Add the Schedule C affiliated-group roster — md_ingest_sched_c

An organisation that has made the §501(h) lobbying election and belongs to an affiliated group must list, in Schedule C Part II-A, every member of that group with name, address, EIN, expenses and share of excess lobbying expenditures. AffiliatedGroupSchedule/AffiliatedScheduleGrp in the XML.

Usage

md_ingest_sched_c_affil(con = NULL, path = NULL, version = "v2009-2013")

Arguments

con: A connection from md_db().
path: Path to a derived sched-c-affiliated-group-*.csv. NULL looks in the version source directory, preferring the published tables.
version: Pinned efile_sched_c_affil version, e.g. "v3_1".

Value

Row counts, invisibly.

It is a different concept from group exemption, and must stay one

A §501(h) affiliated group is defined by interlocking governance under §4911 - one organisation bound by another's decisions on legislative issues, or enough board overlap to control them. That is a control test, not the group exemption process, so these edges must never enter the group_id components (GROUP-STRUCTURE-FINDINGS.md §2).

They land as related_entity (§4a), the same value Schedule R Part II uses, and the measured overlap is the argument: 834 of 886 edges (94%) already exist as a Schedule R Part II edge. A looser coordination tie - a coalition in the §4a sense - would not overlap Schedule R that heavily. The word "affiliated" in the statute names a control relationship, not a federated chapter structure.

It is reciprocal, which is why the self-loops matter

Every electing member reports the whole group, so a five-member group produces five rosters that each name all five. In TY2013, 78 of 123 filers list their own EIN, and 652 rows collapse to 339 distinct member EINs. Self-references are dropped; the rest are stored in the disclosed direction (filer -> member) like every other relation, and a consumer wanting the undirected view reads the pair both ways.

It is nearly redundant, and is loaded anyway for the provenance

Measured on TY2013, 394 of 430 undirected pairs (91.6%) already exist as Schedule R Part II edges. This source adds tens of pairs, not thousands.

It earns its place on the evidence tag rather than the coverage: an edge corroborated by both a Schedule R disclosure and a §501(h) affiliated-group filing is a stronger claim than either alone, and only this ingest can say so. Loading it to inflate the row count would be the wrong reason.

Two inputs

the published table, or the old extract: From e-file v2_3 on, this is the published table SC-P02-T01-AFFILIATED-GROUP-<year>.CSV (21,091 rows for 2009-2025 in v3_1), read when sources/efile_sched_c_affil/<version>/ holds it. Otherwise the function falls back to the earlier local extract of raw XML, described below. Through TY2012 the element is AffiliatedGroupSchedule/AffiliatedSchedule with AddressUS/; from TY2013 it is AffiliatedScheduleGrp with USAddress/. The extract matches on field suffix to serve both.


  2009  355 rows /  58 filings
  2010  543 rows /  90 filings
  2011  554 rows /  98 filings
  2012  761 rows / 117 filings
  2013  652 rows / 123 filings
  ---------------------------------
       2,865 rows / 486 filings / 459 distinct member EINs