Ramesh Johari's course on Platform and Marketplace Design, starts tomorrow

Take the class, or you can also scroll down and see the syllabus with links to the papers.

MS&E 336: Platform and Marketplace Design

Course time

Tuesdays and Thursdays, 10:00-11:50 AM

Instructors

Ramesh Johari

Associate Professor

Management Science and Engineering

Electrical Engineering (by courtesy)

Huang Engineering Center, Room 311

E-mail: ramesh.johari@stanford.edu

Office hours: TBA, Huang 311

Additional office hours by appointment

Nick Arnosti (TA)

Management Science and Engineering

E-mail}: narnosti@stanford.edu

Office hours: TBA

Course website

The course website will be accessible through CourseWork.

Catalog course description

The last decade has witnessed a meteoric rise in the number of online markets and platforms competing with traditional mechanisms of trade. Examples of such markets include online marketplaces for goods, such as eBay; online dating markets; markets for shared resources, such as Lyft, Uber, and Airbnb; and online labor markets. We will review recent research that aims to both understand and design such markets. Emphasis on mathematical modeling and methodology, with a view towards preparing Ph.D. students for research in this area.

Detailed course description

Markets are an ancient institution for matching the supply for a good or service with its demand. Physical markets were typically slow to evolve, with simple institutions governing trade, and trading partners generally facing a daunting challenge in finding the “right” partner. The information technology revolution, however, has generated a sea of change in how markets function: now, markets are typically complex platforms, with a range of mechanisms involved in facilitating matches among participants.  Recent trends point to an unprecedented level of control over the design, implementation, and operation of markets: more than ever before, we are able to engineer the platforms governing transactions among market participants.  As a consequence, market operators or platforms can control a host of variables such as pricing, liquidity, visibility, information revelation, terms of trade, and transaction fees. The decisions made by the platform and the market participants interact, sometimes in intricate and subtle ways, to determine market outcomes.

This course is intended to prepare students for research on online and platform markets.  The course was inspired by the following observation: in the last decade, a wide range of graduates with quantitative backgrounds have been put into positions where they are effectively designing markets every day.  Often this is a side effect of being thrust into a software engineering, product development, or regulatory role: for example, a new hire might be asked to change how users browse through search results on eBay or Airbnb.  As is immediately apparent to a market designer, small changes to that basic infrastructure---the search engine---can radically alter the behavior of the market itself.  The goal in the class is to prepare students to be able to think conceptually about these market design challenges.

With that motivation in mind, we have three main goals for the quarter:

1. Problems.  The first goal is to use the quarter to identify open research directions that have risen to the forefront with the rise of online platforms.  We live in an exciting time for market design, with great interest in the fundamentals, as well as a rich set of applications that motivate research directions, and provide data and testbeds for validation.  A key emphasis in this part of the class is to focus on “levers” that affect the information that market participants obtain about each other in a variety of ways.
2. Tools. The second goal is to ensure students have access to a basic set of tools with which to reason about such markets.  The course assumes students have already had prior exposure to game theory and economic modeling.  In this course, we will focus on a set of tools that have specifically proven helpful in studying platform markets, and the effects of design interventions on these markets.  A key emphasis is on large market models.
3. Applications.  Along the way, we hope to learn about how the research questions we identify and the tools we learn are relevant to specific marketplaces.  This will be through a mix of mathematical modeling, empirical research, as well as anecdotal evidence.

The course will be taught using a mix of lecture format and seminar-style guided discussion. Much of what we will discuss is active research, so the reading material will be drawn from relevant papers in the literature; this material will be available from the course website. The focus will be on encouraging discussion of both open theoretical questions and modeling issues. This is particularly important since the course content draws from a range of disciplines (operations research, computer science, economics, etc.). The course should provide a unique forum for a lively exchange of ideas across these boundaries.

Evaluation

Your grade in the course will be based on two components.

1. Participation in lecture [ 40% of course grade ]. You will be expected to read papers and actively participate in lectures.  To help make sure this happens, for at least four of the weeks of the quarter, you will have to choose one paper that you need to read and prepare a “mini-review”.  The mini-review consists of answers to the following three questions, in 100 words or less each:
1. What is the paper about?
2. What are the strengths of the paper?
3. What are the weaknesses of the paper?

Each mini-review will be graded credit/no-credit, i.e., you will receive full credit for this

component if you satisfactorily complete each mini-review.

2. Course project [ 60% of course grade ]. A course project is the main evaluation component of the class.  The course project is meant to get you thinking actively about research problems in the market design problems represented by the course material.  The project will culminate in a presentation to your fellow students, and a written report (due by March 20, 2015).
   I will distribute more details on the project in the first week of lecture.

Course outline

Note: The content described on the course outline below will take up the first 14 lectures of the quarter.  (This is why each lecture is 110 minutes, instead of the usual 75 minutes.)  The remainder of the quarter will be used for guest speakers and discussion and presentation of course projects.

