Competitive Strategy: Dynamic Rivalry Notes

Competitive Strategy: Dynamic Rivalry

Experience Effects

  • As firms increase production, they improve their production capabilities.
  • This is summarized by the "experience curve".
  • The experience curve was developed by Bruce Henderson at the Boston Consulting Group in the 1960s.
  • The formula for the experience curve is: C<em>n=C</em>1naC<em>n = C</em>1n^a, where:
    • C1C_1 is the unit cost of the first unit.
    • CnC_n is the unit cost of the nth unit.
    • nn is the cumulative output to that point.
    • a < 0 is the elasticity of cost with respect to output.
  • An "85% experience curve" means that a doubling of total output reduces unit costs by 15%.
  • This is a useful way to think about competitive cost advantages over time.

Experience Effects (Continued)

  • Learning By Doing.
  • Experience effects can arise from:
    • Improved labor efficiency.
    • Standardization.
    • Technology-driven learning.
  • Firms are incentivized to compete for current market share, as increased production improves their future cost position.
  • Low-producing firms may fall behind.

Cooperation

  • Dynamic rivalry questions often focus on price competition.
  • It is difficult to motivate firms to jointly behave like a single monopolist because firms that cut prices steal business from other firms.
  • Price wars reduce total industry profits, so cooperation can be beneficial.
  • Price competition is only one aspect of interactions. Firms also:
    • Design products
    • Set standards
    • Advertise
    • Complement other firms’ activities
    • Lobby for or against regulations

Elements of Cooperation

  • Mutual interests.
  • Equitable sharing of value added.
  • Punishment:
    • Price wars.
    • Holdups.
  • Credibility of deterrents.
  • Returning to a cooperative equilibrium.
  • Uncertainty leads to error.

The Nash Reversion Strategy

  • Consider N identical Bertrand oligopolists repeatedly playing a one-stage pricing game, discounting the future at rate δ.
  • Marginal costs are constant, and capacities are large, leading to zero profits in the one-shot game.
  • Cooperation yields monopoly profit πMπ_M, split evenly among the firms.
  • Firms cooperate until one cuts price, then price at marginal cost in subsequent periods.

Equilibrium

  • Cooperating forever yields a payoff of πMN11−δ\frac{π_M}{N} \frac{1}{1 - δ}.
  • Cutting price today yields a payoff of πM+0(δ1−δ)π_M + 0(\frac{δ}{1 - δ}).
  • Cooperating yields a higher payoff if δ≥N−1Nδ ≥ \frac{N - 1}{N}.

Characteristics of the Equilibrium

  • Cooperation is "easier" with fewer firms.
  • Other strategies can also be consistent with this equilibrium. The "cooperative" outcome does not need to be the monopoly price, lower prices work as well.
  • Strategic variables can be more complex than just price.
  • Division of territory, like the coordination game played early in the course, is also possible.

Tit for Tat

  • "Tit for tat" strategy (2 firms): cooperate until rival cuts price, then price at marginal cost until rival raises price.
  • This does not satisfy subgame perfection.
  • Tit for tat offers the hope of returning to cooperation. Cooperative pricing can reintroduce cooperation if the rival plays tit for tat.
  • In 1980, Robert Axelrod organized tournaments where people submitted strategies to play in repeated prisoner’s dilemma games.
  • Tit for tat, submitted by Anatol Rapoport, dominated.

Price Leadership

  • Observed in competition between John Fairfax and Sons (Sydney Morning Herald, Sun) and Rupert Murdoch’s News Limited (Daily Telegraph, Daily Mirror).
  • There were seven price increases between 1941-74.
  • Fairfax moved first in four cases, and News Limited followed. They changed prices simultaneously in three cases.
  • In 1975, Fairfax raised the price, News Limited did not follow, and Fairfax lowered its price back.
  • News Limited made 1.6 million with this move; Fairfax lost 1.3 million.
  • News Limited became the price leader.

Collusion

  • "People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices," - Adam Smith, The Wealth of Nations, 1776.
  • Collusion is common.
  • Collusion is illegal in the United States, with harsh penalties such as fines and/or jail time.

Collusion: If it is illegal, why study it?

  • Studying overt collusion helps understand what makes dynamic cooperation easier or harder.
  • Tacit cooperation is not illegal.
  • The cooperative equilibrium (Nash Reversion Strategies) can arise from unilateral conduct, mimicking colluding firms.
  • Consult an attorney to know where the line is.
  • Firms should be aware that their suppliers might collude.
  • Machine learning algorithms that screen for anomalous price/bid patterns are an emerging tool.

Collusion: What Does it Look Like?

  • Collusion: Prices high with low variation.
  • Competition: Prices low with high variation.

Collusion: Who Gets Convicted?

  • Highway and Street Construction: 33%
  • Electrical Contracting: 25%
  • Furniture Wholesaling: 5%
  • Water and Sewer Construction: 5%
  • Motion Picture Theaters: 5%
  • Refuse Systems: 3%
  • Source: Jon Joyce (1989) via McAfee.
  • Firms convicted tend to be owner/operator firms.

Collusion: Recent Large Cartels

  • Auto Parts: 1999-2017, 70-80 markets, 3.2-5.0 trillion in affected commerce, 0.6-1.0 trillion in injuries.
    • Mostly bid rigging.
    • Began in the late 1990s in Japan, with a breakdown in the car manufacturer loyalty to specific suppliers.
  • Banking: 1990-2013, 88 markets, over 1.5 trillion in affected commerce.
    • Price fixing, bid rigging, and conspiracies to misrepresent price indices (e.g., LIBOR).

Collusion: How is it done?

