Bank Financial Model: How to Model a Commercial Bank Step by Step
Why Banks Are Different
A standard three-statement model doesn’t work for banks. Here’s why:
Debt is the raw material. For a manufacturer, debt is financing. For a bank, deposits and borrowed funds are the product — the bank takes in deposits at low rates and lends at higher rates. You can’t separate “operating” from “financing” the way you do for a non-financial company.
Working capital doesn’t apply. Banks don’t have inventory or accounts receivable in the traditional sense. Their “inventory” is money itself. Instead of working capital, you model loan growth, deposit growth, and the composition of the balance sheet.
Leverage is the business model. A typical bank operates at 10–12x leverage (assets/equity). Regulatory capital ratios — CET1, Tier 1, Total Capital — replace the debt/EBITDA covenants you’d see in a corporate model.
Key Components of a Bank Model
| Component | What It Covers | Corporate Equivalent |
|---|---|---|
| Net Interest Income (NII) | Interest earned on loans minus interest paid on deposits | Revenue / gross profit |
| Net Interest Margin (NIM) | NII ÷ average earning assets | Gross margin % |
| Non-Interest Income | Fees, trading revenue, wealth management, service charges | Other revenue streams |
| Provision for Credit Losses (PCL) | Expected losses on loans | No direct equivalent (closest: warranty reserve) |
| Non-Interest Expense (NIE) | Salaries, technology, branches, compliance | SG&A / operating expenses |
| Efficiency Ratio | NIE ÷ Total Revenue | Operating margin (inverse logic) |
| Capital Ratios | CET1, Tier 1, Total Capital ratios | Debt/EBITDA leverage covenants |
Step-by-Step: Building a Bank Financial Model
Step 1 — Model the Balance Sheet First
Unlike corporate models where you start with the income statement, bank models start with the balance sheet. Project total loans (by category: commercial, residential, consumer, CRE), investment securities, and total earning assets on the asset side. Project deposits (by type: demand, savings, time), borrowed funds, and equity on the liability side.
Step 2 — Calculate Net Interest Income
NII = Interest Income − Interest Expense. Model it by applying average yields to each asset category and average costs to each liability category. The net interest margin (NIM) = NII ÷ Average Earning Assets. NIM typically ranges from 2.5%–4.0% for US commercial banks.
Step 3 — Forecast Non-Interest Income
Non-interest income includes service charges, wealth management fees, mortgage banking revenue, trading gains, and insurance income. Model each line separately — fee income as a % of AUM or deposits, trading revenue based on market conditions, mortgage revenue based on origination volume.
Step 4 — Model the Provision for Credit Losses
This is the most judgment-intensive line. The provision reflects expected losses on the loan portfolio. Model it as a percentage of average loans — typically 0.2%–0.5% in good times, 1%–3%+ in recessions. Under CECL (Current Expected Credit Losses), banks must provision for lifetime expected losses upfront.
The provision flows through the income statement, building the allowance for loan losses on the balance sheet. Net charge-offs (actual losses) reduce the allowance. Model the allowance rollforward: Beginning + Provision − Net Charge-Offs = Ending.
Step 5 — Forecast Non-Interest Expense
Compensation is typically 50–60% of non-interest expense. Model it based on headcount × average comp, or as a percentage of total revenue. Technology and occupancy costs are semi-fixed. The efficiency ratio (NIE ÷ Revenue) measures cost discipline — top banks run at 50–55%, while struggling banks may exceed 70%.
Step 6 — Calculate Pre-Tax Income and Net Income
Total Revenue (NII + Non-Interest Income) − Provision − Non-Interest Expense = Pre-Tax Income. Apply the effective tax rate to get net income. Then calculate EPS and return on equity (ROE).
