EconTutor – Best Economics tutors | Microeconomics, Macroeconomics, Econometrics

Take any firm you know — a pizza kitchen, a bicycle workshop, a car plant — and strip away everything except one question: how much output can it get from the inputs it hires? The production function is the answer, and it is where intermediate microeconomics stops being about markets and starts being about firms. This page teaches production functions and returns to scale properly; it is also the place to start if you are searching for the best economics tutors at university level.

1 · A production function is a boundary, not a description

Write it as Q = f(K, L): output Q as a function of capital K and labour L. Read it carefully. It gives the maximum output each input combination can produce, not merely an output it might produce. Technical efficiency is baked into the definition — waste puts you below the function, never on it.

The function also fixes the vocabulary for time. The short run is any period in which at least one input cannot be adjusted — usually capital. The long run is the period in which everything can be adjusted. Neither is a number of months. A pizza kitchen can change staffing next week but needs a new lease to add an oven, so its short run is defined by the oven, not by the calendar. Hold one input fixed and you are doing short-run analysis; free them all and you are doing long-run analysis. Every result on this page belongs to one of those two boxes, and examiners penalise answers that mix them.

2 · Total, average and marginal product

Fix K and vary L. Three curves describe what happens.

Total product is just Q itself as L rises. Average product, AP_L = Q/L, is output per worker. Marginal product, MP_L = ∂Q/∂L, is the extra output from the last unit of labour. Geometrically: MP_L is the slope of the total product curve at a point, and AP_L is the slope of a ray from the origin to that point.

One relationship between them does most of the exam work. The marginal pulls the average. If the last worker produces more than the current average, the average rises; if less, the average falls. Your course grades work the same way — a mark above your average lifts it, a mark below drags it down. The consequence is exact: MP_L crosses AP_L at the maximum of AP_L. When you sketch these curves, that crossing is the first thing a marker checks.

3 · Diminishing marginal returns

The law of diminishing marginal returns says: holding at least one input fixed, successive units of the variable input eventually add less and less to output. Eventually — not immediately. The second worker in an empty kitchen may well add more than the first, because two people can specialise: one on dough, one on the oven. But the capital is fixed. Keep hiring and each new worker shares the same two ovens with more colleagues. At some point MP_L peaks and starts to fall. That point is the inflection of the total product curve — where it stops steepening and starts flattening.

Be precise about what the law is not. It is not about worker quality; every worker is identical in this model. It is not a long-run claim; it holds because something is fixed. And falling MP_L does not mean falling output — output keeps rising as long as MP_L stays positive. Marking schemes are strict about all three.

4 · The three stages of production

Diminishing returns carve the short run into three stages, with boundaries you can compute.

• Stage I runs from zero labour to the maximum of AP_L. Here MP_L lies above AP_L, so every extra worker raises average output per worker. The fixed capital is underused.
• Stage II runs from the maximum of AP_L to the maximum of total product — the point where MP_L hits zero. Marginal product is positive but falling.
• Stage III is everything beyond. MP_L is negative: extra workers now crowd the fixed capital so badly that total output falls.

A rational firm produces only in Stage II. Stage III is easy to dismiss — you would not hire a worker who reduces output even at a wage of zero. Stage I is subtler: while average product is still rising, expanding the workforce raises productivity across the board, so stopping there leaves the fixed input doing less than it could. The profit-maximising choice of L — where the wage equals the marginal revenue product — always lands inside Stage II. The stages tell you where the answer lives before you compute it.

5 · Returns to scale — the long-run question

Now free every input and ask a different question: if the firm scales all inputs by the same factor t > 1, what happens to output?

• Constant returns to scale (CRS): output scales by exactly t. f(tK, tL) = t·f(K, L).
• Increasing returns to scale (IRS): output scales by more than t — think specialisation and indivisible machinery.
• Decreasing returns to scale (DRS): output scales by less than t — think managers drowning in coordination.

For the Cobb–Douglas family, Q = A·K^α·L^β, there is a one-line test. Scale both inputs by t:

f(tK, tL) = A·(tK)^α·(tL)^β = t^α+β·A·K^α·L^β = t^α+β·Q.

