Upload
thaddeus-bence
View
212
Download
0
Tags:
Embed Size (px)
Citation preview
Confounded by the Field:
WTP in Food Auctions whenPrices are Increasing
John C. Bernard Na He
University of DelawareDepartment of Food & Resource Economics
Research supported by USDA/NRI grant 2003-35400-13812
WTP Auction Experiments
Use of auction experiments in determining WTP has increased greatly in past decade
WTP for food products a is major area: GM / non-GM foods Food safety / health information Steaks / Bison meat
Advantages: Non-hypothetical nature Auction mechanisms are incentive compatible
Should be better at gauging WTP than hypothetical survey methods
Field Price Issue
Non-hypothetical implies: Product exists in the field or Close substitute exists in the field
Field prices of substitutes could influence subject WTP and their bidding
Problem could be worse under times of field price fluctuations
Objectives
Overall objective was to see if results from auction experiments are confounded by prices in the field
Do field prices act as caps on bidding? How do bids change when field prices are changing? How do field prices influence the expectations of field
prices of substitute products? Can confounding be reduced by examining bid
differences instead of bid levels?
Field Price Censoring
Theory: Rational consumer should not be expected to bid above field price regardless of their value
Let Vi be subject i’s value, Bi be their bid, and P be the field price: Bi = min[Vi ,P]
Could bid over field price dependent on transaction costs. Let Ti be subject’s subjective notion of their transaction costs to get the good outside the experiment Bi = min[Vi ,P+Ti ]
Expected Field Price Censoring
Field price knowledge may be imperfect orNeed to judge unknown field price of a good by the
known field price of a substitute
Let Ei[P] be subject i’s expectation of the field price, then:
Bi = min[Vi , Ei[P]+Ti ]
Therefore can expect range of prices for good based on how consistent expectations are of field prices Censoring value is different for each individual
Literature
Harrison, Harstad, and Rutström (2004) Re-examined earlier study Found failure to account for field price censoring influence
results
Cherry, Frykblom, Shogren, List, and Sullivan (2004) Induced value experiments Subjects shaved bids toward price of outside option
Corrigan and Rousu (2008) Field experiment in supermarket Bids from consumers planning to purchase the product were not
significantly different from their field price expectations
Experiment Sessions
183 subjects professionally recruited from Delaware and vicinity
Payments approximately $50
First set of sessions: May 2007 139 subjects
Second set of sessions November 2007 44 subjects
Experimental Design
Description and practice auctions with Vickrey’s sealed-bid fifth price auction
Food auctions Definitions of attributes: conventional and organic Multiple food items, single trial Conventional price given One auction binding
Post-questionnaires Demographics Collect expected field prices for organic versions
Example: All bid their value except for subject C
Sub. Value Bid
A .60 .60B .55 .55C .50 .62D .65 .65
C bids morethan their value
Profit
0 0- .10 .05
Version Definitions
Conventional ‘Typical’ supermarket version Definitely would NOT meet organic standards
Organic No synthetic pesticides No hormones/antibiotics No irradiation No GM ingredients No petroleum/sewage sludge fertilizers
First SessionsFirst Sessions Second SessionsSecond Sessions
Milk (2%, half gallon) $2.35
Chicken Breasts (1 lb.) $3.99
Chocolate Chip Cookies $3.99
Tortilla Chips (14.5 oz.) $3.29
Milk (2%, half gallon) $2.77 (up 18%)
Chicken Breasts (1 lb.) $5.49 (up 37.5%)
Chocolate Chip Cookies $3.99
Tortilla Chips (14.5 oz.) $3.29
Food Products and Conventional Prices
Demand Curve Comparisons
Demand curves constructed following methodology of Lusk and Schroeder (2006)
Plot inverse cumulative density functions of the WTP bids
These regarded as demand curves with assumption each subject consumes at most one unit
Bids for Conventional Versions
Bids for Organic Versions
Tests and Hypotheses
Compare average bids across session sets Average bids expected higher for those facing increased
field prices
Examine percent of subjects bidding equal to (or over) respective field price Should be consistent across session sets
Compare field price expectations across session sets Expectations for organic field price should be higher
when conventional field price higher
Compare premiums for organic over conventional across session sets
Methodology for Comparisons
Test distribution of the bids for normality All bid series were non-normal
Brown and Forsythe tests for homogeneity of variances When variances homogeneous: Wilcoxon-Mann-
Whitney tests used When variances differ: Fligner-Plicello tests
performed
Changes in Bids Between Sets of Sessions
Comparison of Average Bids between First and Second Sessions
Mean of Session H0: Equal
Version Food First Second p-value
Conventional Milk 2.107 2.143 0.0005
Chicken 3.535 4.407 0.0001
Cookies 3.257 2.989 0.0817
Tortilla Chips 2.814 2.589 0.0229
Organic Milk 2.545 2.340 0.1774
Chicken 3.957 4.865 0.0001
Cookies 3.746 3.085 0.0024
Tortilla Chips 3.244 2.812 0.0175
Comparing Bids and Field Prices
Percentage of Bids Equal or Greater than Field Price Across Sessions
Percentage of BidsPercent Greater
Sessions Food Equal Greater than First Set FP
First Set Milk 38.35% 9.02%
Chicken 45.86% 11.28%
Cookies 35.34% 6.02%
Tortilla Chips 37.59% 3.01%
Second Set Milk 21.43% 7.14% 59.52%
Chicken 21.43% 7.14% 76.19%
Cookies 19.05% 7.14%
Tortilla Chips 19.05% 2.38%
Changes in Field Price Expectations
Comparison of Price Expectations between Sets of Sessions
Mean of Session H0: Equal
Food First Second p-value
Milk (2%) 3.220 3.417 0.0134
Chicken Breasts 4.517 6.142 0.0001
Chocolate Chip Cookies 4.079 3.847 0.0600
Tortilla Chips 3.604 3.440 0.1300
Changes in Premiums for Organic
Comparison of Percent Premiums Organic over Conventional
Mean of Session H0: Equal
Food First Second p-value
Milk (2%) 21.52% 17.42% 0.2147
Chicken Breasts 14.80% 20.46% 0.3197
Chocolate Chip Cookies 19.25% 11.98% 0.1081
Tortilla Chips 17.79% 16.22% 0.2996
Conclusions and Implications
Bids appear influenced by field pricesPrice expectations and bids change with
changes in field pricesResearchers need to recognize the
confounded effects of the field on bid levels
Examining premiums / differences in bids appears to remove the field price censoring element