Key implication: Mortgage growth is no strategy if unit economics deteriorate. ANZ’s profit decline, despite expanding its loan book, underscores a broader reality for lenders in 2024–25: scale must be matched by pricing power, capital efficiency and operational productivity. Otherwise, volume simply amplifies cost and capital consumption.

Market context: heat at the edges, moderation ahead

ANZ has leaned into demand pockets, with just under a fifth of its new mortgages written in Western Australia and a further 8% in South Australia and the Northern Territory, according to sector reporting. That geographic skew tracks today’s hottest housing markets—Perth remains exceptionally tight—with faster deal velocity and strong broker activity. Yet ANZ Research expects capital city prices to rise 6–7% in 2024, slowing to 5–6% in 2025 as population growth eases and supply lifts. In short: momentum persists, but the peak tailwind is easing.

This timing matters for banks whose profitability relies on net interest margin (NIM) stabilisation and credit quality. As volumes grow into a cooling price cycle, lenders must resist the temptation to buy market share through deeper rate discounting, which can compress margins precisely when funding costs and regulatory overheads remain elevated.

The unit economics problem: NIM, costs and capital

Why can a bigger mortgage book coincide with weaker profits? Three forces dominate the P&L:

• Margin compression: Deposit competition has intensified, raising the cost of sticky funding, while wholesale funding remains structurally higher than the pre-pandemic era. Aggressive discounting in the broker channel further erodes front-book margins.
• Operating and compliance costs: Ongoing regulatory settlements, remediation and restructuring charges hit the income statement and distract management focus. Compliance isn’t cyclical; it’s a permanent cost of doing business in Australian banking.
• Capital intensity: Mortgages consume risk-weighted assets under APRA’s framework, tying up capital that could be deployed into higher-return segments. ANZ itself notes that, despite recent home-lending growth, it still has more capital invested in New Zealand businesses and agriculture than in home loans—an implicit signal on relative returns.

Add it up, and the return-on-equity of incremental mortgages can fall below hurdle rates if lenders compete primarily on price and speed. Volume, absent productivity and pricing discipline, becomes self-defeating.

Competition: broker-led battleground and the speed premium

Australia’s mortgage broking market remains fiercely competitive, with lenders fighting for visibility and shelf space in aggregator panels. The market commentary is consistent: the broking industry has intensified, with rapid sales cycles in hotspots like Perth compressing decision windows. Cashbacks may have faded, but sharp pricing and fast approvals still win mandates—often at the expense of NIM.

This dynamic rewards lenders with superior workflow orchestration: instant verification of income and identity, automated valuation models, and streamlined credit decisioning that preserve margin by lowering cost-to-serve rather than relying on price undercutting. Those still stitching together legacy systems will continue to pay through the front-book.

Execution reality: an AI-enabled cost reset—done safely

Productivity is the lever management can still pull. The question is how, and how quickly. Generative and predictive AI can lift throughput across underwriting, fraud detection, collections, and customer service. The Australian Taxation Office’s AI governance approach—articulated in public documentation—offers a blueprint: define model risk tiers, establish clear accountabilities, enforce auditability and human-in-the-loop controls. That matters because Australia’s AI Ethics Principles set expectations for AI that is safe, secure and reliable, especially in credit decisioning where bias and explainability risks are acute.

Practical priorities for banks include: straight-through processing for low-risk applications, broker onboarding with real-time compliance checks, dynamic credit-policy engines tuned by outcomes data, and AI-assisted customer remediation. These initiatives can take 12–24 months to materially impact the cost-to-income ratio. Boards should demand a technology roadmap that pairs near-term automation savings with medium-term platform simplification—retiring duplicative systems that inflate run costs.

Geographic mix and risk posture: WA strength, cyclical sensitivity

Heavier exposure to Western Australia brings upside from robust labour markets and constrained housing supply, but it also concentrates cyclicality tied to commodities. Lenders should pressure-test portfolios for a two-speed economy scenario: if mining-related incomes soften, what happens to serviceability and arrears in WA-heavy cohorts? The reality is that Australian borrowers have borne significant rate increases; serviceability buffers and prudential settings have helped, but variable-rate sensitivity remains a watchpoint as fixed-rate roll-offs wash through.

Portfolio hygiene therefore matters: dynamic LVR monitoring, granular segmentation by borrower type and region, and early-warning triggers that couple transaction data with employment indicators. Growth shouldn’t outpace the institution’s ability to observe and act on emerging risk.

Outlook: fewer quick wins, more grind

Market strategists have cautioned that bank earnings face a transition year: fewer one-offs to the upside, more grind on costs. ANZ does not publish quarterly profit figures, limiting short-term visibility, but the medium-term drivers are clear. Base case: housing activity stays resilient while price growth moderates; NIM stabilises as deposit repricing peaks; credit losses remain contained. Downside: stickier funding costs and a turn in arrears clip earnings. Upside: a decisive cost reset, re-pricing discipline and capital rotation into higher-return assets.

For now, growth at any price is out. Growth at the right price—underpinned by productivity and governance—still creates value.

Strategic playbook for leaders

For banks and non-bank lenders:

• Re-price with intent: Codify minimum economic thresholds for front-book lending and enforce pricing exceptions tightly in the broker channel.
• Accelerate cost take-out: Target straight-through processing on defined segments within 12 months; link executive incentives to cost-to-income milestones.
• Govern AI like a financial model: Adopt model risk policies aligned to Australia’s AI Ethics Principles; validate fairness, explainability and stability before scaling.
• Optimise capital: Recycle capital from sub-economic mortgage cohorts into higher-ROE business lending where risk-adjusted returns justify allocation.
• Manage geographic concentration: Set exposure limits for WA and other fast-growing regions, backed by stress scenarios tied to commodity cycles.

For corporates beyond banking, the lesson travels well: volume rarely fixes a margin problem; regulation is a durable cost; and AI, without governance and a platform simplification agenda, won’t deliver the structural productivity lift boards are banking on.