Data
Download the data
Everything the site shows is in these files, rebuilt with every refresh.
Data version 2026.11.18, data as of November 18, 2026.
Files
-
site.jsonJSONEvery person’s latest figure, starting figure, change and rank, plus the office medians, the reference figures and what each form counts. The site’s charts and tables read this file.
-
person/<id>.jsonJSONOne file per person: every snapshot, every filing on record with its link, the starting figure, published estimates, rank and that person’s updates. Example:
person/aaron-bean.json. -
people.csvCSVOne row per person: the columns of site.json, flattened.
-
One row per person per snapshot: every figure the history charts draw, with its filing and how it was read.
-
filings.csvCSVOne row per filing on record, read or not, with its link and why it was or was not used.
-
updates.jsonJSONThe dated log shown on the Updates page.
-
sources.jsonJSONThe source registry shown on the Sources page.
-
forms.jsonJSONWhat each office’s form includes and leaves out, as on the Method page.
-
credits.jsonJSONCredits and licenses for the portraits.
What the fields mean
Dollars are whole dollars as reported unless a field says otherwise. Change fields are in 2025 dollars. Lowest, middle and highest are written low, mid and high in the JSON files. An empty cell or a null means the value does not exist for that row, not zero. When open is true, high is only the known part.
Debt is subtracted in two parts: other debts, and the counted part of the loans on real estate. In snapshots.csv, lowest is assets_low minus debt_subtracted_for_lowest, and highest is assets_high minus debt_subtracted_for_highest. The mortgage debt left out of both is in the two mortgage_debt_not_counted columns.
site.json42 fields
| Field | Meaning |
|---|---|
meta.generated | Date the data was built. |
meta.data_version | Version stamp, the build date as YYYY.MM.DD. |
meta.site_name | Site name. |
meta.cpi_base_year | The year whose dollars every inflation-adjusted figure is in. |
meta.cpi | The price index used: base_year, index (annual average by year), source and any other notes on the series. |
meta.counts | Counts every page quotes: people (everyone on the ledger), with_figure, no_figure and filings (disclosure reports on record, read or not). Also seated (people in office, sitting or acting), elect (elected and not yet in office), former (people who have left office), ranked (people with a rank), unranked_with_figure (a figure but no rank), no_figure_seated (in office with no figure) and no_reliable_figure (in office with no rank). people is seated plus elect plus former. |
meta.offices[] | Each office: id, label, form and n (people in office or elected to it, not former officials). |
meta.occupations[] | Prior-occupation categories: id, label and description. |
meta.benchmarks[] | Reference figures: id, label, value, year, source, url, caveat and retrieved. |
meta.medians | Per office and for all: low, mid, high and n, over the ranked people only. |
meta.forms | Per office: form_name, who_files, values_as_of, debts_reported_as, open_ended, includes, leaves_out and notes. |
people[] | One row per person. |
people[].id | The person’s id. |
people[].name | Name. |
people[].sort | Name in last, first order. |
people[].office | Office id of the current or most recent role. |
people[].title | Title of that role. |
people[].state | State code, where the office has one. |
people[].district | House district, 0 for at large. |
people[].party | D, R, I or null. Null for Cabinet heads and justices, whose offices are not won under a party label. |
people[].status | sitting, acting, elect or former. |
people[].since | Start of the current role. |
people[].left | For a former official, the day they left their last office, as YYYY-MM-DD. Null otherwise. |
people[].entered | Start of the first tracked office. |
people[].run_start | Start of the current unbroken run in the tracked offices, as YYYY-MM-DD. The run goes back through earlier roles until a gap of more than 366 days out of office. |
people[].born | Year of birth only, as a number, or null. |
people[].img | Portrait file name, or null. |
people[].occ[] | Prior-occupation category ids. |
people[].latest | The latest figure: year, as_of, low, mid, high, open, open_low, grade, kind, source, filed, url, nothing_listed (true when the filing lists no assets and no debts of any kind) and read (how the filing was read: e-filed, scan-two-readings, keyed-by-hand, courtlistener-scan, courtlistener-tables or state-form). Null when there is none. |
people[].nofig | Why there is no reliable figure, as a reason id, or null. |
