Assumption Audit
Every measurement smuggles in premises. These are the ones this report depends on, ranked by what breaks if they are wrong.
Load-bearing assumptions
An assumption is load-bearing if the report's conclusions collapse when it fails. These are listed first because they are where scrutiny is most productive.
Dollars are a valid proxy for influence
OCDI-Gov treats money as the unit of "control". But influence also flows through endorsement, organised labour, media attention, religious networks, incumbency and party structure — none of which are denominated in dollars. If non-monetary influence is large and differently distributed, an all-money measure will systematically mislocate power.
If false: OCDI-Gov measures the distribution of money, not of power, and the normative conclusions do not follow.
Financial concentration transmits into governing outcomes
The step from "money is concentrated" to "government is captured" requires a transmission mechanism. H2 tested the most direct observable one — vote-switching — and did not find it.
Status: the assumption is not refuted outright, but its most testable form failed. Any continued reliance on it must name a different mechanism and measure it.
The electorate has coherent shared interests to be worked against
The framing of "voters as a cohesive group whose interests money opposes" assumes the electorate has a determinate collective interest. Social choice theory gives strong reasons to doubt this: Arrow's impossibility theorem and Condorcet cycles show that coherent collective preferences need not exist even when every individual is perfectly rational.
If false: "money working against the public interest" has no stable referent, and divergence measures cannot be given that interpretation. This does not undermine measuring divergence — only the moral gloss placed on it.
Structural assumptions
Equal formal authorisation is the right baseline
OCDI-Gov compares finance against one-person-one-vote. That baseline is stipulated by law, but it is a normative choice: it treats legal equality as the standard from which deviation is meaningful. An alternative baseline (taxes paid, population including non-voters, affected parties including future generations) would give different scores.
Party unity is a meaningful measure of independence
Low party unity is treated as evidence of independent judgement. It could equally reflect constituency pressure, ideological extremity, or strategic position-taking on votes whose outcome is already determined.
FEC disclosures are substantially accurate
The report takes filings at face value. They are legally compelled and audited, but undisclosed 501(c)(4) transfers into Super PACs are a known gap — money whose origin is structurally unobservable.
Assumptions that would be convenient to leave unstated
- That the 2024 cycle is representative. One cycle, in a presidential year, with unusual features. Nothing here establishes a trend.
- That federal patterns generalise to state and local government. Untested; state races have far lower disclosure thresholds and far less competition.
- That the categories in FEC transaction codes mean what their labels suggest. Code 24A ("independent expenditure against") is taken at face value.
- That measuring capture is itself politically neutral. Choosing to measure money rather than, say, administrative discretion or bureaucratic autonomy is already a theoretical commitment about where power resides.
This report was requested in terms that named its expected conclusion. That framing is itself an assumption — that the answer is known and the work is illustrative. It has been treated here as a hypothesis rather than a premise, which is why H2 is published as refuted.