Difficulty: Intermediate
Prerequisites: Parts 1 and 2 of these notes (Sections 6.1 to 6.4). Understanding of the Cobb-Douglas production function and isoquants.
Section 6.5 zooms out from one-input-at-a-time thinking to ask: what happens when the firm scales everything up at once? The answer, returns to scale, helps explain why some industries are dominated by a few giants while others are populated by small firms. Section 6.6 addresses differences in productivity across firms and over time, including technical change. These ideas bridge into Chapter 7's cost analysis, because the type of returns to scale directly determines the shape of long-run average cost curves.
Returns to scale describe what happens to output when all inputs increase proportionally. Constant returns to scale (CRS) means output rises in the same proportion; increasing returns (IRS) means it rises more; decreasing returns (DRS) means less. For a Cobb-Douglas function, the exponents (a + b) tell you which case applies. Firms can also differ in productivity even when using the same inputs, due to differences in technology, management or innovation.
Constant returns to scale (CRS)
A production function exhibits CRS if doubling all inputs exactly doubles output: f(2L, 2K) = 2f(L, K). More generally, f(xL, xK) = xf(L, K) for any positive scalar x. Think of it as building an identical second factory and getting exactly double the output.
Increasing returns to scale (IRS)
Doubling all inputs more than doubles output: f(2L, 2K) > 2f(L, K). Think of it as gaining extra efficiency from going bigger, often due to specialisation of labour or equipment in a larger plant.
Decreasing returns to scale (DRS)
Doubling all inputs less than doubles output: f(2L, 2K) < 2f(L, K). Think of it as the managerial headaches of running a bigger operation: coordination and communication become harder.
Scale elasticity (g)
For a Cobb-Douglas function q = AL^aK^b, the scale elasticity is g = a + b. If all inputs increase by 1%, output increases by g%. CRS when g = 1, DRS when g < 1, IRS when g > 1.
Homogeneous of degree g
A production function is homogeneous of degree g if f(xL, xK) = x^g f(L, K). CRS is homogeneity of degree one.
Relative productivity
The ability of one firm to produce more output than another from the same inputs. Differences arise from management quality, technology, regulations or organisational structure.
Technical change (technical progress)
A change in the production function that allows more output from the same inputs. The entire production function shifts, meaning isoquants move closer to the origin.
Neutral technical change
Technical progress that increases output without changing the ratio of marginal products at any given input combination. The isoquants shift inward but keep the same shape.
Non-neutral technical change
Technical progress that alters the relative productivity of inputs, changing the MRTS at a given input ratio. One input becomes relatively more productive than the other.
Returns to scale asks: if a firm increases all inputs by the same proportion, does output increase by the same proportion, more or less?
This is a long-run concept (all inputs vary). Do not confuse it with marginal returns, which involve changing one input while holding others fixed.
Constant returns to scale (CRS): f(2L, 2K) = 2f(L, K). Building an identical second plant doubles output. The linear (potato salad) production function q = x + y is a simple example.
Increasing returns to scale (IRS): f(2L, 2K) > 2f(L, K). Reasons include greater specialisation of labour and capital in larger plants. A single large factory may be more efficient than two small ones.
Decreasing returns to scale (DRS): f(2L, 2K) < 2f(L, K). Often attributed to the difficulty of managing, coordinating and integrating activities as the firm grows. In effect, the management input is not being scaled up proportionally.
For q = AL^aK^b, define g = a + b.
If g = 1: CRS. If g < 1: DRS. If g > 1: IRS.
g is the scale elasticity: a 1% increase in all inputs raises output by g%.
When inputs are doubled, output becomes 2^g times the original. So if g = 0.51 (U.S. tobacco), doubling inputs only raises output by a factor of 2^0.51, which is roughly 1.42.
CRS: Japanese synthetic rubber (a = 0.5, b = 0.5, g = 1.00), Japanese beer (0.60, 0.40, g = 1.00).
DRS: U.S. tobacco (0.18, 0.33, g = 0.51), Bangladesh glass (0.27, 0.45, g = 0.72).
IRS: Bangladesh fabricated metal (0.98, 0.28, g = 1.26), New Zealand mining (0.69, 0.45, g = 1.14).
Many production functions exhibit IRS at small scales (gains from specialisation), CRS at moderate scales and DRS at large scales (management difficulties).
The spacing of isoquants on a diagram reflects returns to scale. The closer together the isoquants as you move along a ray from the origin, the greater the returns to scale.
Even if every firm in an industry produces efficiently, some firms may be more productive than others.
Differences arise from better management, superior technology, access to new inventions, or fewer institutional restrictions (union work rules, regulations, discrimination).
A more productive firm's production function lies above a less productive firm's: it gets more output from the same inputs.
Technical change shifts the production function so that the same inputs produce more output. Isoquants move closer to the origin.
Neutral technical change: output rises, but the MRTS at any given input ratio stays the same. Isoquants shift inward uniformly, keeping their shape.
Non-neutral technical change: the relative productivity of inputs changes. One input becomes relatively more useful, altering the MRTS at each input combination. The shape of isoquants changes.
Examples of technical progress include new inventions, improved organisational methods, and better managerial practices.