people[].base | The starting figure: basis (entry or earliest), anchor (entry, return or earliest), anchor_date (the date it was measured from, null for earliest), heading (the heading printed above the comparison), label, year, low, mid, high, open, open_low, nothing_listed and url. Null when there is none. |
people[].chg | The change: verdict, multiple, multiple_at_least, d_low, d_mid, d_high (in base-year dollars), years and unknown. Null when there is none. |
people[].chg.unknown[] | The ends of the two figures that are not known, as codes: start or latest, then high or low, then open (an item with no ceiling) or unvalued (an item with no stated value or amount). For example start-high-unvalued. When two ranges do not overlap and the verdict is still unclear, one of these is the reason. |
people[].chg.no_call_reason | Why a change is unclear although the ranges may not overlap: nothing-listed (one of the two reports lists no assets and no debts) or nothing-valued (one lists rows, none with a value). Null otherwise. |
people[].earlier_office | When the latest figure is from a report for an earlier office: office, as_of, years (its age at the ledger's date) and stale (true when more than 3 years old). Null otherwise. |
people[].changed_after_report | For a Cabinet head whose holdings changed after the report behind the figure: kind (certificate-of-divestiture or periodic-transaction-report), date (the first such filing), n (how many) and pending (a newer report pending at the Office of Government Ethics: label and date, or null). Null otherwise. |
people[].changed_after_report.pending | A newer report pending at the Office of Government Ethics: label (its type and year, such as Annual (2026)) and date (the date OGE gives it). |
people[].nochg | Why there is no change, as a reason code, when chg is null: one-snapshot (one figure only), latest-not-comparable (latest figure graded D), latest-machine-read-scan (latest figure is a machine reading of a scan), no-cpi-for-year (no inflation figure for the year yet), no-cpi (the inflation series is missing), no-comparable-snapshot (no starting figure) and no-snapshot (no figure at all). |
people[].pub | A published estimate of the official’s own or household wealth: value, by, year, url, scope and scope_note. Null when there is none. |
people[].rank | Rank by the middle: all and office. Null when not ranked. |
people[].unranked_reason | Why a person with a figure has no rank: latest-graded-d, not-yet-in-office, not-in-office (a former official), newer-filings-unread (the newest filing on record is not read yet and was filed a year or more after the one behind the figure) or earlier-office-figure-old (the figure is from a report for an earlier office and more than 3 years old). Null otherwise. |
people[].hist[] | History rows: [year, low, mid, high, open (1 or 0), grade], in dollars as reported. |
person/<id>.json33 fields
| Field | Meaning |
|---|---|
person | The person: id, name, sort_name, bioguide, wikidata, born, party, entered, run_start, portrait, bio, roles and current. |
person.born | Year of birth only, as a number, or null. No file gives a full date of birth. |
person.run_start | Start of the current unbroken run in the tracked offices, as in site.json. |
person.portrait | The portrait: file, credit (who made it), license, license_url and source_url (the page the image came from). Null when there is none. |
person.roles[] | Every tracked role: office, title, department, state, district, start, end, status, evidence_url and checked. A governor's or Cabinet head's role that ended because a successor took the same state or department also carries end_basis, set to successor-start. Roles in Congress and on the Court never carry it. |
person.current | The current role: office, title, department, state, district, since and status. |
person.bio | summary, occupations, education and sources. Null when there is none. |
person.bio.education[] | Each degree: school, degree and year. |
person.bio.sources[] | Where the bio comes from: label and url. |
snapshots[] | One figure per date, oldest first: as_of, year, filing_id, kind, source, url, filed, read (how the filing was read, as in site.json), low, mid, high, open, open_low, nothing_listed (true when the filing lists no assets and no debts of any kind), grade, grade_reason and flags. |
snapshots[].assets | low, high, open, n (rows), n_valued, n_open and n_unvalued. |
snapshots[].liabilities | Debts other than loans on real estate: low, high, open, n, n_open and n_zeroed (debts the filer says were repaid by year end). |
snapshots[].home_debt | Mortgage debt not counted in low, mid or high: low, high, open and n. It holds home loans and any real-estate debt above the listed real estate. n counts home-loan rows only. |
