Quant Memo
Core

Normalized Earnings and Mid-Cycle Margins

The most recent year's earnings are often the wrong starting point for a valuation because they carry one-off items and cyclical extremes — normalizing strips those out to find a sustainable base to grow from.

Prerequisites: FCFF vs FCFE: Which Cash Flow to Discount

A retailer just reported its best year ever: a one-time insurance payout after a warehouse fire, a competitor's temporary shutdown that pushed customers its way, and a tax refund from an old dispute. Plug that year's net income straight into a DCF as the starting cash flow and next year's forecast inherits all three windfalls, none of which will repeat. The valuation ends up wildly optimistic for reasons that have nothing to do with the business.

Normalized earnings are what a company would earn in a typical year under typical conditions, with one-off items removed and cyclical extremes replaced by a mid-cycle average — it's the starting point a forecast should actually grow from, not whatever number the most recent income statement happens to show.

What gets adjusted out

Two different problems get folded into "normalization," and it helps to keep them separate.

One-off items are genuinely non-recurring: litigation settlements, asset write-downs, restructuring charges, a factory fire, a one-time tax adjustment. These get added back (or subtracted, if they flattered results) because next year's business won't repeat them.

Cyclical extremes are different — they're real, recurring parts of the business, but the current year happens to sit at an unusually high or low point in a cycle that will mean-revert. A homebuilder's margin in a housing boom, an airline's margin during a fuel-price collapse, a miner's margin at a commodity peak — these are not one-off items to strip out entirely, but the current level shouldn't be extrapolated forward either. The fix is to replace the current margin with an average margin across a full cycle (often 5–10 years, covering at least one peak and one trough).

mid-cycle avg this year's actual margin time, across a full cycle
Extrapolating this year's peak margin forward overstates normal earnings; the mid-cycle average is the defensible base case.

Worked example

A steel producer reports the following operating margins over a six-year commodity cycle: 4%, 6%, 9%, 14% (this year, a price spike), 7%, and 5% for the two trough years before that. Revenue this year is $800 million.

  1. Naive approach. Using this year's 14% margin: operating income = 800 \times 0.14 = \112$ million. Forecasting growth off this base assumes steel prices stay at cycle highs forever.
  2. Mid-cycle margin. Average the six years: (4+6+9+14+7+5)/6=45/6=7.5%(4+6+9+14+7+5)/6 = 45/6 = 7.5\%.
  3. Normalized operating income. Apply the 7.5% mid-cycle margin to current revenue: 800 \times 0.075 = \60$ million — nearly half the naive figure.

A DCF built on $112 million of "sustainable" operating income would be discounting a number the business is unlikely to repeat for several years, if ever, without a further price spike.

What this means in practice

Analysts normalize earnings by rebuilding the income statement from revenue down: adjusting margins to mid-cycle or peer-typical levels, removing identified one-off items line by line (never a single blanket haircut), and checking that the resulting figure is consistent with the company's own long-run reported average and with what comparable companies earn through a full cycle. The normalized number, not the trailing reported number, becomes the base year that the explicit forecast period grows from.

Normalizing is not the same as smoothing away real deterioration. If a company's margins have structurally declined — new competition, a technology shift, permanent cost inflation — using an old mid-cycle average overstates normal earnings just as badly as extrapolating a peak does. Check whether the cycle is actually cyclical (mean-reverting) or whether the business has permanently changed before choosing which years to average.

Related concepts

Practice in interviews

Further reading

  • Damodaran, Investment Valuation (ch. 'Estimating Earnings')
ShareTwitterLinkedIn