Interconnection queues are a pricing failure, and faster studies will not clear them

Grid interconnection is argued about as though the only question were how much transmission exists. I take the other side. The queue in front of the American grid is a pricing failure, and it is a particular one. The charge a developer will face for the network upgrades its project triggers is not knowable at the moment the decision to build is made. It depends on which other applications are studied first, and on which of them withdraw. A charge settled that way is a draw, not a price.

The obvious objection is that connection is already priced, and priced dearly, because the developer is charged for the upgrades its project triggers. That is true and it is beside the point. A price does its work before the commitment, by telling one party to proceed and another to go elsewhere. A charge revealed only after the study, and sized by the composition of a queue the applicant cannot see, arrives too late to sort anybody. It transfers money. It does not allocate capacity.

What is arriving at the fence

Something large is being built and it will ask to be connected. Construction spending on manufacturing plant ran at an annual rate of 75 billion dollars in 2020, 82 in 2021, 125 in 2022, 202 in 2023 and 244 in 2024, before falling to 215 in 2025. Nonresidential construction spending went from an annual rate of 856 billion dollars in 2020 to 1,309 in 2025, a rise of 453 billion, or 52.9 per cent. Both series are stated at a seasonally adjusted annual rate, so each point is a yearly rate and not one month's outlay.

Neither series counts a connection. Neither records an application, a study, a withdrawal or a day of waiting, and this brief holds no dataset that does. What the two establish is the scale of the thing arriving: structures that will draw power, put up across six years in which the cost of money changed completely.

The rate rise priced patience on one side only

The federal funds effective rate averaged 0.38 per cent in 2020, 0.08 in 2021, 1.68 in 2022, 5.02 in 2023 and 5.14 in 2024, easing to 4.21 in 2025. From the 2021 average to the 2024 one that is a climb of 5.06 percentage points. The ten-year Treasury constant maturity rate ran 0.89 per cent in 2020 and 4.29 in 2025, a move of 3.4 points.

For a sponsor with land bought, equipment ordered and a construction loan drawn, delay is now expensive in a way it was not when the policy rate averaged 0.08 per cent. For an applicant holding a queue position and little else, a year costs roughly what it always did. Dearer money therefore does not press on a queue evenly. It presses on the applicants who have already committed capital, which are the ones most able to build, and hardly at all on those who have committed nothing. A first-come rule cannot favour that group, because being ready is not what it ranks on.

An input with a published price

The crews who build and energise connections are bought in a market, and what they cost is published. In the occupational wage estimates for 2025, Maine reports 1,160 electrical power-line installers and repairers out of 640,610 jobs in the state, which the file states as 1.81 for every thousand. Their annual median wage is 83,090 dollars against a state median of 51,430, a gap of 31,660 dollars, or 61.6 per cent.

That row carries a location quotient of 2.15, meaning Maine's share of jobs in the occupation is a little over twice the national share. Maine is also the only jurisdiction in which the occupation appears in this file at all, because the file keeps the twelve occupations each state is most concentrated in and drops the rest. Nothing there describes the national workforce, and a wage above a state's median across all occupations is a price rather than a measure of shortage. The narrow point survives both caveats. The price of that labour is published and can be read before a commitment is made; the charge for a connection is settled after one.

What planning can and cannot do

Planning is not the weaker argument because it is wrong. Transmission is lumpy, shared and slow, and no set of individual price signals will assemble a network across separate interconnections and many jurisdictions. Somebody has to build ahead of demand, and only a planner does that. Concede all of it. It still does not reach the queue. Capacity is a question of how much room exists. A queue is a question of who enters it and in what order. Add wires without changing the rule and the order is unchanged, and the new room is claimed by whoever filed first.

Most of what is called interconnection reform is service-rate reform. Clustering studies, imposing deadlines and adding engineers all raise the rate at which applications are processed. None of them changes the rate at which applications arrive, and arrivals respond to the cost of applying. Raise throughput while entry stays cheap and the backlog re-forms at the new throughput. That is an argument about mechanism. It is not a measured claim about any particular reform, and the files this brief works from would not support one.

What follows

Three things follow. Post the charge before the study, as a published price for capacity at each point of interconnection, firm for a stated period, with the network owner carrying the risk that it judged wrong. Make a position cost something to hold and something to surrender, through deposits that scale with the capacity requested and rise the longer a position is held, and through transferability, so that a position has a visible price and the applicant who cannot build sells to the one who can.

The third is measurement, and it is the one that would settle the argument I have just made. Withdrawal rates and time in queue, published per interconnection and per study cycle, would show whether a reform changed the composition of the queue or only its speed. Nobody arguing this question, on either side, can point at that series today. I have made the pricing case without it, and a reader is entitled to hold that against the case.

My position is that the wires are a planning problem and the queue is a price problem, and that treating them as one problem is the mistake. Build the wires. Then decide what a place in front of them costs, before the study rather than after it. Otherwise the same queue forms in front of the new capacity, with the same people at the front of it.