Returns to Scale Explained: Increasing, Constant and Decreasing

Returns to scale describe what happens to output when a firm increases every one of its inputs by the same proportion: if doubling all inputs more than doubles output, the firm has increasing returns to scale; if output exactly doubles, it has constant returns to scale; and if output less than doubles, it has decreasing returns to scale. It is a long-run idea about the production function itself, which is what separates it from economies of scale (a cost idea) and from diminishing marginal returns (one input changes while another stays fixed).

This guide runs the test on a table and on a formula, step by step, then draws the exact line between the three ideas. For the one-line definition, see returns to scale.

The three cases at a glance

The test always has the same shape. Multiply every input (all the labor, all the capital, all the materials) by the same factor, then compare how much output grew with how much the inputs grew.

Any common factor works, not just 2: if every input rises by 50 percent, constant returns means output rises by exactly 50 percent. For the last column, think of one "bundle" as a fixed recipe of inputs, say 10 workers plus 4 machines.

How to test for returns to scale

1. Write down the starting inputs and output. For example, 10 workers and 4 machines make 100 units.

2. Scale every input by the same factor. If the factor is 2, that means 20 workers and 8 machines. If labor doubles but capital triples, the input mix changed, and this is no longer a returns to scale question.

3. Find the new output and divide by the old output. That ratio is the output multiplier.

4. Compare the output multiplier with the input multiplier. Bigger means increasing, equal means constant, smaller means decreasing.

The returns to scale calculator walks through the same four steps and adds the elasticity of scale, a single number that summarizes the result.

Worked example 1: a table that passes through all three cases

A furniture workshop grows by adding whole bundles of 10 workers and 4 machines. Its output at each size:

Work each row.

From 1 to 2 bundles. Workers go from 10 to 20 (20 / 10 = 2) and machines from 4 to 8 (8 / 4 = 2), so every input doubled. Output goes from 100 to 240, and 240 / 100 = 2.4. Since 2.4 is greater than 2, this stretch shows increasing returns to scale. Check it a second way: constant returns would have produced 2 x 100 = 200 chairs, and 240 beats 200.

From 2 to 3 bundles. Workers go from 20 to 30 (30 / 20 = 1.5) and machines from 8 to 12 (12 / 8 = 1.5). Output goes from 240 to 360, and 360 / 240 = 1.5. The multipliers match, so this stretch shows constant returns to scale. Second check: 1.5 x 240 = 360 exactly.

From 3 to 4 bundles. Workers go from 30 to 40 (40 / 30 = 1.33) and machines from 12 to 16 (16 / 12 = 1.33). Output goes from 360 to 420, and 420 / 360 = 1.17. Since 1.17 is less than 1.33, this stretch shows decreasing returns to scale. Second check: constant returns would have given 360 x 4/3 = 480 chairs, and 420 falls short of 480.

Two lessons hide in this table. First, returns to scale can change as a firm grows, so always compare each step with the step before it, not just the first and last rows. Second, output per bundle tells the same story: 100 / 1 = 100, then 240 / 2 = 120, then 360 / 3 = 120, then 420 / 4 = 105. It rises, holds steady, then falls.

Worked example 2: testing a production function

Many intermediate courses give you a formula instead of a table. The method is identical: replace every input with t times that input and see what happens to Q.

Q = K^0.5 x L^0.5. Start with K = 16 and L = 4. Then Q = 16^0.5 x 4^0.5 = 4 x 2 = 8. Double both inputs to K = 32 and L = 8. Then Q = 32^0.5 x 8^0.5 = (32 x 8)^0.5 = 256^0.5 = 16. Output went from 8 to 16, exactly double, so constant returns to scale.

Q = K^0.6 x L^0.6. Start with K = 1 and L = 1, so Q = 1 x 1 = 1. Double both: Q = 2^0.6 x 2^0.6 = 2^1.2, which is about 2.30. Output rose 2.30 times while inputs rose 2 times, so increasing returns to scale.

Q = K^0.3 x L^0.5. Start with K = 1 and L = 1, so Q = 1. Double both: Q = 2^0.3 x 2^0.5 = 2^0.8, which is about 1.74. Output rose 1.74 times while inputs rose 2 times, so decreasing returns to scale.

Notice the shortcut. For any function of the form Q = A x K^a x L^b, scaling both inputs by t multiplies output by t^(a + b). So you only need the sum of the exponents: above 1 is increasing, exactly 1 is constant, below 1 is decreasing. The constant A has no effect on the answer. This family is called the Cobb-Douglas production function.

