contract theory
play

Contract Theory: A New Frontier for AGT Part I: Classic Theory Paul - PowerPoint PPT Presentation

Contract Theory: A New Frontier for AGT Part I: Classic Theory Paul Dtting London School of Economics Inbal Talgam-Cohen Technion ACM EC19 Tutorial June 2019 Plan Part I (Inbal): Classic Theory Model Optimal Contracts


  1. Contract Theory: A New Frontier for AGT Part I: Classic Theory Paul Dütting – London School of Economics Inbal Talgam-Cohen – Technion ACM EC’19 Tutorial June 2019

  2. Plan • Part I (Inbal): Classic Theory • Model • Optimal Contracts • Key Results • Break (5-10 mins) • Part II (Paul): Modern Approaches • Robustness • Approximation • Computational Complexity *We thank Tim Roughgarden for feedback on an early version and Gabriel Carroll for helpful conversations; any mistakes are our own

  3. 1. What is a Contract? 3

  4. An Old Idea Les Mines de Bruoux, dug circa 1885 4

  5. Purpose of Contracts • Contracts align interests to enable exploiting gains from cooperation • “What are the common wages of labour, depends everywhere upon the contract usually made between those two parties, whose interests are not the same .” [Adam Smith 1776] 5

  6. Classic Contract Theory “ Modern economies are held together by innumerable contracts ” [2016 Nobel Prize Announcement] Laureates Oliver Hart and Bengt Holmström 6

  7. Classic Applications • Employment contracts • Venture capital (VC) investment contracts • Insurance contracts • Freelance (e.g. book) contracts • Government procurement contracts • … → Contracts are indeed everywhere 7

  8. New Applications Classic applications are moving online and/or increasing in complexity • Crowdsourcing platforms • Platforms for hiring freelancers • Online marketing and affiliation • Complex supply chains • Pay-for-performance medicare → Algorithmic approach becoming more relevant 8

  9. Basic Contract Setting [Holmström’79] • 2 players: principal and agent • Familiar ingredients: private information and incentives • Let’s see an example… 9

  10. Example • Website owner (principal) hires marketing agent to attract visitors • Two defining features: 1. Agent’s actions are hidden - “ moral hazard ” 2. Principal never charges (only pays) agent - “ limited liability ” 10

  11. Moral Hazard “ Well then, says I, what ’ s the use of you learning to do right when it ’ s troublesome to do right and ain ’ t no trouble to do wrong, and the wages is just the same? ” Mark Twain, Adventures of Huckleberry Finn 11

  12. Limited Liability Typical example: an entrepreneur and a VC • The entrepreneur builds the company • The VC diversifies the risks and has deep pockets 12

  13. Timing Time Action’s Principal offers agent Agent takes Principal Agent outcome pays agent a contract costly, accepts (parties have rewards the according hidden (or refuses) symmetric info) principal to contract action 13

  14. 2. Connection to AGT 14

  15. Relation to Other Incentive Problems [Salanie] Uninformed player Informed player has the initiative has the initiative Private Mechanism Signaling information is design (screening) (persuasion) hidden type Private information is Contract design - hidden action 15

  16. New Frontier • Economics and computation – lively interaction over past 2 decades • Especially true for mechanism design and signaling Can we recreate the success stories of AGT in the context of contracts? • Are insights from CS useful for contracts? Is contract theory useful for AGT applications? In Part II: A preliminary YES to both 16

  17. Already Building Momentum • Pioneering works: • Combinatorial agency [Babaioff Feldman and Nisan’12,…] • Contract complexity [Babaioff and Winter’14,…] • Incentivizing exploration [Frazier Kempe Kleinberg and Kleinberg’14] • Robustness [Carroll’15,…] • Adaptive design [Ho Slivkins and Vaughan’16,...] • Recent works: • Delegated search [Kleinberg and Kleinberg’18,…] • Information acquisition [Azar and Micali’18,…] • Succinct models [Dütting Roughgarden and T.- C.’19b,…] • EC’19 papers: • [Kleinberg and Raghavan’19, Lavi and Shamash’19, Dütting Roughgarden and T.- C.’19a] 17

  18. The Algorithmic Lens • Offers a language to discuss complexity • Has popularized the use of approximation guarantees when optimal solutions are inappropriate • Puts forth alternatives to average-case / Bayesian analysis that emphasize robust solutions to economic design problems More on this in Part II But first, let’s cover the basics 18

  19. 3. Formal Model 19

  20. Contract Setting • Parameters 𝑜, 𝑛 • Agent has actions 𝑏 1 , … , 𝑏 𝑜 • with costs 0 = 𝑑 1 ≤ ⋯ ≤ 𝑑 𝑜 (can always choose action with 0 cost) • Principal has rewards 0 ≤ 𝑠 1 ≤ ⋯ ≤ 𝑠 𝑛 • Action 𝑏 𝑗 induces distribution 𝐺 𝑗 over rewards (“technology”) • with expectation 𝑆 𝑗 Recall two • Assumption: 𝑆 1 ≤ ⋯ ≤ 𝑆 𝑜 defining features • Contract = vector of transfers Ԧ 𝑢 = 𝑢 1 , … , 𝑢 𝑛 ≥ 0 20

  21. Example Contract: 𝑢 1 = 0 𝑢 2 = 1 𝑢 3 = 2 𝑢 4 = 5 No visitor General visitor Targeted visitor Both visitors 𝑠 1 = 0 𝑠 2 = 3 𝑠 3 = 7 𝑠 4 = 10 Low effort 𝑆 1 = 1. 1.3 0.72 0.18 0.08 0.02 𝑑 1 = 0 Medium effort 𝑆 2 = 5 5.2 0.12 0.48 0.08 0.32 𝑑 2 = 1 High effort 𝑆 3 = 7. 7.2 0 0.4 0 0.6 𝑑 3 = 2 21

