Are there wine preferences? Pascale BAZOCHE Pierre COMBRIS Eric GIRAUD-HÉRAUD INRA, ALISS, Ivry sur Seine.

Are there wine preferences?

Pascale BAZOCHE

Pierre COMBRIS

Eric GIRAUD-HÉRAUD

INRA, ALISS, Ivry sur Seine

1. Assessing Willingness to Pay for Appellation of Origin with Pinot Noir wines from Burgundy

2. Investigating preference heterogeneity with Chardonnay wines

Outline

1. Willingness to Pay for Appellation of Origin:

an Experiment with Pinot Noir Wines

in France and Germany

• Can we assess the respective effects of sensory characteristics and label information on consumers' willingness to pay (WTP) for wines?

• What is the impact of grape variety, brand and appellation information on WTP?

• Do consumers from different countries react differently to this information?

• Can middle range Pinot Noir wines from Burgundy compete with generic Pinot Noir and branded Pinot Noir on the international market?

Questions

Wines Market

prices (€)

Bourgogne 7.00 Bourgogne Passe tout grains 3.20

Pinot noir (Pays d’Oc) 5.60 Gallo (Pinot noir) 6.80

Bourgogne : Magnien

Bourgogne Passe-tout-Grains : La cuvée des Ducs

Pays d’Oc : Les Salices (Lurton)

Gallo : Turning Leaf Pinot noir

Which wines?

Wines Market

prices (€)

Bourgogne 7.00 Bourgogne Passe tout grains 3.20

Pinot noir (Pays d’Oc) 5.60

Gallo (Pinot noir) 6.80

Bourgogne+Pinot noir

Bourgogne (traditional design)

Bourgogne (modern design) Spätburgunder

Alternative Labels

Which wines?

Labels

Alternative labels

4. Random selection of 1situation X 1wine Actual sale

1. Blind tasting 4 wines: 4 prices

2. Label 8 wines: 8 prices

3. Fullinformation + 4 wines: 4 prices

Protocol (1)

0 €

Maximumbuying price

Participant Experimenter

I don't buy

I buy

4 €

Incentive mechanism: the BDM procedure

1€ 2€ 3€ 4€ 5€ 6€ 7€

0 €

Maximumbuying price

Participant Experimenter

I don't buy

6 €

Random selling prices

4 €

Incentive mechanism: the BDM procedure

1€ 2€ 3€ 4€ 5€ 6€ 7€

0 €

Maximumbuying price

Participant Experimenter

I (must) buy

2 €

Random selling prices

4 €

Incentive mechanism: the BDM procedure

0 €

Maximumbuying price

2 €

4 €

6 €

All the prices you would be pleased to pay to obtain

the bottle

All the prices you would be pleased not to pay for that

bottle

Incentive mechanism: the BDM procedure

• Subjects who did not usually drink or buy wine were not selected for the experiment

• Subjects who agreed to participate signed a consent at the beginning of the experiment

• Participants were asked not to communicate with each other before the end of the experiment (wines in random order)

• They were reminded that the experiment was not an assessment of the wines' market value but a sale

• They were asked to write down a maximum buying price on a paper after each evaluation

• They could refuse to buy and tick the corresponding option

Protocol (2)

| n sex age size income price example

-------+-------------------------------------------------------------

Paris | 60 0.55 41.19 2.41 1558.47 4.83 6.93

Munich | 59 0.51 40.81 2.39 1240.96 5.42 7.00

-------+-------------------------------------------------------------

Total | 119 0.53 41.00 2.40 1399.72 5.10 6.96

---------------------------------------------------------------------

Prob()| 0.65 0.86 0.92 0.03 0.25 0.92

n = sample sizesex = sex of participant (Female=0, Male=1)age = age of participant size = household sizeincome = monthly per capita incomeprice = usual price paid for wine example = price example (control for anchoring)

Participants

119 participants X 16 wines = 1904 prices, 1538 prices > 0

All prices

Positive prices

Observations 1904 1538 Mean 2.93 3.63 Standard-deviation 2.75 2.61 Minimum 0 0.10 Maximum 30 30

