Intermarket Interactions and Efficiency in Electronic Markets: Eretailing Vs' Eauction Market

1
Inter-market Interactions and Efficiency in
Electronic MarketsE-retailing Vs. E-auction
Market

• Byungtae Lee
• Associate Professor
• Graduate School of Management
• Korea Advanced Institute of Science and
  Technology
• btlee_at_kgsm.kaist.ac.kr
• Based on joint work with Kumar Mehta

2
Agenda

• Motivation
• Analytical Model
• Empirical Model
• Empirical Results and Implication
• Future Research Direction

3
Economics of Information Systems Cybermics
Research

• E-xciting Digital Economics of eBusiness and
  eCommerce,
• Technology-driven (Pulled) new business/market
  models
• Economic Principles are not changing (Shapiro and
  Varian) but still out-of-box phenomena
• Theories Microeconomics, Macroeconomics,
  Industrial Organization, Institutional Economics
• Tools Mathematical Economics and Econometrics

4
Cybermics (Continued)

• Digitized (Agent, Business Process, Goods)
• Economic Characteristics Change
• Transaction Costs Change
• Cyber-mediaries to deal with new transactional
  costs
• New Market and Economic Organizations

5
Cybermics (Continued)

• B2B EC IOS, EDI, SCM
• Smart Cyber-Shopping Shopbots, PCA
• E-Payment States vs. Federal Government
• E-Lemon Market
• E-Library
• Component-ware Technology
• E-Posted Price vs. E-Auction vs. E-Reverse
  Auction

6
Cybermics (Continued)

• Surf or Ride?
• Build or Rent? Portal Service
• GIS-based DSS for EC Management
• .

7
E-Channels

• Want to buy a ticket? Where?
• Travelocity.com
• eBay.com
• Priceline.com
• Spot Market, Auction Market, Reverse Auction
  Market

8
Introduction

• Rise of parallel market mechanisms
• Increasing Popularity of Internet Auctions
• Alternative to Internet Retailing
• Mode for Dynamic Pricing
• Claims and counter claims of bargain buys

9
What Mechanisms and Prices?

• Are they different?
• What mechanism should we use? As a buyer or a
  seller
• E-Channel Management
• E-Channel Design (Portfolio) Decision
• E-Channel Yield Management
• E-retailing vs. E-Auction Comparison

10
E-Auction

• In 2002, 129 billion Internet Auction
• 20 of transactions of 13B B2C e-Commerce in
  1998
• B2C Auctions Constitute 33 of Auction Market
• Potentially very large market with growth
  exceeding the retailing

11
Electronic Auction

• Traditional Auctions vs. Electronic Auctions
• Hard to price items vs. Readily available items
• Competition from other auctions and retailers
• Small number of bidders vs. Global market
• Short duration vs. flexible duration
• Access to information
• Scale of operations
• Diverse Bumping Rules

12
(No Transcript)
13
Smart Use of E-Auction Bidding

• Individual price discrimination
• Costless demand curve estimation
• Global garage sale
• Core Process for e-Procurement

14
Research Issues

• Firms Strategy for Optimal Choice of Market
  Mechanism Auction vs. Retail
• Correction of Price Bias from Winners Curse
• Impact of multiplicity of market mechanisms
• Other Determinant of Bid Dispersion

15
Price and Bid Dispersions I
16
Why Prices Disperse?

• Stigler (1961)
• Homogenous goods, Rational consumers
• Price dispersion as a result of search costs
• Grossman and Stiglitz (1980)
• Arbitrage model, two types of agents
• Market failure when at extremes of informational
  efficiency
• More Rothschild (1973)
• Market with Incomplete Information

17
Price Dispersion and Market Efficiency

• Brynjolfsson and Smith (1999)
• Internet vs. Conventional Posted-Prices
• Internet posted-prices 6-7 lower
• Smith and Brynjolfsson (1999)
• Market heterogeneity, Branding, etc. as a source
  of increased price dispersion
• Lee (1997) Prices for used-cars on AUCNET higher
  than conventional market
• Clemons et al. (1998) Airline ticket prices vary
  by as much as 20
• Seidman et al. (1999)
• E-Auction beat e-Retailing on Average

18
Brick-n Mortar vs. E-channel?

• Are E-Channel Prices from Equilibrium?
  (Borenstein and Saloner, 2001)
• Are they same? May be different transaction
  costs and customers
• Why and How Different?

19
Price and Bid Dispersions II
20
Why Over-bidding?

• Not Rational?
• Entertainment Values?
• Winners Curse?
• Search Costs on Cyberspace?