  • Archer Daniels Midland colluded with several firms producing lysine during the 1990s.
  • ADM employee-turned-informant Mark Whitacre recorded hotel-room conversations.
  • Documentary "Fair Fight In the Marketplace"
  • Feature Film "The Informant!"

Collusion: Market Division

  • If buyers have identical demand for a product, the optimal monopoly price is the same regardless of how many are being sold to.
  • 10 buyers, each with demand qi=10−Pq_i = 10 - P. Total demand is Q=100−10PQ = 100 - 10P.
  • Inverse demand:
    • Single buyer: P=10−q1P = 10 - q_1
    • All buyers: P=10−Q10P = 10 - \frac{Q}{10}
  • If marginal cost is c < 10, the optimal monopoly price is P∗=c+102P^* = \frac{c+10}{2} when selling to any subset of buyers.
  • Assigning sellers to buyers ("no poach" agreement) generates the "right" pricing incentives.
  • Movie theaters have been known to collude by assigning distributors to theaters, so that each theater buys from only the assigned distributor.

Market Division

  • A firm that deviates can only do so by stealing business with price cuts.

Sustaining Collusion: Leniency Programs and Confessions

  • Lufthansa employees confessed to fixing passenger and cargo fuel surcharges with British Airways, Korean Air, and Air France-KLM during 2000-05.
  • The US Department of Justice offers immunity to the first to confess, which influenced Lufthansa.
  • Many cartel cases are initiated based on confessions.
  • It is difficult for firms to write contracts compensating employees for illegal activity.
  • Trust is an issue.

Sustaining Collusion: Differentiated Products

  • Firms need to agree on price, but what if the quality is different?
  • OPEC faced this problem. Nigerian heavy sour crude oil is different than “Saudi light.”
  • Suppose the Saudis believe 2 difference per barrel is appropriate, but the Nigerians believe it should be 3.
  • Refineries have different technologies, so some buyers will be indifferent with a 2 difference, and others with a 3 difference.
  • OPEC has this problem even though they talk. It would be far harder to execute this complicated coordination if talking is illegal.

Sustaining Collusion: Reaction Time

  • Collusion is harder to sustain if there is a long time between sales because it is harder to punish the cheater.
  • Milk producers might have to wait until next year’s school contracts before they can retaliate.
  • The discount factor δ reflects reaction time. When reaction time is slow, δ is low.

Sustaining Collusion: Random Demand

  • If it is hard to distinguish between a drop in demand and a price cut by your rival (all you see is you sold less), then it is harder to sustain collusion.
  • A price war of some fixed duration may be necessary.
  • The price war deters cheating. The fixed duration enables returning to collusive payoffs.

The Joint Executive Committee (JEC)

  • The JEC set market shares for rail shipments from Chicago to East Coast cities prior to the Sherman Antitrust Act of 1890.
  • The number one product shipped was grain (73%).
  • Prices for other goods reflected grain prices.
  • Price wars ensued if market shares strayed from predetermined levels.
  • Robert Porter estimated a model of cartel behavior over 1880-86 and found 10 price wars (1 of every 3 weeks).

Sustaining Collusion: Trading with the government

  • Government officials are harder to buy off, so it is easier to collude against the government.
  • Many price-fixing or bid-rigging conspiracies involve government contracts.
  • The large turbine generator case is one example.

Overcoming Random Allocations: Rings

  • Antiques dealers have been known to run cartels that bid for new pieces.
  • One dealer will be designated to “win” the auction at a lowball price.
  • Afterward, the dealers get together for a “knockout” auction where the auction revenue is shared by the dealers.
  • A knockout auction can be (approximately) efficient.
  • Side payments are necessary, which increases the likelihood of detection.

Price Fixing and Entry

  • Cartelized industries are attractive to entrants.
  • There is typically “too much” entry.
  • Real estate agents have historically split 5-6% commissions and have not competed fiercely on price.
  • The 5-6% commission level has been a contractual requirement for listing homes on the Multiple Listing Service.
  • The National Association of Realtors and other organizations lost a 1.8 billion jury verdict (for this exact practice) in October 2023.
  • Barriers to entry are very low to be a realtor, so there are lots of them. But the best agents receive nearly all of the benefits, as they avoid price competition.

Solutions to Tacit Cooperation Problems

  • Exclusive Territories.
  • Industry Associations.
  • Published Price Lists.
  • Capturing Regulators.
  • Multimarket Contact.

Industry Associations

  • Provide a basis for cooperation on different dimensions.
  • Build-the market advertising.
  • Changes to the regulatory structure.
  • Research on the future direction of the industry.
  • Provide a reason for executives to network with each other, making coordination easier.

Published Price Lists

  • Permit firms to see what their rivals are doing.
  • A list of prices for all products is made publicly available.
  • When goods are differentiated, a substitution matrix identifying which products are good substitutes for one another is necessary.
  • Firms can still discount off the price list.

Capturing Regulators

  • On March 13, 2001, George W. Bush reversed course on regulating carbon dioxide.
  • According to Stigler (1971), firms will often “capture” regulators and promote regulations that benefit insiders.
  • Regulations lead to the erection of entry barriers or restrictions on pricing that soften competition.
  • Consider recent regulations of the healthcare and financial services industries.

Multimarket Contact

  • In the airline industry, firms compete against other firms in many “markets” simultaneously.
  • A market is defined as an origin-destination pair of cities.
  • Delta and American competed against each other in 1150 markets in 2007.
  • Deviating from cooperative pricing may be more costly, as retaliation would hit more markets.
  • Firms' incentive constraints are pooled with multi-market contact, and the set of sustainable allocations is at least as large.
  • Ciliberto and Williams (2010) found that firms with high multi-market contact achieve near-perfect cooperation in setting fares (i.e., fares internalize business-stealing effects), and that cooperation increases with the level of contact when that level is relatively low.