Automotive Cases 2026: Plant Utilization Arithmetic

The one series that makes an automotive case concrete

Most writing on automotive cases asserts that the industry is cyclical. The Federal Reserve measures it, monthly, with a long benchmark beside every reading, and the series is the best single input an automotive case can borrow.

Table 7 of the G.17 release carries a line for motor vehicles and parts, NAICS 3361 to 3363, with a weight of 6.61 in the index. Four numbers on that line matter here: a 1972 to 2025 average of 74.4 percent, a 1994 to 1995 high of 87.7 percent, a 2009 low of 33.0 percent, and a preliminary August 2026 reading of 69.0 percent, checked 30 September 2026 at federalreserve.gov.

Sit with the 2009 figure. A capital intensive sector ran at less than half its normal operating rate for a year. No other published industrial series in that table falls that far: utilities bottomed at 78.1 percent in the same year, and total manufacturing at 63.5 percent.

Table 3 of the same release gives the volumes. US motor vehicle assemblies ran at 10.43 million units in August 2026 at a seasonally adjusted annual rate, split into 1.13 million autos and 9.29 million trucks, of which 9.00 million were light and 0.29 million medium and heavy, published at federalreserve.gov. That last figure is the one to remember if a case hands you a commercial vehicle prompt, because the heavy segment is about one thirtieth the size of the light truck segment.

An automotive case prompt, costed at each of those rates

Here is the prompt, and the cost figures in it are ours, posed as assumptions. A candidate should announce their own the same way.

Northfield Motors runs one assembly plant and wants to know at what volume it stops losing money.

1. Capacity. 240,000 units a year at nameplate.
2. Fixed cost. $900 million a year, covering tooling amortization, salaried labor, plant engineering and allocated overhead.
3. Transfer price. $28,000 a unit to the dealer network.
4. Variable cost. $22,600 a unit, covering materials, purchased parts, hourly line labor and outbound freight.
5. Hurdle. Cover the fixed cost in the year it is incurred.

Contribution per unit is $28,000 minus $22,600, so $5,400. As a share of price that is 19.29 percent, call it roughly 19 percent.

Break-even volume is $900,000,000 divided by $5,400, which is 166,667 units rounded to the nearest whole vehicle. Against 240,000 units of capacity, that is 69.44 percent utilization.

Read the two outer rows against each other. Between the 2009 low and the 1994 to 1995 high, volume moves by a factor of 2.66. Profit moves from a $472,320,000 loss to a $236,592,000 profit, a range of $708,912,000, or roughly $709 million. Compare that range with the profit at the top of it: $708,912,000 over $236,592,000 is 2.996, so the plant's cyclical swing is almost exactly three times the profit of its best observed year. Against the theoretical nameplate profit of $396,000,000 it is still 1.79 times.

What the August 2026 row licenses in a case interview

Be careful with the sentence that comments on that row, because it is easy to overclaim. The plant's break-even sits at 69.4 percent utilization. The Federal Reserve's reading for the sector in August 2026 was 69.0 percent. Those two numbers are four tenths of a point apart.

What that licenses you to say is narrow and useful: on these assumptions, a plant of this cost structure would have been running marginally below break-even at the sector's August 2026 operating rate. It does not say that the industry lost money in August 2026. Real plants have different fixed costs, different contribution margins and different product mixes, and the published figure is an index across the whole sector rather than any one plant.

That distinction is the case skill. The published rate is a benchmark you calibrate your assumptions against, and it becomes evidence only once you have said what cost structure you assumed.

Why one point of price is worth a quarter of the profit

Contribution is 19.29 percent of price, so 80.71 percent of every dollar of revenue is already committed to variable cost before any fixed cost is covered. That ratio is what makes price the most dangerous lever on the page.

Take the 87.7 percent row, the plant's best realistic year, at 210,480 units and $236,592,000 of profit. Cut the transfer price by one percent, which is $280 a unit. Total revenue given up is 210,480 times $280, so $58,934,400.