21
Multiplicity of e-Mechanisms

• From E vs. Bricks and Mortar to E vs. e
• E and e Interact or Inform each other
• E-Mechanisms affects bidding dispersion beyond
  firms differentiation strategies

22
Research Framework
Auction Participants can observe retail markets!
Posted prices distribution
Min. observed prices
23
Search Cost Determinants

• Complexity of product description- Confidence on
  getting what you want
• Price dispersion in retail get smaller as product
  information diffuses
• Uncertainty increase in product generation change
• Search efficiency of surfers
• Expected marginal gain by search

24
Bidding Determinant

• Observed (or Belief) Price distribution
• Number of Participants
• Mix of Surfing Expertise
• Information Spill-over in the Market

25
Theoretical Model

• Homogenous consumers
• Implications similar to Stigler (1961)
• Search will increasingly yield diminishing
  benefit
• Optimal searching marginal benefitsearch cost
• Frictionless ? Market Failure
• Implications for auction of posted-price goods
• Winning auction bids reveal degree of friction
• Revenue increasing with increased no. of bidders
• Increasing search cost beneficial for retailers

26
Theoretical Model (Continued)

• Heterogeneous consumers
• n1 Experts (informed) vs. n2 non-experts
  (uninformed)
• proportion of cost disadvantage a
• price uniformly distributed (0,1)
• optimal sampling by both
• Population mix l experts and 1-l non-experts
• Population size N
• Winning bid b

27
Implications

• Probability of non-expert winning
• increases as with decrease in l
• increases with search cost disadvantage (a)
• Mixed results for population size, N - depends on
  level of l

28
Summary of Search Theory

• Different products have price dispersions
• Different product attracts informed and
  uninformed customer distributions
• Different auction mechanisms determine winning bid

29
Hypothesis

• H1 Market Friction (Seemingly Winners curse)
  exists
• H2 Degree of Over-bidding decreases as Market of
  The product matures
• H3 Degree of Over-bidding decreases for More
  Informed Customers
• H4 Discontinued products will not over-bidding

30
Data

• Brand New Computer Products (Scanners, Digital
  Cameras, Drives, Printers, Motherboards)
• Technical complexity used to categorize products
  in expert/non-expert
• Age New Release, Mature, Discontinued
• Winning Bids
• Time and Day of auction closing (High/Low
  Traffic)
• Auctions closing before or after 400 p.m.
• Auctions closing on weekday or weekend
• Retail prices for the exact same product on the
  same day
• Informed Customers Homogenous (experts) and
  Heterogeneous (mix)

31
Data Descriptive Statistics

• Total Observations 448 separate auctions
• Product Category
• Printers 157, Drives 68, Scanners 68
• Digital Cameras 117, Motherboards 38
• Product Type
• Non-expert 368, Expert 80
• Retail Price
• Closing Time
• After 4 PM 177
• On Weekend 71

32
Data Collection

• Bid Dispersion is collected from only ONE
  e-Auction site
• Price Dispersion was collected from multiple
  sites with helps of Shopbots and PCA
• At the same date with minimum delay

33
Empirical Results
34
Regression Results
Number of Observations 448 Adjusted
R-Square 0.761 F-Value 238.603 Coefficient
Value t-statistic
p-value Marginal Return, -0.251
-10.508 0.000 Price Dispersion (new),
0.179 6.917 0.000 Information
Subsidy (old), -0.563
-22.120 0.000 Increased Traffic (after 4
pm) 0.189 8.063 0.000 Search
Efficiency, 0.238
9.611 0.000 Weekend,
0.184 7.689 0.000 Preference
between Channels 0.265 7.463 0.000
35
Conclusions

• Search cost / Market friction, manifests in form
  of Over-bidding in the auction market
• E-auction is a very useful mechanism for sellers
• Periodically estimating demand curves for pricing
• Liquidation of inventory at the highest possible
  price
• Some bidders do get bargain buys
• E-auction market indicates that E-commerce is
  largely efficient in terms of price discovery

36
Implications for Auctioneer

• Timing closing of auctions
• Weekdays
• Weekends
• What mix of products?
• New Release, Mature, Liquidation
• High priced vs. low priced
• Bargain buys on liquidation items can serve
  promotional purpose
• Profit Margin or Fast Liquidation

37
Future Work

• Bidder Data from Ubid
• Auction vs. Reverse Auction Comparison
• Asian Market Vertical Channel Control

38
Agent-Modeling

• Prototypes for Market Mechanism Design and
  Evaluation Toolkits?
• Auction, Power Market Design?
• Seminar (Research-Oriented) Econ and IS Papers?

39
Thank You