P50 vs P70 vs P90: how BSEE estimates decommissioning costs

A lease file that says “decommissioning cost: $14 million” implies a level of precision that does not exist. BSEE does not model a single number for what it will cost to plug a well, remove a platform and clear a site. It models a distribution — a range of plausible outcomes with different probabilities attached. The number that ends up on a bonding demand, a lease file or a GOMDecom record is a single point pulled off that curve, and which point gets pulled is a regulatory choice, not a physical fact.

That choice has a name: the P-value. P50, P70 and P90 are shorthand for different percentiles of BSEE’s cost-estimate distribution, and the percentile BOEM selects for its financial assurance program determines how much security an operator has to post. Move the percentile, and you move billions of dollars of required bonding without changing a single well.

Two bureaus, one number

It is worth being precise about who does what, because the two roles get conflated. The Bureau of Safety and Environmental Enforcement (BSEE) builds the probabilistic cost estimates. The Bureau of Ocean Energy Management (BOEM) — BSEE’s sister bureau, both descended from the old Minerals Management Service — administers the leasing and financial assurance program that decides how much security an operator must post, and BOEM’s rule is what actually tells a Regional Director which BSEE percentile to use. BSEE prices the risk; BOEM prices the bond.

The underlying model is not theoretical. Since a rule BSEE finalized in December 2015, operators have had to report their actual expenditures for decommissioning wells, platforms and other OCS facilities rather than estimates. BSEE uses that reported spending data — collected since 2016 for wells and facilities and since 2017 for pipelines, according to the agency’s public materials on the program — to build a probabilistic cost model for every facility on the OCS. The reporting rule closed the loop: real invoices, not vendor guesses, now feed the P-value that eventually shows up in a bonding letter.

Reading the percentile

BOEM’s own rulemaking gives the cleanest definition of what a P-value means in practice. Describing the P70 standard it adopted in 2024, BOEM stated that a supplemental financial assurance figure set at P70 means:

“Based on the uncertainty and risk applied by BSEE to its model, there is a 70% probability of covering the decommissioning cost of the facility (and therefore a 30% probability of exceeding it).”

Apply the same logic across the curve and the percentiles read like this:

• P50 — the amount BSEE’s model expects to cover the cost of decommissioning about half the time. Half of comparable facilities will cost less than this figure; half will cost more. It is the median of the distribution, not a conservative estimate.
• P70 — the amount that covers the cost roughly 70% of the time, leaving a 30% chance the real bill runs higher. This has been the regulatory default for supplemental bonding since 2024.
• P90 — the amount that covers the cost about 90% of the time. It is the most conservative of the three commonly cited values and produces the largest bonding number.

None of these is “the real cost.” All three describe the same underlying distribution of possible outcomes; they differ only in how much cushion against overrun the government has decided to demand. A P50 figure and a P90 figure can describe the identical facility and differ by a wide margin, purely because of where on the curve each one sits.

From percentile to bonding demand: 30 CFR 556.901

The 2024 overhaul of BOEM’s financial assurance rules put this mechanism directly into the Code of Federal Regulations. 30 CFR 556.901, governing base and supplemental financial assurance, states:

“Where BSEE-generated probabilistic estimates are available, BOEM will use the estimate at the level at which there is a 70 percent probability that the actual cost of decommissioning will be less than the estimate (P70).”

That single sentence is the operative link between BSEE’s statistical model and the dollar figure a lessee is told to post. The regulation applies P70 specifically when a lessee or co-lessee does not meet BOEM’s financial-strength criteria — broadly, an investment-grade credit rating threshold under the 2024 rule. A financially strong operator can avoid supplemental bonding on the strength of its balance sheet; an operator that cannot clear that bar has its exposure priced at the P70 mark on BSEE’s curve.

The rule also builds in a second, related use of the same percentile: the proved-reserves exemption. Under the 2024 framework, BOEM will not require supplemental financial assurance where a lease’s proved oil and gas reserves are valued at least three times the BSEE decom cost exposure estimate associated with that lease’s facilities — with the estimate, again, set at P70. Raise or lower the percentile used for that comparison and you directly change how many leases clear the 3-to-1 bar without posting a bond.

BOEM’s rule does not use the P-value to decide whether an operator is liable. Operators remain responsible for the full cost of decommissioning regardless of which percentile was used to size the bond. The P-value only sets how much security the government collects up front.

That last point is easy to lose in the mechanics: a P50-sized bond does not mean the operator only owes 50% of the eventual bill. It owes 100% of it. P50 versus P70 versus P90 is purely a question of how much of that full liability is secured in advance — and how much residual risk sits with the government, the surety, or a future insolvency proceeding if the actual cost lands above the posted figure.

Why the percentile choice moves real money

The 2024 final rule — Risk Management and Financial Assurance for OCS Lease and Grant Obligations, effective June 29, 2024 — replaced a subjective five-factor financial-strength test with objective credit-rating and reserves criteria and codified the P70 standard described above. BOEM estimated the rule would require roughly $6.9 billion in new supplemental financial assurance from industry, with a substantial share of that falling on small and mid-sized operators.

That figure is downstream of the percentile choice, not independent of it. Moving from a lower percentile to P70 pulls the bonding demand higher across every affected lease simultaneously; moving back down would pull it lower the same way. The 2024 rule was BOEM’s answer to a gap the Government Accountability Office had documented starkly: as of June 2023, BOEM held about $3.5 billion in supplemental bonds against a GAO-cited decom cost exposure range of $40 billion to $70 billion for Gulf of Mexico infrastructure. GOMDecom covered that backlog and the bonding gap in more detail in our earlier analysis; the point relevant here is narrower — the size of that gap is exactly the kind of exposure the P-value choice is meant to close, and the same gap is exactly what shrinks if the percentile used to compute it moves down the curve.

