Machinery and capital goods: cycles, dealers, and aftermarket

What sits inside machinery and capital goods

Machinery is the sub-sector most people picture when they hear "industrials," and the one interviewers reach for to test whether you actually understand cyclicality or have only memorized the word. The bucket covers construction and mining equipment (excavators, loaders, haul trucks, crushing and screening plant), agricultural equipment (tractors, combines, planters, sprayers, plus the precision-agriculture electronics bolted onto them), industrial and process machinery, compressors, pumps, valves and the broader flow-control category, material handling (lift trucks, conveyors, warehouse automation), and factory automation (robots, motion control, sensors, controllers). Where these sit relative to aerospace, transportation, and building products is laid out in the industrials sub-sectors map.

The label that matters is "capital goods." A capital good is bought as an investment, not as a consumable. A contractor buying an excavator is underwriting a return: the machine earns revenue per hour of use over some service life, and the purchase competes against every other use of that contractor's capital. That single fact is the root of everything else here, because an investment decision can be deferred. When end markets look uncertain, a fleet owner runs the existing machine another season, defers the trade-in, and spends on maintenance instead. Nobody stops working; they just stop buying. That is why machinery revenue swings far harder than the activity it serves, and why a candidate who says "construction spending fell, so equipment sales fell proportionally" gets a follow-up.

The installed base and the replacement cycle

The central concept here is the installed base: every unit the manufacturer has ever sold that is still in service, sitting in customer fleets and on job sites. New unit sales in a given year are small relative to that base. Most demand in a normal year is replacement demand, driven by machines aging out of economic service, not by fleets getting bigger.

So the analysis you want is not "how much construction is happening" but "how old is the fleet, how hard is it working, and what is a used one worth." Three signals do most of the work:

Fleet age. Every machine has an economic service life, after which maintenance cost per operating hour, downtime, and residual value all move against the owner. When average fleet age rises above its normal range, replacement need is accumulating whether or not orders are being placed yet.

Utilization. Rental fleet utilization and operating hours per machine tell you whether the installed base is working hard or sitting idle. High utilization ages equipment faster and pulls replacement forward; idle fleets push it out, and any recovery in activity gets absorbed by existing capacity before a single new order appears.

Used-equipment values. Auction and dealer resale prices are the cleanest real-time read on the balance between supply of and demand for capacity. Firm used prices usually mean end users cannot get new machines fast enough. Falling used values mean capacity is being released back into the market, and they compress the trade-in economics that fund a large share of new purchases.

Here is the point interviewers actually want. Deferred replacement is not lost demand, it is demand that moved. A hypothetical fleet of 10,000 machines with a ten-year service life needs roughly 1,000 replacements a year in steady state. Two soft years at 600 units each do not delete 800 machines of need; they park it, and leave the fleet measurably older, which makes the eventual catch-up sharper than the activity level alone would suggest. The mechanism runs in reverse too: a boom that pulls replacement forward borrows from the next several years, which is why a company posting record deliveries is sometimes a worse forward setup than one posting weak ones.

Contrast that with the consumer world, because candidates coming from a retail frame get it backwards. A consumer purchase that does not happen this quarter often simply never happens; the occasion passed. A machine purchase that does not happen leaves a physical asset in the field getting older and more expensive to run every month, and that accumulates as pent-up replacement need with a hard physical constraint at the end of it. Here the whole discipline is reading the age and condition of an installed base.

The dealer channel and the sell-in versus sell-through trap

Most machinery reaches the end customer through a dealer network, and in construction and agriculture those dealers are frequently independent businesses, not owned outlets. The manufacturer recognizes revenue when it ships a machine to the dealer. The dealer recognizes a sale when it moves that machine to a contractor or a farmer. Two different events, sometimes several quarters apart.

The language is sell-in (manufacturer to dealer) versus sell-through (dealer to end user). Reported manufacturer revenue is sell-in. Real end demand is sell-through. They converge over a full cycle and diverge violently within it.

The mechanism is dealer inventory, usually discussed in months of supply. When dealers believe demand is strengthening they build inventory ahead of it, so the manufacturer ships more than end users are buying and reported revenue overshoots real demand. When dealers get nervous, or floor-plan financing costs bite, they stop ordering and sell down what is on the lot. That is destocking, and during it reported revenue can fall dramatically while end-user purchases are only modestly softer. Restocking does the same in reverse and makes a modest recovery look like a boom.

So the channel does not just transmit the cycle, it amplifies it both ways. Reported results are the end-demand cycle plus a channel-inventory derivative layered on top.

This is a favorite interview trap, usually delivered as a simple-sounding prompt: "Machinery company revenue is down 20 percent. What happened?" The weak answer treats reported revenue as end demand. The strong answer separates the two and names what it would check: dealer months of supply versus normal, retail sell-through where the industry publishes it, used-equipment values, and orders versus deliveries. It is also why management teams talk constantly about "underproducing retail demand," code for shipping fewer units than end users are buying to bring channel inventory back to target.

Aftermarket: the part that sets the multiple

Every machine in the installed base consumes parts, service, and eventually a rebuild. That aftermarket business (wear parts, filters, field service, remanufactured units, extended warranties, and increasingly telematics and subscription condition monitoring) behaves nothing like the new-equipment business attached to it.

Aftermarket revenue carries structurally higher margins, because parts pricing is set against the cost of downtime rather than against a competitive bid on a new unit. It is far less cyclical, because a working machine still needs parts and a machine kept in service longer needs more of them. And critically, it scales with the installed base, not with new orders. When new-unit sales fall and customers extend equipment life, aftermarket demand often holds flat or rises, putting a floor under earnings exactly when the cyclical business is at its worst.

