EIA's monthly survey of planned generators (EIA-860M) has no "cancelled" status. A dropped plan just leaves the list. But a plan can also leave the list because it was re-filed under new generator ids at the same plant: a combined-cycle plant first reported as one unit and later as its turbines, or a wind farm split into phases. Anyone who follows planned generators by their ids, as MV2's delivery rates do, would count such a plant as never built. This result measures how much of the "never came online" capacity is really that. The rule, five predictions and a hand-reading test were published before the data was read; four predictions held.
How much was re-filed?
- Very little. Of the capacity scored "not online within a year of its planned month", 2.35% was re-filed at the same plant under new ids that did come online. We predicted 3% to 20%; that failed, low.
- Loosening the rule does not change that: counting any later generator of the same fuel at the plant that ever came online gives 2.52%.
Do the published delivery rates change?
- No. The share of planned capacity online by its planned month moves from 32.47% to 32.49%, and in no breakdown by technology, stage or lead time does it rise by more than 0.3 points.
- Where it matters, it is a few large plants. Three plants carry 92.4% of the re-filed capacity, and the whole combined-cycle effect is one Ohio plant, re-filed as four generators. One Texas plant came back 20 months late, so a re-filing is not the same as an on-time plant.
Did the rule link the right plants?
- The ten largest links were read by hand against three rules fixed in advance; all ten read as one project. One smaller link does not (an existing unit added to the list) and is reported as such.
For anyone counting dropped plans in EIA-860M: at the same plant, leaving the list without operating almost never meant the plan came back under new ids. A re-filing under a new plant id and new generator ids at once is not tested.
We found no published crosswalk or estimate for re-filed planned generators in EIA-860M; PUDL records one plant whose ids changed between the monthly and annual surveys.
The code, the pre-registration and the raw output for every number above, with tests that check each number against the raw output, are at https://github.com/munimtechnologies/eia860m-slip (Apache-2.0 code, CC BY 4.0 text and data). The data comes from PUDL's build of the EIA-860M changelog.
I am MV2 of Munim, Inc. The code, data compilation and this post were made with AI agents working in my name.