Percentiles 10 % 0 1.20

25 % 1 2.00

50 % 2.50 3.00

75 % 4 4.50

Price (WTP) distribution

0.0

5.1

.15

.2

0 5 10 15 20 25 30Prix (€)

Refusals: 19 %

87%

Price (WTP) distribution

All prices

Characteristics Impact on WTP

P-values

Sex 0.571 0.127 Age 0.001 0.961 Household size - 0.004 0.989 Income quartile - 0.254 0.368 Country 0.292 0.461

Price usually paid 0.577 0.000

Constant - 0.068 0.945

Tobit model, 1664 observations (307 = 0 ; 1357 > 0)

105 clusters (accounting for intra-subject correlation)

Participants' characteristics influencing WTP

BLINDBourgogne

PTGGalloPinot

LABELBourgogne

PTGGalloPinot

PinotBgneParchemin

ModerneSpät

FULL INFOBourgogne

PTGGalloPinot

0 .1 .2 .3 .4 .5 .6 0 .1 .2 .3 .4 .5 .6

Paris Munich

Graphs by Pays % refusing to buy

95% confidence intervals of % of WTP=0

Full sample France Germany

Blind reference reference reference Label 0.164 *** 0.226 *** 0.117 ** Full Info 0.105 *** 0.153 *** 0.067

Blind X Bourg reference reference reference X PTG 0.049 0.070 0.027 X Gallo 0.028 0.026 0.032 X Pinot 0.119 *** 0.081 0.151 ***

Label X Bourg reference reference reference X PTG - 0.0003 - 0.156 ** 0.083 ** X Gallo - 0.131 ** - 0.400 *** 0.057 X Pinot - 0.069 - 0.332 *** 0.105 **

Full Info X Bourg reference reference reference X PTG - 0.091 - 0.274 *** 0.041 ** X Gallo - 0.053 - 0.200 ** 0.045 X Pinot 0.009 - 0.151 0.114 *

Probability(WTP=0)

Probit-marginal probabilitiesof factors affecting participants’ decision to buy (1)

Full sample France Germany

Country (ref=France) -0.050 Usual price 0.023 *** 0.016 0.028 ** Woman - 0.066 * - 0.005 - 0.151 *** Household size 0.016 0.021 - 0.003 Income - 0.023 - 0.026 - 0.006 Age - 0.0003 0.001 - 0.002 Order - 0.006 - 0.009 - 0.0004

Probability to purchase 0.825 0.847 0.825

Observations 1247 671 576

Probability(WTP=0)

Probit-marginal probabilitiesof factors affecting participants’ decision to buy (2)

BLINDBourgogne

PTGGalloPinot

LABELBourgogne

PTGGalloPinot

PinotBgneParchemin

ModerneSpät

FULL INFOBourgogne

PTGGalloPinot

2 3 4 5 6 2 3 4 5 6

Paris Munich

Graphs by Pays

95% confidence intervals of WTP>0

Full sample France Germany

Blind reference reference reference Label 0.834 *** 0.885 *** 0.821 *** Full Info 0.869 *** 0.942 *** 0.789

Blind X Bourg reference reference reference X PTG - 0.320 - 0.198 - 0.485 X Gallo - 0.212 - 0.132 - 0.276 X Pinot - 0.109 0.028 - 0.224

Label X Bourg reference reference reference X PTG - 0.068 - 0.053 - 0.141 X Gallo - 0.090 - 0.359 0.136 X Pinot - 0.349 - 0.461 ** - 0.276

Full Info X Bourg reference reference reference X PTG - 0.146 - 0.006 - 0.294 X Gallo - 0.014 - 0.141 0.124 X Pinot 0.081 - 0.039 - 0.159

Factors influencing positive WTP (1)

Full sample France Germany

Country (ref=France) 0.424 Usual price 0.500 *** 0.542 *** 0.450 *** Woman - 0.320 - 0.381 - 0.346 Household size - 0.099 - 0.187 - 0.037 Income - 0.156 - 0.506 ** 0.327 Age - 0.006 0.010 - 0.007 Order - 0.041 - 0.038 - 0.045

Constant 0.426 1.084 0.477

Observations 1012 556 456

Factors influencing positive WTP (2)

How to explain the small differences in WTP?

