Session 11: Cost Curves, Market Structure, and the Logic of Profit Maximisation

Part of the Microeconomics Knowledge System. This post covers Session 11 — the bridge between cost theory and market structure analysis.

Session 11 does two things. It closes out cost theory with two ideas that most courses underexplain — economies of scope and the envelope theorem. Then it opens market structure analysis by giving you the universal engine that runs through every market form: π = R − C, the MR = MC rule, and the HHI as a measurement tool.

These are not separate topics that happen to share a lecture slot. They are the same argument at different levels of abstraction. Once you see how cost constraints shape capacity decisions, and how capacity decisions shape competitive outcomes, the logic of market structure becomes a lot less arbitrary.

The Big Picture Before the Details

Here is how this session fits into the overall arc:

Production Function
      ↓
Cost Minimisation (isoquant/isocost logic)
      ↓
Short-Run Cost Curves ←—→ Long-Run Cost Curves
(capacity constrained)      (envelope of SR curves)
      ↓                              ↓
   Economies of Scope         Envelope Theorem
      ↓
Profit = Revenue − Cost
      ↓
Profit Maximisation: MR = MC
      ↓
Market Structure Analysis (Sessions 12–15)

Every session from 12 onwards reuses this cost structure. If you cannot reason about SR vs LR costs quickly, you will struggle with shutdown decisions, entry/exit logic, and merger analysis — all of which appear repeatedly.

Part 1: Economies of Scope

What it actually means

Economies of scope exist when producing two goods together is cheaper than producing them separately.

Formally: C(q₁, q₂) < C(q₁, 0) + C(0, q₂)

This is not the same as economies of scale. Scale is about producing more of one thing. Scope is about producing a portfolio of things jointly.

Why joint production is cheaper — the actual mechanisms

There are four main mechanisms. Know at least two well:

1. Shared fixed costs The same asset is used across multiple outputs. A logistics network, a compliance team, a data infrastructure — once it exists, the marginal cost of applying it to a second product is low. The fixed cost is spread across more units and more products simultaneously.

2. Shared capabilities and knowledge Some capabilities are hard to replicate but easy to redeploy. A brand, a distribution relationship, a regulatory approval — once you have it for Product A, using it for Product B costs far less than building it from scratch.

3. By-products One production process creates inputs or outputs that feed into another. A steel plant produces heat as a by-product. A financial institution's risk data for one product is valuable for pricing another. Nothing is wasted.

4. Cross-utilisation and demand smoothing Assets sitting idle during off-peak periods for one product can serve another product during those same periods. Hotels applying this logic across leisure and corporate segments is a classic example.

Why this matters for strategy

The economic logic of scope economies is the economic logic behind:

• Bundling (not just a pricing trick — often genuinely cheaper to produce together)
• Platform businesses (one infrastructure, many products — the scope economy is the moat)
• Conglomerates (viable when the shared asset is real — destroyed when it is just financial engineering)
• Adjacency expansion in strategy consulting advice (the right answer to "should we expand into X?" depends on whether scope economies with the existing business are real)

The trap most managers fall into: assuming scope economies exist just because two businesses look related. The test is ruthlessly concrete — does joint production actually lower cost, or does coordination complexity wipe out the savings? Many diversification failures come from this mistake.

Case: Indian private business schools (~2010)

The professor used this example to illustrate scope economies going wrong. In the mid-2000s, hundreds of private MBA programs opened across India. Most failed within a few years.

The economic explanation is not just "too much supply." It is about scope:

• Placement credibility is the core asset for an MBA program. It is built slowly through relationships with recruiters.
• Industrial groups (Tata, Birla, Reliance) that ran business schools had a genuine scope economy — they could create a pipeline from education to employment within their own conglomerate.
• Standalone programs without industrial backing had no such shared asset. Their placement credibility had to be built from scratch, at full cost.
• When supply expanded rapidly, programs without genuine scope economies could not compete on placements. Without placements, demand collapsed.

This is scope economies as a survival condition, not just a cost advantage.

Part 2: The Envelope Theorem and SR vs LR Costs

The mathematical idea (stripped down)

The envelope theorem answers a specific question: given a family of curves parameterised by some variable c, what is the curve that is tangent to every member of that family?

The procedure:

1. You have a family: g(x, y, c) = 0
2. Write the envelope condition: ∂g/∂c = 0
3. Eliminate c between these two equations
4. The result is the envelope — the boundary curve that touches each member of the family at exactly one point

The professor showed this with circles as a mathematical example. The economics translation is what matters.

The economic translation

In microeconomics, the "family of curves" is the set of short-run cost curves, each corresponding to a different fixed capital level K.

• Fix K at some level → derive the SR cost curve for that plant size
• Fix K at a different level → derive a different SR cost curve
• Do this for every possible K → you get a family of SR cost curves

The long-run cost curve is the envelope of this family.

