Returns to scale

Topic: Production Function, Returns and Costs · NCERT: Class 12, Ch 3 "Production and Costs"

Meaning

Returns to scale describe how output changes when a firm raises all its inputs (for example, both workers and machines) in the same proportion. This can happen only in the long run, the period in which every input, including plant size, can be changed.

• Formula: if q = f(x₁, x₂) and both inputs are multiplied by t (t > 1):
• f(tx₁, tx₂) = t·f(x₁, x₂) → constant returns to scale (CRS)
• f(tx₁, tx₂) > t·f(x₁, x₂) → increasing returns to scale (IRS)
• f(tx₁, tx₂) < t·f(x₁, x₂) → decreasing returns to scale (DRS)

Returns to scale explain why the long-run cost curve is U-shaped. They also show whether bigger firms produce more cheaply. They are the basis of growth accounting and the measurement of productivity.

Explanation

How it works: the doubling test

• Take a firm. Double both its labour (L) and its capital (K), so t = 2. Then compare the new output with the old output.
• Output exactly doubles → CRS
• Output more than doubles → IRS

• Output less than doubles → DRS

• Factor proportions stay the same. The ratio of workers to machines does not change. Only the size of the operation changes.

• Method for any pair of input bundles:
• Find t (new input ÷ old input).

• Compare the new output with t × old output. If it is greater, the firm has IRS. If it is equal, CRS. If it is smaller, DRS.

• NCERT Table 3.1 (L, K → output):

• (1,1) → 1 and (2,2) → 10. Inputs doubled, output rose 10 times → IRS.
• (2,2) → 10 and (4,4) → 50. Inputs doubled, output rose 5 times → IRS.
• (3,3) → 30 and (6,6) → 60. Inputs doubled, output exactly doubled → CRS.

The cost twin: returns to scale and LRAC

• LRAC (long-run average cost) is the cost per unit of output when the firm can change all inputs. Assume input prices stay fixed.

Type	What happens	LRAC	Name
IRS	Cost doubles, output more than doubles	Falls	Economies of scale
CRS	Cost and output both double	Constant (flat)	—
DRS	Cost doubles, output rises by less	Rises	Diseconomies of scale

• Why the LRAC is U-shaped: a typical firm moves through IRS → CRS → DRS as it grows.
• Early growth brings IRS, so the LRAC falls.
• In the middle range the firm has CRS, so the LRAC is flat.
• A very large firm faces DRS, so the LRAC rises.

What causes IRS and DRS

• Causes of IRS:
• Specialisation and division of labour. In a big unit, each worker does one narrow task and gets faster at it.
• Indivisible large machines. Some machines, such as a blast furnace or a rolling mill, come only in large sizes. A small firm cannot buy half of one. The machine is fully used only at large output.

• Dimensional economies. Capacity depends on area or volume, but material cost depends on surface or length.

  • If you double a pipe's diameter, its cross-section area (πr²) becomes about 4 times larger.
  • The metal needed for its wall only doubles.
  • So carrying capacity more than doubles while material cost only doubles. This is IRS.

• Causes of DRS:

• Limits on management and coordination. A very large firm has more layers of managers. Decisions are slower and supervision is weaker.
• Scarce natural inputs. Good land, mineral deposits and water cannot always grow in the same proportion. Firms then have to use poorer-quality inputs.

Cobb-Douglas: reading returns to scale from the exponents

• Form: q = x₁^α · x₂^β, often written Q = A·L^α·K^β.
• A = level of technology or efficiency.

• α and β are positive constants.

• Scaling test: (tx₁)^α (tx₂)^β = t^(α+β) · q₀

• α + β = 1 → CRS. Britannica confirms that exponents adding up to 1 give constant returns to scale [6].
• α + β > 1 → IRS

• α + β < 1 → DRS

• α and β are output elasticities. An output elasticity is the % change in output when one input rises by 1% and the other inputs stay fixed.

• Example: if α = 0.75, a 10% rise in labour raises output by about 7.5%.

• The sum of the elasticities gives the returns to scale. The literature writes this as ε = ε_L + ε_K, where ε = 1 means CRS [5].

• Worked examples:

• CRS (NCERT Ex. 28): Q = 5L^½K^½ at L = 100, K = 100 gives 5 × 10 × 10 = 500. At (200, 200), Q = 5 × 200 = 1,000. Output exactly doubles.
• IRS (NCERT Ex. 29): Q = 2L²K² at (5, 2) gives 2 × 25 × 4 = 200. At (10, 4), Q = 2 × 100 × 16 = 3,200 = 2⁴ × 200. Output rises 16 times, because the exponents add up to 4.
  • At (0, 10), Q = 0. In a multiplicative form, every input is essential. If any input is zero, output is zero.

• DRS: q = L^0.3K^0.3 has α + β = 0.6. Doubling both inputs raises output only 2^0.6 ≈ 1.52 times, an increase of about 52%.

