The Dividend Discount Model: Building a Three-Stage DDM

Introduction

The dividend discount model (DDM) is the primary intrinsic valuation methodology for banks and most financial institutions. Where non-financial companies are valued using unlevered DCF (projecting free cash flows to the firm and discounting at WACC), financial institutions require equity-level valuation because debt is an operating input, not a financing choice. The DDM values equity directly by discounting expected future dividends at the cost of equity. For banks, dividends represent the true distributable cash flow: the earnings that can be returned to shareholders after meeting regulatory capital requirements, funding loan growth, and maintaining adequate buffers. The three-stage DDM (explicit forecast, transition, terminal value) is the standard implementation used in FIG equity research, pitch books, and M&A analysis, and it connects directly to the justified P/BV framework and P/TBV multiples that dominate relative bank valuation.

The DDM works for banks because it solves the fundamental problem that breaks traditional DCF: it operates entirely at the equity level. The discount rate is cost of equity (not WACC), the cash flows are dividends or distributable earnings (not unlevered free cash flow), and the output is equity value per share (not enterprise value). Since bank leverage is an operating necessity (deposits and borrowings fund the loan portfolio that generates net interest income), isolating operating cash flows from financing cash flows is impossible. The DDM sidesteps this problem by focusing on what shareholders actually receive: dividends.

The model also captures the regulatory reality of bank capital management. Banks cannot distribute all of their earnings as dividends. Regulatory capital requirements (CET1 minimums, stress capital buffers, G-SIB surcharges) set a floor on retained capital. The DDM explicitly models this constraint through the payout ratio, which determines how much of each dollar of earnings reaches shareholders. A bank with a 17% ROE but a 30% payout ratio distributes only 5.1 cents per dollar of book value as dividends, retaining the rest to fund growth and maintain capital ratios.

The Dividend Discount Model (DDM)

The DDM values equity as the present value of all future expected dividends. In its simplest form (the single-stage or Gordon Growth Model), the formula is: \( P_0 = D_1 / (r - g) \), where \( P_0 \) is the current stock price, \( D_1 \) is next year's expected dividend per share, \( r \) is cost of equity, and \( g \) is the perpetual dividend growth rate. For banks, \( D_1 = BV_0 \times ROE \times (1 - b) \), where \( BV_0 \) is current book value per share and \( (1 - b) \) is the payout ratio. Dividing both sides by book value yields the justified P/BV ratio: \( P/BV = (ROE - g) / (r - g) \). The single-stage DDM assumes constant growth forever, which is unrealistic for banks with above-average or below-average ROE that will eventually converge toward industry norms. The three-stage DDM relaxes this assumption by allowing different growth rates across distinct periods.

The three-stage DDM divides the bank's future into three periods, each with different assumptions about earnings growth, payout ratios, and return on equity.

Stage 1: Explicit Forecast Period (3-5 Years)

In Stage 1, the analyst projects earnings per share, ROE, and dividends per share explicitly for each year. This is where bank-specific fundamental analysis drives the valuation. The key drivers include net interest margin trajectory (how the rate environment affects the spread between earning asset yields and funding costs), provision for credit losses (normalized versus cyclically elevated provisions, the impact of CECL on provision timing), non-interest income growth (fee revenue, trading, advisory), efficiency ratio improvement (operating leverage from technology investment and scale), and the payout ratio (constrained by regulatory capital targets).

Each year's dividend is calculated as EPS multiplied by the payout ratio, then discounted at the cost of equity: \( PV = D_t / (1 + K_e)^t \). The explicit period captures the analyst's highest-conviction fundamental views and is where differentiated analysis creates valuation edge.

Stage 2: Transition Period (5-10 Years)

Stage 2 bridges the explicit forecast and the terminal value. During this period, dividend growth linearly converges from the Stage 1 growth rate to the long-term sustainable growth rate. ROE gradually moves toward the cost of equity as competitive advantages erode and the bank matures. The payout ratio typically increases as reinvestment needs decline (a bank growing at 3% needs less retained capital than one growing at 8%).

Each year in the transition is still explicitly calculated: the growth rate for year \( t \) is interpolated between the Stage 1 exit growth rate and the terminal growth rate. This produces a smooth deceleration rather than an abrupt shift from high growth to terminal growth, which would distort the valuation.

Stage 3: Terminal Value (Perpetuity)

Terminal value captures the value of all dividends beyond the explicit and transition periods, calculated using the Gordon Growth Model:

Where \( D_{T+1} \) is the first dividend in the terminal period, \( K_e \) is cost of equity, and \( g_L \) is the long-term sustainable growth rate. Terminal value typically represents 60-80% of total intrinsic value, making the terminal assumptions (cost of equity and sustainable growth) the most important inputs in the model.

The terminal growth rate is bounded by the sustainable growth formula: \( g = ROE \times (1 - \text{payout ratio}) \). If terminal ROE is 10% and the payout ratio is 60%, sustainable growth is 4%. Analysts typically cap terminal growth at 2-4%, approximating long-term nominal GDP growth. A terminal growth rate above the long-run nominal GDP growth rate implies the bank will eventually become larger than the economy, which is unsustainable.

