Skip to main content

The H-Model for Dividends

The H-Model, developed by Russell Fuller and Chi-fu Huang in 1992, offers an elegant solution to a common valuation challenge: companies whose dividend growth rates decline smoothly and linearly over time. Rather than explicitly forecasting dividends for 10–20 years (as multi-stage models require), the H-Model uses a closed-form formula that assumes growth rates converge from an initial high rate to a stable perpetual rate in a linear fashion. This mathematical shortcut preserves accuracy while dramatically reducing computational overhead, making it popular among professional portfolio managers and financial analysts.

The H-Model is particularly useful for valuing mature high-dividend-yield companies, dividend-focused funds, and businesses transitioning from growth to stability at a measured pace. Its primary strength lies in combining formula-based elegance (like Gordon Growth) with the realism of acknowledging that growth doesn't switch abruptly from high to low, but declines gradually.

Quick Definition​

The H-Model formula is:

Stock Value = D₀ × (1+gₗ) / (r - gₗ) + D₀ × H × (gₛ - gₗ) / (r - gₗ)

Where:

  • D₀ = most recent dividend (current or last paid)
  • gₛ = short-term (initial) dividend growth rate
  • gₗ = long-term (perpetual) dividend growth rate
  • r = discount rate (cost of equity)
  • H = half-life of the high-growth period (in years)

The first term represents a stable perpetual dividend growing at gₗ. The second term is a premium for the extra growth during the decline period, scaled by H. Intuitively, H is half the time it takes for growth to decline from gₛ to gₗ.

Key Takeaways​

  • Linear Decline Assumption: Growth rates decline in a straight line from gₛ to gₗ, not in discrete phases like two-stage models.
  • Closed-Form Elegance: Unlike explicit multi-stage models, the H-Model produces a single formula, avoiding year-by-year projections.
  • Half-Life Parameter: H captures how quickly the growth transition occurs; longer H means slower convergence to perpetual growth.
  • Computational Efficiency: Ideal for portfolio managers comparing dozens of dividend stocks; fast calculations without spreadsheet complexity.
  • Realistic Transitions: Captures real-world dividend growth patterns better than abrupt single-to-double-stage switches.
  • Sensitivity to H: The half-life parameter dramatically affects valuation, so justifying it is critical.

The Logic Behind the Formula​

The H-Model rests on a mathematically clever insight: if dividend growth declines linearly from gₛ at year 0 to gₗ at year 2H, you can split the valuation into two components:

1. Base Perpetuity: Assume dividends grow forever at rate gₗ. This gives:

Value₁ = D₀ × (1 + gₗ) / (r - gₗ)

This is the Gordon Growth Model applied to the perpetual rate.

2. Growth Premium: Layer on a premium for above-perpetual growth during the decline period. Since growth averages (gₛ + gₗ) / 2 during the initial years, the premium is:

Value₂ = D₀ × H × (gₛ - gₗ) / (r - gₗ)

Sum them:

Total Value = Value₁ + Value₂

This is the H-Model. The elegance is that it approximates the present value of an infinite dividend stream with linearly declining growth using a simple algebraic formula, avoiding the need to project and discount dividends year-by-year.

Interpreting the Half-Life Parameter (H)​

H is the half-life of the high-growth period: the time it takes for growth to decline from gₛ to the midpoint between gₛ and gₗ.

Example:

If gₛ = 8%, gₗ = 3%, and H = 5 years:

Growth rate at Year 0:   8.0%
Growth rate at Year 2.5: 5.5% (midpoint: (8+3)/2)
Growth rate at Year 5:   3.0% (perpetual)
Growth rate at Year 10:  3.0% (perpetual forever)

The dividend growth declines from 8% to 3% linearly over 10 years (2H), with the midpoint reached at H = 5 years.

Longer H → slower decline to perpetuity, higher valuations. Shorter H → faster convergence, lower valuations. This makes H a critical but often underappreciated parameter.

Step-by-Step Valuation Example​

Company: Gradual Maturity Inc.

Given:

  • Current dividend paid (D₀): $1.50
  • Initial growth rate (gₛ): 7%
  • Perpetual growth rate (gₗ): 3%
  • Half-life (H): 4 years
  • Cost of equity (r): 8%

Step 1: Calculate base perpetuity component.

Value₁ = $1.50 × (1 + 0.03) / (0.08 - 0.03)
       = $1.50 × 1.03 / 0.05
       = $1.545 / 0.05
       = $30.90

Step 2: Calculate growth premium component.

