Discounted Earnings Model

Type: Intrinsic Value Method

Overview

The Discounted Earnings Model estimates what a bank should be worth by looking at how much money it is expected to earn in the future. Each year's projected earnings are adjusted to reflect the fact that a dollar earned five years from now is worth less than a dollar earned today. Adding up all those adjusted future earnings gives you the bank's estimated intrinsic value.

The model projects a bank's net income over a period of 5 to 10 years, then adds a terminal value that captures all earnings beyond the projection period. Each year's projected earnings are discounted by the cost of equity, which represents the return that investors require for holding the stock.

For banks, this earnings-based approach is preferred over the standard discounted cash flow (DCF) analysis that works for most companies. Free cash flow is not a meaningful concept for financial institutions. Banks don't have capital expenditures in the traditional sense, and the line between their operating activities and financing activities doesn't apply cleanly since taking deposits is both a funding source and a core business activity. Earnings, after adjusting for credit loss provisions and capital retention needs, are the right measure of what a bank produces for its shareholders.

The Discounted Earnings Model is most valuable when a bank's current earnings don't reflect its long-term earning power. If provisions are temporarily elevated, if a bank is mid-turnaround, or if strategic changes are reshaping the earnings profile, point-in-time metrics like P/E ratios can be misleading. Projecting earnings forward and discounting them back gives a more complete picture of value.

Formula

Intrinsic Value = Sum of [Projected Earnings_t / (1 + r)^t] + Terminal Value / (1 + r)^n

For each projected year t, estimate net income available to common shareholders. This is the bank's bottom-line profit after all expenses, provisions for credit losses, and taxes.

Discount each year's earnings by the cost of equity (r) raised to the power of the corresponding year. The cost of equity is typically 10% to 13% for banks, depending on size and risk profile. At a 12% cost of equity, $100 of earnings five years out is worth roughly $56.70 today.

The terminal value captures all earnings beyond the explicit projection period. It is calculated using either a terminal price-to-earnings (P/E) multiple applied to the final projected year's earnings, or the Gordon Growth formula: Terminal Value = Final Year Earnings x (1 + g) / (r - g), where g is the long-term sustainable growth rate (typically 2% to 4% for banks).

Sum the present values of all projected earnings and the discounted terminal value, then divide by shares outstanding to get intrinsic value per share.

How to Apply

  1. Establish a base earnings level by normalizing the bank's most recent annual earnings. Adjust for unusual items such as outsized provision releases or charges, securities gains or losses, legal settlements, and restructuring costs. Pre-provision net revenue (PPNR) is a useful starting point, to which a normalized provision for credit losses and tax rate are applied. For example, if a bank reported $500 million in net income but benefited from a $75 million provision release that inflated earnings above sustainable levels, the normalized starting point might be closer to $425 million.
  2. Project earnings for 5 to 10 years. The key assumptions driving this projection are: - Loan growth rate, typically tied to nominal GDP growth (4% to 5%) or the bank's historical growth trajectory - Net interest margin (NIM) trajectory, considering the expected direction of interest rates and the bank's asset-liability positioning - Fee income growth, based on the bank's business mix and strategic initiatives - Expense growth, incorporating expected efficiency improvements or investment spending - Provision normalization, returning to a through-cycle average provision-to-loan ratio (typically 0.25% to 0.50% for well-run banks) Each assumption should be grounded in the bank's historical performance, management guidance, and macroeconomic expectations. Avoid simply extrapolating a single year's results forward.
  3. Apply capital constraints to growth projections. A bank cannot grow assets faster than its capital base allows without raising external equity. The formula for maximum sustainable asset growth is: ROE x Retention Ratio / Equity-to-Assets Ratio. For a bank with 12% ROE (return on equity), a 60% retention ratio, and a 9% equity-to-assets ratio, sustainable asset growth caps out at roughly 8%. If your projected loan growth exceeds this rate, the model needs to account for one of three outcomes: the bank raises new equity (which dilutes per-share value), it reduces its dividend payout to retain more capital, or it allows its capital ratios to decline toward minimum regulatory requirements.
  4. Select an appropriate cost of equity for discounting. For most banks, 10% to 13% is a reasonable range. Factors that push toward the higher end include smaller size, weaker asset quality, more volatile earnings, geographic concentration, and higher commercial real estate exposure. The Capital Asset Pricing Model (CAPM) provides a structured estimate: Cost of Equity = Risk-Free Rate + Beta x Equity Risk Premium. Bank betas typically range from 0.8 to 1.3. Alternatively, a build-up method starts with the risk-free rate and adds premiums for equity risk, size, and bank-specific factors. The cost of equity has a large impact on the final valuation, so testing a range of values is more useful than picking a single point estimate.
  5. Calculate terminal value using either a terminal P/E multiple or the Gordon Growth formula. For the multiple approach, apply a P/E ratio of 8x to 14x to the final projected year's normalized earnings. Higher-quality banks with above-average ROE, clean asset quality, and strong franchises warrant multiples toward the upper end. For the Gordon Growth approach, use: Terminal Value = Final Year Earnings x (1 + g) / (r - g), where g is the long-term sustainable growth rate, typically 2% to 4% for banks. Discount the terminal value and all projected earnings to present value. Sum the components and divide by shares outstanding to arrive at intrinsic value per share. Because the terminal value often accounts for 50% to 70% of total intrinsic value, the assumptions behind it deserve careful scrutiny.

