H-Model Valuation

The H-Model is a dividend valuation framework designed for companies transitioning from rapid growth to stability. Rather than assuming constant growth forever (as in a perpetuity) or making sudden jumps (as in typical two-stage models), the H-Model posits that growth declines linearly over a specified horizon. This captures the real arc of maturing businesses and reduces the sensitivity to arbitrary “high-growth” vs. “stable-growth” cutoff dates.

The problem it solves

Standard dividend discount models typically use one of two structures. The constant-growth model (Gordon Growth) assumes dividends grow at rate g forever. Simple, elegant—but wildly wrong for young, fast-expanding firms. The two-stage model assumes high growth for n years, then abrupt switch to stable growth. More realistic, but the cutoff is arbitrary: why exactly year 5 for this company?

The H-Model offers a middle path. It acknowledges that growth doesn’t switch on a dime. A company growing 20% annually doesn’t suddenly drop to 4% in year 6 and stay there. Instead, growth slows gradually as the company matures, faces larger base effects, and loses the ability to reinvest capital at sky-high returns.

The “H” refers to the half-life of the decline. If you model growth decaying linearly over 10 years, H = 5 (half the period). If over 20 years, H = 10. This parameter anchors how long the transition takes.

The formula and its components

The H-Model value per share is:

P = D₀ × [(1 + gₗ)/(r – gₗ) + H × (gₛ – gₗ)/(r – gₗ)²]

where:

• D₀ = current dividend per share
• gₗ = initial (long-term or high) growth rate
• gₛ = stable, perpetual growth rate (typically 2–3%)
• r = cost of equity (discount rate)
• H = half-life of the linear decline (years)

The first term, (1 + gₗ)/(r – gₗ), looks like a perpetuity formula—the value if growth stayed at gₗ forever. The second term, H × (gₛ – gₗ)/(r – gₗ)², is the adjustment for the fact that growth will eventually decay to gₛ.

Interpreting H (the critical parameter)

H is the linchpin. It controls the slope of the growth decline. A large H means the transition is long and gentle; a small H means growth collapses quickly.

For a startup or young company with competitive advantages, H might be 8–15 years: growth is high but you expect it to stay elevated for a decade or so before settling. For a mature company undergoing modest cyclical upswing, H might be 2–3 years: you expect normalisation soon.

The choice of H often reflects industry dynamics and company age. A software-as-a-service (SaaS) company with strong network effects and low churn might justify H = 10; a cyclical industrial firm might warrant H = 3. The model is sensitive to H, so varying it in a range—say, H = 5 to 10—is wise.

Why linear decline is more realistic than a hard cutoff

Imagine a company growing 15% annually. A traditional two-stage model might say: “15% for five years, then 3% forever.” By year 6, growth has dropped from 15% to 3% overnight. That’s mechanically discontinuous.

The H-Model allows growth to decline smoothly: 15% in year 1, 13% in year 2, 11% in year 3, and so on, reaching 3% by year 11 (if H = 5). This tracks real-world patterns more closely. Businesses rarely hit a wall; instead, competitive pressures, market saturation, and reinvestment constraints gradually erode returns.

This smoothness also reduces the model’s sensitivity to the arbitrary “forecast horizon” problem. In a standard two-stage model, moving the cutoff from year 5 to year 7 can meaningfully change value. The H-Model is more forgiving because the transition is continuous.

Practical application: valuing a growth-stage company

Suppose a dividend-paying tech company currently yields $2.00 per share and is growing dividends at 12% annually. You believe that over time, it will mature and settle at 3% growth. The cost of equity is 9%.

Using the H-Model, you might assume:

• gₗ (initial growth) = 12%
• gₛ (stable growth) = 3%
• r (cost of equity) = 9%
• H = 7 (you expect the transition to take roughly 14 years, with half-life of 7)
• D₀ = $2.00

Plugging in:

P = $2.00 × [(1.12)/(0.09 – 0.12) + 7 × (0.03 – 0.12)/(0.09 – 0.12)²]

The math: (1.12 / –0.03) = –37.33 (perpetuity with negative denominator means high value), and the adjustment term adds incremental value from the expected dividend growth during the transition phase. The result might be, say, $95 per share.

(Note: the numerator being negative is odd here; this example illustrates why r must exceed gₗ for the model to work smoothly—cost of equity must be higher than initial growth.)

Comparison with other models

The H-Model sits between pure constant-growth dividend models—which ignore growth decay—and free cash flow DCF approaches. It’s simpler than a full multi-stage DCF because it requires fewer explicit forecast years, yet more nuanced than a single-rate perpetuity.

It’s also a dividend-focused model, so it’s most naturally applied to dividend-paying stocks. For non-dividend payers, use free cash flow or earnings as a proxy for payable cash, or resort to a full DCF.

When the H-Model works best

The H-Model excels when:

• The company pays (or is expected to pay) steady, growing dividends
• Growth is elevated today but expected to normalize over a medium horizon (5–15 years)
• The industry is mature enough that you can estimate a stable long-run growth rate
• You want to avoid the mechanical sharp-cutoff assumption of traditional two-stage models

It works less well when:

• Growth is unpredictable or volatile
• The company is in a nascent, uncertain market where the long-term stable rate is unknowable
• Dividends are irregular or reinvestment needs are opaque

Sensitivity and stress-testing

Because the H-Model depends on gₗ, gₛ, r, and H, varying each reveals how robust your valuation is:

• If initial growth drops from 12% to 10%, does the value change by 5% or 50%? (High sensitivity means you must be confident in growth assumptions.)
• If the cost of equity rises from 9% to 10%, how much does value decline? (This tests the discount rate assumption.)
• If H shifts from 7 to 5 years, does the conclusion hold? (Short-term uncertainty around the transition pace.)

A responsible H-Model valuation presents a range: base case, upside, and downside, each with defensible assumptions.

See also