Value₂ = $1.50 × 4 × (0.07 - 0.03) / (0.08 - 0.03)
       = $1.50 × 4 × 0.04 / 0.05
       = $0.24 / 0.05
       = $4.80

Step 3: Sum the components.

Stock Value = $30.90 + $4.80 = $35.70

At $35.70, the stock is fairly valued if trading at that price; below indicates undervaluation.

H-Model vs. Two-Stage Model: When to Use Each​

Both models handle growth transitions, but with different approaches:

Two-Stage Model: Best when there's a clear transition point—a company completing a major acquisition, exiting a high-growth market phase, or achieving a milestone. Example: a fintech company growing 12% annually for 6 years as it scales, then settling to 4%.

H-Model: Best when growth is expected to decline smoothly and continuously—a mature utility gradually increasing dividend growth from 2% to 3% as customer base expands, or a maturing tech company whose growth moderates gradually from 10% to 4% over 10 years.

Choosing the Half-Life Parameter​

H is the most subjective input in the H-Model. How do you estimate it?

Method 1: Industry Lifecycle

Research how long companies in the industry historically take to transition from growth to stability. If utility companies typically take 8–10 years for dividend growth to stabilize, set H = 4–5 years.

Method 2: Competitive Position

Strong competitive positions sustain above-market growth longer. A company with network effects or brand moats might justify H = 6 years. A company in intense competition might justify H = 2–3 years.

Method 3: Management Guidance

If management projects dividend growth explicitly for several years, let that inform H. If they're vague, use industry norms.

Method 4: Peer Analysis

Compare your company to similar mature peers. If peer companies' dividend growth rates converged to 3% by year 8, your H should reflect a similar timeline.

Common H values:

  • Early-stage growth companies: 6–8 years
  • Mid-market transitional companies: 4–6 years
  • Mature, stable companies: 2–4 years
  • Very mature, slow-change companies: 1–2 years

Sensitivity Analysis: Impact of H​

Doubling H from 4 to 8 years increases value by ~10%. This underscores the importance of carefully justifying H rather than using arbitrary values.

Realistic Application: Dividend Aristocrats​

Dividend aristocrats—companies with 25+ years of consecutive dividend increases—often fit the H-Model well. Their growth is neither perpetual at high rates nor abruptly stable; it's a smooth, measured increase year-over-year. Using the H-Model to value them captures this pattern elegantly.

Example: Consumer Staples Giant

A major consumer staples company with $3.00 annual dividend, historically growing dividends 5.5% annually, faces slowing growth as markets mature. Analysts expect initial growth to continue at 5.5% but gradually decline to 3% (GDP growth) over the next decade.

Using H-Model:

  • D₀ = $3.00
  • gₛ = 5.5%
  • gₗ = 3.0%
  • H = 5 years (half-life suggests 10-year full transition)
  • r = 7.5% (appropriate for a stable, dividend-focused consumer staples firm)
Value = $3.00 × 1.03 / (0.075 - 0.03) + $3.00 × 5 × (0.055 - 0.03) / (0.075 - 0.03)
      = $3.09 / 0.045 + $0.375 / 0.045
      = $68.67 + $8.33
      = $77.00

At $77 per share, the stock is fairly valued given those assumptions.

Advantages and Limitations​

Advantages:

  • Computational simplicity: One formula beats 10+ years of explicit projections.
  • Realistic decay: Linear decline from gₛ to gₗ mirrors real-world dividend patterns better than abrupt shifts.
  • Portfolio-friendly: Analysts can value dozens of dividend stocks quickly and compare valuations systematically.
  • Transparent assumptions: The H parameter is explicit and debatable.
  • Mathematical elegance: A closed-form solution that connects intuition to algebra.

Limitations:

  • Linear assumption weakness: Not all companies' growth declines linearly. Some accelerate, then decelerate. Some plateau for years, then drop abruptly.
  • H-parameter estimation challenge: Choosing H requires judgment; errors cascade through valuation.
  • Limited to dividend-paying stocks: Doesn't value non-dividend payers or those with erratic payouts.
  • No explicit stage modeling: Unlike two-stage models, you can't model discrete strategic shifts (acquisitions, market exits) as distinct phases.
  • Terminal value concentration: Like all DDM variants, heavily dependent on perpetual growth assumptions (gₗ and r), which are subject to estimation error.

Common Mistakes​

  1. Underestimating H: Setting H too low implies dividends converge to perpetual growth too quickly, undervaluing the stock. Conservative defaults (H = 3–4) might be safer than optimistic high-H estimates.

  2. Ignoring gₛ ≥ r constraint: If initial growth exceeds discount rate, the H-Model formula breaks down. Ensure gₛ < r and gₗ < r always.