Why? Because in the long run, the firm is free to choose K. For any target output level Q, it will choose the K that minimises cost. That minimum, traced across all output levels, is the LR cost curve — and it sits at or below every SR curve because the SR curves are the LR curve with the additional constraint that K is fixed.

The tangency result — what to remember cold

• LRAC is the lower envelope of the family of SRAC curves
• Each SRAC curve touches LRAC at exactly one point — the output level for which that particular capital configuration is cost-minimising
• Away from the tangency point, SRAC > LRAC — because you are now using the wrong plant size for the output you are producing

The intuition in one sentence: the short run is the long run with one hand tied behind your back.

Graph logic (how to read this mentally)

Picture the LRAC as a smooth U-shaped curve. Now picture a family of SRAC curves, each one touching the LRAC at a single point and lying above it everywhere else. The SRAC curves are "bowl-shaped" and each one sits inside the LRAC envelope at its tangency point.

For exam purposes:

• LRAC lies below every SRAC except at the tangency
• The tangency is not necessarily at the minimum of the SRAC (only at the minimum of LRAC does it happen that the SRAC is also at its minimum)
• To the left of the tangency: firm is overinvested (too much capital for that output)
• To the right of the tangency: firm is underinvested (not enough capital, pushing against capacity)

The Cobb-Douglas worked example

Setup:

• Production: Q = √(KL) = (KL)^0.5
• Cost: C = rK + wL
• This function has constant returns to scale (scaling both inputs by t scales output by t)

Short run (K fixed at K̄):

From Q = √(K̄ · L), solve for L:

• L = Q²/K̄

Substitute into cost:

• C_SR(Q) = rK̄ + w(Q²/K̄)

The fixed part is rK̄ (capital cost, unavoidable). The variable part is w(Q²/K̄), which rises with Q at an increasing rate. This is why SR cost curves are convex — as you push more output through a fixed plant, variable costs accelerate.

Long run (choose K optimally):

Apply the envelope condition — take the SR cost function and minimise over K:

• ∂C_SR/∂K = r − w(Q²/K²) = 0
• Solve: K* = (w/r)^0.5 · Q

Substitute back:

• C_LR(Q) = 2√(wr) · Q

The LR cost function is linear in Q — a feature of this CRS Cobb-Douglas setup. Unit cost is constant because the optimal input mix scales perfectly with output.

What this tells you:

• In the short run, cost is convex (capacity constraint bites)
• In the long run, cost is linear (optimal redesign is possible)
• The gap between them is the cost of operating under a capacity constraint

In reality, many cost functions are cubic (not linear) because there are initial economies followed by eventual diseconomies. The Cobb-Douglas gives us the method; the real world gives us the shape.

Part 3: Market Structure — The Foundation

Profit: the universal identity

Every market structure analysis starts here:

π = R − C

Where:

• R = Total Revenue = P × Q
• C = Total Cost (including opportunity cost)
• π = Economic profit (not accounting profit)

What changes across market structures is how R behaves with Q — specifically, how marginal revenue (MR) relates to price ℗. The cost side follows the same logic everywhere.

What "market" actually means

A market is defined by product type AND geography. This is not a pedantic distinction — it has real strategic and regulatory consequences.

A firm might be:

• A monopolist on a specific airline route
• One of many competitors in the national airline industry

These are different markets. The pricing power in the first case is dramatically higher. When assessing your own firm's competitive position, always ask: "In which specific product + geography do we have pricing power?" The answer is almost never the same as the broad industry definition.

Regulators care about this intensely during merger review. Misdefining the market is the most common error in antitrust analysis.

Why profit maximisation is the right assumption

The MR = MC rule comes from maximising π = R − C with respect to Q:

• dπ/dQ = dR/dQ − dC/dQ = 0
• MR − MC = 0
• MR = MC

The decision logic behind this:

• If MR > MC: the next unit adds more to revenue than to cost → produce it
• If MR < MC: the next unit costs more than it earns → don't produce it
• At MR = MC: no further gain from changing Q → this is the optimum

The second-order condition (SOC) — do not ignore this:

MR = MC is necessary but not sufficient. You need a maximum, not a minimum. The SOC requires:

dMR/dQ < dMC/dQ

In plain terms: the MR curve must be falling faster than (or the MC curve must be rising faster than) at the intersection point. This is typically satisfied when MC is upward-sloping at the optimum. When MC is downward-sloping, a MR = MC intersection can be a profit minimum — a classic exam trap.

Always state both FOC and SOC when answering profit maximisation questions.

Shutdown vs exit — a distinction that matters

The shutdown decision is a short-run decision. The exit decision is a long-run decision.

Short-run shutdown rule:

• Produce if P ≥ AVC (price covers variable costs)
• Shut down if P < AVC (price does not even cover variable costs)
• Logic: fixed costs are sunk in the short run — irrelevant to the produce/not-produce decision. The only question is whether revenue covers the costs you can actually avoid by not producing.