So the sum of the exponents decides everything: α + β = 1 gives CRS, α + β > 1 gives IRS, α + β < 1 gives DRS. Check it with numbers. Take Q = 10·K^0.5·L^0.5 with K = 4 and L = 9: Q = 10 × 2 × 3 = 60. Double both inputs to K = 8, L = 18: Q = 10·√(8 × 18) = 10·√144 = 120. Exactly double — CRS, as α + β = 1 predicts. If the exponents summed to 1.2 instead, doubling inputs would scale output by 2^1.2 ≈ 2.30; if they summed to 0.8, by 2^0.8 ≈ 1.74.

Here is the most-tested distinction in this topic: diminishing marginal returns and decreasing returns to scale are different claims. Diminishing returns is a short-run statement about one input against a fixed input. Returns to scale is a long-run statement about all inputs moving together. The two coexist happily: Q = 10·K^0.5·L^0.5 has constant returns to scale, yet MP_L = 5·(K/L)^0.5 falls as L rises with K fixed. If an exam answer treats the two as the same thing, the marks are gone.

Worked example — a pizza kitchen with two ovens

A pizza kitchen has two ovens and one prep counter on a fixed lease — capital is locked, so this is the short run. Output depends on the number of workers per shift: Q = 12L² − L³ pizzas per day.

Step 1 — Tabulate total product. Substitute L = 1, …, 9:

L (workers)	1	2	3	4	5	6	7	8	9
Q (pizzas/day)	11	40	81	128	175	216	245	256	243

Output climbs steeply, then slowly, then falls. The three stages are already visible in the numbers.

Step 2 — Derive the marginal and average product. Differentiate for MP_L = dQ/dL = 24L − 3L². Divide by L for AP_L = 12L − L². At L = 3, for instance: MP_L = 72 − 27 = 45 and AP_L = 36 − 9 = 27. The marginal worker beats the average, so the average is rising.

Step 3 — Find where diminishing returns begin. Maximise MP_L: dMP_L/dL = 24 − 6L = 0 gives L = 4, where MP_L = 96 − 48 = 48. Up to four workers, specialisation dominates. Beyond four, crowding at the two ovens sets in and each extra worker adds less than the one before.

Step 4 — Find the Stage I/II boundary. Maximise AP_L: dAP_L/dL = 12 − 2L = 0 gives L = 6, where AP_L = 72 − 36 = 36. Check the crossing rule: MP_L(6) = 144 − 108 = 36. Marginal equals average exactly at the average’s peak. ✓

Step 5 — Find the Stage II/III boundary. Set MP_L = 0: 24L − 3L² = 3L(8 − L) = 0 gives L = 8. Total product peaks at Q = 12(64) − 512 = 256 pizzas. Stage II is the range 6 ≤ L ≤ 8.

Step 6 — Perturbation. The owner hires a ninth worker anyway. Output falls from 256 to Q(9) = 972 − 729 = 243 — thirteen pizzas a day lost, and MP_L(9) = 216 − 243 = −27. Nine people around two ovens get in each other’s way. This is Stage III.

Step 7 — Resolve. The ninth worker goes. The rational hiring range is Stage II: between six and eight workers. Where exactly depends on the wage and the price of pizza — the firm hires until the wage equals the marginal revenue product — but no wage, however low, justifies leaving Stage II.

Step 8 — Interpretation. Notice what the fixed ovens did. They created the whole shape: early specialisation gains (Stage I is worth pushing through), then crowding (Stage II, where the firm lives), then congestion so severe that labour destroys output (Stage III). In the long run the kitchen can add a third oven — and then the question changes from diminishing returns to returns to scale.

Comfortable with the algebra but not the diagram — or the other way round? Closing that gap is what the best economics tutors do in a first one-on-one session: your module’s problem sets, worked until the curves and the calculus are one model. Book a trial session.

Practice

Q1. A workshop’s short-run production function is Q = 18L² − L³.
(a) At what L do diminishing marginal returns set in?
(b) Where does Stage II begin and end?
(c) What is the maximum output?

Q2. Classify the returns to scale of: (a) Q = 4·K^0.3·L^0.5; (b) Q = 2·K^0.5·L^0.5; (c) Q = K^0.8·L^0.7. For (b), verify numerically with K = 9, L = 16 and then both inputs doubled.