Example Calculation

A bank earned $500 million in the most recent year. After normalizing for an unusually low provision (the bank released $50 million from reserves), normalized earnings are estimated at $450 million. Earnings are projected to grow at 5% annually for 5 years, producing projected earnings of $450M, $472.5M, $496.1M, $520.9M, and $547.0M.

A terminal P/E of 10x is applied to Year 5 earnings: Terminal Value = $547M x 10 = $5.47 billion.

Using a 12% cost of equity, the present value of the 5 years of earnings is approximately $1.74 billion, and the present value of the terminal value is approximately $3.10 billion. Total intrinsic value = $1.74B + $3.10B = $4.84 billion. With 200 million shares outstanding, intrinsic value per share is approximately $24.20.

Notice that the terminal value ($3.10 billion) represents about 64% of the total intrinsic value. This is typical for discounted earnings models and illustrates why the terminal value assumptions matter so much. If the terminal P/E were 12x instead of 10x, the terminal value would jump to $6.56 billion, raising intrinsic value per share to roughly $28.40. That is a 17% increase from changing just one assumption.

Strengths

  • Provides a forward-looking valuation that explicitly models the bank's future earnings trajectory. Unlike P/E or P/B ratios that reflect a snapshot of current conditions, the Discounted Earnings Model incorporates expected changes in net interest margin, credit costs, efficiency, and growth over a multi-year horizon.
  • Allows explicit modeling of credit cycle dynamics. Earnings projections can incorporate a normalization of provisions from current levels to through-cycle averages, capturing value that point-in-time metrics may miss. A bank with temporarily elevated provisions and depressed current earnings can be modeled toward its normalized earning power, revealing value hidden by the current cycle position.
  • Adaptable to different scenarios. Multiple earnings paths can be modeled (base case, optimistic, pessimistic) to produce a range of intrinsic values. This scenario analysis helps investors understand how sensitive the valuation is to assumptions about credit costs, rate movements, and loan growth rather than relying on a single estimate.
  • Incorporates capital constraints naturally. The projection of earnings growth can be bounded by the bank's ability to generate capital internally through retained earnings, reflecting the regulatory reality that capital adequacy requirements limit how fast a bank can grow without raising new equity.

Limitations

  • Highly sensitive to assumptions about future NIM, credit costs, and growth. Small changes in these inputs compound over the projection period and can produce large changes in intrinsic value. A half-percentage-point change in assumed NIM growth, for instance, can shift the final valuation by 15% to 25% or more over a ten-year projection.
  • Terminal value often represents 50% to 70% of total intrinsic value, meaning the valuation is heavily dependent on assumptions about long-term earnings power and the terminal multiple or growth rate. A model where two-thirds of the value comes from a single assumption about what the bank is worth a decade from now has inherent precision limits.
  • Does not explicitly capture the value of excess capital (unlike the Excess Capital Return Model) unless the projection incorporates capital return assumptions such as buybacks or special dividends. A bank sitting on significant excess capital above regulatory minimums may be undervalued by a straight earnings projection if that capital is not factored into shareholder returns.
  • Provision normalization is one of the most critical and subjective assumptions. Estimating the through-cycle provision rate requires judgment about future credit conditions, underwriting quality, and the loan mix. A bank that underwrites conservatively may have a structurally lower through-cycle loss rate than its peers, but distinguishing between genuinely better underwriting and simply not having been tested yet is difficult.
  • Substituting earnings for free cash flow introduces conceptual imprecision because not all earnings are distributable. Some portion must be retained to support growth and maintain regulatory capital requirements. The model can overstate value if it implicitly treats all projected earnings as available to shareholders when, in practice, a meaningful share will be retained in the business.