That is 24.91 percent of the operating profit, call it roughly a quarter. One point of price is a quarter of the profit, and the volume it would take to recover the loss makes it worse. After the cut, contribution falls to $5,120 a unit and break-even volume rises to $900,000,000 divided by $5,120, which is 175,781 units rounded to the nearest whole vehicle. Break-even utilization moves from 69.44 percent to 73.24 percent, so almost four points of the plant's cushion disappear for one point of price.

Name the distinction out loud, because the two figures get confused. The margin given up is one percent of price. The additional volume needed to stand still is 5.5 percent, since 175,781 over 166,667 is 1.0547. Those are different quantities and calling the first the second is the standard error on this case type.

Talbot Trucks, and the ceiling a discount rate moves

McKinsey publishes an eTruck case that reframes the pricing question entirely. Talbot Trucks is a European truck maker asking whether to build electric trucks, and its third question asks "what is the maximum price Talbot Trucks can charge for its eTruck so that the total cost of ownership is equal to that of a diesel truck?", checked 30 September 2026 at mckinsey.com. Two assumptions sit under it: the case states that "Diesel trucks cost €100,000 and have no residual value at the end of their lifetime", and that the team concluded total cost has to match for customers to consider the eTruck.

The comparison exhibit is an image we could not read, and the model answers sit behind a control that returns no text. So the operating saving below is ours, and we label it as an assumption rather than a figure from the firm.

Assume the eTruck saves €12,000 a year in energy and maintenance over an eight year life. Undiscounted, the saving is 8 times €12,000, so €96,000, and the price ceiling is €100,000 plus €96,000, which is €196,000.

Now discount at 8 percent. The eight year annuity factor is one minus 1.08 to the power of minus eight, divided by 0.08. Since 1.08 to the eighth is 1.85093, the factor is 5.74664. The present value of the saving is €12,000 times 5.74664, so €68,960 rounded to the nearest euro. The ceiling becomes €168,960, call it roughly €169,000.

The gap between the two answers is €27,040, which is 27.04 percent of the diesel truck's own price. A candidate who never names a discount rate has not made an arithmetic error. They have left the single largest term in the answer undeclared, and the interviewer cannot tell which of the two numbers they meant.

Two automotive cases, two ways to decide the same thing

Deloitte publishes an automotive case with a different convention. LeadAuto is an Asian carmaker moving into buses, and the firm writes that "the client is struggling to understand profitability per bus due to the number of option combinations offered to the customer." Its second question asks "What is the operating profit margin per bus unit for the Base and Luxury configurations?", and the suggested answer includes "Recommend the base configuration", checked 30 September 2026 at usrecruiting.deloitte.com. The data sheet is an image, so we reconstruct no figures from it.

Put the two conventions side by side. McKinsey asks for the customer's lifetime cost and derives a price ceiling from it. Deloitte asks for the manufacturer's margin per unit and picks the configuration that wins on margin. Both are automotive investment cases published by major firms, and they measure success on different sides of the transaction.

The practical consequence for a candidate is one question asked before computing anything. Are we deciding what the customer will pay, or what we earn per unit? A structure built for one answers the other correctly and scores badly.

Where the points go on a plant case

What follows is our own marking scheme, read from the rubric code in src/lib/grading. No consulting firm publishes one of its own, and the weights below are ours rather than any firm's.

What nobody publishes about plant economics

We looked for a first-hand candidate account of an automotive case and found nothing citable. Reddit refuses connections from this network, and the session had no video transcript tooling. McKinsey publishes two first-hand recruiting blog posts and neither mentions a sector case. Deloitte publishes a tips video on its case page and no account of a specific case. Competitor preparation blogs cover automotive cases at length and are not citable sources.

Four things are not public, and each one matters to this page. The Talbot Trucks cost comparison is an image, so every operating saving above is our assumption rather than the firm's figure. The LeadAuto data sheet is likewise an image, and we reconstruct nothing from it. No carmaker publishes fixed cost per plant or contribution per vehicle, so the $900 million and $5,400 lines are ours. And the Federal Reserve publishes a weighted sector index rather than the operating rate of any particular plant, which the release itself makes clear.