Firms ›
Halifax Limited
Group: Lloyds Banking Group
Beyond banking
The figures above cover this firm’s banking, credit and mortgage decisions only. The Ombudsman also decided complaints about it in other sectors, which behave differently and are kept separate rather than averaged in.
| Sector | Decisions | Uphold rate |
|---|---|---|
| Banking, credit and mortgages — shown above | 818 | 17.4% |
| Payment protection insurance | 74 | 29.7% |
| Investments and pensions | 48 | 12.5% |
| Insurance (excluding PPI) | 9 | — |
What this firm loses on
| Issue | Decisions | Upheld | Uphold rate |
|---|---|---|---|
| Mortgage administration | 175 | 36 | 20.6% |
| Complaint handling | 119 | 29 | 24.4% |
| Credit file reporting | 114 | 28 | 24.6% |
| Charges and fees | 110 | 23 | 20.9% |
| Service failure | 78 | 20 | 25.6% |
| APP scam refund refused | 71 | 8 | 11.3% |
| Treatment in financial difficulty | 71 | 13 | 18.3% |
| Product mis-sold | 62 | 5 | 8.1% |
| Card dispute or chargeback | 61 | 9 | 14.8% |
| Account closed or refused | 46 | 6 | 13.0% |
| Unauthorised transaction | 46 | 4 | 8.7% |
| Interest and rates | 36 | 5 | 13.9% |
| Irresponsible lending | 36 | 14 | 38.9% |
| Support for vulnerable customers | 31 | 8 | 25.8% |
| Payment delayed or failed | 29 | 7 | 24.1% |
| Data and privacy | 28 | 4 | 14.3% |
| Funds blocked or frozen | 27 | 7 | 25.9% |
| Identity checks and access | 13 | 4 | 30.8% |
Context
Current account switching
Net customers gained or lost through the switching service.
| Quarter | Gains | Losses | Net |
|---|---|---|---|
| Q1-2026 | 2,669 | 24,538 | -21,869 |
| Q4-2025 | 2,607 | 28,236 | -25,629 |
| Q2-2025 | 3,669 | 18,421 | -14,752 |
| Q1-2025 | 4,026 | 19,733 | -15,707 |
Open Banking API availability
Mean of the unweighted figure across endpoint categories, as published by Open Banking. Unweighted takes the worst endpoint per day, so it reflects what a customer hits rather than a volume-weighted average that healthy endpoints can carry.
2026-07: 100.00%
Decisions are attributed to a firm only where the decision text names it unambiguously. Where a name could refer to more than one firm, the decision is left unattributed rather than guessed, so these counts are a floor, not a total. Method.