HookThe £7.3bn merger the regulator refused
In April 2019 the Competition and Markets Authority blocked Sainsbury's £7.3bn takeover of Asda. The combined group would have held a bigger share of British groceries than Tesco, and the regulator concluded shoppers would face higher prices and worse quality. Then the market answered the question the lawyers had argued over: Aldi overtook Morrisons in September 2022 to become Britain's fourth-largest supermarket, and by 2024 the two German discounters held nearly a fifth of the market between them — competition doing from below what the CMA feared the merger would undo from above.
Theme 3 is the theory underneath that fight. Why firms grow, and why some deliberately break themselves up. What firms actually maximise — because it is not always profit. How revenue and cost curves decide output and the shut-down call. How market structure, from thousands of price-taking farmers to a single water company, determines prices, profits and efficiency. And then the same toolkit applied to the labour market, where the 'price' is your wage — before the regulators walk back in to referee all of it.
ModelFirms: why small survives, and the principal–agent problem
British business is a handful of giants standing on millions of small firms, and smallness persists for good reasons: niche markets too thin for scale economies, personal service that size destroys, owners who value independence over expansion, and finance constraints that cap growth whether owners like it or not. Alongside the private sector sit public sector organisations delivering non-market output, and not-for-profits: the John Lewis Partnership is owned by its staff, and Nationwide — a mutual owned by members — handed them £100 'Fairer Share' payments in 2023 rather than paying external shareholders.
Once a firm lists its shares, ownership and control divorce. Shareholders (the principals) want profit; the managers they hire (the agents) may prefer scale, salary, empire or a quiet life — and they hold better information about the business than its owners do. That is the principal–agent problem, and it explains real behaviour: why executive pay is loaded with shares and options (to re-align incentives), why activist investors buy stakes and agitate, and why takeover threats discipline drifting management. It is also your first evaluation tool in any 'firms will profit-maximise' essay: only if the people running them are made to.
MechanismGrowing and un-growing: integration, constraints, demergers
Organic growth means expanding from within — Aldi grows by opening stores, not buying rivals: slower, cheaper, culturally safer, but sometimes too slow for the opportunity. External growth is merger and takeover, classified by direction. Horizontal integration: same industry, same stage — Morrisons buying Safeway in 2004 for market share and purchasing power. Vertical integration: a different stage of the same chain — Tesco's £3.7bn purchase of the wholesaler Booker gave it control over supply into the convenience sector. Conglomerate integration: unrelated markets, spreading risk the way Unilever spans Marmite to Dove. Growth hits four constraints: the size of the market itself, access to finance, owner objectives, and regulation — the CMA exists precisely to stop some mergers happening.
Firms also un-grow. A demerger splits a company into separately owned businesses: GSK spun off its consumer-health arm Haleon (Sensodyne, Panadol) in July 2022 — at around £30bn, London's biggest listing in over a decade — so that pharma and consumer businesses could be managed, valued and financed on their own terms. Demergers sharpen focus and can unlock shareholder value, but they surrender synergies; workers face restructuring; consumers may gain from sharper management or lose cross-subsidised investment.
ModelObjectives: what firms actually maximise
The default assumption is profit maximisation: produce every unit that adds more to revenue than to cost, and stop where MC = MR. But the spec demands the alternatives, each with its own output level. Revenue maximisation: produce until MR = 0 — beyond that, extra sales cut total revenue. Sales (output) maximisation: grow as far as break-even allows, where AC = AR — the strategy of platforms buying market share. Satisficing — Herbert Simon's word — means earning enough profit to keep shareholders quiet while pursuing other aims: managerial comfort, staff pay, ethical commitments.
The distinctions are not academic. Amazon reported near-zero profit for years while revenue compounded, because market share today buys pricing power tomorrow; a corner shop satisfices the moment the owner values Sunday mornings over marginal revenue. Every objective implies a different price and quantity on the same cost-and-revenue diagram — profit max at MC = MR, revenue max at MR = 0, sales max at AC = AR — and comparing them on one diagram is among the highest-yielding analytical moves in Paper 1.
ModelRevenue, costs and the firm's arithmetic
Revenue first. Total revenue = price × quantity; average revenue = TR ÷ Q = price, which is why the AR curve is the demand curve. Marginal revenue is the change in TR from one more unit — and for any firm that must cut price to sell more, MR sits below AR, because the price cut cheapens every unit, not just the last one. MR reaches zero exactly where PED = 1, tying Theme 3 straight back to elasticity: a price-cutting strategy only raises revenue while demand remains elastic.
Costs next. Fixed costs don't vary with output (rent, licences); variable costs do (ingredients, hourly labour). Average cost = TC ÷ Q; marginal cost = the cost of one more unit. In the short run at least one factor is fixed, so the law of diminishing marginal productivity bites: cramming more bakers around one oven adds less and less extra output per baker, so marginal cost eventually rises — which is why MC curves slope up and why MC cuts AC at AC's minimum (while your marks are cheaper than the average, you drag the average down; once dearer, you pull it up).