snapshots[].mortgage | Every loan on real estate as reported: low, high, open, n (every real-estate loan), n_home (home loans), under_cap (the loans held to the value of the listed real estate), real_estate (that listed real estate), counted_for_lowest and counted_for_highest (what was subtracted for each end) and capped (true when the limit held an amount below the loans). The debt subtracted for the lowest figure is liabilities.high plus counted_for_lowest. For the highest figure it is liabilities.low plus counted_for_highest. |
filings[] | Every filing on record, newest first: filing_id, kind, period_year, as_of, filed, url, source, format, status, read (how it was read, null when not read), low, mid, high, open, open_low, grade, is_snapshot, superseded_by and note. |
latest_detail | For the latest figure: by_class (low, high, count per asset class) and top_assets (up to ten counted rows: class, owner, low, high and name). In a name, a street address with a house number, an account number and the name of the beneficiary of a college savings or custodial account are taken out or replaced by a note that each was withheld. The name of a private person other than the official is withheld, marked (name withheld), and a real-estate item keeps its town and state but not its street, marked (street withheld), or (number withheld) for a lot or unit number. |
published[] | Published estimates: publisher, value, as_of, url, title, scope (filer, household or spouse) and scope_note. |
baseline | The starting figure: basis, anchor, anchor_date, heading, tier, label, as_of, filing_id, url, low, mid, high, open, open_low, nothing_listed and real (in base-year dollars), as in site.json base. Null when there is none. |
change | verdict, basis, d_low, d_mid, d_high, multiple, multiple_at_least, unknown, years, cpi_base_year, cpi_provisional and latest_real (the latest figure in base-year dollars). Null when there is none. |
change.unknown[] | The ends of the two figures that are not known, as in site.json chg.unknown. |
change.no_call_reason | nothing-listed, nothing-valued or null, as in site.json. |
earlier_office | office, as_of, years and stale, as in site.json. Null otherwise. |
changed_after_report.pending | label and date, as in site.json. |
changed_after_report.events[] | Each filing after the report: kind, date and url (the filing at the Office of Government Ethics). |
changed_after_report | kind, date, n and pending, as in site.json, plus events: each filing at the Office of Government Ethics after the report, with kind, date and url. Null otherwise. |
no_change_reason | Why there is no change, as a reason code, when change is null. The codes are those of site.json nochg. |
no_baseline_reason | took-office-before-e-filing, no-figure or null. |
rank | all and office. Null when not ranked. |
rank.all | Among all ranked people: rank (by the middle), best and worst (the rank band the ranges allow) and of (how many are ranked). |
rank.office | The same within the office: rank, best, worst and of. |
unranked_reason | Why a person with a figure has no rank: latest-graded-d, not-yet-in-office, not-in-office (a former official), newer-filings-unread (the newest filing on record is not read yet and was filed a year or more after the one behind the figure) or earlier-office-figure-old (the figure is from a report for an earlier office and more than 3 years old). Null otherwise. |
nofig | reason, detail and evidence_url when the person has no figure. |
updates[] | This person’s entries in the updates log: date and text. |
people.csv56 columns
| Column | Meaning |
|---|---|
id | The person’s id. Their profile is at /p/<id>/ and their full record at data/person/<id>.json. |
name | Name as shown on the site. |
sort_name | Name in last, first order. |
office | Office id: house, senate, governors, court or cabinet. |
title | Title of the current or most recent role. |
state | Two-letter state or territory code, where the office has one. |
district | House district. Empty for other offices, 0 for an at-large seat. |
party | D, R or I. Empty for justices. |
status | sitting, acting, elect or former. |
since | Start of the current role. |
left | For a former official, the day they left their last office. Empty otherwise. |
entered | Start of the first tracked office. “Since taking office” is measured from here. |
run_start | Start of the current unbroken run in the tracked offices. A run is broken only by a gap of more than a year out of office. Same as entered for most people. |
occupations | Prior-occupation categories, separated by a vertical bar. |
latest_year | Year of the latest figure. |
latest_as_of | Date the latest figure is valued at. |
latest_kind | Report type: annual, candidate, nominee, new-filer, termination or amendment. |
latest_source | Source id: house, senate, oge, judicial, or state: and a state code. |
