Mid-Cycle Normalization in Valuation

Investors using mid-cycle normalization in valuation replace a company’s actual reported earnings with an average profit level that the business typically earns at a neutral point in the economic cycle. This smooths out distortions caused by commodity booms, busts, or temporary demand shifts, revealing what a company is “really” worth when stripped of temporary windfalls or headwinds.

Why raw earnings can mislead

Consider a gold miner that reported these earnings over seven years:

If you valued this business using only 2021 earnings at a 15× price-to-earnings-ratio, you would arrive at a valuation of $4.5 billion ($300M × 15). But gold prices in 2021 were at a cyclical peak—an unlikely baseline for perpetuity.

If you used 2018 earnings (the trough) at the same 15× multiple, you would value it at just $450 million ($30M × 15), seemingly absurd. The company is the same; only the commodity cycle has shifted.

Mid-cycle normalization asks: What is the true earning power of this miner when gold prices are neither depressed nor booming? By averaging the seven years, you get $157M in normalized earnings, producing a valuation of $2.36 billion ($157M × 15). This figure sits between the extremes and reflects the business’s “normal” profitability.

The logic: protecting against cyclical extremes

Most industries cycle. Commodity cycles, credit cycles, business cycles, and real estate cycles swing between expansion and contraction. A company’s earnings spike during booms (high prices, peak demand, cheap capital) and crater during busts (low prices, demand collapse, credit tightening).

A naive investor using raw reported earnings risks two errors:

1. Overvaluing at peaks: Assuming peak earnings are sustainable leads to bubble valuations. “The company earned $500M last year—at a 20× multiple, it is worth $10 billion!” Yet if peak earnings are 5× normalized earnings, this valuation is absurdly high.

2. Undervaluing at troughs: Assuming depressed earnings are the new normal leads to passing up bargains. “The company only earned $50M—it is not worth more than $1 billion.” Yet if trough earnings are 1/3 of normalized earnings, the valuation is far too low.

Mid-cycle normalization sidesteps both traps by anchoring to a long-term average. The business still has the same risk and growth profile; the cycle is merely at a different phase.

How to calculate normalized earnings

The most common approach is a simple average of the past 5–7 years of a key metric. The metric choice depends on the business:

For capital-intensive cyclicals (mining, oil, utilities): Use EBITDA or operating cash flow. These filter out the noise of depreciation schedules and focus on pure operating power.

For financial companies (banks, insurance): Use normalized net income, adjusting for loan-loss provisions and realized losses to reflect a mid-cycle credit environment. A bank’s 2019 earnings (before COVID) and 2008 earnings (during the crisis) are both distorted; an average smooths it.

For cyclical manufacturing (autos, machinery, cyclical consumer goods): Use earnings from mid-cycle years, or a 3–5 year average, sometimes with manual adjustments for one-time charges.

Worked example: A mining company

Earnings over six years: $200M, $120M, $80M, $60M, $100M, $250M.

Simple average: ($200 + $120 + $80 + $60 + $100 + $250) / 6 = $135M normalized earnings.

Some analysts weight more recent years heavier, or exclude the single highest and lowest years to avoid outliers. The choice depends on whether you believe the cycle is repeating or shifting.

Identifying what is “mid-cycle”

The trickiest part of mid-cycle normalization is defining where the cycle actually is. A few clues:

Commodity prices: Where is oil, copper, or gold relative to its 10-year average? Peak or trough prices signal cyclical extremes.

Capacity utilization: Is the industry running at 95% capacity (boom) or 65% (bust)? Mid-cycle typically means 75–85%.

Credit spreads and borrowing costs: During peaks, interest rates are low and money is cheap. During troughs, spreads are wide and credit is tight. Mid-cycle credit costs are between these extremes.

Industry revenue and order books: Growing order backlogs signal expansion; shrinking backlogs signal contraction.

Profit margins: Companies enjoy fat margins in peak times and thin margins in troughs. Mid-cycle margins are between the two.

These metrics together suggest where you are in the cycle. A cautious analyst uses multiple to triangulate.

Challenges and pitfalls

Structural change: Suppose a company’s normalized earnings were $100M from 2010–2019, but a regulatory change or technological disruption permanently reduced addressable demand. Averaging in pre-disruption earnings produces a misleading “normal.” You must identify whether the change is cyclical or structural.

Anchoring: Once you choose a number, it is psychologically hard to update. If you settled on $135M normalized earnings and the business has since shifted, you may cling to that anchor rather than recalculating.

Cycle length: Is the cycle 3 years, 5 years, 7 years, or 10 years? Averaging over the wrong horizon biases your result. An oil company’s 5-year average may differ sharply from its 10-year average.

Multiple selection: Averaging earnings is only half the battle. You still must decide what P/E multiple to apply. A company with normalized earnings of $100M might reasonably trade at 8×–12× depending on growth, leverage, and return on capital. Choosing 20× because the market is irrational does not follow from the normalization logic.

When to use normalized earnings—and when not to

Use mid-cycle normalization if:

• The business is clearly cyclical (mining, energy, construction, automotive, banking).
• Multiple full cycles of data exist (at least 5–7 years).
• The industry structure has not fundamentally changed.
• You are comparing current price to this normalized valuation to identify whether the market is mispricing cyclical extremes.

Do not use it if:

• The company is in a secular growth phase (tech, biotech, high-growth SaaS) and cycle-smoothing is irrelevant.
• You are valuing an undercovered small-cap with only 2–3 years of data.
• A structural change (regulatory, technological, competitive) has altered the earning power permanently.
• The most recent year’s earnings are materially different from the historical pattern; this suggests the cycle is shifting.

Normalized earnings as a contrarian signal

The contrarian often wins by buying cyclicals when normalized earnings suggest the price is cheap—even when current earnings are depressed. A mining company trading at $1 billion when its normalized earnings are $150M (a 6.7× multiple) might be a bargain if historical multiples are 10×–12×. The market is pessimistic; the cycle will turn.

Conversely, a peak-cycle company trading at 20× normalized earnings is expensive by historical standards, even if this year’s earnings justify it. As the cycle cools, multiples compress, creating a drawdown risk.

Alternative approaches

Scenario-based valuation: Instead of averaging, create three cases—bull, base, bear—assign probabilities, and weight the valuations.

Sum-of-the-parts: For conglomerates, value each division separately using its own cycle normalization, then sum.

Discounted cash flow with conservative assumptions: Project normalized free cash flow over 10 years, discount at a risk-adjusted rate, ignore the cycle by using flat sustainable numbers.

These alternatives avoid the mechanical averaging, but mid-cycle normalization has the advantage of simplicity and an explicit tie to historical reality.

See also