Bank-Specific Considerations

The Discounted Earnings Model exists because the standard DCF framework that works for most companies falls apart when applied to banks. Understanding why requires looking at what makes bank finances different.

Why Free Cash Flow Doesn't Work for Banks

Industrial companies invest in factories, equipment, and working capital. Subtracting those investments from operating cash flow gives you free cash flow, the money available to investors. Banks don't work this way. A bank's primary investment is growing its loan book, and that growth is funded by deposits and borrowings rather than retained cash flow. Taking in deposits is simultaneously a core business activity and a source of funding, so the line between operations and financing that DCF relies on doesn't exist cleanly.

Earnings as the Value Proxy

Earnings (net income available to common shareholders) serve as the best proxy for value creation at a bank. However, not all earnings are free to be returned to shareholders. Regulatory capital requirements dictate that a bank must retain enough earnings to support its asset base and meet minimum capital ratios. Distributable earnings are always less than total earnings for a growing bank, and dividend capacity is subject to regulatory approval.

The Provision Challenge

The provision for credit losses adds a layer of cyclicality to bank earnings that doesn't exist for non-financial companies. In good years, provisions are low and earnings are inflated relative to what the bank can sustain through a full credit cycle. In bad years, provisions spike and earnings collapse. Any meaningful discounted earnings analysis must normalize provisions to reflect what the bank's earnings look like across an entire credit cycle, not just where they happen to be at the moment the analysis is performed.

When to Use This Method

The Discounted Earnings Model is most appropriate when a bank's current earnings are significantly above or below normalized levels, making point-in-time ratios like P/E unreliable. If provisions are temporarily elevated following an economic downturn, current earnings understate the bank's long-term earning power. If provisions are unusually low during a benign credit environment, current earnings overstate it. The model lets you project through those distortions.

It is also the right tool when a bank is undergoing strategic transformation. A bank that is exiting a low-return business line, integrating an acquisition, or investing heavily in technology will have a different earnings profile three to five years from now than it does today. Modeling that trajectory explicitly produces a more meaningful valuation than applying a multiple to current results.

For quick screening or for banks with stable, normalized earnings, simpler approaches like P/E or price-to-book comparisons are usually sufficient and less assumption-dependent.

Method Connections

The Discounted Earnings Model is conceptually similar to the Dividend Discount Model (DDM), but uses total earnings rather than just the portion paid out as dividends. If you apply the retention ratio to the earnings projection and only discount the dividends, the Discounted Earnings Model converges to the DDM. The two approaches should produce similar results for a bank with a stable payout policy.

The Gordon Growth Model is a simplified single-stage version of the Discounted Earnings Model that assumes constant earnings growth forever. The terminal value in a multi-year discounted earnings projection is often calculated using the Gordon Growth formula, making the Gordon model a building block within the broader framework.

The Excess Capital Return Model complements the Discounted Earnings approach by separately valuing capital above regulatory requirements. Where the Discounted Earnings Model may undervalue a bank holding significant excess capital (because it focuses on operating earnings rather than balance sheet value), the Excess Capital Return Model captures that value explicitly. Using both together produces a more complete valuation.

Common Mistakes

Ignoring Capital Constraints

The most common error is projecting earnings growth without checking whether the bank has sufficient capital to support that growth. If projected asset growth requires the bank to retain most of its earnings, the effective distributable earnings are lower than headline earnings suggest. A bank earning 10% ROE with a 9% equity-to-assets ratio and a 40% payout ratio can grow assets at roughly 6.7% per year. Projecting 10% loan growth without accounting for this cap produces an unrealistic earnings trajectory.

Using a Single Year's Provisions

Failing to normalize provisions is equally damaging. Using a single year's provision as the basis for all projected years ignores the credit cycle and produces overly optimistic or pessimistic projections depending on where in the cycle the analysis begins. A bank that reported a 0.10% provision-to-loan ratio during an unusually benign credit year should not be projected at that rate indefinitely. A through-cycle average of 0.30% to 0.50% is more appropriate for most banks.

Applying Industrial DCF Mechanics

Applying industrial-company DCF mechanics to banks is a fundamental methodological error. Using operating free cash flow, WACC (weighted average cost of capital) discounting, or enterprise value calculations designed for companies with defined capital expenditure cycles produces misleading results when the subject is a financial institution. Banks should be valued using equity-based methods with the cost of equity as the discount rate.

