Methodology

This page explains how we build our original data products — the Federal Lobbying Intelligence Index and The Lobbying Ledger. The Ledger publishes real figures today, and the sections below describe the methodology behind them. The Index does not: its page is a clearly-labeled format preview containing sample data only, and its finalized method will appear here in full before it goes live with real figures.

Primary source

Our source for federal lobbying data is the filings made public under the federal Lobbying Disclosure Act (LDA) — the law requiring registered lobbyists and lobbying firms to file reports naming their clients, the general issue areas they lobbied on, and the income or lobbying expenses they reported for the period. These filings are public records maintained by the Clerk of the House and the Secretary of the Senate, and we retrieve them from the official lda.gov filing database.

Published Lobbying Ledger editions are built entirely from those filings, and report only what the filings disclose. The Index is not: the table on its page today is illustrative sample data, labeled as such, and is not derived from any filing. If and when the Index publishes real figures, they will be drawn from the same official filings, under the finalized methodology described below — which will be published here before any real Index figure appears.

What The Lobbying Ledger covers

The Ledger is a recurring, dated publication. Each edition covers one reporting period — a single Monday-through-Sunday week in Washington, D.C. time — and reports:

• New lobbying registrations — original registrations posted during the period. Registration amendments are excluded, and are not counted anywhere in an edition.
• Termination filings — quarterly reports filed to terminate a registrant-client lobbying engagement, including those reporting no lobbying activity for the period. Amendments to termination filings are excluded.
• Disclosed issue areas — a tally of the official LDA general issue codes named across that period’s new registrations. One registration can disclose several issue areas, so those counts are not mutually exclusive and generally sum to more than the number of registrations.

The first published edition covers July 27–August 2, 2026. Editions are dated by the reporting period they cover.

How the reporting window is defined

A filing belongs to an edition based on when it was posted to the public LDA database — the official posting timestamp (dt_posted) the database records for it — not the date the underlying lobbying activity occurred, and not the quarter the filing covers. A termination filed in July for the second quarter therefore appears in a July edition, with the quarter it reports on shown alongside it.

We apply that Monday-to-Sunday cutoff in our own code, against each filing’s own timestamp, rather than relying on the date filters we send to the lda.gov API. The API filter is used only as a coarse net that deliberately fetches a slightly wider range; the exact boundary is enforced afterwards, per filing, and is correct across daylight-saving transitions.

How each edition’s figures are computed

Each edition is built from a dated, archived copy of the filings retrieved from lda.gov, and every figure in it is computed from that archive by code — never hand-tallied, estimated, or entered by hand. The build refuses to produce an edition if a filing turns out to be a different filing type than the one it was retrieved as, or if the same filing appears twice. Because each edition is built from a fixed archive, its figures are stable: a published edition is a record of what the database showed for that period at the time it was retrieved, and every edition states that retrieval timestamp.

How we treat termination filings

A termination filing reflects the termination of one specific registrant-client lobbying engagement, as reported under the LDA. That is all it establishes, and it is all we report. It does not necessarily mean an organization has no other federal lobbying representation — a client may retain other registrants, or file a new registration later. We do not describe a termination as a firm being “fired,” and we do not infer why an engagement ended, including when a filing reports a dollar figure alongside a “no activity” status. Where a filing reports an amount, we present it as source data and nothing more.

Income and expenses are not the same figure

The LDA collects two different money figures from two different kinds of filer: outside lobbying firms report the income they received from a client, while organizations that lobby for themselves report the expenses of their own in-house lobbying. These are not comparable measures of the same thing. We show each amount labeled with the reporting type it came from, and we do not add them together, average them, or rank them against one another in a combined figure.

Official data versus our own analysis

In every Ledger edition, the counts, registrations, terminations, and issue-area tally are official LDA data: they report what the filings disclose, with no inference or editorializing beyond the counts each section describes. Editorial selection and commentary from us appears only in “Filings We’re Watching,” which is labeled as Federal Policy Monitor analysis wherever it appears — and is omitted entirely from an edition when we have not selected any filings, rather than filled with placeholder commentary.

Provenance

Every filing we publish keeps its LDA filing UUID and a link to its official filing document on lda.gov, so any figure can be checked against the primary record. Every edition also carries a source block naming the official source, the reporting period, and the timestamp the underlying data was retrieved from lda.gov.

What the Index will measure

The Federal Lobbying Intelligence Index is intended to aggregate LDA filing activity into a comparable, sector-level view: how much lobbying activity is occurring in a sector or on a specific issue, and how that is changing over time. It has not published any real figures: the table on its page today is sample data, labeled as such. The finalized scoring formula, update cadence, and exact sector definitions will be published here in full before the Index goes live with real data — we do not consider a scoring method final until it is public and reproducible in principle from the same public filings we cite.

Corrections

If you believe something we’ve published is inaccurate, contact us at benjamin@cottoncomms.com. Corrections are reflected in the piece with an updated date.