latest_filed | Date the report was filed. |
latest_url | Link to the official copy of the report. |
latest_read | How the latest filing was read: e-filed, scan-two-readings (a scan read by two independent readings), keyed-by-hand, courtlistener-scan (CourtListener’s reading of a scan), courtlistener-tables or state-form. |
lowest | Lowest possible net worth, in dollars as reported. |
middle | Middle of the range, in dollars as reported. Open-ended items count at their floor. |
highest | Highest possible net worth, in dollars as reported. When open_top is 1, only the known part. |
open_top | 1 when an asset has no ceiling, so the true highest is unknown. |
open_low | 1 when a debt has no ceiling, so the true lowest is unknown. |
grade | Data-quality grade, A to D. |
no_figure_reason | Why there is no reliable figure, when there is none. |
start_basis | entry (since taking office, or since a later start after time away) or earliest (since the earliest filing read). |
start_anchor | What the starting figure is measured from: entry (first taking office), return (the start of the current unbroken time in the tracked offices, after more than a year away) or earliest (the earliest usable filing). |
start_anchor_date | The date the starting figure was measured from: entered for entry, run_start for return. Empty for earliest. |
start_heading | The heading the site prints above the comparison, such as “Since taking office in 2019”. |
start_label | Which report the starting figure comes from. |
start_year | Year of the starting figure. |
start_lowest | Lowest possible at the start, in dollars as reported. |
start_middle | Middle of the range at the start, in dollars as reported. |
start_highest | Highest possible at the start, in dollars as reported. |
start_open_top | 1 when the starting figure has no ceiling. |
start_open_low | 1 when the starting figure has no floor. |
verdict | up, down or unclear. Empty when there is no change to report. |
no_change_reason | Why there is no change to report, as a reason code, when there is none. The codes are listed under site.json, people[].nochg. |
change_lowest | Smallest possible change, in dollars of the base year (see real_dollar_year). |
change_middle | Change in the middle of the range, in dollars of the base year. |
change_highest | Largest possible change, in dollars of the base year. |
multiple | Latest middle divided by starting middle, in real dollars. Only when the starting middle is at least $250,000, both are above zero and both ends of the starting range are known. |
multiple_at_least | 1 when the latest figure has no known ceiling, so the multiple is a floor. |
unknown_ends | The ends of the two figures that are not known, separated by a vertical bar. Each names the figure (start or latest), the end (high or low) and why (open: an item with no ceiling, unvalued: an item with no stated value or amount). An unknown end is why two ranges that do not overlap can still read unclear. |
years | Years between the starting figure and the latest one. |
real_dollar_year | The base year of the inflation-adjusted columns. |
published_estimate | An outside estimate, kept separate. Never used for rank or verdict. |
published_by | Who published the estimate. |
published_year | Year of the estimate. |
published_url | Link to the estimate. |
rank_all | Rank by the middle of the range among sitting and acting officials with a figure graded A to C. |
rank_office | The same rank within the office. |
unranked_reason | Why a person with a figure has no rank: latest-graded-d, not-yet-in-office, not-in-office (a former official), newer-filings-unread (the newest filing on record is not read yet and was filed a year or more after the one behind the figure) or earlier-office-figure-old (the figure is from a report for an earlier office and more than 3 years old). Empty otherwise. |
snapshots.csv35 columns
| Column | Meaning |
|---|---|
person_id | The person’s id, as in people.csv. |
as_of | Date the figure is valued at. |
year | Year of the figure. |
filing_id | Source and the source’s own document number. |
kind | Report type. |
source | Source id. |
filed | Date the report was filed. |
url | Link to the official copy of the report. |
read_as | How the filing was read: e-filed, scan-two-readings (a scan read by two independent readings), keyed-by-hand (keyed by hand from the document), courtlistener-scan (CourtListener’s reading of a scan), courtlistener-tables (CourtListener’s tables) or state-form (a state disclosure form). |
lowest | Lowest possible net worth, in dollars as reported. Equals assets_low minus debt_subtracted_for_lowest. |
middle | Middle of the range, in dollars as reported. |
