Microeconomics › Section G · Production, costs, and industry structure › Topic G.5
Costs in the long run and economies of scale
In the long run nothing is fixed, so the firm faces no inherited plant - it chooses one. The long-run average cost curve records, for every output, the cheapest plant available, and its U-shape answers a question the short run cannot even ask: how big should this firm be?
No fixed costs, all choices
In the long run, every cost is variable: leases expire, equipment can be bought or sold, whole plants can be built or closed. With nothing fixed, the firm faces two linked choices - how much output to make, and how to make it: which plant size and which production technology. Economists use 'technology' broadly here: any method of combining inputs, not a specific invention. The long-run cost curves summarise the results of all those choices - and their most important use is as a planning curve: today's short-run costs are the residue of yesterday's long-run decisions.
The long-run average cost curve
Long-run average cost (LRAC) is the lowest cost per unit at which each output can be produced when the firm is free to choose the best plant for it. Picture the family of short-run ATC curves, one per plant size. The LRAC is the envelope of them: the lower boundary traced by taking, at each output, whichever plant does that output cheapest. Each short-run curve touches the long-run one at the output it is best suited for and sits above it elsewhere. A firm in the short run is stuck on its one inherited curve; a firm planning for the long run slides along the envelope.
Economies of scale
Where LRAC falls as output rises, the firm enjoys economies of scale: bigger is cheaper per unit. The sources are the increasing returns of the previous topic, now in cost terms - finer specialisation at large volume, indivisible assets (a furnace, a logistics network, one accounting system) spread over more output, and bulk buying power over inputs. 'Bigger is cheaper' is not a slogan but an arithmetic fact in this range: the same list of forces explains why a national supermarket chain buys groceries for less than a corner shop can.
Diseconomies of scale
Past some size, LRAC turns upward: diseconomies of scale, the cost of bigness. Coordination multiplies - layers of management, slow decisions, costly monitoring - and motivation thins as each worker's contribution shrinks toward invisibility in the total. Between the two stretches there is often a flat middle: constant returns to scale, where the firm replicates its proven recipe, and unit costs hold steady. The whole journey - falling, flat, rising - gives LRAC its typical shallow U, usually flatter at the bottom than a short-run ATC curve.
Worked example: LRAC across five plant sizes
A bakery chain plans its scale. Five possible plants and their long-run total costs: output 100 for $220; 300 for $480; 600 for $720; 900 for $1,170; 1,200 for $1,800. 1. Unit costs: divide - \(2.20, 1.60, 1.20, 1.30, 1.50\) dollars. 2. Read the shape: LRAC falls to 600 (economies of scale), bottoms at $1.20, then rises (diseconomies). 3. The efficient plant size is 600 loaves a day - the minimum efficient scale: the smallest output at which unit costs hit their floor. 4. Check: from 600 to 900 costs creep up only slowly (1.20 to 1.30); beyond 900, coordination bites harder. 5. Plan: to bake 1,200 loaves a day, two 600-loaf plants ($1.20 a loaf, $1,440 in all) beat one 1,200-loaf plant ($1.50, $1,800). Assumptions: the table already reflects the best technology at each scale.
Minimum efficient scale and market structure
The minimum efficient scale (MES) - the smallest output at which unit costs bottom out - quietly shapes entire industries. If MES is small relative to total market demand, many firms can operate at efficient size, and the industry hosts crowds: cafés, hairdressers, plumbers. If MES is huge, only a handful of firms fit the market efficiently - aircraft engines, chip fabrication - and in the extreme, where one firm's MES exhausts the whole market, a natural monopoly emerges, the subject of a later section. Cost curves do not just describe a firm; they predict how many firms an industry can hold.
