q45 · askamerica
How much does campaign money cluster in time around when federal rules get finalised?
Answer
Very little, at least for the case this corpus can actually test (EPA/energy-and-auto-sector
rulemaking, 2019-2025). Two tests:
- Volume test — monthly count of EPA final rules vs. monthly individual-contribution dollars
from donors self-reporting an oil/gas/energy/coal/utility/petroleum/electric employer:
corr ≈ 0.09-0.10, r² ≈ 0.01, n = 75-83 months. Essentially no relationship — months with
more EPA rulemaking activity are not months with more energy-sector giving.
- Landmark-rule test — the 12 highest-salience EPA/DOT rules in the window (Clean Power Plan
repeal, both WOTUS redefinitions, the 2020/2024 oil-and-gas methane rules, CAFE/light- and
heavy-duty GHG standards, the 2025 CAFE reset): energy-sector giving in the month one of
these landed averaged ~26% above the non-event-month average ($6.46M vs $5.13M). That gap
is not trustworthy as a rule-timing effect — 6 of the 12 landmark rules happen to fall in 2020
or 2024, the two presidential-election years that already dominate the series (Aug/Sep/Oct 2020
alone: $21-40M/month vs. a $2-6M typical month). The apparent bump is much better explained by
election-cycle seasonality than by rule finalization.
Net finding: at the resolution this corpus can measure, campaign money tracks the election
calendar, not the regulatory calendar. That is itself informative given the question's premise.
Method
fedregister.fr_documents (doc_type='RULE', agency EPA) for finalization dates; observed
coverage 2019-2026 (declared window is 2010-2026 but not actually backfilled before 2019).
fec.individual_contributions.employer ILIKE energy/auto keywords, transaction_date filtered
to 2019-2025 (a handful of corrupted outlier dates outside this range were excluded).
- Monthly counts/sums joined and correlated with SQL
CORR/REGR_*.
- A day-level event study (distance-to-nearest-landmark-rule bucketed daily contribution totals)
was attempted but is not usable: the connector's
query tool caps results at 500 rows, and
a GROUP BY transaction_date over ~2,500 distinct dates returned only an arbitrary 500-date
subsample — not a real time series. I did not present numbers from that truncated pull.
fec.committee_contributions.transaction_date (PAC-to-candidate money, the more
legislatively-targeted channel) is corrupted table-wide — EXTRACT(year ...) returns values
like 27438, 29765, 32640; no plausible-date filter returns any rows. Reported via
report_issue (subject: "fec.committee_contributions.transaction_date corrupted table-wide").
This blocked testing the channel most likely to show tight timing around rule/legislative
events, so this analysis is limited to itemized individual contributions.
Why the design is this shape
Federal rules are finalized by agencies, not Congress, and campaign contributions go to
candidates/committees, not agencies — so any "clustering" has to run through an indirect channel
(industry giving to legislators who oversee/could override the agency). Web search for direct
academic literature on contribution-timing-vs-rule-finalization came up empty; the closest
established finding (Fouirnaies & Hall, "Financial Incumbency Advantage" / access-oriented PAC
giving) is that industry contributions cluster around legislative power — committee
assignments, session timing — not administrative rule dates. That is consistent with what this
corpus shows: no rule-count-driven clustering, only the well-known election-cycle pattern.
Caveats
- Single-sector case study (EPA/energy+auto); other regulator-industry pairs (e.g., SEC/finance,
HHS/pharma) were not tested and could show a different pattern.
- "Landmark rule" list was hand-picked by title keyword, not from an official
economically-significant-rule flag (none exists in this schema).
- committee_contributions (PAC money) unusable due to the date-corruption defect above — the
channel most plausibly timed around legislative oversight of rulemaking wasn't testable.
- Day-level event study blocked by the 500-row query cap; only month-level and window-level
aggregates could be computed reliably.
Sources
- Fouirnaies & Hall, "The Financial Incumbency Advantage: Causes and Consequences" (Journal of
Politics) — access-oriented PAC giving tracks legislative power, not rule dates.
- fedregister.fr_documents, fec.individual_contributions, fec.committee_contributions (this
corpus).