PBT collapsed 45.3% as one-offs and provisions swamped 11.1% revenue growth

Management-disclosed underlying NPAT of $46.9m, which strips out acquisition-related regulatory assurance costs, one-off staff exit costs, and hedge de-designation impacts, represents the more relevant measure of run-rate performance, and this figure met the stated guidance floor of at least $45m.

The statutory collapse was concentrated in the first half: 1H2025 NPAT was just $3.6m, with the implied second half recovering strongly to $35.2m. This skew reflects the deliberate reset of selected New Zealand lending portfolios in 1H2025, which depressed statutory earnings before a second-half recovery.

Total assets contracted 6.9% to $8.6b, with equity marginally lower at $1.2b. The gross borrowings comparison ($825m versus $8b in FY24) reflects a change in classification methodology — FY24 included total deposits and other borrowings — rather than a genuine deleveraging event.

The $8.1m difference between underlying NPAT ($46.9m) and statutory NPAT ($38.8m) is explained by disclosed one-offs: acquisition costs tied to Heartland Bank Australia, staff exit costs, and hedge-related fair-value items. The effective tax rate rose from 28.7% to 32.2%, adding a modest additional drag on top of the PBT decline, but this is secondary to the operating-level impairment and cost pressures.

The NZ portfolio reset creates an unresolved credit-quality question. The NZ banking segment contributed $21.9m versus Australian banking's $27.2m on NIM of 3.87% and 3.01% respectively, but the prior-period comparatives are unavailable in segment form, making it impossible to assess whether the portfolio reset has fully normalised or whether arrears and provision levels remain elevated. This matters because if credit costs remain structurally higher, the second-half recovery will not sustain.

ROE has more than halved to 3.2%. Against equity of $1.2b, a 3.2% ROE (versus 6.0% in FY24) signals the Australian acquisition has yet to generate a return commensurate with the capital deployed. The payout ratio against statutory NPAT reached 96.6% for FY25 versus 30.5% in FY24, reflecting that the full-year dividend of NZ$0.04 per share consumes nearly all reported earnings at current statutory profit levels — though this is sustainable at the underlying earnings run-rate.

The result met that floor at $46.9m underlying NPAT. The strongly second-half-weighted statutory earnings profile — with $35.2m of NPAT implied in 2H2025 against just $3.6m in 1H2025 — suggests the business normalised materially following the portfolio reset, which is modestly supportive of the FY2026 guidance basis.

Whether the second-half run-rate is a reliable base depends on whether NZ credit quality has genuinely stabilised and whether the Australian banking segment can continue contributing at or above its FY2025 level. The absence of stated targets beyond the underlying NPAT floor limits the ability to assess how quickly ROE can recover toward a level that justifies the capital raised for the Australian acquisition.

The 11.1% revenue growth is a genuine positive: net operating income expansion of this magnitude in a challenging NZ credit environment indicates the business is growing its loan book and maintaining spread. However, with no segment-level prior comparatives disclosed, it is not possible to decompose how much of the improvement in 2H2025 came from reduced impairment charges versus volume and margin recovery.

Operating cash flows of $673.2m represent a meaningful step down from FY24's $928.4m, which is consistent with the lower statutory profitability and the portfolio dynamics of a finance company where loan origination and repayment flows directly affect operating cash. Capex fell sharply to $4.4m from $28.1m — the FY24 figure included acquisition-related expenditure — so the reduction is structural rather than a sign of under-investment.

This briefing cannot assess the credit quality trajectory of the NZ portfolio, the sustainability of the Australian segment contribution, or the pace of ROE recovery, as segment-level prior comparatives and arrears data were not available in the supplied materials.