Utility Pressure Map

How the need-to-plan ratio works

The estimate answers one question for each utility: does the capital plan it has published cover what its own pressures imply it needs to spend over the next five years? The need side is built without reference to the plan. The plan side is the utility's stated figure, adjusted for two things regulators and courts have already taken off the table.

The ratio

1.0 means the independent estimate equals the stated plan. Above 1.0 the plan is short and something is deferred. Below 1.0 the plan contains spending this model does not price, such as a specific undergrounding commitment, a renewable build-out required by statute, or generation for a single contracted customer.

Capital capacity (the denominator)

• Stated plan: the utility's published capital plan, scaled to five years when the published window is different, with the generation and storage share removed. The need side of the ratio prices delivery and asset condition, not supply, so the plan side has to be measured the same way. For vertically integrated utilities this share can be 30% to 40% of the plan and it is estimated where the utility does not break it out.
• Regulatory reduction: the share of a plan assumed to be cut by regulators at an affordability score of 5, applied only when the plan is still pending a decision. Default 10%, from a 2025 California general rate case decision that came in about 10% below the request. Scaled by the utility's affordability score. Set to zero for plans that already reflect a commission decision.
• Known unrecovered liabilities: amounts already committed elsewhere, such as wildfire settlements not covered by a state fund.

Capital need (the numerator)

Capacity need, for serving new load:

• Contracted GW: large load under signed agreements, counted in full.
• Pipeline GW: announced but uncontracted load. Where a utility publishes a late-stage or high-probability figure, that is used; the long list of interconnection requests is not.
• Share forecast to materialize: default 60%. Reference point: one large Mid-Atlantic holding company's high-probability pipeline fell about 40% after it required collateral from applicants.
• Delivery capital per GW: transmission and distribution to serve one gigawatt, default $1.2B. Generation is outside the ratio on both sides.
• Organic growth: electrification and customer growth not tied to a specific large-load contract. Each utility carries an editorial rating of 0 (flat), 1 (typical) or 2 (high or electrification-driven), applied to a per-customer amount with a default of $1,500 over five years at the high rating. This term exists because several utilities have large distribution capacity programs driven by electrification policy rather than data centers.

Reliability need, for the replacement backlog:

• Asset age index = (0.70 − net plant ÷ gross plant) ÷ 0.25, held between 0 and 1. Net and gross plant in service come from FERC Form 1 (Account 101 less Account 108). A fleet at 0.70 scores 0; a fleet at 0.45 scores 1. Not yet loaded for any utility.
• Routine outage index = (SAIDI excluding major event days − 60) ÷ 180, held between 0 and 1. Minutes per customer per year from the EIA-861 reliability file, averaged over 2020 to 2023. 60 minutes scores 0; 240 scores 1. The national average is about 120 and has been flat since 2013. Loaded for 11 utilities.
• When neither measured input is loaded, the editorial score alone is used and the detail panel says so.

Resilience need, for hazard exposure:

• Major event index = SAIDI attributable to major event days ÷ 500, held between 0 and 1, averaged over 2020 to 2023. The 2020 to 2024 national five-year average was about 320 minutes, which scores 0.64. Loaded for 11 utilities.

Gas need, for pipeline integrity:

• Vintage pipe index = share of distribution mains that are cast iron, wrought iron, bare steel or unprotected steel ÷ 0.10, held between 0 and 1. From PHMSA gas distribution annual reports. A system with 10% or more vintage main scores 1. Not yet loaded for any utility.

Requested but not approved, the most direct evidence of unfunded need:

• When a utility asks for a number and the commission funds less, the difference is the utility's own estimate of need that did not get funded. It is added to need dollar for dollar and shown as its own hatched segment. The largest examples on the list are a multi-billion dollar reduction in a 2025 California general rate case and a distribution system plan in Colorado approved at a fraction of the request. Where the figure is a revenue requirement rather than capital it is used as a floor, and the detail panel says so.
• This is also why the regulatory reduction is not applied to a plan that already reflects a decision. Applying both would count the same cut twice.

Baseline, for work that happens regardless of the scores:

Triage

When need exceeds capacity, baseline is funded first and the remaining categories are funded in the selected order until capacity is exhausted. The default order is capacity, gas safety, resiliency, reliability. The reasoning: new load usually has a counterparty contributing to cost and a regulator supportive of growth; gas integrity and wildfire hardening are typically mandated; distribution reliability is the most discretionary and the slowest to recover. The order is a selector because that reasoning is contestable.

What is loaded and what is not

Loaded for every utility: the stated capital plan, contracted and pipeline large load, customer counts and known unrecovered liabilities, all from public disclosures. The detail panel marks any of these that are estimates rather than published figures.

Loaded for 11 of the 19 electric utilities: reliability history from the EIA-861 reliability file, 2020 through 2023, as SAIDI with and without major event days, averaged over the four years. These feed the routine outage index and the major event index. Each detail panel shows whether history is loaded and which operating company it came from. For holding companies the largest operating company is used. The other eight run on the editorial score alone.

Not loaded for any utility: asset age from FERC Form 1 and vintage pipe share from PHMSA. The fields exist in the data file and the formulas use them when present. The sources, if you want to fill them:

• FERC Form 1, Comparative Balance Sheet: Account 101 (electric plant in service) and Account 108 (accumulated provision for depreciation). Investor-owned utilities only.
• EIA-861, Reliability file: SAIDI with and without major event days, by utility, 2013 onward. Covers investor-owned, public power and cooperatives.
• PHMSA Gas Distribution Annual Report, Part B: miles of main by material and installation decade, by operator.

What a ratio near 1.0 can hide

The denominator is the stated plan. A stated plan is often already the constrained number: it is what remained after a commission cut the request, after the company decided not to issue equity, and after a liability started consuming the balance sheet. A ratio near 1.0 can therefore mean the plan is balanced, or that the triage already happened before the plan was published. The requested-but-not-approved term recovers part of this where a cut is documented. The financing note in each detail panel is there for the rest, and it is informational rather than in the math.

Calibration and limits

The per-customer amounts were set so the largest investor-owned utilities land near their published plans. For those utilities the ratio is partly circular; for everyone else it is informative. Customer count stands in for system size, which favors dense urban systems. Unit costs vary widely by territory. Load pipelines are the noisiest input, which is why contracted and pipeline load are separated and the conversion rate is a slider. The ordering across utilities is more stable than any individual value.