Part 1 (2 lectures): Introduction to platforms

We introduce platforms, and cover some of the relevant economics literature that defines and analyzes two-sided platforms.

Topics of interest:

1. What is a (two-sided) platform?
2. What is the objective of the platform operator,
   what information does she possess, and
   what tools does she have to acheive these objectives?
3. What are the objectives of platform participants,
   what information do they possess,
   and what actions are available to them?

Papers:

1. Rochet and Tirole (2003), Platform competition in two-sided markets
2. Rochet and Tirole (2006), Two-sided markets: A progress report
3. Armstrong (2006), Competition in two-sided markets
4. Caillaud and Jullien (2003), Chicken & egg: Competition among intermediation service providers
5. Parker and van Alstyne (2005), Two-sided network effects: A theory of information product design

Part 2 (2 lectures): Operational details of platforms -- pricing

We consider what the introductory papers might have missed, focusing on pricing strategies.  The emphasis is on operational details of platform behavior.

Topics of interest:

1. Pricing usage
1. Membership fees and subscriptions
2. Usage-based fee with flat fee per transaction/match
3. Usage-based fee with volume-based fee per transaction/match
2. Pricing visibility: paying for preferential access to the other side of the market
3. Pricing transaction risk: paying for reduced uncertainty of trade

Papers:

1. Gomes and Pavan, Many-to-Many Matching and Price Discrimination
2. Einav, Faronato, Levin, Sundaresan, Sales mechanisms in online markets: What happened to Internet auctions?
3. Celis, Lewis, Mobius, Nazerzadeh, Buy-it-Now or Take-a-Chance: Price Discrimination through Randomized Auctions

Part 3 (3 lectures): Operational details of platforms -- reputation and feedback

We study the role of reputation systems used by online platforms to help participants judge trading partners they have never met.

Topics of interest:

1. Examples of reputation and feedback systems
2. What incentives do particular systems provide to market participants?
3. How do we design systems that incentivize honest feedback?
4. How should the platform use the feedback system as a “lever” to improve market performance?

Papers:

1. Resnick, Kuwabara, Zeckhauser, Friedman (2000), Reputation systems
2. Cabral and Hortacsu (2010), The dynamics of seller reputation: Evidence from eBay
3. Dellarocas and Wood (2007), The Sound of Silence in Online Feedback: Estimating Trading Risks in the Presence of Reporting Bias
4. Bolton, Greiner, and Ockenfels (2013), Engineering Trust: Reciprocity in the Production of Reputation Information
5. Horton and Golden, Reputation Inflation in an Online Market
6. Fradkin, Grewal, Holtz, Pearson, Bias and Reciprocity in Online Reviews: Evidence from Field Experiments on Airbnb
7. Nosko and Tadelis, The Limits of Reputation in Platform Markets: An Empirical Analysis and Field Experiment

Part 4 (3 lectures): Operational details of platforms -- search

We discuss how the platform can mediate matches by directing the search effort of each side of the market.

Topics of interest:

1. How do market participants cope with the search frictions of finding trading partners?
2. What information should the platform share with market participants about potential matches?
3. What mechanisms can the platform provide to participants to improve the signals they send each other?

Papers:

1. Coles et al. (2010), The job market for new economists: A market design perspective
2. Lee, Niederle, Kim, Kim, Propose with a rose? Signaling in Internet dating markets
3. Halaburda and Piskorski, Competing by Restricting Choice: The Case of Search Platforms
4. Arnosti, Johari, and Kanoria, Managing Congestion in Dynamic Matching Markets
5. Tadelis and Zettelmeyer, Information Disclosure as a Matching Mechanism: Theory and Evidence From a Field Experiment
6. Horton and Johari, At What Quality and What Price? Inducing Separating Equilibria as a Market Design Problem

Part 5 (4 lectures): Modeling tools

In this part of the course we will cover some tools that have proven helpful in modeling and analyzing operational aspects of two-sided platforms.  We emphasize the use of large market models.

Topics of interest:

1. Large market models of static markets
1. Directed search and decentralized matching
2. Double auctions
2. Large market models of dynamic markets
1. Repeated (dynamic) auctions
2. Dynamic matching models

Papers of interest:

1. Rogerson, Shimer, Wright, Search-theoretic models of the labor market: A survey
2. Burdett, Shi, Wright (2001), Pricing and Matching with Frictions
3. Eeckhout and Kircher (2010), Sorting and decentralized price competition
4. Iyer, Johari, Sundararajan, Mean field equilibria of dynamic auctions with learning
5. Bodoh-Creed, Optimal Platform Fees for Large Dynamic Matching Markets
6. Besbes, Balseiro, Weintraub, Repeated Auctions with Budgets in Ad Exchanges: Approximations and Design
7. Duffie, Malamud, Manso, Information percolation with equilibrium search dynamics
8. Arnosti, Johari, and Kanoria, Managing Congestion in Dynamic Matching Markets