How the money data is built

Where it comes from

• Contributions — the Federal Election Commission's own bulk filings. Every PAC contribution to every candidate, as filed. The link to a member is exact rather than name-matched: the unitedstates project records each legislator's FEC candidate identifiers, so there's no guessing about who received what.
• Lobbying — quarterly disclosures filed under the Lobbying Disclosure Act. Registered lobbyists must list the specific issues they worked on, and they frequently name bills outright. Every claim on the site quotes the filing verbatim and links to it.

What counts as a donation

Four different things arrive in the FEC's committee-to-candidate file under the same transaction code, and only the first is a PAC donating money:

• A PAC's own contribution. Capped at $5,000 per election, so $10,000 a cycle. This is the only thing counted under “from PACs”.
• Money an organization carried. Individuals write checks earmarked for a named candidate and the organization passes them on. It is the individuals' money, it is not capped, and it is shown as its own figure. Reading these as PAC contributions is a real error we made and corrected — see the changelog below.
• Joint fundraising distributions. A committee set up to raise money for several campaigns and split the proceeds. Also individuals' money, also shown separately.
• Transfers between a candidate's own committees. A House campaign fund moved to a Senate run. Nobody donated anything, so it is counted as nothing.

Independent expenditures are reported separately again, because they are not gifts to the candidate — and because money spent against a member would otherwise appear as though it were backing them. Getting that wrong would invert the meaning of the number entirely.

Individual donors are included, on member pages: totals, the largest givers, and the employers that appear most often. The FEC records employers as free text, so those are grouped by a normalized form of the name and the raw strings are shown alongside — the grouping is a convenience, not a claim about corporate structure.

The line we don't cross

VoteWire never infers what a donor wanted. It would be easy to map industries to likely positions and generate a striking percentage for every member — and it would be the only number on this site that nobody could check.

So a position is recorded only when the organization stated one: “support H.R. 1340” in its own filing, or a public statement with a source. Most filings say they are monitoring a bill, which is a genuine non-position, and we report it as exactly that. The result is thin coverage — a handful of bills per organization rather than hundreds. That's the correct trade.

Why “voted with donors” is not a bribery statistic

The single most important caveat on this site. PACs overwhelmingly give to members who already vote their way — that's what the money is for, from the giver's point of view. A member who takes money from a teachers' union and then votes for education funding was very probably going to do that anyway, which is precisely why the union chose them.

So a high alignment figure mostly measures sorting — who picked whom — rather than influence. It is a starting point for a question, not the answer to one. Where it's genuinely interesting is the exceptions: a member who takes an industry's money and then votes against it, or who reverses a long-held position shortly after a new donor appears.

Known gaps

• Organization names are matched between two federal datasets that share no identifier. We only accept exact matches after normalising, so some real connections are missed. Missing a link is recoverable; inventing one is not.
• Lobbying filings are quarterly, so a bill voted on early in a quarter may not appear until later.
• Only organizations that donate to sitting members are checked for lobbying — the full database is far larger, and the rest isn't relevant here.
• Dark money is by definition absent. Nothing on this page captures spending that isn't disclosed.