• Original study (1928): Charles Cobb (a mathematician) and Paul Douglas (an economist) proposed the function [6]. They fitted it to US manufacturing data for 1899–1922.

• Labour's exponent was about 0.75 and capital's about 0.25.
• The sum is 1, so the result showed CRS. It also matched labour's actual share of about three-quarters of income at that time.

In India

• Returns to scale are the basis of productivity measurement.
• Growth accounting splits output growth into two parts: growth from more inputs, and growth from better use of inputs. The second part is called TFP (total factor productivity), which is output growth that input growth cannot explain [4].
• Formula (under CRS): g_Q = α·g_L + β·g_K + g_A, where g_A is TFP growth, also called the Solow residual.
• The Solow method needs only CRS and perfect competition [5].

• Illustrative example: output grows 7%. Labour grows 2% with a share of 0.6, and capital grows 8% with a share of 0.4.

  • The inputs contribute 0.6 × 2 + 0.4 × 8 = 1.2 + 3.2 = 4.4%.
  • TFP growth = 7 − 4.4 = 2.6%.

• India KLEMS database (RBI):

• KLEMS stands for Kapital, Labour, Energy, Materials and Services. These are the inputs it tracks for each industry [3].
• It measures labour productivity and TFP for Indian industries and the whole economy from 1980-81 onwards. It uses data from CSO, NSSO, ASI and Input-Output tables [3].
• TFP is measured as a residual: what is left after removing the growth of labour, capital and intermediate inputs [3].

• The 8 July 2024 update covers 27 industries for 1980-81 to 2022-23, with a Data Manual 2024. It also gives a provisional economy-wide estimate for 2023-24 [2].

• Indian examples of IRS: steel, cement and electronics assembly. In these industries small plants have higher unit costs than large ones.

Don't confuse with

• Returns to a factor (Law of Variable Proportions):
• This is a short-run idea. One input varies while the others stay fixed, so factor proportions change.

• Returns to scale are a long-run idea. All inputs vary together in the same proportion.

• Diminishing marginal product (MP) does not mean DRS. MP is the extra output from one more unit of an input, with other inputs held fixed.

• In Q = 5L^½K^½, keep K = 100 and raise L from 100 to 400. MP_L falls from 2.5 to 1.25, and output only doubles although labour rose 4 times.

• Yet the same function has CRS.

• Economies of scale are the cost side of IRS. Returns to scale describe the physical input–output relation. Economies of scale describe the falling LRAC that follows from it when input prices are fixed.

• TFP growth is not returns to scale. TFP is a shift in the whole production function (a rise in A). Returns to scale describe movement along the function when inputs grow.

Prelims Hooks

• Returns to scale mean all inputs change in the same proportion. The idea applies only in the long run, while the Law of Variable Proportions is a short-run idea.
• Cobb-Douglas q = x₁^α x₂^β: α + β = 1 → CRS; > 1 → IRS; < 1 → DRS. α and β are output elasticities. Under CRS with perfect competition, they also equal factor income shares [5].
• Trap: "Diminishing marginal product implies decreasing returns to scale" is false. Q = 5L^½K^½ has a diminishing MP_L but CRS.
• Cost twin: IRS ↔ falling LRAC (economies of scale); CRS ↔ flat LRAC; DRS ↔ rising LRAC.
• Cobb and Douglas (1928) used US manufacturing data for 1899–1922. They found labour ≈ 0.75 and capital ≈ 0.25, which gives CRS.
• The India KLEMS database is hosted by the RBI. The July 2024 update covers 27 industries for 1980-81 to 2022-23 [2]. The Solow residual (TFP growth) assumes CRS + perfect competition [5].

Mains Points

• Scale and Indian manufacturing (GS-III, industrial policy):
• Where IRS exists (steel, cement, electronics assembly), small plants have higher unit costs. They lose to larger foreign rivals.

• Policies that help firms grow can lower unit costs and raise export competitiveness. Examples are easier labour and land rules, infrastructure for industrial clusters, and incentives linked to output.

• Input-driven vs TFP-driven growth:

• Growth that comes only from adding more capital and labour runs into diminishing returns.
• Growth that lasts needs TFP gains from technology, skills, better allocation of resources and good institutions. RBI's KLEMS data (27 industries, up to 2022-23, provisional for 2023-24) shows which sectors are becoming more productive [2][3].

• Caveat: exponents equal factor shares only under CRS and perfect competition [5]. When firms have market power or IRS exists, TFP may be mis-measured. This matters in debates that use a falling labour income share as evidence of inequality.

• DRS and natural limits (GS-III, agriculture and resources):

• Scarce land and water limit how far farm output can be scaled up by using more of every input.
• This supports productivity led by technology (better seeds, more efficient irrigation) over simply expanding the area under farming.

Related concepts

• Increasing returns to scale
• Constant returns to scale
• Decreasing returns to scale
• Cobb-Douglas production function

Read more

• Returns to scale and the Cobb-Douglas production function