Telematics has sharpened this. Connected machines report operating hours, fault codes, fuel burn, and location, which lets the manufacturer capture service revenue that used to leak to independent shops and sell monitoring as recurring subscription revenue.

For a banker, aftermarket mix is the single most useful explanation for why two machinery companies with similar growth and similar end markets trade at meaningfully different multiples. The one with a large, high-margin, recurring aftermarket stream has less earnings volatility and a better return profile through the trough, and the market pays for that. If you are asked why Company A trades at a premium to Company B, aftermarket mix should be one of the first two things out of your mouth, alongside the through-cycle margin record. How that translates into an actual valuation is covered in how industrials companies are valued.

Short-cycle and long-cycle order books

Machinery is not one cycle, it is at least two, and interviewers use the distinction to separate candidates who have thought about the sector from those who have read a summary of it.

Short-cycle businesses (components, bearings, fluid power, standard pumps and valves, sensors) ship weeks after the order arrives. Backlog is thin and tells you almost nothing. Orders are the signal, close to a real-time read on industrial production, because customers place them against near-term operating needs. Short-cycle turns first in both directions, which is why these names are watched as an early indicator for the rest of the sector.

Long-cycle capital equipment (large process machinery, engineered compression trains, mining plant, complex automation systems) is quoted, engineered, and built over quarters or years. Backlog is deep, and revenue in any period is largely determined by orders taken earlier. Here book-to-bill (orders divided by revenue in the period) is the leading indicator and reported revenue is the lagging one. A long-cycle business can post growing revenue out of backlog for a year after orders have already rolled over, which is exactly the setup that catches people out. Reading book-to-bill, backlog quality, and cancellation risk properly is covered in backlog, book-to-bill, and cyclicality.

Operating leverage and decremental margins

Machinery manufacturing carries a heavy fixed cost base: plants, tooling, engineering staff, and a service footprint that cannot scale down proportionally with volume. That produces steep operating leverage, and interviewers test it constantly with some version of "what happens to margins if volumes fall 20 percent."

Work it with contribution margin and clean hypothetical numbers. Take 1,000 of revenue, variable costs of 650 (a 35 percent contribution margin), and 200 of fixed cost. EBIT is 150, a 15 percent margin. Now cut volume 20 percent with no price change. Revenue is 800, contribution is 280, fixed cost is unchanged at 200, EBIT is 80. Margin went from 15 percent to 10 percent, and EBIT fell 47 percent.

The ratio bankers quote is the decremental margin: change in EBIT divided by change in revenue. Here that is 70 divided by 200, a 35 percent decremental, which is simply the contribution margin because nothing else moved. Run it upward and you get the incremental margin. Management teams guide to both, and the useful nuance is that reported decrementals usually land better than the raw arithmetic, because cost actions (reduced overtime, plant shift reductions, lower incentive compensation) partially offset the fixed base. A team claiming it can hold decrementals to 25 percent is claiming it will take out cost fast enough to beat its own contribution margin. Whether it did that in a previous downturn is worth checking.

Two things fall out of this worth saying aloud. Mid-cycle earnings matter more than trailing earnings, since peak and trough EBIT on the same asset base can differ by a multiple. And the balance sheet has to be sized for trough cash flow, not average cash flow.

Where the deal flow comes from

Three patterns generate most of the mandates in this sub-sector.

Financial sponsors are persistent buyers of aftermarket-heavy and niche flow-control assets, because a large installed base plus recurring parts revenue produces the through-cycle cash flow that supports an acquisition financing. Pure new-equipment assets are much harder to lever, for the trough-cash-flow reason above; the vocabulary in those financing conversations sits in the leveraged finance terms glossary.

Diversified conglomerates separate machinery businesses regularly, either because a sub-scale unit does not fit the portfolio or because a slower cyclical asset is depressing the multiple of a better one next to it. Those separations, and the spin-off versus sale versus carve-out decision behind them, are covered in industrials M&A and spin-offs.

Third are roll-ups: buying parts distributors, service networks, and small component makers as bolt-ons to build the aftermarket density that supports a higher multiple. Small individually, constant in aggregate, and a large share of what an industrials analyst actually works on well before touching a headline transaction.

Practice question

A machinery manufacturer reports revenue down 20 percent year over year. Walk me through what you would want to know before concluding end demand fell 20 percent.

I would not assume reported revenue equals end demand, because this company recognizes revenue when it ships to independent dealers, not when a dealer sells the machine to a contractor. So the first thing I would check is dealer inventory, in months of supply versus its normal range. If dealers are destocking, sell-in falls much faster than sell-through, and part of that 20 percent is a channel adjustment rather than lost end demand. I would also look for management language about deliberately underproducing retail demand to bring channel inventory back to target.

Second, the installed base signals: fleet age, machine operating hours or rental utilization, and used-equipment values. If fleets are still working hard and used prices are firm, customers are deferring replacement rather than shedding capacity, and that deferred need is accumulating rather than disappearing.

Third, the mix. If aftermarket parts and service is roughly flat while new units carry the whole decline, the installed base is intact and being run longer, which is a very different situation from a shrinking one. Finally, I would ask whether price or volume drove it, since a 20 percent volume decline against a fixed cost base implies a much worse margin outcome.

What the interviewer is listening for: Whether you know that machinery revenue is sell-in, not end demand, and that dealer inventory amplifies the cycle in both directions. They also want to hear you reach for installed-base evidence (fleet age, utilization, used values) instead of a general statement about the economy, and to notice that aftermarket behaves differently from new equipment.