Two kinds of cost in every bill
In the last two guides you watched a firm turn inputs into output through its production function, and you met the law of diminishing marginal returns — pile more workers onto the same kitchen and each new hand eventually adds less bread. Now we flip the picture from physical to financial. Every input the firm uses has a price, so behind that falling productivity sits a *rising cost*. The first job is to notice that not all costs behave the same way when you make more.
Split them in two. A [[fixed-and-variable-cost|fixed cost]] does not budge with output: the bakery's monthly rent, the oven lease, the insurance. Bake zero loaves or a thousand, the landlord still wants the same cheque. A variable cost rises and falls with how much you make: flour, yeast, the electricity to run the oven, the hours of the bakers. Make nothing and these vanish; make more and they climb. The whole distinction lives only in the short run — the stretch of time over which at least one input (here, the building and oven) is stuck. Wait long enough and *everything* becomes variable: a lease ends, a bigger oven can be bought. Fixed cost is really just "cost you cannot escape this month."
Adding them up: total cost
Stack the two together and you get [[total-cost|total cost]] — every dollar the firm spends to produce a given quantity. It is simply fixed cost plus variable cost. At zero output, total cost is not zero; it equals the fixed cost, because the rent falls due even on a day the oven stays cold. As output rises, the fixed part sits still while the variable part climbs, so total cost rises too — and crucially, it rises *faster and faster* once diminishing returns set in, because each extra loaf needs more and more added labour to make it.
Let us put real-ish numbers on our bakery, all in words. Suppose rent and the oven lease come to 60 dollars a day no matter what — that is the fixed cost. Each successive batch of loaves takes more labour and ingredients than the last, because of diminishing returns. So the variable cost of reaching 1 loaf is 20 dollars, of 2 loaves is 30 (only 10 more), of 3 is 45 (15 more), of 4 is 65 (20 more), of 5 is 90 (25 more). Add the fixed 60 to each and you have total cost: 80, 90, 105, 125, 150 dollars for 1, 2, 3, 4, 5 loaves. Notice the gaps between those totals — 10, then 15, 20, 25 — widening as you go. That widening is diminishing returns showing up as money.
The cost of just one more
Here is the number that runs the whole show. [[marginal-cost|Marginal cost]] is the extra cost of producing exactly one more unit — the bump in total cost when output rises by one. It is the firm's mirror of the consumer's marginal thinking you met earlier: the chooser asked "what does the next glass add to my satisfaction?"; the producer asks "what does the next loaf add to my bill?" In our table the marginal cost of the 1st loaf is 20 dollars, the 2nd is 10, the 3rd is 15, the 4th is 20, the 5th is 25. Those are exactly the gaps we noticed — and after the second loaf they start climbing, which is diminishing returns biting again.
Why does marginal cost often *dip* before it rises? In the very early stretch, an extra worker or batch can make the whole operation run more smoothly — two bakers can specialise, one kneading while the other shapes — so the next loaf is cheaper than the last. That is why marginal cost falls at first. But the fixed kitchen can only hold so many; soon workers crowd the same oven, and each new loaf demands disproportionately more input. From there marginal cost rises, and rises for good. The dip-then-climb shape of marginal cost is the financial fingerprint of the production function you studied last time.
Average cost and its U
Marginal cost is about the next unit; [[average-cost|average cost]] is about all of them at once. Average (or per-unit) cost is simply total cost divided by quantity — what each loaf works out to on average if you bundle the whole day's spending and split the bill across every loaf. In our bakery: 80 over 1 is 80 dollars a loaf; 90 over 2 is 45; 105 over 3 is 35; 125 over 4 is about 31; 150 over 5 is 30. It tumbles fast, then flattens. Push further (with diminishing returns biting hard) and it would eventually turn back up.
That drop-then-rise gives average cost its famous U-shape. Two forces pull in opposite directions. As you make more, the big fixed cost — that 60 dollars of rent — gets spread over more and more loaves, so its share per loaf shrinks toward nothing; this *spreading-out* effect drags average cost down hard at first. But variable cost per loaf is creeping up as diminishing returns set in; eventually that *crowding* effect overpowers the spreading, and average cost turns back up. The bottom of the U is the firm's most efficient scale — the output where each loaf is as cheap as it can possibly be.
Where marginal cost cuts average cost
Now the elegant part, and it is a true rule, not a coincidence: the marginal cost curve always passes through the very bottom of the average cost curve. Whenever the cost of the *next* loaf is below the *average* so far, making it pulls the average down. Whenever the next loaf costs more than the running average, making it drags the average up. So average cost can only be falling while marginal is below it, and only rising once marginal climbs above it — meaning the average bottoms out exactly where marginal cost crosses it. There is no other place it can turn.
The cleanest way to feel this is the test-score analogy. Imagine your running average grade so far is 80. Sit one more exam — that exam's score is your "marginal" grade. If you score 70 (below your average), your average must fall. If you score 90 (above it), your average must rise. The only score that leaves your average unchanged is exactly 80 — when the marginal equals the average. The average stops moving precisely when the marginal catches up to it; the same logic forces marginal cost to slice through average cost at its lowest point.
Loaves Fixed Variable Total Marginal cost Average cost 1 60 20 80 20 80.00 2 60 30 90 10 45.00 3 60 45 105 15 35.00 4 60 65 125 20 31.25 5 60 90 150 25 30.00 MC (the step in Total) pulls AC down while MC < AC, then pushes AC up once MC > AC -> they meet at AC's low point.
What costs are really for
Why drill into all this? Because marginal cost is the single number that tells a firm how much to make. A profit-seeking firm should keep producing as long as the next unit earns more than it costs — as long as the revenue from one more loaf beats its marginal cost — and stop where the two meet. The next guides turn this into the famous "produce until marginal revenue equals marginal cost" rule. Average cost, meanwhile, answers a different question: not *how much to make*, but *whether the whole venture is worth it* — if the price you can fetch is below your lowest average cost, no quantity will save you. Marginal guides the next step; average judges the whole.
Two honest caveats before we move on. First, these tidy curves assume the firm knows its own costs cleanly and that one input is fixed in the short run — real firms work with messy, lumpy, half-guessed numbers, and the smooth U is a model, not a measurement. Second, the "cost" an economist cares about is the opportunity cost of every resource, including the owner's own time and the money tied up in the business — not just the cash that leaves the bank. That gap between cash-out costs and true economic cost is exactly what the next guides on profit will pry open. Get fixed, variable, marginal, and average straight, and the rest of this rung clicks into place.