Do not apply a defined-benefit funding indicator to every pension scheme
Establish scheme-specific metric applicability before comparing occupational-pension risk tables.
The same pension label can hide different meanings
Defined-benefit and defined-contribution arrangements can require different risk explanations. A funding indicator from one scheme should not automatically carry the same meaning for another's members. Truthound can check a supplied scheme/metric relationship without making the financial judgment.
The October dashboard's distinction
EIOPA published its occupational-pension risk dashboard on October 29, 2026. It distinguishes DB and DC schemes while combining quarterly and annual reporting with market information. Improvement in a particular DB indicator should not become a statement about every scheme.
Create an applicability register first
Link scheme type, metric definition, applicability and reference period. An institution with several arrangements may need more than its institution identifier as a key. Preserve which scheme a value describes before calculating an average or comparison.
Inapplicable is not missing
Inapplicable and unobserved need different states. An absent DB-only indicator in a DC row should not become 0 or automatic failure. Conversely, missing data for an applicable indicator should not be labelled not applicable.
The allowed pairs in Truthound
Truthound Core Python SDK 3.3.4, operating on a prepared Polars table, provides ColumnPairInSetValidator for supplied allowed pairs and InSetValidator for agreed categories. The responsible owner establishes which scheme/metric associations are permitted from official definitions. The validator does not infer that policy.
Keep the existing rule when appropriate
An existing Pandera 0.35.0 pandas predicate can retain the mapping. Compare an out-of-scope metric, unresolved scheme type and applicable row without data. General validation features do not establish equivalence before boundaries and null policy are aligned.
Show who entered the comparison
Show which populations enter each group comparison. Combining DB funding and a different DC risk into one score requires an explicit separate definition and justification. Identical table shapes do not establish identical economic burdens.
Preserve the different observation dates
Quarterly reporting, annual reporting and market cutoffs also remain distinct. Appearing in the same publication does not make every indicator an observation from one date. Handle differently refreshed metrics according to the actual comparison purpose.
Make the explanation accessible
The final explanation should identify scheme, metric and unresolved or inapplicable scope. That makes it harder to pass an unsuitable number to another scheme's users. Visible applicability is the starting point for an accessible risk explanation.
Conditions
An owner supplies applicability after checking the official scheme and indicator definitions.
Alternatives
Retain existing Pandera 0.35.0 pandas predicates for the scheme-specific applicability mapping.
Limitations
Table checks do not determine pension solvency or an individual's financial choices.
Evidence
- Event checked
- Subject checked
- Modified
