Terminal Value in Dividend Models

A terminal value in dividend models represents the capitalised worth of all dividends expected beyond an explicit forecast period. In the dividend discount model, projecting dividends year-by-year only makes sense for a limited horizon—typically 5 to 20 years. Beyond that, dividend streams must be collapsed into a single number, the terminal value, which often dominates the total valuation and demands careful judgment.

Why terminal value matters in valuation

The dividend discount model begins with a simple premise: a stock’s intrinsic value equals the present value of all future dividends. In theory, “all future” means forever. In practice, detailed year-by-year forecasts break down at the horizon—say, year 10 or year 15. Beyond that point, a company’s dividend prospects become too uncertain to forecast directly.

Terminal value solves this pragmatically. It expresses the value of the dividend stream from the horizon onwards as a single number, discounted back to today. For a typical mature company, terminal value often represents 60–80% of total equity value, sometimes more. This concentration means small changes in terminal value assumptions can swing the entire valuation conclusion by 20% or 30%.

The Gordon growth formula

The most widely used terminal value formula rests on perpetual growth logic. If a dividend in the final explicit year (year N) is D_N, and dividends are expected to grow at a constant rate g thereafter, then:

Terminal Value at Year N = D_N × (1 + g) / (r – g)

Here, r is the discount rate (required return) and g is the perpetual growth rate. The numerator, D_N × (1 + g), is the dividend one year after the forecast horizon. The denominator, r – g, is the spread between required return and perpetual growth.

This is the Gordon growth formula: it values an endless stream of dividends growing at a constant rate. For the formula to work, r must exceed g (otherwise it produces a negative or infinite value, a sign that the assumptions are contradictory).

Setting the perpetual growth rate

The perpetual growth rate g is critical and contested. Three approaches dominate:

Nominal GDP growth. Over very long horizons, a company’s dividend cannot grow faster than the economy itself without eventually consuming the entire economy. Long-run nominal GDP growth—say, 2% real growth plus 2% inflation—often anchors g at 3–4%. This approach emphasises conservatism and is broadly defensible.

Company-specific sustainable growth. A company’s dividend can grow faster than GDP if it gains market share or improves returns on invested capital. Sustainable growth is often approximated as the return on equity (ROE) multiplied by the retention ratio (earnings not paid out). A firm retaining 50% of earnings and earning 10% on equity can sustain ~5% dividend growth. This method ties g to fundamental earning power and is used by practitioners who want to reflect company-specific prospects.

Analyst consensus or explicit long-term forecasts. Some valuation models use published long-term growth rate expectations (often available as a 5-year or 10-year consensus from brokerage research). This is shorter-term than perpetual, but serves as a bridge from explicit forecasts to terminal assumptions.

No approach is bulletproof. GDP growth is inherently backward-looking. Sustainable growth depends on forecast profitability, which is itself uncertain. Analyst consensus has proven unreliable over decades. The key is to test sensitivity: run the valuation at g = 2%, 3%, and 4%, and see how much the conclusion shifts. If small changes in g overturn the investment decision, the valuation is fragile.

Horizon choice and two-stage models

How many years should the explicit forecast period run? A longer horizon—say, 15 or 20 years—allows more detailed modelling but requires more forecasting assumptions. A shorter horizon—5 or 10 years—is simpler but compresses more growth into the terminal value.

The two-stage model is common: an explicit high-growth period (years 1–7, say, growing at 8%) transitions abruptly into a perpetual mature-growth phase (growing at 3% forever). This avoids the awkwardness of assuming constant growth from today onwards, which is unrealistic for young or cyclical firms.

In a two-stage model, terminal value represents years 8 to infinity. The firm is assumed to have reached “maturity” by year 7, meaning it no longer re-invests heavily and dividend growth stabilises. The abrupt transition is stylised—in reality, maturation is gradual—but it offers computational simplicity and makes assumptions transparent.

Sensitivity and reasonableness checks

Terminal value is vulnerable to small forecast errors amplified over long horizons. A 1% change in the assumed perpetual growth rate can swing terminal value by 10–20%. This argues for disciplined sensitivity analysis:

• Calculate intrinsic value at g = base case – 1%, base case, and base case + 1%.
• Compare the range to the current market price. If the range straddles the price, the valuation is highly dependent on growth assumptions and should be treated cautiously.
• Stress-test r (discount rate) similarly.

Another check: compare the implied dividend yield at the horizon to market history and peer averages. If terminal value implies a year-10 dividend yield of 0.5%, lower than any mature equity market, reconsider whether g is too high or r too low.

Single-stage (perpetuity) models

Simpler valuations sometimes dispense with an explicit forecast period altogether. If a company is already mature and paying a stable dividend—a utility, say, or a well-established dividend aristocrat—the Gordon perpetuity can be applied to current or near-term dividends directly:

Value = D / (r – g)

where D is the current or next dividend. This collapses all future growth into a single g parameter and requires no horizon. It is elegant but risky: it assumes the company is already at steady state, and any misestimate of r or g cascades through the entire valuation.

Exit multiple alternatives

Some practitioners avoid explicit perpetuity growth and instead apply an exit multiple at the horizon. For instance, assume the stock will trade at 12× earnings per share at year 10, or 1.5× net asset value. The terminal value is then the equity value at the horizon (dividend yield, earnings, or book value) multiplied by the assumed multiple. This method sidesteps the perpetual growth question but introduces another source of assumption risk: the exit multiple itself.

Real vs. nominal dividend streams

Most terminal value calculations use nominal (current-dollar) dividends and a nominal discount rate. However, for very long-horizon valuations (20+ years), some practitioners favour a real (inflation-adjusted) framework, working in constant dollars and using real discount rates and real growth rates. The results converge if consistently applied, but nominal framing is more prevalent and easier to reconcile with market prices.

See also