Across Bank Types

Large, Diversified Banks

For large banks with stable, diversified earnings streams, a shorter projection period of 5 years is often sufficient. Their earnings are more predictable because revenue sources are spread across lending, fee income, trading, and wealth management. The terminal value is more reliable because the baseline earnings are less likely to be distorted by a single loan concentration or geographic exposure.

Community and Regional Banks

Smaller banks with more concentrated portfolios and more volatile earnings typically require longer projection periods of 7 to 10 years. This captures a full credit cycle and reduces the risk of anchoring the valuation to a single favorable or unfavorable period. Pay particular attention to loan concentration risk when building projections for these banks. A community bank with 40% of its loans in commercial real estate will have a very different earnings trajectory through a downturn than one with a diversified loan book.

Turnaround Situations

Banks in turnaround situations require the most care. The earnings trajectory from current depressed levels back to normalized profitability needs to be modeled explicitly, year by year. Simply assuming the bank returns to peer-level ROE in year two or three may be too aggressive. Turnarounds often take longer than expected, and the path is rarely smooth.

Related Valuation Methods

  • Dividend Discount Model — Values a bank stock by estimating what its future dividend payments are worth today, making it particularly applicable to banks with steady payout histories.
  • Price to Earnings Valuation — A method for estimating what a bank stock should be worth by comparing its share price to the earnings it generates per share.
  • Excess Capital Return Model — Values a bank by splitting its capital into two parts: what regulators require it to hold, and the extra capital above that minimum which could be returned to shareholders
  • Gordon Growth Model (Bank Application) — Estimates what a bank stock should be worth based on its expected dividend, the return investors require, and a sustainable growth rate that links profitability, payout decisions, and cost of equity into a single fair value formula

Related Metrics

  • Earnings Per Share (EPS) — How much profit a bank earns for each share of its stock, calculated by dividing net income by the number of shares outstanding.
  • Return on Equity (ROE) — Measures how much profit a bank earns for each dollar of shareholder equity. One of banking's most watched profitability metrics because it captures both operating performance and the effect of leverage in a single number.
  • Return on Average Assets (ROAA) — Measures how much profit a bank earns relative to its total asset base, stripping out leverage effects that can distort equity-based profitability measures like ROE.
  • Net Interest Margin (NIM) — Measures the spread between what a bank earns on loans and investments versus what it pays on deposits and borrowings, expressed as a percentage of earning assets. NIM is the single most important revenue metric for most banks.
  • Pre-Provision Net Revenue (PPNR) — Measures a bank's core earnings power before subtracting the cost of bad loans, showing how much revenue the bank generates from its everyday operations before credit losses reduce the bottom line
  • Provision for Credit Losses to Average Loans — Measures how much a bank spends on expected loan losses relative to its total loans each year, showing the current-period cost of credit risk and the pace of reserve building
  • Non-Performing Loans (NPL) Ratio — Measures the percentage of a bank's loan portfolio that is non-performing (90+ days past due or on non-accrual), making it the primary gauge of credit quality and lending risk
  • Net Charge-Off Ratio — Measures what percentage of a bank's loans were actually lost during a period, after accounting for amounts recovered on previously written-off loans. The most direct measure of what credit risk actually costs a bank.
  • Cost of Funds — Measures the average interest rate a bank pays on all of its interest-bearing funding, including deposits, borrowings, and subordinated debt
  • Efficiency Ratio — Shows how many cents a bank spends to generate each dollar of revenue, with lower values indicating tighter cost control.
  • Equity to Assets Ratio — Shows what percentage of a bank's total assets are funded by shareholders' equity rather than deposits and borrowings, providing a simple measure of capital strength and leverage.
  • Dividend Payout Ratio — Measures the percentage of a bank's earnings distributed to shareholders as dividends, indicating how much profit is returned versus retained to build capital.

Frequently Asked Questions

What is intrinsic value and how do I estimate it for a bank?

The Discounted Earnings Model estimates intrinsic value by projecting future earnings and discounting them to present value, adapted for bank-specific dynamics like provision cycles and capital constraints. Read more →

Why is bank valuation different from valuing other companies?

Banks require earnings-based or book-value-based models rather than traditional DCF because free cash flow is not a meaningful concept for financial institutions. Read more →

How does the dividend discount model work for bank stocks?

The Discounted Earnings Model uses the same discounting framework as the DDM but values total earnings rather than just dividends, making it more flexible for banks with changing payout policies. Read more →

Apply this method using the Bank Screener to evaluate 300+ publicly traded US banks.