DCF Valuation for Cyclical Businesses

Valuing a cyclical business in a downturn using last year’s earnings is like pricing a house after a flood—you capture the damage, not the asset. A DCF for a cyclical company must normalise cash flows across a full business cycle, not anchor to today’s trough or peak, so the discount rate and terminal value reflect sustainable earning power.

Why Standard DCF Fails on Cyclical Businesses

A standard DCF projects explicit cash flows for 5–10 years, then assumes a terminal value based on perpetual growth at a stable rate. The problem with cyclical companies: there is no stability. A mining firm in a commodity downcycle generates near-zero free cash flow today; extrapolate that forward, and the company appears nearly worthless. A homebuilder in a housing bust has thin margins and rising defaults; model it as permanent, and you miss the inevitable recovery.

The flaw is not the method itself—DCF logic is sound—but the assumption of steady state. For cyclical businesses, “steady state” must mean the average level across multiple cycles, weighted by their likely frequency and duration. Otherwise, the valuation is a snapshot of one economic phase, not a measure of true earning power.

Defining the Cycle and Gathering Historical Data

Step one: identify the relevant cycle length and amplitude.

For commodities (oil, metals, agricultural products), cycles often run 3–5 years from trough to peak and back. A mining company’s earnings can swing 2–3x across the cycle; margins compress in a glut and widen in a shortage.

For residential real estate, the cycle is typically 7–10 years: supply depletes, prices rise, new building accelerates, oversupply emerges, construction stops, and the cycle restarts.

For automotive, cycles are shorter—2–4 years—driven by consumer credit availability and replacement demand.

For commercial real estate and industrial REITs, cycles vary by property type and tenant profile; office may be 5–7 years while logistics is flatter.

To quantify normalised earnings, you must gather at least one full cycle of historical data—ideally two. If the company or sector has been through a 2008-style crisis, include that. Look for:

• Net income or free cash flow
• EBITDA and EBITDA margin
• Return on invested capital or return on assets
• Leverage ratios (debt-to-equity, net debt)
• Capital expenditure intensity and timing
• Working capital swings (inventory, receivables, payables)

Plot these against an external cyclical indicator—commodity prices, housing starts, credit spreads, unemployment, industry capacity utilization—to confirm the firm’s cycle moves with the sector.

Calculating Normalised or Average Cash Flow

Three practical methods:

Method 1: Simple average. Calculate free cash flow or earnings for each year in the last two cycles (or 7–10 years if a full cycle is not clear), then average them. If an oil company earned $100M, $50M, $30M, $80M, and $120M across five years, the normalised number is $76M. This is easy but can be distorted by outliers.

Method 2: Weighted average. If you believe one phase of the cycle is more common (e.g., “this industry spends 60% of time in normal mode, 20% in boom, 20% in bust”), weight each year’s result accordingly. Trough years get lower weight; peak years get lower weight too, if the peak is rare.

Method 3: Cycle-phase modelling. Forecast a stylised cycle in your explicit projection period (e.g., years 1–7 cover a full boom-bust-recovery arc), then use a normalised figure for the terminal value. This is more sophisticated and anchors the forecast to your actual assumptions about industry dynamics.

In practice, Method 1 or 2 is most transparent for investor communication. Show the historical earnings table, highlight the normalised number, and explain why it’s the right anchor.

Adjusting Capital Expenditure and Working Capital Across the Cycle

Cyclical businesses don’t just earn differently in downturns—they invest and finance differently too.

Capital expenditure often lags the cycle: firms cut capex hard when profits fall, so trough years show artificially high free cash flow. Conversely, when demand recovers, capex surges to expand capacity, suppressing free cash flow despite rising earnings. To normalise:

• Calculate average capex as a % of revenue or EBITDA over the cycle.
• In trough years, use the cycle-average capex ratio, not the actual (depressed) number.
• Acknowledge that peak years may show elevated capex; adjust accordingly.

Working capital swings are another lever. In an upswing, inventory and receivables balloon; cash tied up grows. In a downturn, firms liquidate inventory and collect receivables more aggressively; working-capital is a cash source. A normalised forecast should assume working capital stays broadly neutral (as a % of growth), unless the company chronically carries excess inventory or has worsening receivables.

Debt capacity also cycles. In boom years, firms load up on leverage, confident in earnings stability. In busts, they deleverage to survive. Your DCF should use a normalised leverage ratio for the terminal value—not the distressed level of today—and calculate interest expense accordingly.

Building the Explicit Projection: Past Trough through Near Cycle

A robust DCF for a cyclical company often uses 7–10 years of explicit projection, not the standard 5 years. This allows you to model:

• Years 1–3: recovery from trough (or downturn from peak, if timing is that way)
• Years 4–6: normalisation and mid-cycle operations
• Years 7–10: late-cycle upswing or flattening toward perpetuity

Within each phase, use your industry knowledge to forecast:

• Revenue growth (constrained by capacity, demand outlook, competitive supply)
• Operating margins (should trend toward historical average, with some volatility)
• Capex and working-capital needs (elevated in expansion, suppressed in contraction)
• Debt issuance/paydown (conservative in downturns, opportunistic in booms)

Example: a copper miner in 2024, exiting a strong cycle:

The terminal value is built on year 7–10 normalised figures, not trough. This captures the long-term earning power.

Terminal Value: The Linchpin for Cyclical Firms

Because cyclical firms spend so much of their life away from “normal,” the terminal value is especially sensitive. A terminal growth rate of 2–3% (GDP-like growth) is usually defensible; a perpetuity multiple should be near or slightly below the historical average EV/EBITDA, not the trough or peak multiple.

If a miner typically trades at 6–8x EV/EBITDA in normal times, using 4x (a distressed multiple) for the terminal value will severely undervalue. Conversely, using 10x (a bull-market multiple) will overvalue. The most disciplined approach: use 5–6x as a normalised multiple and stress-test downward if you expect structural industry headwinds (e.g., long-term demand decline due to electrification).

Alternatively, use a perpetuity formula:

Terminal value = Normalised FCF × (1 + g) / (WACC − g)

Here, “normalised FCF” is the cycle-adjusted number from years 7–10, and g is a realistic long-term growth rate (usually 2–3%). This is theoretically cleaner but requires conviction on the normalised level, which is where discipline matters.

Sensitivity Analysis and Scenario Stress

For cyclical companies, sensitivity tables are essential. Show what happens to valuation if:

• Cycle length shifts: trough arrives earlier or later than expected.
• Cycle amplitude widens: a deeper recession, longer boom, or wider margin swings.
• Terminal-value assumption changes: the perpetuity multiple or growth rate shifts by ±100bps.
• Capex efficiency varies: the company invests more or less aggressively in recovery.

Also run explicit scenarios:

• Base case: cycle-normalised, reasonable assumptions.
• Downside: longer, deeper trough; normalised margins 200bps lower.
• Upside: shorter trough; structural demand tailwinds lift the cycle.

This transparency forces you (and investors) to acknowledge the range of outcomes and the key drivers of value.

When a Cycle Breaks

Occasionally, a cycle does not return as expected. Structural demand shifts—such as coal mining facing long-term coal demand collapse—may render historical cycles meaningless. In such cases:

1. Acknowledge the structural change openly.
2. Model a “new normal” lower than historical levels, or a gradual decline.
3. Consider whether a DCF is even the right tool; liquidation value or scenario-based approaches may be more honest.

The DCF is a tool for valuing going concerns. If you believe the business is structurally impaired, say so.

See also