Returns to scale test for Cobb-Douglas
q = A L^a K^b \quad \Rightarrow \quad g = a + bIf both inputs are multiplied by x, new output = x^g times original output.
g = 1: CRS
g < 1: DRS
g > 1: IRS
Homogeneity
f(xL, xK) = x^g f(L, K)A production function is homogeneous of degree g. CRS corresponds to g = 1 (homogeneity of degree one).
Returns to scale in various industries: Japanese beer and synthetic rubber exhibit CRS (g = 1.00). U.S. tobacco has strong DRS (g = 0.51), meaning doubling inputs raises output by only about 42%. Bangladesh fabricated metal has IRS (g = 1.26), so doubling inputs raises output by about 139%.
Varying returns to scale: many firms experience IRS when small (specialisation gains), CRS at moderate size, and DRS when large (coordination costs). This pattern helps explain why most industries have firms of a characteristic size range.
Students often confuse returns to scale with diminishing marginal returns. Returns to scale is about scaling all inputs simultaneously (a long-run concept). Diminishing marginal returns is about adding one input while holding others fixed (a short-run concept). A production function can exhibit diminishing marginal returns to labour and still have increasing returns to scale.
Students sometimes think DRS means the firm is inefficient. It does not. DRS is a property of the production function, not a sign of waste.
Students often forget the "all else equal" condition for returns to scale. If you double labour but hold capital fixed, that is not a test of returns to scale.
⚠️ The Cobb-Douglas returns-to-scale rule (g = a + b) is one of the most frequently tested items in this chapter.
⚠️ Be able to distinguish returns to scale (all inputs vary proportionally) from diminishing marginal returns (one input varies, others fixed).
⚠️ Know the isoquant spacing interpretation: closer together along a ray from the origin means greater returns to scale.
⚠️ Understand neutral vs. non-neutral technical change and how each affects isoquants.
⚠️ Be ready to calculate what happens to output when inputs are doubled for a given Cobb-Douglas function (multiply original output by 2^g).
Fill in the blank: For a Cobb-Douglas production function q = AL^aK^b, the returns to scale are determined by g = ______. (a + b)
True or False: Increasing returns to scale means that the marginal product of labour is always increasing. (False. IRS is about scaling all inputs proportionally, not about the marginal product of a single input.)
Fill in the blank: A production function is homogeneous of degree one if it exhibits ______ returns to scale. (constant)
True or False: Neutral technical change alters the MRTS at any given input combination. (False. Neutral technical change shifts isoquants inward but preserves the MRTS at each input ratio.)
Fill in the blank: If a Cobb-Douglas production function has g = 0.72 and a firm doubles its inputs, output rises by a factor of ______. (2^0.72, which is approximately 1.65)
Q: A Cobb-Douglas production function is q = L^0.69 K^0.45. Does it exhibit increasing, constant or decreasing returns to scale? If the firm doubles both inputs, what happens to output?
A: g = 0.69 + 0.45 = 1.14, which is greater than 1, so the production function exhibits increasing returns to scale. Doubling inputs multiplies output by 2^1.14, which is approximately 2.20. Output more than doubles.
Q: Explain why a production function can have diminishing marginal returns to labour and increasing returns to scale at the same time.
A: Diminishing marginal returns to labour means that adding one more worker (holding capital fixed) eventually yields a smaller increase in output. Increasing returns to scale means that scaling up all inputs together (both labour and capital) raises output more than proportionally. These are different experiments: one holds other inputs constant, the other scales everything. A Cobb-Douglas function with a < 1, b < 1, but a + b > 1 has both properties.
Q: A firm currently produces 100 units of output. After neutral technical change, it can produce 120 units from the same inputs. Has the MRTS at any given input combination changed?
A: No. Neutral technical change increases output uniformly without altering the MRTS at a given input ratio. The isoquants shift inward but keep the same shape and slope at each K/L ratio.
Q: Why might a firm experience increasing returns to scale at small output levels but decreasing returns to scale at large output levels?
A: At small scales, specialisation of workers and equipment generates efficiency gains that make output rise more than proportionally to inputs (IRS). At large scales, the difficulty of coordinating, communicating and managing a bigger organisation outweighs those gains, causing output to rise less than proportionally (DRS). In effect, management becomes a bottleneck that is not fully scalable.
Returns to scale directly determine the shape of long-run average cost curves in Chapter 7. CRS produces a flat LRAC, IRS produces a falling LRAC, and DRS produces a rising LRAC.
Technical change connects to the broader theme of economic growth in macroeconomics. The same production-function framework (often Cobb-Douglas) is used in growth accounting.
The varying-returns-to-scale pattern (IRS then CRS then DRS) helps explain industry structure: why some industries have many small firms and others are dominated by a few large ones.
returns to scale, constant returns to scale, CRS, increasing returns to scale, IRS, decreasing returns to scale, DRS, scale elasticity, homogeneous production function, Cobb-Douglas exponents, varying returns to scale, relative productivity, technical change, technical progress, neutral technical change, non-neutral technical change, isoquant spacing, specialisation, coordination costs, Perloff Chapter 6, intermediate microeconomics, ECN 302