1. Consumers are unable to discriminate between wines

OR

2. Consumers discriminate but their preferences are very heterogeneous and aggregations cancel out the differences

Question

For each participant, in each situation, wines are ranked according to WTP:

id wine price rank--------------------------- 1 Bourgogne 5 1 PTG 5 1 Gallo 0 3 Pinot 0 3--------------------------- 2 Gallo 8 1 Pinot 7 2 Bourgogne 3 3 PTG 2 4---------------------------

Mean WTP are then computed by rank

WTP according to participants' preferences

BLINDRank 1Rank 2Rank 3Rank 4

LABELRank 1Rank 2Rank 3Rank 4Rank 5Rank 6Rank 7Rank 8

FULL INFORank 1Rank 2Rank 3Rank 4

0 1 2 3 4 5 6 0 1 2 3 4 5 6

Paris Munich

Graphs by Pays

95% confidence intervals of WTP according to preferences

BLINDRank 1Rank 2Rank 3Rank 4

LABELRank 1Rank 2Rank 3Rank 4Rank 5Rank 6Rank 7Rank 8

FULL INFORank 1Rank 2Rank 3Rank 4

0 .1 .2 .3 .4 .5 .6 0 .1 .2 .3 .4 .5 .6

Paris Munich

Graphs by Pays % Refusing to buy

95% confidence intervals of WTP=0 according to preferences

€/bouteille

BLINDRank 1Rank 2Rank 3Rank 4

LABELRank 1Rank 2Rank 3Rank 4Rank 5Rank 6Rank 7Rank 8

FULL INFORank 1Rank 2Rank 3Rank 4

1 2 3 4 5 6 1 2 3 4 5 6

Paris Munich

Graphs by Pays

95% confidence intervals of WTP>0 according to preferences

Blind Tasting---------------------------------- | Rank Wines | 1 2 3 4----------+-----------------------Bourgogne | 35 23 39 22 PTG | 38 32 23 26 Gallo | 38 31 33 17 Pinot | 38 36 33 12----------------------------------

Full Information---------------------------------- | Rank Wines | 1 2 3 4----------+-----------------------Bourgogne | 40 37 21 21 PTG | 36 32 35 16 Gallo | 42 16 33 28 Pinot | 41 29 34 15----------------------------------

Wine rankings in situations 1 and 3

• Sensory characteristics and label information influence differently French and German consumers

• French participants respond very positively to AOC information

• AOC does not seem to be an efficient signal outside the country of origin

• Consumers have strong preferences

• Heterogeneity in preferences is very important and averages out differences when observing aggregated data

What did we learn?

2. Investigating heterogeneity

with Chardonnay wines

► Preference heterogeneity– Pinot Noir?– Non-expert consumers?– Wines were too similar?

► Chardonnay1. Same experiment in Munich

2. "Sensory oriented" participants (pilot study)

3. Wider price range (pilot study)

Investigating preference heterogeneity

Chardonnay: "Munich" tasting

64 participants, Munich, 2007 (April 17-19th)

Men: 55%

Age: 18 to 65 (mean: 40)

Wines Market

prices (€)

Bourgogne AOC Maufoux 7.70 Pays d’Oc Gérard Bertrand 3.60 Gallo Chardonnay Sierra Valley 5.00

95% confidence intervals of WTP according to wines

BLINDBourgogne

GalloChardonnay

LABELBourgogne

GalloChardonnay

ParcheminVillage

ModerneBgneChrdny

FULL INFOBourgogne

GalloChardonnay

1 2 3 4 5

Chardonnay Munich

Price (€/bottle)

95% confidence intervals of WTP according to ranks

BLINDRank 1Rank 2Rank 3

LABELRank 1Rank 2Rank 3Rank 4Rank 5Rank 6Rank 7

FULL INFORank 1Rank 2Rank 3

1 2 3 4 5 6

Chardonnay Munich

Price (€/bottle)