highest | Highest possible net worth, in dollars as reported. Equals assets_high minus debt_subtracted_for_highest. |
open_top | 1 when an asset has no ceiling. |
open_low | 1 when a debt has no ceiling. |
assets_low | Assets at the low end of their brackets, each item counted once. |
assets_high | Assets at the high end of their brackets, each item counted once. |
assets_unvalued | Items listed with no stated value. They add nothing to the sums. |
debt_subtracted_for_lowest | All the debt subtracted to reach the lowest possible figure: other debts at the high end of their brackets, plus the counted part of real-estate debt for this end. |
debt_subtracted_for_highest | All the debt subtracted to reach the highest possible figure: other debts at the low end of their brackets, plus the counted part of real-estate debt for this end. |
other_debts_low | Debts other than loans on real estate, at the low end of their brackets. A debt the filer says was repaid by year end counts as zero. |
other_debts_high | The same debts at the high end of their brackets. |
real_estate_debt_low | Every loan on real estate as reported, home loans included, at the low end of their brackets. |
real_estate_debt_high | The same loans at the high end of their brackets. |
real_estate_loans | How many loans on real estate the filing lists. |
home_loans | How many of those loans the filing places on a home it does not list. These are never counted. |
listed_real_estate_low | The real estate the filing lists, at the low end. Real-estate debt is counted only up to this value. |
listed_real_estate_high | The same real estate at the high end. When an item has no ceiling, only the known part. |
real_estate_debt_counted_for_lowest | The part of the real-estate debt subtracted for the lowest possible figure. |
real_estate_debt_counted_for_highest | The part of the real-estate debt subtracted for the highest possible figure. |
real_estate_debt_capped | 1 when the limit held the real-estate debt below its reported amount. |
mortgage_debt_not_counted_low | Mortgage debt left out of every figure, low end: home loans, plus any real-estate debt above the listed real estate. |
mortgage_debt_not_counted_high | The same debt left out, high end. |
grade | Data-quality grade, A to D. |
grade_reason | Why the grade was given. |
flags | Notes on the figure, separated by a vertical bar, such as open-top, unvalued-items or home-debt-excluded. |
filings.csv19 columns
| Column | Meaning |
|---|---|
person_id | The person’s id, as in people.csv. |
filing_id | Source and the source’s own document number. |
kind | Report type. |
period_year | Year the report covers. |
as_of | Date the values describe, when known. |
filed | Date the report was filed. |
source | Source id. |
format | The document as published: pdf-text, pdf-scan, html, paper, csv or manual (keyed by hand). |
status | ok, partial (only part could be read) or unparsed (not read). |
read_as | How it was read, as in snapshots.csv. Empty when it was not read. |
lowest | Lowest possible net worth from this filing. Empty when it was not read or is held for review. |
middle | Middle of the range from this filing. |
highest | Highest possible net worth from this filing. |
open_top | 1 when an asset has no ceiling. |
grade | Grade, when the filing is a snapshot. |
is_snapshot | 1 when this filing gives the figure for its date. |
superseded_by | The filing that replaced this one, when one did. |
note | Why a filing is listed but not used, when it is not. |
url | Link to the official copy of the report. |
How each number is worked out is on the Method page.
Reusing the data
You may copy, publish and build on these files. Credit The Disclosure Ledger and link to the filing a figure came from, which every row names.
The figures are drawn from reports filed under the Ethics in Government Act, and federal law limits how those reports may be used. The limits apply to anyone who reuses data drawn from them, not only to this site. Under 5 U.S.C. 13107(c), it is unlawful to obtain or use these reports:
- for any unlawful purpose,
- for any commercial purpose, other than by news and communications media for dissemination to the general public,
- for determining or establishing the credit rating of any individual, or
- for use, directly or indirectly, in the solicitation of money for any political, charitable or other purpose.
The Attorney General can bring a civil action against anyone who breaks these rules, and a court can impose a penalty. Do not use these files to screen anyone for credit, lending, housing or a job, or in any appeal for money. Governors’ figures come from state forms, and state law may set limits of its own.
The figures are ranges read from what officials report. They are not an appraisal or a statement of anyone’s actual net worth.