There's another factor involved in cost: volume.
Tooling up to make 10,000 car sets a day, of chassis components or of anything, costs a fortune, but has a huge impact on the part cost, once the tools are amortized.
{ Ten thousand cars per working day, by the way, times 200 working days per year, would give you 2 million cars per year, roughly. That would be a great year for any single brand, but it's not enough to support a GM or a Ford all by itself. They need multiple factories, which they already have, and they'd love to keep them all that busy. It has been done. }
Take a rough guess at say, a lower control arm.
Making exactly one requires some press tooling, not to mention the press itself.
Figure a million dollars for one transfer press, used, and one set of dies.
So that one part would cost you a million dollars.
Of course you can make a similarly functional part from tig-welded tubing and simple bent parts for a lot less, but not the same exact part you would use in a production car, and you still have tooling to pay for somehow.
What you can't do with simple parts in simple cheap tooling is have the next part fall out of the die ten seconds after the first one did, and so on for as long as you're feeding stock and power.
When you're making that production style lower control arm from coiled strip, pretty much as fast as your press can go, you're only paying for a couple of pounds of steel, a little labor, and substantial electricity to run the press, so maybe a couple of dollars per unit, as it comes out of the press. You have to wash it, maybe punch a couple extra holes in it, paint it and so on in order to make a part you can ship, but if you're making trainloads of parts every day, every operation doesn't add a lot to the cost.
I don't think brand F would mind too much if I reveal that in 1967, when I worked there, they were paying about a quarter of a dollar for a good quality shock absorber, and a very small fraction of a penny for the rolpin that holds the diff pin in place, and their computers tracked part costs to the millionth of a dollar. Those little rolpins came in big rectangular steel containers that held about half a cubic yard; buy that many at a time, of anything, and you can get a very attractive price.
One of the more boring but interesting parts of my job as a manufacturing engineer was to estimate the cost of making a given part on a drawing, in a given quantity, using the tooling we had, along with a budgetary estimate for new tooling or adaptations that would be necessary. Every few weeks we would get a new plan set to estimate. The parts might be future production stuff for our own products, or special racing parts, or it might be competitors products; we were never told which was which, until a particular plan set was released for production.
The actual cost estimators somehow back-correlated the costs we reported with each and every tiny little feature that appeared in each and every part we looked at, so they were able to infer the cost of moving an axle bearing by a millimeter, or using a larger or smaller bearing.
I'm sure there were some nonlinearities in that estimating process that showed up from time to time.
For example,
We made a dozen-ish axle assemblies for the then next generation of Econoline vans, with the offset driveshaft, by removing alignment pins, shifting big machine tool parts around with some precision, running our special parts, then restoring the tools to their original configuration, in the middle of the night, for each step in the operation of all the affected parts. Great fun.
If those machine tools had not been designed, in some cases decades earlier, with some excess capacity and rapid and reversible adaptability, the offset axles for Econolines probably would never have appeared, because it would have required some entirely new tools, the expense of which could not be justified based on the anticipated/guessed sales volume.
I agree with Greg that _nobody's_ cost estimators are very accurate, but after seeing what goes into the process, I'm amazed that they do as well as they do. Mass production is _way_ more difficult than most people imagine.
< I left to work on products with volumes of one or two a year, or less, and found that was no picnic, either; just different. >
Mike Halloran
Pembroke Pines, FL, USA
OTOH, whenever I change organizations I always make a point of sitting down with the purchasing managers to understand how material, process, and other engineering choices impact pricing as this can vary significantly between companies.