Cyclical research evaluates earnings and cash flow across a cycle rather than extrapolating the latest period. Estimate a defensible normalized level, test whether the balance sheet survives the trough, and identify whether the price already anticipates a recovery or a decline.
A low multiple can describe peak earnings
When supply is tight, prices and margins can rise sharply. Dividing price by those earnings can make a cyclical look cheap just as new capacity arrives. Losses near a trough can hide future earning power while debt limits the ability to survive until recovery.
Read a long history of volumes, prices, capacity, margins, inventories and capital spending. Decide whether the cycle is broad, company-specific or structurally changing. Normalization is an analyst's assumption; label the years and the reasoning behind it.
Make survival part of the valuation
Build downside cases for liquidity, maturities, working capital and the cash needed through a long trough. A recovery case can be right about demand and still fail for shareholders through refinancing or dilution.
Special situations add conditions: the value that could be released, who can act, the timetable and what failure looks like. Forced selling and complexity are hypotheses about mispricing, not reasons to ignore liabilities.
Read the cycle as a story in Amten
Amten's pictures for this are the margin wave (a company's margin cycle with today's dot) and the peak meter (how close today is to the record). The balance sheet is read per $100 of assets, cash against debt. A Sector rotation story shows where the sector sits among leading, weakening, lagging and improving, with a five-step trail.
Then make the cycle a what-if: "What if industrial orders fall 15%?" carries the shock through customers and suppliers and redraws the three endings. Keep the price outcome separate from whether normalized earnings were reached; following the story shows each new quarter against the path.
Common questions
Does a low P/E prove a cyclical is cheap?
No. The earnings may be near a temporary peak. Test normalized earning power and survival through the trough.
Are normalized earnings reported facts?
No. They are estimates, and should be shown beside the actual history and the assumptions behind them.