The degenerate-input contract defines how all EventStudy return models behave when the estimation data is degenerate. Two modes are supported:
Lenient (default): The model sets is_fitted = FALSE,
emits exactly one warning() per (event_id, firm_symbol)
per fit call, and propagates NA through all abnormal returns
and downstream statistics. No event is silently dropped; zeros are
never substituted for NA.
Strict: The model raises a descriptive stop() error
naming the component, event_id, and firm_symbol, plus
the specific reason for the degeneracy.
Configuration:
Via
ParameterSet:ParameterSet$new(degenerate_handling = "strict")Via package option:
options(EventStudy.degenerate_handling = "strict")
The ParameterSet field takes precedence over the package option; if
neither is set, the default "lenient" mode is used.
Degenerate conditions covered:
Fewer than 2 finite observations in the estimation window (insufficient observations for OLS).
Zero or near-zero variance in index returns (
sd < .Machine$double.eps), making OLS estimation undefined.Single-event group (relevant for multi-event statistics; applied in Phase 2).
NApropagation from upstream pipeline steps.
NA propagation semantics:
When is_fitted = FALSE, model$abnormal_returns() returns
a tibble with abnormal_returns = NA_real_ for all rows. All
downstream test statistics that depend on fitted models then receive
NA inputs and propagate NA to their outputs. The event
is retained in the output tibble – it is never silently dropped.
Multi-event exclusion (2026-09-24 re-evaluation, item A2):
A single-event's NA propagation above is not enough for
cumulative (CAR-based) multi-event statistics (CSectT, PatellZ, Sign,
GeneralizedSign, BMP, KolariPynnonen): an event with ZERO finite
event-window abnormal returns must be EXCLUDED from every such
statistic's group, not merely NA-propagated, because a cumulative sum
that coalesce()s a fully-degenerate event's abnormal returns to
0 would silently understate the true cross-sectional dispersion. The
exclusion is applied once per group (see .exclude_all_na_events)
before any statistic is computed, and reported exactly once: silently
if the excluded event's model already emitted the one contract warning
guaranteed by CONTRACT-04 (an unfitted ModelBase instance), or
via a single additional warning naming the excluded event_id(s)
otherwise. Strict mode errors instead. This exclusion rule leaves the
pre-existing STATS-03 convention for a PARTIAL gap unchanged: an event
with at least one finite event-window abnormal return, but a missing
value on some individual day, still contributes 0 to that event's CAR
on the missing day(s) via coalesce(abnormal_returns, 0).