A price-taking artisan bakery sells sourdough at the market price of £2.00, so AR = MR = £2.00 per loaf. Marginal cost rises with output: the 100th loaf costs £1.20 to make, the 200th £1.50, the 300th £2.00, the 400th £2.60. Profit-maximising output is 300 loaves — the last one adds exactly what it costs (MC = MR); the 400th would lose 60p. If average cost at 300 loaves is £1.70, supernormal profit = (£2.00 − £1.70) × 300 = £90 a day. Now the shut-down check: if average variable cost were £2.10, price would not even cover each loaf's ingredients and labour, so close immediately; at AVC = £1.60 the bakery survives the short run even if total average cost (say £2.20) exceeds price, because every loaf sold contributes 40p towards rent that must be paid anyway. In the long run, price must cover full average cost or the firm exits.
ModelScale: economies, diseconomies and the profit vocabulary
In the long run all factors are variable, and growing can cut unit costs — internal economies of scale: purchasing (Tesco's buying terms versus a corner shop's), technical (one large plant outperforming five small ones), managerial, financial (big borrowers pay lower rates), marketing and risk-bearing. Push size far enough and diseconomies arrive: coordination failures, communication layers, vanishing motivation. The LRAC curve therefore falls to the minimum efficient scale — the lowest output at which average cost bottoms out — runs flat, then may rise. External economies of scale cut costs for a whole industry in one place: Cambridge's life-sciences cluster and the City's legal-and-finance talent pool make every member firm cheaper to run.
The profit vocabulary is where marks leak. Normal profit is the minimum return that keeps the entrepreneur's resources in this industry — it is counted inside costs, so a firm earning it is doing fine, not failing. Supernormal profit is anything above that. The shut-down rules follow from cost structure: in the short run, stay open if revenue covers variable costs (P ≥ AVC), since fixed costs are owed regardless — Cineworld kept screens running through its 2022 bankruptcy protection because ticket sales covered running costs even while its debts were unpayable. In the long run, price must cover average cost in full.
ModelEfficiency, perfect competition and monopolistic competition
Four efficiency lenses judge every market structure. Allocative efficiency: price equals marginal cost, so output matches what consumers value. Productive efficiency: production at minimum average cost. Dynamic efficiency: innovation over time — funded, awkwardly, by the supernormal profit static efficiency condemns. X-inefficiency: cost padding that survives only where competitive pressure is weak.
Perfect competition — many buyers and sellers, identical products, perfect information, free entry and exit — makes every firm a price taker facing a horizontal AR = MR line. Short-run supernormal profit attracts entry, market supply shifts right, price falls until only normal profit remains: the long run is allocatively and productively efficient, but with no supernormal profit there is little to fund R&D. Currency markets and some agricultural commodities come close; most of the high street does not. Monopolistic competition keeps the low barriers but differentiates the product — every town's takeaways, barbers and cafés. Each firm faces a gently downward-sloping demand curve, earns supernormal profit only until imitators arrive, and settles at long-run normal profit with mild inefficiency: price a little above marginal cost, output short of minimum average cost — the price society pays for variety.
DataOligopoly: concentration, collusion and the game
An oligopoly is a market dominated by a few interdependent firms behind high barriers — and interdependence is the defining word: each firm's best move depends on its rivals' responses. Measure the structure with an n-firm concentration ratio: the combined market share of the largest n firms.
Interdependence pushes two ways. Towards collusion: overt cartels are illegal — the CMA fined ten construction firms nearly £60m in 2023 for rigging demolition bids — so firms drift into tacit coordination, matching a price leader without a word exchanged, the pattern petrol forecourts are repeatedly accused of. Towards competition: but game theory explains why price wars are rare. In the prisoner's dilemma, both firms holding price high is the jointly best outcome, yet each is tempted to undercut; if both cut, both end up worse off. One round of that lesson teaches oligopolists to compete on everything except price — loyalty schemes (Clubcard and Nectar member pricing), range, delivery, brand — where advantages cannot be matched by Friday afternoon.
Kantar's 2024 grocery data put Tesco at roughly 27.6%, Sainsbury's 15.3%, Asda 12.6% and Aldi 10.0%. CR4 = 27.6 + 15.3 + 12.6 + 10.0 = 65.5% — two-thirds of the market in four sets of hands: a textbook oligopoly. Then critique the number, because that is where the marks are: CR4 says nothing about churn within the top four (Aldi displaced Morrisons in 2022, and that entry from below disciplines prices more than a static 65.5% suggests), and nothing about local concentration — a town with one supermarket faces a local monopoly however competitive the national figure looks.
ModelMonopoly, price discrimination, monopsony, contestability
Pure monopoly is a sole seller; UK competition law treats 25% market share as the working threshold for monopoly power — Google's share of UK search, around nine in ten queries, shows power without literal monopoly. A price-making monopolist restricts output where MC = MR, prices above marginal cost, and stands accused on three counts: allocative inefficiency (P > MC), productive inefficiency, and X-inefficiency behind its barriers. The defence: scale economies can make one supplier genuinely cheapest — a natural monopoly like water mains, where duplication wastes resources and regulation beats break-up — and patent-protected supernormal profit funds the drugs and technologies dynamic efficiency requires. Third-degree price discrimination — same product, different prices to groups with different PEDs, separable and unable to resell (peak versus off-peak rail, 16–25 Railcards) — transfers consumer surplus to the producer, yet can raise total output and keep marginal services running.
