Vulnerable customers (FG21/1)

FG21/1 (“Guidance for firms on the fair treatment of vulnerable customers”, February 2021) sets the FCA’s expectations on identifying vulnerability and ensuring outcome equivalence for vulnerable customers. For principal firms supervising AR networks, the guidance is operationalised at the AR level (where the customer interaction happens) but evidenced at the principal level (where the regulated firm is accountable).

This page is about the roll-up problem the product solves: turning AR-level vulnerability data into principal-level MI that defends the FG21/1 position.

The four drivers of vulnerability

FG21/1 sets four drivers. A customer can be vulnerable on one driver, several drivers, or transiently. Vulnerability is dynamic, not a label.

Driver	Examples
Health	Physical or mental health conditions affecting capacity, mobility, or stress tolerance
Life events	Bereavement, divorce, redundancy, sudden caring responsibilities
Resilience	Low income, low savings, unstable employment, problem debt
Capability	Limited financial literacy, language barriers, digital exclusion

The file-review rubric across MCOB, ICOBS, and CONC includes vulnerability-identification items mapped to these drivers. The rubric does not require an exhaustive intake interrogation; it requires evidence of attentive practice.

The outcome-equivalence expectation

FG21/1 expects firms to evidence outcome equivalence for vulnerable customers. Vulnerable customers should achieve outcomes as good as those achieved by non-vulnerable customers in comparable circumstances. The expectation is statistical, not anecdotal: firms should be able to demonstrate that across their book, vulnerable cohorts are not systematically receiving worse outcomes.

Outcome equivalence is the principal’s defence position under FG21/1. The product is built around making that defence reconstructible from AR-level evidence.

The roll-up problem

The principal does not see most customer interactions directly. ARs do. AR-level evidence has to be aggregated to the principal level to satisfy the outcome-equivalence test. Four indicators feed the roll-up:

1. Identification rate. What proportion of customers each AR identified as potentially vulnerable. Captured per file review and per complaint.
2. Recorded adjustments. Whether the AR recorded specific accommodations made for vulnerable customers (extra time, alternative communication channel, escalation to specialist staff).
3. Complaint outcomes for vulnerable cohorts. Upheld rate, redress level, time-to-final-response, broken down by whether the customer was flagged as vulnerable.
4. Product-fit indicators. Whether the products distributed to vulnerable customers were appropriate for their circumstances; flagged via file-review findings and complaint root cause.

These four indicators aggregate into a vulnerable-customer MI tile on the principal home page.

Variance between ARs as a risk-score input

Identification rate is meaningful only relative to a baseline. The product’s interpretation: an AR with very low identification rate relative to its peers in the same vertical is more likely to be under-recording than to be serving a low-vulnerability customer base.

Variance between ARs on identification rate is therefore itself a risk-score input. The composite risk score weights “identification rate divergence from cohort mean” as one of the components in the file-review-derived signal. An AR identifying vulnerability at less than half the network mean is flagged for additional file-review sampling.

The same logic applies to recorded adjustments. An AR with high vulnerability identification but very low recorded adjustments suggests practice that recognises vulnerability but does not act on it.

What the AR-side surface captures

On the AR-user side, the product captures vulnerability indicators through:

• The file-review process: when the principal samples the AR’s cases, the rubric prompts the reviewer to assess whether vulnerability was identified and how it was handled.
• The complaint object: at intake, the AR flags whether the customer was identified as potentially vulnerable, on which driver, and what adjustments were made.
• The MI return: aggregate counts of vulnerability identification and adjustments by quarter.

The principal-level MI tile

The vulnerable-customer MI tile on the principal home page renders:

• Network-wide identification rate (rolling 12 months) with trend.
• Per-AR identification rate, with cohort-mean comparison.
• Recorded adjustment ratio (adjustments per identified case) per AR.
• Complaint outcome equivalence: upheld-rate ratio between vulnerable and non-vulnerable cohorts, with confidence interval.
• Outliers list: ARs flagged on any of the four indicators.

Cross-links

• Consumer Duty for the four Consumer Duty outcomes that interlock with vulnerability evidence.
• DISP 1 for the complaint object’s vulnerable-customer outcome breakdown.
• Annual self-assessment for the per-AR vulnerable-customer attestation in the annual review packet.