliability obligation, no penalty for being unavailable at the worst possible moment.^[16]

Second and more important one, PPAs can't solve the free-rider problem and this is where electricity economics gets genuinely interesting. Imagine two utilities serving adjacent territories: Utility A, which is responsible and contracts heavily for capacity, and Utility B, which cuts corners and buys the bare minimum.

A brutal heat wave hits both regions simultaneously. Demand spikes everywhere. Utility B's customers start going dark. The grid operator looks at the situation and does what it always does: dispatches generation from wherever it exists on the interconnected grid. Electrons don't check contracts before they flow.^[17]

Utility A's carefully contracted generators bail out Utility B's customers. Utility B's ratepayers — who paid nothing toward that reliability — benefit just as much as Utility A's ratepayers who funded it entirely. This is the textbook definition of a public good: something that is non-excludable (you can't stop Utility B's customers from benefiting) and non-rival (Utility A's reliability doesn't diminish when Utility B uses it). Once enough capacity exists on the interconnected grid to prevent blackouts, every utility in that footprint benefits — whether they paid for it or not.^[16]

Why Bilateral Cost-Shifting Doesn't Fix It

The natural response is: can't Utility A just bill Utility B after the fact?

Yes. Real-time settlement mechanisms do exactly this for the energy that flows. But this misses the deeper problem entirely. You cannot retroactively fund a power plant that already retired. The sequence is fatal:^[10]

Utility B under-contracts for years, saving money
Marginal generators, starved of revenue, retire 3–5 years before the crisis
The heat wave hits; there's now a physical shortage, not just a paperwork gap
Utility B tries to compensate Utility A — but there's nothing left to buy

By the time cost-shifting would kick in, the capacity has already vanished from the system. And even setting timing aside, getting dozens of utilities to voluntarily negotiate their fair share of a collective reliability obligation — without a single holdout blowing up the arrangement — is a coordination problem private markets have never solved at scale.^[17]

The mandatory, centrally-enforced capacity market is the solution to that coordination problem. It's the only mechanism that forces every utility to pay for its share of grid reliability before the plants retire, not after the lights go out.

The Honest Critique: Capacity Markets Are Imperfect

None of this means capacity markets are a clean solution. They're widely criticized, and fairly so:

They overpay existing plants. Because the clearing price is paid to all winning bidders, cheap nuclear and hydro plants that would run regardless receive the same windfall as an expensive new gas peaker that actually needed the revenue^[18]
They overprocure. Grid operators tend to overestimate peak demand, resulting in billions in excess payments that consumers bear^[19]
They're politically fragile. States increasingly subsidize favored generators outside the market framework, distorting auction results and undermining the price signal^[9]
PJM's December 2025 capacity auction fell 5.2% short of its reliability requirements for the 2027–2028 delivery year — the first capacity shortfall in PJM's history — despite record-high clearing prices, driven by data center load growth and accelerating generator retirements^[20]

Texas (ERCOT) takes the opposite approach: no capacity market at all, just a pure energy-only market with very high price caps (~$5,000/MWh) that let scarcity pricing do the work. It's theoretically cleaner. But Winter Storm Uri in 2021 showed what happens when that system fails — prices hit the cap for four straight days, millions lost power, and political pressure to intervene exploded immediately anyway.^[9]

The Bottom Line

Electricity markets are layered precisely because electricity is physically unlike every other commodity. The spot market handles right-now supply and demand. The day-ahead market handles tomorrow's plan. LMP makes prices location-aware. And the capacity market solves the forward reliability problem that voluntary contracting cannot.

The capacity market isn't elegant. It is a workaround for political constraints that prevent scarcity pricing from working properly, and a coordination mechanism for a public goods problem that private markets cannot self-organize around. Reformers have proposed cleaner alternatives: targeted procurement: buying specific reliability attributes like winter firmness or multi-day storage rather than paying every resource the same flat rate regardless of what the grid actually needs; or re-regulation: returning generation ownership to utilities, which several states are already pursuing quietly by subsidizing favored plants outside the market framework. Both ideas have merit. Neither has consensus.^{[28][29][30][31]}

Until society is willing to accept $9,000/MWh electricity bills during heat waves and the political consequences that come with them, the capacity market remains the least-bad coordination mechanism the industry has found. The design debate has been running for two decades. The grid, however, is not waiting: PJM's 2027–2028 auction already fell short of reliability requirements for the first time in its history, and the gap between theoretical market reform and immediate physical need has never been wider.