Flip the power to the buyer's side and you get monopsony: the NHS employs well over a million people and dominates the market for nursing labour; supermarkets so dominate grocery buying that the Groceries Code Adjudicator was created in 2013 to police their treatment of suppliers. A monopsonist pays below the competitive wage or price and buys less. Finally, contestability reframes everything: what disciplines incumbents is not the number of firms but the credibility of entry. Where sunk costs are low, hit-and-run entrants can raid any supernormal profit, so even a lone incumbent prices close to normal profit. Digital licensing lowered banking's entry barriers enough for Monzo — founded 2015 — to pass nine million customers by 2024. Policy lesson: sometimes lowering entry barriers beats breaking anyone up.
MechanismThe labour market: derived demand, mobility, minimum wages
Demand for labour is derived demand — baristas are hired because flat whites sell. It rises with labour productivity and the product's price, falls with the wage, and bends with substitutability: self-checkouts replaced till staff as the relative cost of capital fell. Labour supply to an occupation responds to the wage and to everything else — conditions, status, flexibility (post-2020 remote working reshuffled whole occupations' attractiveness), training length. Two frictions stop workers flowing where wages point: geographical immobility (housing-cost gaps lock workers out of high-wage cities) and occupational immobility (no redundant steelworker becomes a radiographer by Friday). Where the market is competitive, the wage settles where supply meets demand; where one employer dominates, monopsony power drags it lower.
Governments intervene on both sides. The National Minimum Wage arrived in 1999 at £3.60 an hour; the National Living Wage for over-21s reached £12.21 from April 2025 — and the Low Pay Commission's repeated finding is that it achieved large pay rises with little clear evidence of significant job losses, exactly what economics predicts when employers hold monopsony power. Maximum wage rules are rarer (the bankers' bonus cap came and went); in the public sector, pay review bodies set pay for millions — monopsony management by committee. How a wage floor plays out always turns on the elasticities of labour demand and supply: inelastic demand means higher pay bills, not fewer jobs.
CaseRegulating the giants — and the referee's limits
Intervention to control market power runs through the CMA and the sector regulators. Merger control: Sainsbury's–Asda blocked in 2019; Microsoft's takeover of Activision Blizzard — worth around £55bn — blocked in 2023, then cleared only after cloud-gaming rights were sold to Ubisoft — a UK regulator restructuring a global transaction. Price regulation caps what natural monopolies can charge, RPI − X style: the cap rises with inflation minus an efficiency factor, forcing cost savings to reach consumers. Profit regulation is the alternative and invites padding — if allowed profit is a mark-up on costs, costs mysteriously grow. Quality standards and performance targets (rail punctuality, water leakage) close the corner-cutting gap price caps open. To promote competition: privatisation, deregulation (buses outside London in 1986), competitive tendering for public contracts. To protect the weaker side: the Groceries Code Adjudicator for suppliers, minimum-wage enforcement with HMRC publicly naming underpayers for workers.
The limits are Theme 1's government failure wearing a micro suit. Regulatory capture: a regulator dependent on the firm's own data starts seeing the world the firm's way — Ofwat approved years of dividends while Thames Water accumulated a £19bn debt pile the public now worries about. Asymmetric information: firms will always know their costs better than their regulator. Enforcement is expensive; targets get gamed. So the honest evaluation compares tools — regulate the monopoly, lower its entry barriers, or change its ownership — and picks per market, because each answer fails somewhere.
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
Theme 3 sits in Paper 1 alongside Theme 1, and returns in Paper 3. Its currency is the cost-and-revenue diagram, so drill it to 90 seconds: axes labelled costs/revenue (£) and output, MC cutting AC at AC's minimum, output dropped from MC = MR, price read up to AR, and the profit rectangle shaded between AR and AC at that output. For comparisons — monopoly versus perfect competition, profit max versus revenue max — one diagram with both positions marked is faster and scores the same analysis (AO3) as two.
Quantitative skills are guaranteed marks: calculate revenues, costs, profit and market shares from the extract's numbers, show the working, state the units. In data responses, application (AO2) means the firm on the page — its actual market share, objective and constraint — not a generic firm. Chains of reasoning stay king: entry barriers fall → hit-and-run entry credible → incumbent cuts price towards normal profit → allocative efficiency improves, every link explicit.
Essays carry 16 marks for knowledge, application and analysis and 9 for evaluation (AO4). The evaluation levers that fit almost every Theme 3 question: contestability (behaviour depends on entry threat, not firm count), time horizon (short-run supernormal profit versus long-run entry), the firm's actual objective (a satisficer won't exploit power fully), and regulation as counterweight. Then judge: answer the exact question, commit to a side, and condition it — 'the merger is likely to raise prices unless entry from discounters stays credible' is a justified judgement; a list of 'howevers' is not.