2026: the same mechanism, a lower percentile

BOEM did not leave the P70 standard in place for long. On March 9, 2026, it published a proposed rule that would revise several elements of the 2024 framework, citing Executive Order 14154 and a stated goal of freeing up industry capital for OCS exploration and production. The proposal’s central mechanical change is exactly the one this piece has been describing: it would revise “the level of BSEE’s probabilistic estimates of decommissioning costs used for determining the amount of supplemental financial assurance… from P70 to P50.”

That is not a technical footnote — it is the entire lever. A few of the proposal’s other changes compound the effect of the lower percentile:

• Credit-rating threshold lowered. The 2024 rule’s investment-grade cutoff (BBB-/Baa3) would fall to BB-/Ba3, letting more sub-investment-grade lessees clear the financial-strength test without posting supplemental assurance at all.
• The 3-to-1 reserves exemption gets easier to clear. Because the exemption compares proved reserves value to the BSEE decom cost exposure estimate, moving that estimate from P70 to P50 lowers the bar a lease must clear on the cost side of the ratio, independent of any change to reserves value.
• Predecessor financial strength returns to the analysis, and Regional Directors could waive a supplemental assurance demand where decommissioning is scheduled within a year and backed by an acceptable third-party contract.

BOEM’s own estimate of the compliance-cost effect, cited in law-firm analysis of the docket, is that the proposal would save industry on the order of $484 million per year in financial assurance compliance costs relative to the 2024 rule. The comment period on the proposal, after an extension, closed May 15, 2026. As of this writing, the rule remains proposed. The 2024 P70 standard, including 30 CFR 556.901 as written, stays in force unless and until a final rule is published — and a phase-in schedule of that kind should be treated as the standard timeline BOEM has published, since case-specific extensions, disputes and processing delays can move an individual company’s date.

What this means for reading a cost figure on a lease

For a BD or commercial team using BSEE and BOEM data to assess Gulf decommissioning exposure, three things follow from all of the above.

The dollar figure attached to a lease is a modeled decom cost exposure, not a tender value, an opportunity value or a contract budget. It reflects BSEE’s statistical model at whatever percentile is in force at the time, built from other operators’ historical actual expenditures on comparable scope. It says nothing about what a specific operator will actually pay a specific contractor for a specific campaign.

The percentile in force changes what “large” or “small” means when comparing figures across time. A decom cost exposure figure pulled from BSEE’s model in 2025 under the P70 standard is not directly comparable to one pulled after a hypothetical shift to P50 — the second number would be smaller for the identical physical scope, purely because of where BOEM chose to sit on the curve, not because the remaining observable scope shrank.

A cost estimate cannot tell you whether work has started, been contracted, or been abandoned. BSEE’s model prices scope; it does not track procurement. If nothing has been reported against a well or structure, that is best read as scope not observed in monitored sources — never as proof no contractor exists, and never as proof the work has been shelved. Confirming what remains outstanding on a lease depends on the same regulatory chain GOMDecom tracks elsewhere: idle status, filings, approvals and, ultimately, an End of Operations Report. An EOR is what proves an item is done; a missing one proves nothing beyond the fact that nothing has been reported yet, and the remaining observable scope should be read accordingly. Our guide to reading BSEE’s signal chain covers that distinction in more detail, and the methodology page explains how GOMDecom keeps a modeled cost figure separate from an observed procurement signal.

The percentile debate playing out between the 2024 rule and the 2026 proposal is, at bottom, an argument about how much of the Gulf’s decommissioning risk the federal government wants secured in advance versus left exposed if an operator fails. For a contractor reading a BSEE-derived cost number, the practical discipline is the same regardless of which percentile is currently in force: treat it as a modeled exposure figure tied to a specific regulatory choice, not as a price. See a modeled cost figure read this way, next to the filings behind it, in a sample Opportunity Brief.

Sources

• Electronic Code of Federal Regulations, 30 CFR 556.901 — Base and supplemental financial assurance.
• Federal Register, Risk Management and Financial Assurance for OCS Lease and Grant Obligations (final rule, Apr. 2024).
• Federal Register, Risk Management and Financial Assurance for OCS Lease and Grant Obligations (proposed rule, Mar. 2026).
• Bureau of Safety and Environmental Enforcement, BSEE Decommissioning Costs Reporting Rule Finalized (Dec. 2015).
• Bureau of Ocean Energy Management, Risk Management and Financial Assurance Rule Implementation Timeline.
• Bureau of Ocean Energy Management, Financial Assurance Requirements for the Offshore Oil and Gas Industry (2026 rulemaking status and comment deadline).
• US Government Accountability Office, GAO-24-106229: Offshore Oil and Gas — Interior Needs to Improve Decommissioning Enforcement and Mitigate Related Risks (Jan. 2024).
• Bracewell LLP, BOEM Finalizes Long-Awaited Overhaul of Offshore Oil and Gas Financial Assurance Regulations (2024) and BOEM Proposes to Loosen Supplemental Financial Assurance Requirements (2026).
• Liskow & Lewis, BOEM Proposed Rule Includes Significant Changes to 2024 Financial Assurance Regulations (2026).
• GOMDecom, the Gulf's overdue iron: what the decommissioning backlog means for contractors in 2026 and the decommissioning signal chain: how BSEE filings reveal work before execution.