Why returns to scale differ

Increasing returns often come from physical geometry and from splitting work into narrow tasks. Doubling the height and diameter of a storage tank, with the same wall thickness, multiplies the material in its walls by about 4 but its capacity by 8, so capacity per unit of material doubles. A larger operation can also give each worker one specialized job.

Constant returns have a simple logic called replication: if a firm can build an exact copy of its plant, staffed the same way, it should get exactly twice the output. That is why constant returns is the natural benchmark.

Decreasing returns usually appear when some input cannot truly be doubled. The most common candidate is management attention: twice the workers and twice the machines still answer to one leadership team, and coordination slips.

Returns to scale vs economies of scale

Returns to scale is about physical quantities: inputs in, output out. Economies of scale are about cost: what happens to long-run average cost as output grows. The economies of scale guide covers the long-run average cost curve and its causes in detail, so here the focus is only on how the two ideas connect.

They line up when input prices stay fixed. Suppose the workshop pays $200 per worker per week and $100 per machine per week. One bundle (10 workers and 4 machines) costs 10 x $200 + 4 x $100 = $2,000 + $400 = $2,400.

Check each average: 2,400 / 100 = 24; 4,800 / 240 = 20; 7,200 / 360 = 20; 9,600 / 420 = 22.86. Increasing returns gave falling average cost (economies of scale), constant returns gave flat average cost, and decreasing returns gave rising average cost (diseconomies of scale).

They come apart when input prices change with size. Say that at 3 bundles the workshop is big enough to win a bulk deal that cuts the machine rental to $75. Total cost becomes 30 x $200 + 12 x $75 = $6,000 + $900 = $6,900, and average cost is 6,900 / 360 = $19.17. Average cost fell from $20.00 to $19.17 between 2 and 3 bundles, an economy of scale, even though the production function showed constant returns over that step. The reverse can happen too: a firm that grows large enough to push up local wages can see average cost rise with constant returns. So increasing returns to scale is one source of economies of scale, not a synonym for it. The side-by-side returns to scale vs economies of scale comparison puts the two definitions next to each other.

Returns to scale vs diminishing marginal returns

The second confusion is with diminishing marginal returns, which the diminishing returns guide covers in full. The difference is which inputs change.

Go back to the workshop at 2 bundles: 20 workers, 8 machines, 240 chairs. Now hold the 8 machines fixed and change only labor. With the 8 machines fixed, 10 workers make 150 chairs, 20 workers make 240 and 30 workers make 290 (the 150 and 290 are assumed figures for illustration). The second group of 10 workers added 240 minus 150 = 90 chairs, and the third group added only 290 minus 240 = 50. Each extra group adds less while capital stays frozen. That is diminishing marginal returns. Yet when the owner scaled both inputs together to 3 bundles, output rose by exactly 50 percent. The same firm faces diminishing marginal returns in the short run and constant returns to scale in the long run at once, and nothing about that is contradictory.

The time frames matter because the short run is defined by having a fixed input. The short run vs long run guide explains that line and what firms can and cannot change on each side of it.

Where this fits in AP Microeconomics

AP Microeconomics Unit 3 covers long-run production costs, where the long-run average cost curve falls, flattens and rises. Many textbooks label those stretches economies of scale, constant returns to scale and diseconomies of scale, mixing cost and production language on one graph. That works because, with fixed input prices, each stretch of the cost curve mirrors a stretch of the production function. The production costs module develops the full set of curves, and the production costs sandbox lets you change output, fixed cost and the wage and watch MC, ATC and AVC respond.

See it move

Make it move

The short-run cost curves, with marginal cost through the minimum of AVC and ATC. Tap it to drag the curves, open the full version, or put it on your own site free, or turn it into a five-minute class activity.

Common mistakes

Scaling only one input. If capital stays fixed, you are measuring marginal returns, not returns to scale. Every input must rise by the same factor.

Comparing the wrong multipliers. Compare the output multiplier with the input multiplier, not with 1. Output rising from 100 to 180 when inputs doubled is growth, but it is decreasing returns to scale because 1.8 is less than 2.

Treating "decreasing returns to scale" and "diminishing marginal returns" as the same phrase. One is long run with all inputs moving; the other is short run with one input moving.

Calling returns to scale and economies of scale synonyms. Returns to scale is physical; economies of scale is cost. Input price changes can separate them.

Adding the exponents of only some inputs. In Q = K^0.3 x L^0.5 x M^0.4, the sum is 0.3 + 0.5 + 0.4 = 1.2, so returns are increasing. Leaving out M gives the wrong answer of 0.8.