These curves cause more confusion than they should because intermediate microeconomics draws them on top of each other and expects you to keep track. The clean way to think about short-run vs. long-run cost curves is this: in the short run something is fixed; in the long run nothing is. That single distinction is what makes the long-run average cost curve the envelope of many short-run curves.
The Short Run: One Thing Held Fixed
The "short run" in microeconomics is not a literal calendar. It is the period during which at least one input cannot be changed — usually capital. The firm cannot install a new factory or scrap an old one; it can only vary labor, materials, and other variable inputs.
That fixed input gives the short-run total cost two parts. Total fixed cost (TFC) is the bill for the input you cannot change — building rent, equipment lease, debt service. It does not move with output. Total variable cost (TVC) is everything else.
A short-run average total cost (SRATC) curve plots ATC against output for a given size of plant. It is U-shaped: average fixed cost (AFC) falls as output spreads the fixed cost over more units, and average variable cost (AVC) eventually rises because the fixed input limits how productively you can stack on more workers (the law of diminishing marginal returns). The minimum of SRATC is the most efficient output level for that specific plant size.
The Long Run: Every Input Is a Choice
In the long run, every input can be adjusted — including the plant. The firm can build a bigger factory, downsize, or relocate. That freedom is what lets it pick the plant size that minimizes cost at whatever level of output it expects to produce.
Long-run average cost (LRAC) is built by asking, for each possible output Q, what is the lowest average cost achievable when the firm is free to choose the best plant? The answer at each Q traces out the LRAC.
There is no "fixed" component in the long run, because there is no input the firm has to live with. That is the only structural difference, and every other contrast between SRATC and LRAC follows from it.
Why LRAC Is the Envelope of the SRATC Curves
Picture three plant sizes — small, medium, and large — each with its own SRATC. Each curve is U-shaped, and each has a low point at the output level it was designed for. Now the firm chooses which curve it is on by choosing the plant.
For low outputs, the small plant has the lowest ATC. For middling outputs, the medium plant does. For high outputs, the large plant does. The LRAC follows the lowest piece of whichever curve is best at each output. Stack many plant sizes and the LRAC becomes the smooth boundary that lies below or just touching every SRATC. That is exactly what "envelope" means in this context.
A consequence: every point on the LRAC is tangent to some SRATC curve — but only the bottom of the LRAC is tangent to the bottom of an SRATC curve. Everywhere else, the LRAC touches an SRATC on its way down or up, not at its minimum. That is a favorite test trap: students assume the firm always operates at the minimum of an SRATC. It only does at one output — the bottom of the LRAC.
A Worked Example With Three Plants
Suppose a firm faces three possible plant sizes with these SRATC at three output levels:
- Plant Small: ATC at Q=10 is $20; ATC at Q=20 is $18; ATC at Q=40 is $26
- Plant Medium: ATC at Q=10 is $24; ATC at Q=20 is $15; ATC at Q=40 is $19
- Plant Large: ATC at Q=10 is $30; ATC at Q=20 is $22; ATC at Q=40 is $14
The LRAC at each output is the minimum across rows:
- Q=10: Small wins at $20.
- Q=20: Medium wins at $15.
- Q=40: Large wins at $14.
So the LRAC passes through $20, $15, and $14 at those outputs. With more plant sizes filling in, the LRAC becomes the smooth lower envelope.
What the Shape of LRAC Tells You
The slope of LRAC is the textbook home of three named effects:
- Economies of scale: LRAC slopes down as Q rises — bigger plants are cheaper per unit because of specialization, bulk input pricing, and spreading certain costs more thinly.
- Constant returns to scale: LRAC is flat — doubling all inputs doubles output, so average cost stays the same.
- Diseconomies of scale: LRAC slopes up — bigger gets worse per unit, typically from management complexity, coordination costs, and bureaucratic drag.
The output where LRAC reaches its minimum is the minimum efficient scale (MES). A market where MES is large relative to total demand will support only a few firms (the recipe for natural monopoly or oligopoly); a market where MES is small can support many small firms.
The Practical Takeaway
Choose the curve by the time horizon of the decision. For "should we hire two more people for next month?" the firm is on its SRATC — capital is fixed. For "should we build a second plant?" the firm is choosing along the LRAC. The walkthrough on marginal cost vs. average cost shows how the marginal cost curve relates to ATC; the same crossing logic applies to either short-run or long-run curves.
Conclusion
Short-run vs. long-run cost curves separate by what is fixed: SRATC sits on top of a fixed input and is U-shaped because of diminishing returns; LRAC has nothing fixed because the firm chooses the plant, and it ends up as the lower envelope of every possible SRATC. Read the LRAC for scale economies, read the SRATC for what is happening right now in an existing plant, and remember that the two curves coincide only at the LRAC's minimum.
