Form N-PX: test actual voting against the stated policy
A voting policy describes intended behavior; a Form N-PX record supplies evidence about particular votes within its reporting scope. Compare the policy in force, the exact ballot item and the shares actually voted before judging consistency. Keep abstentions, split votes and shares loaned without recall distinct. An absence of votes is not automatically a vote against management.
Establish whose votes are covered
The SEC Form N-PX instructions distinguish registered management investment companies’ proxy-voting reporting from institutional investment managers’ reporting of specified executive-compensation votes under Exchange Act Sections 14A(a) and (b). Do not assume every manager reports every ballot item. Check filer type, reporting period, covered series or accounts and the relevant reporting arrangement before treating missing records as missing votes.
Use the form as the primary reference for the fields and scope, not as proof of a particular organization’s policy compliance. The following figures and policy are hypothetical. This is an evidence comparison, not a legal conclusion about a filer or a recommendation on how investors should vote.
Reconstruct a ballot before assigning a label
Assume an invented fund has 100 share units in a simplified economic-position population for one item. It reports 60 voted for, 20 against and 10 abstaining. Another 10 were loaned and not recalled for voting. Assume equal voting rights for the voted shares and a complete record for this example. The 100-unit position is not a claim that all 100 units retained voting entitlement.
| Status | Share units | Treatment in the worksheet |
|---|---|---|
| For | 60 | Cast in favor |
| Against | 20 | Cast against |
| Abstain | 10 | Distinct voting instruction |
| Loaned, not recalled | 10 | Not a cast against vote |
The shares voted sum to 90. Support as a fraction of shares voted is 60 / 90, or about 66.7%, if the chosen definition includes abstentions in that denominator. Support as a fraction of the assumed 100-unit economic position is 60%, a different comparison rather than a voting-eligibility rate. Neither calculation establishes the proposal’s official passage result, whose rules and total voting population may be different.
Match the policy version to the decision
Suppose the fictional policy generally opposes a compensation arrangement unless a specified performance safeguard exists. Research whether the safeguard was present for this exact ballot, and retain the policy version applicable on the vote date. A later policy revision cannot automatically explain an earlier instruction. A broad marketing statement may be less specific than the operative voting guidelines.
A vote for the proposal can be consistent with the exception, inconsistent with the rule, or unresolved because the exception evidence is missing. Do not choose among those possibilities from the direction alone. Save the stated rationale and separate what the organization said from what the available documents independently support.
Investigate split votes without erasing them
Different series, accounts or decision mandates may produce different instructions. Aggregate support can conceal that distribution. Preserve the reporting entities and share amounts attached to each direction, then test whether the aggregate reproduces the underlying records. Repeated records across reporting arrangements also require deduplication before summing.
Loaned shares raise a separate question about recall decisions and policy, not merely vote direction. An uncast share may reflect lending, eligibility or other circumstances. Determine what the actual record supports before interpreting intent. Do not assume all economic holdings had identical voting rights or were eligible at the same record date.
Report consistency at the right level
Build a table linking policy clause, ballot facts, vote direction, share count and rationale. Use working outcomes such as consistent with documented exception, apparent inconsistency or insufficient evidence. These labels should remain specific to the tested item and should not become a blanket score for every action taken by the manager.
If comparing years, preserve changes in covered accounts, proposal categories and reporting scope. A shift in aggregate support can come from ballot composition rather than a changed philosophy. The final research judgment should explain both the observed action and the boundary of the evidence used to interpret it.
FAQ
Does no recorded vote mean a vote against?
No. Establish reporting scope, eligibility and the reason for any uncast shares before interpreting the absence.
Does every institutional manager report every proposal on N-PX?
No. Filer type and the applicable reporting requirements matter; do not generalize across those scopes.