Q3. A firm has Q = 5·K^0.4·L^0.6.
(a) Compute Q at K = 32, L = 32.
(b) Both inputs double. Use the homogeneity result to find the new output without recomputing the powers.
(c) Show that MP_L falls as L rises even though returns to scale are constant.

Answers: Q1 (a) MP = 36L − 3L² peaks at L = 6; (b) Stage II runs from the AP maximum at L = 9 (AP = MP = 81) to the TP maximum at L = 12; (c) Q(12) = 864. Q2 (a) 0.3 + 0.5 = 0.8 — decreasing; (b) 1 — constant: Q = 2·3·4 = 24, and at K = 18, L = 32, Q = 2·√576 = 48, exactly double; (c) 1.5 — increasing. Q3 (a) 32^0.4·32^0.6 = 32, so Q = 160; (b) α + β = 1, so doubling scales Q by 2¹: Q = 320; (c) MP_L = 3·(K/L)^0.4, which falls as L rises with K fixed — CRS and diminishing marginal returns coexist.

Key takeaways

• A production function gives the maximum output from each input bundle; the short run fixes at least one input, the long run frees them all.
• MP_L is the slope of the total product curve; AP_L is the slope of the ray from the origin. The marginal pulls the average, so MP_L crosses AP_L at AP_L‘s maximum.
• The three stages: Stage I ends at maximum AP_L, Stage II ends where MP_L = 0 (maximum total product), Stage III has negative MP_L. Rational firms operate in Stage II only.
• For Cobb–Douglas Q = A·K^α·L^β, returns to scale are read off α + β: equal to 1 constant, above 1 increasing, below 1 decreasing.
• Diminishing marginal returns (short run, one input) and decreasing returns to scale (long run, all inputs) are different concepts — a CRS function still has diminishing MP_L.

Why students choose our economics tutoring

• One-on-one format: every session is private and built around your module — your problem sets, your lecture notes, your department’s notation.
• Intermediate-level specialists: our tutors teach calculus-based micro the way your course does, from product curves through cost functions and beyond.
• Online across the UK: live online sessions serve students in Bristol, Nottingham, Manchester, Cardiff, Lancaster, Surrey and Reading — the same tutor every week, no travel.
• Exam-first preparation: sessions work through past papers with marking schemes in view, because distinctions like diminishing returns versus returns to scale are exactly where marks are won and lost.

FAQ

Q: What is a production function in economics?
A: It is the relationship giving the maximum output a firm can produce from each combination of inputs, written Q = f(K, L). It summarises the firm’s technology and is the starting point for all cost and supply analysis.

Q: What is the difference between diminishing returns and decreasing returns to scale?
A: Diminishing marginal returns is a short-run idea: add more of one input against a fixed input and its marginal product eventually falls. Decreasing returns to scale is a long-run idea: scale all inputs together and output rises less than proportionally. A function can have constant returns to scale and diminishing marginal returns at the same time.

Q: How do you find the three stages of production?
A: Compute AP_L and MP_L from the production function. Stage I ends where AP_L is maximised (which is where MP_L = AP_L), Stage II ends where MP_L = 0 (where total product is maximised), and Stage III is everything beyond.

Q: Why do firms only produce in Stage II?
A: In Stage III extra labour reduces output, so no wage justifies hiring there. In Stage I average product is still rising, so the fixed input is underused and expanding is always worthwhile. That leaves Stage II as the only range where the profit-maximising choice can sit.

Q: How do you check returns to scale for a Cobb–Douglas function?
A: Add the exponents. For Q = A·K^α·L^β, scaling both inputs by t scales output by t^α+β. So α + β = 1 means constant, greater than 1 increasing, less than 1 decreasing returns to scale.

Q: Do I need calculus for production theory?
A: At intermediate level, yes — marginal product is a partial derivative, and finding stage boundaries means setting derivatives to zero, as in the worked example above. If your course is intermediate micro, expect to differentiate; tell us your module when booking and sessions will match its level.

Book the best economics tutors for intermediate micro

Whether product curves clicked today or you are still staring at the cubic, a one-on-one session turns this topic from formulas you memorise into a model you can drive. Tell us your university and module, and we will match you with the right tutor this week.