
Maintenance
Part of Field guide to estimate methodology: the five-step pipeline
A closer look at estimate methodology estimate in 2027
Estimate methodology inputs ranked by strength: a working hierarchy of evidence for financial claims, from adversarial court records down to repeated assertion.
Not all inputs are equal, and the useful skill is telling them apart quickly. An estimate methodology estimate is the figure a documented method produces from named inputs, with the range around it. In 2027 one hard limit applies: official data describes years that have closed, and the 2027 figures arrive on a fixed schedule.
What to take away
- Evidence strength depends on why the document was created, not on who published it.
- Documents made under obligation outrank documents made for publicity.
- Repetition sits at the bottom, and most published figures rest there.
- For a 2027 estimate, the release calendar binds before the evidence does.
The hierarchy
1. Adversarial findings. Facts established in a court where both sides had an interest in contesting them. This is the strongest input available anywhere in the subject, because somebody was motivated to disprove it and failed.
The judiciary explains what is publicly available in its guide to court records. Docket pages cost $0.10 each on PACER, capped at $3.00 per document, and fees are waived when quarterly use stays at or under $30.
Evidence Strength Hierarchy
- Adversarial findingsstrongest
- Legal filingsstrong, narrow
- Official statisticspopulation
- Registration recordsclaims only
- Published agreementsgroup terms
2. Filings made under legal obligation. Disclosures a company or an individual must make, in a prescribed format, with consequences for inaccuracy. Retrievable through the SEC's EDGAR full text search, which covers filings from 2001 onward. Strong within scope, and the scope is narrow.
3. Official statistics with a published design. Surveys that define their population, state their sampling method and publish their own limitations. The Bureau of Labor Statistics documents its survey design in the Handbook of Methods. These describe populations, never individuals, and that restraint is why they are trustworthy.
4. Registration records. Trademark and copyright registers, company registers, property records. They establish claims, ownership assertions and transactions on dates. They do not establish value or income.
5. Published agreements. Collective bargaining agreements, published scales, event prize structures. Real documents setting terms for categories of people rather than for individuals.
6. First-person statements. What a person or their representatives say about their own finances. Better informed than any outside guess, unverified, and made by someone with an interest.
7. Reported figures from briefings. What an outlet publishes after being told by an interested party. Accurate transcription of an interested summary.
8. Repetition. A figure that appears in several places because it was copied. This is not evidence at all, and it is where most published numbers in this genre sit.
The test that produces the order
The ordering is not about prestige. It follows one question: why was this document created, and did anyone have a reason to get it right?
Why Documents Get Checked
Input
- Court finding
- Resolving a dispute
- Filing
- A legal obligation
- Official statistic
- Measuring a population
- Registration
- Claiming a right
- Published agreement
- Governing a relationship
- First-person statement
- Persuasion or correction
Created for
- Court finding
- An opposing party
- Filing
- Regulator and filer liability
- Official statistic
- Methodology and peer review
- Registration
- Examining office
- Published agreement
- Both negotiating sides
- First-person statement
- Nobody
Who checked it
- Court finding
- Filing
- Official statistic
- Registration
- Published agreement
- First-person statement
| Input | Created for | Who checked it |
|---|---|---|
| Court finding | Resolving a dispute | An opposing party |
| Filing | A legal obligation | A regulator, and the filer's own liability |
| Pay-versus-performance table | An SEC disclosure rule | The filer's own controls |
| XBRL-tagged financial statement | An SEC filing requirement | Automated validation at filing |
| Official statistic | Measuring a population | A published methodology and peer review |
| Registration | Claiming a right | An examining office, within its remit |
| Published agreement | Governing a relationship | Both sides who negotiated it |
| First-person statement | Persuasion or correction | Nobody |
| Briefing | Placing a story | Nobody |
| Repetition | Filling a page | Nobody |
The line falls between published agreements and first-person statements. Above it, somebody had a reason to be accurate. Below it, nobody did.
What an estimate methodology estimate looks like in 2027
A 2027 estimate is usually a 2026 fact restated. The IRS publishes inflation adjustments for the next tax year in a Revenue Procedure each October, so the tax-year-2027 brackets, standard deduction and related limits arrive in October 2026.
The Social Security Administration announces the 2027 cost-of-living adjustment and the taxable wage base in the same month, from third-quarter CPI-W data. Neither figure is knowable before that announcement, and every 2027 benefit estimate uses both.
Wage benchmarks lag further. The Bureau of Labor Statistics publishes Occupational Employment and Wage Statistics for May of a year in the following spring, so the May 2026 estimates land in spring 2027 and the May 2027 set arrives in 2028.
Census income and poverty estimates keep the same rhythm. American Community Survey one-year estimates for 2026 are released in September 2027, and the five-year estimates lag a year behind that.
Company disclosures for calendar 2026 arrive earlier in the year. A large accelerated filer's 10-K is due 60 days after its fiscal year closes, so a December 31, 2026 year end means a filing by early March 2027.
Court records lag the most. A dispute over a 2027 transaction rarely produces a final judgment before 2028, so rung one is close to empty for the year in question.
Government and nonprofit data trails even that. IRS Form 990 filings for 2026 reach public databases such as ProPublica's Nonprofit Explorer a year or more after the tax year ends.
The Federal Reserve publishes its annual bank stress test scenarios in the first quarter, so the 2027 scenarios land around February 2027 and set the shocks banks model that year.
How to use the hierarchy on an article
Read for documents rather than for claims. Note every source the article names, with issuer and date, and place each one on the list. Then ask which rung the article's main assertion actually rests on. For a 2027 claim, add one more question: could this document exist yet?
Almost every net worth article lands on rung eight. Occasionally one reaches rung seven. Articles about company officers sometimes reach rung two for part of the claim and then continue past it without marking the transition, which is the pattern described on executive net worth.
Why the top rungs never reach a personal balance
Even perfect evidence at rung one or two is narrow. A court finding settles the matter that was litigated. A filing describes one relationship with one company. Neither contains personal liabilities, and no rung on this ladder does, because nothing requires an individual to publish what they owe.
That is why the honest outputs here are documented slices, wide ranges, or refusals, as set out on estimate methodology. What a list compiler actually holds is on richest rankings, and how the same hierarchy applies to pay claims is on contracts and salaries.
Common questions
Is a well-known publication a better source than a small one?
The rung depends on the document, not the masthead. A large outlet reporting a briefing is at rung seven, and a small site linking a court filing is at rung one. Judge the input.
Where do industry databases sit?
Usually at rung seven or eight, since they aggregate reported figures and drop the caveats. Some are useful indexes to primary documents, which is a different job and a valuable one.
What if an article cites nothing at all?
Then it is at rung eight by default. An absence of named documents is itself the finding, and it takes seconds to establish.
Can a 2027 estimate ever reach rung one?
Only for matters already decided. A 2027 filing can sit at rung two within days of submission, but a court finding on 2027 conduct usually waits until 2028 or later.