Long-run exit rule:

• Stay if P ≥ ATC (price covers total costs including fixed)
• Exit if P < ATC (cannot recover full costs even in the long run)

The zone AVC < P < ATC is economically interesting: a firm is losing money in total but it is rational to keep producing in the short run because shutting down loses even more (fixed costs are still incurred). In the long run, this is not sustainable.

Common mistake: Treating shutdown and exit as the same decision. They are not — they operate on different time horizons and compare price against different cost benchmarks.

Part 4: Measuring Market Concentration — HHI

What the HHI measures

The Herfindahl-Hirschman Index is a numerical measure of market concentration. It tells you, in a single number, how spread out or concentrated market power is across firms.

Formula: HHI = Σ sᵢ²

Where sᵢ is the market share of firm i (expressed as a fraction between 0 and 1).

Why squaring the shares?

Squaring does two things:

1. It gives larger firms disproportionately more weight — a firm with 50% share contributes 0.25 to HHI, while ten firms with 5% each contribute only 0.025 each
2. It makes HHI sensitive to inequality of market shares, not just the number of firms

This means HHI captures what matters economically: a market with one dominant firm and many small ones is far more concentrated than a market with the same number of firms but equal shares.

The range and benchmarks

US antitrust practice (referenced in lecture): HHI above 0.18 (or 1800 if shares are in percentage points) typically raises significant merger concerns.

HHI is only as good as your market definition

This is the critical caveat. The number is easy to compute. What is hard is deciding which firms to include — i.e., defining the relevant market.

Define the market too narrowly: a firm looks like a monopolist even if it faces real competitive pressure from substitutes. Define it too broadly: a dominant firm looks like just another competitor.

The HHI debate in any real merger is almost always a debate about market definition, not about the arithmetic.

The scale-efficiency tradeoff — competition is not always better

A fragmented market (low HHI) is not automatically welfare-maximising. Fragmentation can:

• Prevent firms from reaching efficient scale
• Reduce investment in R&D (insufficient returns)
• Create globally uncompetitive firms in industries where scale matters

The professor used Indian banking as the case. RBI's stated goal has been to consolidate the sector into 4–5 globally competitive banks. Indian banking HHI (based on total assets) has historically been around 0.05–0.09 — very low by international standards. The argument for consolidation is scale efficiency and global competitiveness. The argument against is market power and the political economy of bank mergers (union resistance, regional political concerns).

The right question is not "is concentration high or low?" but "what level of concentration maximises welfare given the technology, scale economics, and governance of this specific industry?"

Concept Connections

• Economies of scope → explains bundling strategy, platform economics, conglomerate logic
• Envelope theorem → formally justifies why LRAC is the minimum of SRAC curves
• SR vs LR costs → drives the shutdown (SR) vs exit (LR) distinction
• π = R − C → universal profit identity; reused in every market structure
• MR = MC → universal optimisation rule; implications differ by market structure
• HHI → measurement tool for market power; foundation for merger analysis

Each of these reappears in Sessions 12–15. The time you spend getting them right here compounds across the rest of the course.

What the Professor is Likely to Test

Based on lecture emphasis, the highest-probability exam questions from this session are:

1. Economies of scope — definition + application "Define economies of scope. How does it differ from economies of scale? Give two examples from a real industry."

The trap: giving scale examples. The test is C(q₁, q₂) < C(q₁, 0) + C(0, q₂). Make sure your examples actually involve joint production.

2. SR vs LR cost relationship "Explain why the LRAC curve is the envelope of SRAC curves."

Must include: constraint story (SR has fixed K), tangency at optimal Q, diagram description, managerial translation.

3. HHI computation + interpretation "Three firms have market shares of 50%, 30%, and 20%. Compute HHI and interpret."

HHI = 0.25 + 0.09 + 0.04 = 0.38. High concentration by any benchmark. But follow up: what is the market definition? What entry barriers exist?

4. Second-order condition "MR = MC is necessary but not sufficient for profit maximisation. Explain."

Must include: the SOC (dMR/dQ < dMC/dQ), the profit minimum trap (downward-sloping MC), and the upward-sloping MC requirement.

5. Shutdown vs exit "A firm is covering its variable costs but not its total costs. Should it produce?"

Yes — in the short run. Fixed costs are sunk. But this is not sustainable in the long run. Both parts of the answer are required.

Quick Reference

Next: Session 12 — Perfect Competition: the shutdown rule in full detail, producer surplus, industry supply, and long-run competitive equilibrium.

Back to: Microeconomics Master Index

Source: Lecture slides by Prof. Kaushik Bhattacharya, IIM Lucknow (Sessions 11a, 11b). Textbook: Pindyck & Rubinfeld, Microeconomics, Global Edition (2017), Ch. 8.