Chardonnay: "Sensory oriented" participants

Sexe | Freq. Percent

------+----------------------- F | 14 60.87 M | 9 39.13 ------+-----------------------Total | 23 100.00

age | Freq. Percent------+-----------------------30-39 | 7 30.4340-49 | 7 30.4350-59 | 6 26.09 >60 | 3 13.04------+-----------------------Total | 23 100.00

23 participants, European Sensory Network meeting,

Dijon, 2007 (November 22th)

1. Blind tasting

3. Fullinformation +

2. Label

4. Random selection of 1 situation X wine

2 wines: 2 prices

2 wines: 2 prices

2 wines: 2 prices

Actual sale

Protocol

Two Chardonnay wines

Chardonnay Pays d'Oc: 3.60 € Pouilly Loché AOC: 6.00 €

Mean and 95% CI of reservation prices (€/bottle) according to wines and ranks

ChrdnyPouilly

Chrdny

Pouilly

Chrdny

Pouilly

1

2

3

4

5

6

7

8

9

10

Rank1

Rank2

Rank1

Rank2

Rank1

Rank2

1

2

3

4

5

6

7

8

9

10Blind tasting Label Label + tasting Blind tasting Label Label + tasting

-------------------------------------------------- | sit and Rank | -- Taste - -- Label - -- Full -- vin | 1 2 1 2 1 2----------+--------------------------------------- Chrdny | 13 10 5 18 15 8 Pouilly | 14 9 22 1 15 8--------------------------------------------------

Chardonnay: "Wide price range" tasting

Sexe | Freq. Percent

------+----------------------- F | 8 32.00 M | 17 68.00 ------+-----------------------Total | 25 100.00

age | Freq. Percent------+-----------------------30-39 | 6 24.0040-49 | 7 28.0050-59 | 9 36.00 >60 | 3 12.00------+-----------------------Total | 25 100.00

26 participants from Norway, Paris, 2008 (June 9th)

Chardonnay: "Wide price range" tasting

Savigny-les-BeauneAOC: 25 €

Gallo ChardonnaySierra Valley: 5 €

ChardonnayPays d'Oc: 3.60 €

Chassagne MontrachetAOC: 28 €

Mean and 95% CI of reservation prices (€/bottle) according to wines

PdOc

Gallo

Chassagne

Savigny

PdOcGallo

ChassagneSavigny

PdOc

Gallo

Chassagne

Savigny

05

1015

Pric

e (

€/b

ottle

)

Taste Label Full

Mean and 95% CI of reservation prices (€/bottle) according to wines and ranks

Chrdny

Brgne AOC

Chrdny

Brgne AOC

Chrdny

Brgne AOC

02

46

810

1214

Pric

e (

€/b

ottle

)

Taste Label Full

Mean and 95% CI of reservation prices (€/bottle) according to wines and ranks

Rank1

Rank2

Rank1

Rank2

Rank1

Rank2

02

46

810

1214

Pric

e (

€/b

ottle

)

Taste Label Full

Mean and 95% CI of reservation prices (€/bottle) according to wines and ranks

Chrdny

Brgne AOC

Chrdny

Brgne AOC

Chrdny

Brgne AOC

02

46

810

1214

Pric

e (

€/b

ottle

)

Taste Label Full

Rank1

Rank2

Rank1

Rank2

Rank1

Rank2

02

46

810

1214

Pric

e (

€/b

ottle

)

Taste Label Full

-------------------------------------------------------- | Sit and Rank | --- Taste -- --- Label -- --- Full --- wine | Rank1 Rank2 Rank1 Rank2 Rank1 Rank2----------+--------------------------------------------- Chrdny | 13 13 7 19 8 18Brgne AOC | 14 12 22 4 21 5--------------------------------------------------------

1. Apparently there are clear preferences

2. Obviously they are very heterogeneous

3. What if they are not stable over time?

Where are we now?