1 wuyang hu, michele veeman, vic adamowicz dept. of rural economy university of alberta anne...
TRANSCRIPT
1
Wuyang Hu, Michele Veeman, Vic Adamowicz Dept. of Rural Economy University of Alberta
Anne HuennemeyerKFW Group, Germany
Financial assistance from Genome Canada, Genome Prairie, and the Alberta Agricultural Research Institute is greatly
acknowledged
Assessing How Different Genetically Modified Food Labelling Policies May
Affect Consumers’ Choice Behaviour – A Canadian Case Study
International Symposium on Food Safety: Consumer, Trade, and Regulation Issues Hangzhou, Zhejiang, China
October 10-11, 2003
2
Introduction
Transgenic agricultural biotechnology termed Genetic Modification (GM) is fast expanding An annual growth of more than 10% has been
achieved every year for the last six years, since transgenic crops are introduced in 1996.
From 1996 to 2002, the global area of transgenic crops increased 35-fold.
The global area of transgenic crops for 2002 is 58.7 million hectares.
In 2002, for the first time more than half of the world’s population lived in countries where GM crops are grown.
3
Purpose of Study
Apply and examine the power of internet-based choice experiment in marketing research.
Identify consumers’ opinions on various issues surrounding the labelling of GM foods.
Investigate how different GM labelling systems may affect consumers’ behaviour.
4
Information and Labelling
There may be uncertainties about the quality and other features of food produced through GM technology.
In polls consumers tend to “vote for” the right to know the ingredients of their food, reflecting a preference for information.
Increasingly government regulations apply to information that is provided through GM labelling.
Different countries have different GM labelling policies.
5
Currently Announced International Labelling Policies
Countries/Regions Type of Policy GM Content Threshold
Australia/New Zealand Mandatory 1%
China Mandatory n/a
Czech Mandatory 1%
European Union/UK Mandatory 0.9%
Hong Kong Mandatory 5%
Japan Mandatory 5%
Russia Mandatory 5%
Switzerland Mandatory2% or 3% depending on
different situations
Brazil (draft) Mandatory n/a
India (draft) Mandatory n/a
Israel (draft) Mandatory 1%
Malaysia (draft) Mandatory 3%
Korea (draft) Mandatory 3%
Taiwan (draft) Mandatory 5%
Thailand (draft) Mandatory 5%Canada Voluntary 5%USA Voluntary 5%
* Sources: International Service for the Acquisition of Agri-biotech Applications
(http://www.isaaa.org) and the United State Department of Agriculture
(http://www.fas.usda.gov)
6
Preference Elicitation
Revealed Preference Directly observed and convenient to obtain. Reflect consumers’ actual choices Poses difficulties for environmental evaluation
and new product evaluation.
Stated Preference Hypothetical scenarios and highly controlled. Can provide useful information for hypothetical
variables or new products. Can generate unrealistic predictions and
introduce hypothetical bias.
7
Stated Preferences in Marketing Science
Conjoint Analysis Rankings Ratings Pairwise comparisons Adaptive Conjoint Analysis (ACA) Choice-Based Conjoint (CBC)
Discrete Choice Analysis Simulated Auctions
Flexible and consistent with decision making. Suitable for controlled lab environment.
8
Conjoint Models in Marketing
Single ProfilePairwise
Comparison
Full Profile Partial Profile Traditional Survey Methods Dynamic Computer Advantage Exploded Logit Ordered Logit/Probit Multinomial/Conditional Logit Multinomial Probit
CBCRating RankingPairwise
Comparison
ACA
9
Survey Data and Scenarios
Pre-packaged sliced bread Commonly consumed GM product is not yet available
Attributes Brand name (store/national) Type of flour (white/partial whole wheat/whole
wheat/multigrain) Price ($0.99/$1.49/$2.49/$3.49) GM (presence/absence)
Labelling scenarios
10
Survey Design and Implementation
Computerized design Attributes Labelling contexts: mandatory, voluntary and a
mixed scenario for statistical purpose Fractional factorial design with appropriate
blocking Online implementation
Faster and may be less biased than traditional methods in that more realism may be sought
Allows highly interactive designs
11
Sample Choice Set
12
Perceptions of Labelling Issues
S1: The public is sufficientlyinvolved in the regulation of GM foods.S2: Even if food prices were higher, the consumers’ “right to know” warrants mandatory labelling.S3: The decision about introduction of GM foods to Canada should be left to experts.
S4: There is no need for mandatory labelling of GM foods if the final product quality is the same.
S5: Voluntary labelling might be used as a marketing tool rather than providing useful consumer information.
S6: Stricter regulations for approving GM foods are better than a mandatory labelling system for GM foods.
S7: Overall mandatory labelling is preferable to voluntary labelling.
0
10
20
30
40
50
60
70
80
90
100
S1 S2 S3 S4 S5 S6 S7
Statements
Pe
rce
nta
ge
Tend toAgree
Tend toDisagree
Don'tKnow
13
Econometric Model and Variable Definition
Panel conditional logit model (CBC Conjoint approach) Some variables used in the model are:
Variables Definition
MGMO Interaction between mandatory labelling context and GMO
VNOGMO Interaction between voluntary labelling context and NOGMO
BRIGM Interaction between factor score on "bright future" with GMO
BRINOGM Interaction between factor score on "bright future" with NOGMO
HARMGM Interaction between factor score on "could be harmful" with GMO
HARMNOGM Interaction between factor score on "could be harmful" with NOGMO
ANIGM Interaction between factor score on "concerns for animals" with GMO
ANINOGM Interaction between factor score on "concerns for animals" with NOGMO
KNOWGM Interaction between GM knowledge with GMO
KNOWNOGM Interaction between GM knowledge with NOGMO
14
Estimation Results
Variables Coefficient Variables Coefficient
PRICE -0.7293*** VNOGMO -0.1352
BUYNO -2.2253*** BRIGM 0.2352***
STOREB -0.1110*** BRINOGM -0.0793**
WHITE -0.3926*** HARMGM -0.1972***
PARTIAL -0.2134*** HARMNOGM 0.0895***
WHOLE 0.2005*** ANIGM -0.0490
GMO -0.3278*** ANINOGM 0.1005***
NOGMO 0.1822*** KNOWGM -0.128*
MGMO -0.2641** KNOWNOGM 0.0117
* Significant at the 10% significance level **Significant at the 5% significance level
*** Significant at the 1% significance level
15
Conclusions
Conjoint (especially CBC) analysis is well-suited to evaluating consumers’ behaviour in the context of GM foods.
In general, Canadian consumers strongly “vote for” product information associated with labelling.
Mandatory GM labelling raises a “red flag” to consumers.
Voluntary GM labelling does not change consumers’ welfare significantly compared with no labelling.