Choosing a technology
Long-run planning also picks the method, and the best method depends on the scale aimed at. Two ways to print books: an artisanal studio with low fixed costs and high variable cost per book, or an automated line with huge fixed costs and tiny variable cost. Small runs favour the studio - the automated line's setup would bankrupt a print run of hundreds. Huge runs favour the line - its setup spread over tens of thousands of books becomes pennies. The general principle: capital-intensive technologies are cheap only when volume spreads their fixed costs. Input prices matter too: when wages rise relative to machine costs, the least-cost method shifts towards machines, which is why firms in high-wage economies tend to use more capital per worker.
Worked example: two technologies, two volumes
A print studio compares methods for a projected run. Studio method: fixed costs $2,000 plus $3 per book. Automated line: fixed costs $50,000 plus $0.80 per book. 1. At 1,000 books: studio unit cost \((2000 + 3000) \div 1000 = 5\) dollars; line unit cost \((50000 + 800) \div 1000 = 50.80\). The studio wins overwhelmingly - the line's setup overwhelms a small run. 2. At 50,000 books: studio \((2000 + 150000) \div 50000 = 3.04\); line \((50000 + 40000) \div 50000 = 1.80\). The line wins - its setup is spread thin and its low variable cost takes over. 3. The crossover: where \(2000 + 3q = 50000 + 0.8q\), at about 21,800 books, the line's higher fixed cost is exactly offset by its lower variable cost. 4. Conclusion: there is no 'better' technology in the abstract - only the best one for the planned volume. Assumptions: only these two methods, given input prices.
Short-run curves inside the long-run curve
When the plan is executed, the long run crystallises into a short run. The chosen plant becomes the fixed capital; its short-run ATC curve is the one the firm now lives on, with its own U-shape from the spreading-and-congestion forces of the short-run costs topic. If demand drifts from the plant's sweet spot, the firm is stuck paying the difference until the next long-run decision. This nesting explains an everyday business pattern: firms outgrow plants, absorb rising short-run unit costs for a while, then jump to a new curve - and the LRAC is the stitched line of all those jumps.
Common mistakes to avoid
First, confusing economies of scale with diminishing marginal returns: scale is about changing ALL inputs and plant size; diminishing returns is a short-run, one-input story. Second, drawing LRAC as one short-run U: the envelope is flatter and never lies above any of its short-run curves. Third, reading the flat middle as 'no economies': constant returns are still the efficient zone to sit in. Fourth, assuming bigger is always better: diseconomies of scale are real, and the 1,200-loaf plant costs more per loaf than two 600-loaf plants. Fifth, choosing a technology by fixed cost alone: the volume it serves decides the whole comparison.
Key takeaways
In the long run all costs are variable, and LRAC records the cheapest unit cost at each output - the envelope of the short-run ATC curves, a planning curve made of yesterday's choices. Falling LRAC is economies of scale (specialisation, indivisible assets, bulk buying); flat is constant returns; rising is diseconomies of scale (the cost of bigness). The bakery chain's plants gave unit costs of 2.20, 1.60, 1.20, 1.30, 1.50 dollars - minimum efficient scale at 600 loaves, and two efficient plants beating one oversized one. MES predicts industry structure: small MES means many firms, huge MES means few, extreme MES means natural monopoly. Technology choice follows volume: the studio wins at 1,000 books ($5 vs $50.80), the automated line at 50,000 ($1.80 vs $3.04).
Flashcards
Flashcard
The horizon at which no cost is fixed
Test your understanding
Question 1 of 5
A bakery chain's five plant options have long-run total costs of $220, $480, $720, $1,170 and $1,800 for outputs of 100, 300, 600, 900 and 1,200 loaves. What is the minimum efficient scale?
Question 2 of 5
What does a falling long-run average cost curve indicate?
Question 3 of 5
The print studio (fixed $2,000, $3 per book) faces an automated line (fixed $50,000, $0.80 per book). At 50,000 books, which is cheaper per unit?
Question 4 of 5
Why is the long-run average cost curve called the envelope of short-run curves?
Question 5 of 5
Market demand in an industry equals roughly one firm's minimum efficient scale, and costs keep falling over the whole relevant range. What structure does this predict?