  3. Using vague H values without justification: Picking H = 5 arbitrarily, without researching the company's competitive position or industry lifecycle, is reckless. Document your rationale.

  4. Conflating H-Model with two-stage: These models are different. Two-stage assumes growth switches abruptly at year n; H-Model assumes smooth linear decline. Use the appropriate tool for the expected pattern.

  5. Neglecting payout ratio sustainability: The H-Model assumes the company can sustain projected dividend growth through the decline period. If earnings don't support it, growth will plateau earlier than modeled. Cross-check against earnings projections.

  6. Forgetting D₀ is historical: Use the most recent dividend paid, not next year's expected dividend. If you want to use forward dividend, adjust the formula or adjust D₀ accordingly.

Real-World Valuation Comparison​

To illustrate, compare three models on the same company:

Company: Dividend Transition Corp.

Assumptions:

  • Current dividend: $2.00
  • Cost of equity: 8%
  • Perpetual growth: 3%
  • Near-term growth: 6% for 5 years (two-stage)
  • H-Model half-life: 4 years
  • H-Model initial growth: 6%

Single-Stage Gordon Growth (Using g = 3%):

Value = $2.00 × 1.03 / (0.08 - 0.03) = $41.20

Two-Stage (5 years at 6%, then 3% perpetual):

High-growth PV (5 years): ~$8.50
Terminal value at Year 5: $2.52 / 0.05 = $50.40
Discounted terminal: ~$21.90
Total value: ~$30.40

H-Model (gₛ = 6%, gₗ = 3%, H = 4):

Value = $2.00 × 1.03 / 0.05 + $2.00 × 4 × 0.03 / 0.05
      = $41.20 + $4.80
      = $46.00

The H-Model ($46) falls between single-stage Gordon ($41.20) and the two-stage ($30.40) in this example, reflecting a smoother growth transition than two-stage but more generous than pure stable-growth Gordon.

FAQ​

Q: Should I use H-Model or two-stage for most valuations?

A: If growth is expected to decline gradually and smoothly, H-Model is simpler and faster. If there's a discrete transition point or clear phase boundary, two-stage is more intuitive. Many analysts use both and triangulate.

Q: What if the company is in a growth phase, not a decline phase?

A: The H-Model assumes gₛ > gₗ (growth declining). If a company is accelerating, the model doesn't fit. Use a reverse H-Model (gₛ < gₗ), but this is less common and requires modified logic.

Q: How sensitive is H-Model value to gₛ?

A: Moderately sensitive. A 1% change in gₛ affects only the second component of the formula (the premium), not the base perpetuity. For a typical stock, increasing gₛ from 6% to 7% might increase value 5–10%, depending on H and other parameters.

Q: Can I use actual stock prices to back-solve for H?

A: Yes. If a stock trades at $45 and you know D₀, r, gₛ, and gₗ, you can solve the H-Model formula for implied H. This "implied H" reveals what the market expects about growth convergence rates. If implied H is 8 years but you think 4 years is realistic, the stock is overvalued.

Q: Is H-Model better than Gordon Growth for all dividend stocks?

A: No. For purely stable-growth utilities with 50+ year dividend histories, Gordon Growth is cleaner and requires fewer assumptions. H-Model is best for companies in genuine transition—not yet stable, not still accelerating.

Q: How do I determine if linear decline is realistic?

A: Check historical dividend growth rates. If growth has declined roughly linearly over 5+ years, linear assumption is defensible. If growth jumps around (cyclical business) or has plateaued, H-Model is less appropriate.

Q: What discount rate should I use for H-Model?

A: Same as for any dividend valuation: CAPM-derived cost of equity. Utilities typically 6–7.5%; industrial companies 7–8.5%; higher-risk dividend payers 8–10%.

Summary​

The H-Model elegantly captures companies whose dividend growth declines linearly from an initial rate to a stable perpetual rate. By using a closed-form formula rather than explicit year-by-year projections, it provides computational efficiency without sacrificing realism. The model's success hinges on carefully estimating the half-life parameter (H) and the perpetual growth rate (gₗ). When these assumptions are defensible, the H-Model delivers valuations that reflect both near-term growth opportunities and long-term stability.

The H-Model is particularly valuable for portfolio managers and analysts who must value dozens of dividend stocks quickly and systematically. It bridges the gap between the simplicity of Gordon Growth and the precision (but complexity) of explicit multi-stage models, making it a practical tool for professional dividend valuation.

Next: When to Use the Dividend Discount Model​

Continue to chapter 05 to learn when DDM is appropriate, when it fails, and how to combine it with other valuation methods for robust stock analysis.