Consider JPMorgan Chase with the following current parameters: tangible book value per share of approximately $101, ROE of approximately 17%, cost of equity of approximately 10% (beta of 0.98, risk-free rate of 4.3%, equity risk premium of 4.3%), current dividend of $6.00 per share (quarterly $1.50, payout ratio of approximately 30%), and current earnings per share of approximately $20.

Stage 1 (Years 1-5): Project EPS growing at 5-6% annually (driven by modest balance sheet growth, stable NIM, and share buyback accretion). Payout ratio stays at 30%. Year 1 dividend: $6.30. Year 5 dividend: approximately $7.70.

Stage 2 (Years 6-10): Dividend growth rate decelerates linearly from 5% to 3%. ROE gradually compresses from 17% toward 14% as the highest-return growth opportunities are exhausted. Payout ratio increases from 30% to 45%. Year 10 dividend: approximately $10.50.

Stage 3 (Terminal): Terminal growth of 3%, cost of equity of 10%. Terminal dividend (Year 11): approximately $10.80. Terminal value: $10.80 / (10% - 3%) = approximately $154 per share. Discounted back 10 years at 10%: approximately $59.

Sum of Stage 1 PV (approximately $24) + Stage 2 PV (approximately $30) + Terminal PV (approximately $59) = approximately $113 per share DDM value. Compared to JPMorgan's market price of approximately $260+ as of late 2025, this simplified DDM undervalues JPMorgan significantly, reflecting that the market assigns a substantial premium for JPMorgan's competitive moat, scale advantages, and consistently above-average ROE that the convergence assumptions in the model do not capture. An analyst would need to model a longer period of above-cost-of-equity ROE (or use a lower terminal convergence target) to approach market price, demonstrating the model's sensitivity to long-run ROE assumptions.

The DDM's payout ratio assumption is not a free variable for banks. Regulatory capital requirements create a hard floor on retained earnings. A bank's CET1 ratio must remain above its minimum requirement (4.5% under Basel III) plus its capital conservation buffer (2.5% minimum), stress capital buffer (determined by annual stress test results, minimum 2.5%), and G-SIB surcharge (1.0-4.5% for globally systemically important banks). JPMorgan's total CET1 requirement is approximately 11.5%, with an actual CET1 ratio of approximately 15.7%, providing approximately 420 basis points of excess capital.

If a bank's CET1 ratio falls into the buffer zone, automatic restrictions on capital distributions are triggered through the Maximum Distributable Amount (MDA) framework, which limits the percentage of earnings that can be paid as dividends. This means the DDM payout ratio must be calibrated to ensure the bank maintains adequate capital ratios throughout the projection period. In stress scenarios (recession, rising credit losses), projected provisions increase, earnings decline, and the payout ratio must be reduced to preserve capital, reducing the DDM value.

The annual Federal Reserve stress tests directly determine the stress capital buffer, which in turn determines how much capital each bank can distribute. Goldman Sachs's stress capital buffer dropped from approximately 6.4% to 3.4% after the 2025 stress test, freeing up significant capacity for the 33.3% dividend increase and additional buyback authorization. This illustrates how regulatory outcomes directly affect DDM inputs.

The traditional DDM uses dividends per share as the sole cash flow, but modern bank capital management combines dividends with substantial share repurchases. JPMorgan authorized a $50 billion buyback program in 2025. Bank of America announced $40 billion in repurchases. Goldman Sachs returned $16.78 billion to shareholders in 2024 on $17.18 billion in net income, a 97.7% total return ratio. Across the six largest US banks, approximately $100 billion flows to shareholders annually through dividends and buybacks combined.

The augmented DDM (sometimes called the total payout model) uses total capital returned per share (dividends plus the per-share value of buybacks) as the distributable cash flow. This produces higher valuations than the dividend-only DDM because it captures the full economic benefit to shareholders: buybacks reduce share count, increasing EPS and book value per share for remaining holders. For banks with payout ratios of 25-35% on dividends alone but 70-100% total payout ratios (including buybacks), the augmented DDM better reflects actual shareholder economics.

European banks present a contrasting capital return dynamic that is important for cross-border FIG analysis. The ECB prohibited bank dividends entirely from March 2020 to September 2021, creating a pent-up distribution capacity. Since restrictions were lifted, European banks have distributed capital aggressively: UBS estimates European bank total distributions reached approximately EUR 119 billion in 2024 (approximately 10% all-in yield) and EUR 123 billion in 2025. UniCredit distributed EUR 9.0 billion in 2024 (EUR 3.73 billion in dividends, EUR 5.27 billion in buybacks). European banks typically maintain higher cash dividend payout ratios (40-65%) than US peers (26-30%) but have historically returned less through buybacks. The gap is closing: European bank total payout yields now exceed US levels in some cases (approximately 10% versus 6-7% for top US banks), partly because lower European P/TBV multiples mean the same absolute payout produces a higher yield. For DDM purposes, using jurisdiction-appropriate payout ratios and cost of equity assumptions is essential for cross-border comparisons.

The DDM is the intrinsic valuation backbone of FIG analysis, translating a bank's fundamental earning power into a present value estimate through the lens of distributable cash flow. Its direct connection to regulatory capital constraints, the justified P/BV framework, and ROE as the primary return metric makes it the model that ties together the full FIG valuation toolkit.