  22. Contract setting: • 𝑜 actions {𝑏 𝑗 } , costs 𝑑 𝑗 • Expected Utilities 𝑛 rewards 𝑠 𝑘 • 𝑜 × 𝑛 matrix 𝐺 of distributions with expectations 𝑆 𝑗 Fix action 𝑏 𝑗 . Agent • 𝔽 [utility] = expected transfer σ 𝑘∈[𝑛] 𝐺 𝑗,𝑘 𝑢 𝑘 minus cost 𝑑 𝑗 Payoff ≠ payment/transfer Principal • 𝔽 [payoff] = expected reward 𝑆 𝑗 minus expected transfer σ 𝑘 𝐺 𝑗,𝑘 𝑢 𝑘 Utilities sum up to 𝑆 𝑗 − 𝑑 𝑗 , action 𝑏 𝑗 ’s expected welfare 22

  23. Example: Agent’s Perspective Contract: 𝑢 1 = 0 𝑢 2 = 1 𝑢 3 = 2 𝑢 4 = 5 No visitor General visitor Targeted visitor Both visitors 𝑠 1 = 0 𝑠 2 = 3 𝑠 3 = 7 𝑠 4 = 10 Low effort 0. 0.44 0.72 0.18 0.08 0.02 𝑑 1 = 0 Medium effort 1.24 1. 0.12 0.48 0.08 0.32 𝑑 2 = 1 High effort 1.4 1. 0 0.4 0 0.6 𝑑 3 = 2 Exp xpect ected ed tran ansf sfers rs: (0. 0.44, , 2.24, 4, 3.4) for (low, medi dium, m, high) h) 23

  24. Example: Agent’s Perspective Contract: 𝑢 1 = 0 𝑢 2 = 1 𝑢 3 = 2 𝑢 4 = 5 No visitor General visitor Targeted visitor Both visitors 𝑠 1 = 0 𝑠 2 = 3 𝑠 3 = 7 𝑠 4 = 10 Low effort 0.72 0.18 0.08 0.02 𝑑 1 = 0 Medium effort 0.12 0.48 0.08 0.32 𝑑 2 = 1 High effort 0 0.4 0 0.6 𝑑 3 = 2 Exp xpect ected ed tran ansf sfers rs: (0. 0.44, , 2.24, 4, 3.4) for (low, medi dium, m, high) h) 24

  25. Example: Principal’s Perspective Contract: 𝑢 1 = 0 𝑢 2 = 1 𝑢 3 = 2 𝑢 4 = 5 No visitor General visitor Targeted visitor Both visitors 𝑠 1 = 0 𝑠 2 = 3 𝑠 3 = 7 𝑠 4 = 10 Low effort 𝑆 1 = 1. 1.3 0.72 0.18 0.08 0.02 𝑑 1 = 0 Medium effort 𝑆 2 = 5 5.2 0.12 0.48 0.08 0.32 𝑑 2 = 1 High effort 𝑆 3 = 7. 7.2 0 0.4 0 0.6 𝑑 3 = 2 𝑆 3 - exp xpecte cted d tran ansf sfer er = 7 7.2 - 3.4 = 3 3.8 25

  26. A Remark on Risk Averseness • Recall 2nd defining feature: agent has limited liability ( Ԧ 𝑢 ≥ 0 ) [Innes’90] • Popular alternative to risk-averseness • Utility from transfer 𝑢 𝑘 is 𝑣(𝑢 𝑘 ) where 𝑣 strictly concave • Both assumptions justify why the agent enters the contract • Rather than “ buying the project ” and being her own boss 26

  27. A Remark on Tie-Breaking • Standard assumption: If the agent is indifferent among actions, he chooses the one that maximizes the principal’s expected payoff 27

  28. 4. Computing Optimal Contracts 28

  29. Contract Design Goal : Design contract that maximizes principal’s payoff Optimization s.t. incentive compatibility (IC) constraints: • Maximize 𝔽 [payoff] from action 𝑏 𝑗 • Subject to 𝑏 𝑗 maximizing 𝔽 [utility] for agent Related Problems: Implementability of action 𝑏 𝑗 ; min pay for action 𝑏 𝑗 Can all be solved using LPs! 29

  30. First-Best Benchmark • First-best = solution ignoring IC constraints • What principal could extract if actions weren’t hidden • I.e., if could pick action and pay its cost First-best = max 𝑗 {𝑆 𝑗 − 𝑑 𝑗 } • OPT ≠ first-best due to IC constraints 30

  31. Implementability Problem Given: Contract setting; action 𝑏 𝑗 Determine: Is 𝑏 𝑗 implementable (exists contract Ԧ 𝑢 for which 𝑏 𝑗 is IC) LP duality gives a simple characterization! Proposition: Action 𝑏 𝑗 is implementable (up to tie-breaking) ⇔ no convex combination of the other actions has same distribution over rewards at lower cost 31

  32. Implementability LP 𝑏 𝑗 implementable ⟺ LP feasible 𝑛 variables 𝑢 𝑘 (transfers); 𝑜 − 1 IC constraints minimize 0 𝐺 𝑗 ′ ,𝑘 𝑢 𝑘 − 𝑑 𝑗 ′ ∀𝑗 ′ ≠ 𝑗 (IC) s.t. ෍ 𝐺 𝑗,𝑘 𝑢 𝑘 − 𝑑 𝑗 ≥ ෍ 𝑘 𝑘 Agent’s expected 𝑢 𝑘 ≥ 0 (LL) utility from 𝑏 𝑗 given contract Ԧ 𝑢 32

Recommend


More recommend