Step 7 — Model Capital Ratios
Regulatory capital is the binding constraint for banks. Model CET1 Capital, Tier 1 Capital, and Total Capital. Calculate risk-weighted assets (RWA) by applying risk weights to each asset category. The key ratios:
| Ratio | Formula | Minimum Requirement | Well-Capitalized |
|---|---|---|---|
| CET1 Ratio | CET1 Capital ÷ RWA | 4.5% | 6.5%+ |
| Tier 1 Ratio | Tier 1 Capital ÷ RWA | 6.0% | 8.0%+ |
| Total Capital Ratio | Total Capital ÷ RWA | 8.0% | 10.0%+ |
| Leverage Ratio | Tier 1 Capital ÷ Total Assets | 4.0% | 5.0%+ |
Bank Valuation Methods
| Method | Price/Book Value | Dividend Discount Model |
|---|---|---|
| Metric | P/BV or P/TBV | PV of future dividends |
| Key driver | ROE relative to cost of equity | Sustainable dividend payout |
| When to use | Quick relative valuation | Intrinsic value estimation |
| Rule of thumb | P/BV > 1.0x if ROE > COE | Higher dividends = higher value (if sustainable) |
| Limitation | Book value quality varies | Dividend policy may not reflect earnings power |
Note: Standard DCF models using free cash flow to firm don’t work well for banks because you can’t cleanly separate operating and financing cash flows. Use the dividend discount model or residual income model instead.
Interest Rate Sensitivity
Banks are uniquely exposed to interest rate changes. When rates rise, both asset yields and funding costs increase — but not at the same pace. Most banks are asset-sensitive (benefit from rising rates) because loan yields reprice faster than deposit costs. Model this by projecting different NIM scenarios under various rate environments using sensitivity analysis.
Common Bank Modeling Mistakes
Using a corporate DCF. Free cash flow to firm doesn’t work for banks. Use a dividend discount model, excess return model, or P/BV approach instead.
Ignoring credit quality. The provision for credit losses can swing from 0.2% to 3%+ of loans — a massive earnings impact. Your model must capture the credit cycle, not just assume stable provisions.
Overlooking capital constraints. A bank can’t grow loans 15% if its CET1 ratio is already at the minimum. Capital ratios constrain growth — model them explicitly and flag when projected ratios breach regulatory minimums.
Static NIM assumptions. NIM changes with interest rates, competition, loan mix, and deposit behavior. Model NIM dynamically based on the rate environment, not as a flat assumption.
Key Takeaways
- Bank models start with the balance sheet (loans, deposits) and derive income from the spread — the opposite of corporate models.
- Net interest income (NII) and net interest margin (NIM) are the core revenue drivers, supplemented by fee and trading income.
- The provision for credit losses is the most volatile and judgment-intensive line — model it carefully with cycle awareness.
- Capital ratios (CET1, Tier 1) are the binding constraint on growth — always include them and check for regulatory compliance.
- Value banks using P/BV, dividend discount models, or residual income — not standard DCF approaches.
Frequently Asked Questions
Why can’t I use a standard DCF for banks?
A standard DCF calculates free cash flow by separating operating and financing activities. For banks, debt (deposits, borrowings) IS the operating activity — you can’t separate the two. Using FCFF would require subtracting all deposit inflows and adding back all loan disbursements, which doesn’t produce meaningful results. Instead, use equity-based approaches: dividend discount model or residual income model.
What is the efficiency ratio and what’s a good target?
The efficiency ratio = non-interest expense ÷ total revenue (NII + non-interest income). Lower is better — it means the bank spends less to generate each dollar of revenue. Top US banks operate at 50–55%. Community banks often run 60–70%. Anything above 70% signals cost management issues. Think of it as the inverse of an operating margin.
How do interest rate changes affect a bank model?
Rising rates generally help asset-sensitive banks because loan yields reprice faster than deposit costs, expanding NIM. However, higher rates also increase credit risk (borrowers struggle with payments) and may slow loan growth. Model both the NIM benefit and the credit cost of rate changes using scenario analysis.
What is CECL and how does it affect bank modeling?
CECL (Current Expected Credit Losses) is the US accounting standard requiring banks to provision for lifetime expected credit losses at loan origination, rather than waiting until losses are “probable.” This front-loads provisions and creates larger reserves. In your model, the CECL provision depends on loan growth, portfolio composition, and macroeconomic forecasts.
How do I model loan growth for a bank?
Project loan growth by category (commercial & industrial, commercial real estate, residential mortgage, consumer). Use management guidance, historical growth rates, and industry benchmarks. Cross-check against capital constraints — loan growth requires capital, so if the CET1 ratio is tight, the bank can’t grow loans aggressively without raising capital or retaining more earnings.