Vineyard Data Quantification Society "Economists at the service of Wine & Vine" Enometrics XX A Hedonic Analysis of Retail Italian Vinegars Luigi Galletto, Luca Rossetto Research Center for Viticulture and Oenology (Cirve), University of Padova, Italy Vinegar grape 2 nd transformation not a beverage, but a condiment Italy, 1 st Exporter Very limited market research (Mattia, 2004; Radman et al., 2005) No hedonic price estimation Methodology Survey Concluding remarks Summary The Model P j (X) = P(x 1, x 2,, x i,, z n ) where P j is the price of product j and X = x 1, x 2,, x i,, x n is a vector of n observable attributes that completely describe product quality. The partial derivative, P(X)/ x i, of the hedonic price function with respect to the attribute i, can be viewed as the implicit or shadow price of the specific attribute i. Functional forms: linear, log-lin, double log, etc.
The survey Survey area: Verona Province (Veneto) Time: Summer-Fall 2012 Information collected on point of sales Questionnaires: Intrinsic attributes: acidity, row material, aging, production methods, producer s and collective brands Extrinsic attributes: bottle size, bottle closure, back label, bottle color, bottle material Type of point of sale, location visual merchandising (shelf position, facings), Price The survey 46 sale points of different size 21 towns 27 large scale retail LSR labels: local, regional, national or international chain labels beside some no chain stores >> > Famila Group 1036 vinegar bottles: from 4 (discount) to 137 (hypermarket) Modena Balsamic PGI, but no Modena or Reggio Traditional Balsamic PDO Degree of acidity very skew distribution: 5.1%-6.0% range (72%) A quite fragmented vinegar market: 113 different vinegar brands few big producers (>>>Ponti) plus many other small or very small. The survey Two merchandising characteristics: o a) Position (tier) on the shelf o b) Facings: 2 5 (71%) Service quality attributes, which impact the product when is purchased A product has more probability to be chosen if it receive more space or is set on more detectable shelf, since it is more likely to be noticed or attract consumers attention first Eye or hand levels should have a considerably greater probability of being selected Seldom included separately in hedonic price models for food The survey: descriptives (1) Sample composition of categorical variables in percent Outlet location Urban area 47.7 Rural area 52.3 Retail chain Famila Group 35.6 Others 64.4 Outlet type Hypermarket 36.2 Supermarket 43.1 Minimarket 13.6 Discount 7.1 Intrinsic features Geographic indication 48.3 Aged vin. 3.0 Organic vin. 6.6 Apple vin. 15.0 White wine vin. 17.8 Red wine vin. 14.7 Discolored vin. 2.2 : Level zero or baseline in the hedonic regressions Producer's brand Ponti 22.1 Cirio 3.7 Fini 3.5 Monari Federzoni 7.9 Ortalli 8.3 Sasso 3.9 Others 50.6 Extrinsic features Spray cap 3.9 Cork type cap 18.5 Screw cap 77.6 Plastic bottle 4.1 Glass bottle 95.9 Dark bottle 39.9 Transparent bottle 60.1 Back label 54.9 Front label only 45.1
The survey: descriptives (2) Sample composition of categorical variables in percent (2) Shelf position High shelf (uppermost) 21.4 High shelf (eye level) 22.3 Middle shelf (hand level) 40.3 Low shelf (leg level) 6.8 Low shelf (bottom level) 9.2 : Level zero or baseline in the hedonic regressions Descriptive summary of continuous variables Median Mean Stdev Min Max Price per bottle ( ) 1.90 2.84 2.80 0.55 16.90 Price per liter ( ) 3.90 6.74 8.48 0.11 67.60 Bottle size (liters) 0.50 0.56 0.33 0.20 5.00 Acidity (% acetic acid) 6.00 5.97 0.53 4.00 7.50 Facings (no. of bottles) 3.00 3.88 3.44 1.00 44.00 26 (10**) indip. var. hedonic price model Variables Coeff. Constant 3.002 0.441 6.804 0.000 Urban area 0.025 0.043 0.574 0.566 Famila Group -0.042 0.047-0.887 0.375 Hypermarket -0.071 0.049-1.446 0.148 Discount -0.009 0.088-0.102 0.919 Minimarket -0.059 0.065-0.905 0.366 Aged vin. 0.075 0.129 0.582 0.561 Organic vin. 0.024 0.089 0.273 0.785 Apple vin. 0.001 0.088 0.011 0.991 White wine vin. 0.165 0.091 1.811 0.070 Red wine vin. 0.133 0.093 1.433 0.152 Discolored vin. 0.270 0.164 1.643 0.101 Logn-Acidity -0.990 0.242-4.083 0.000 R 2 0.543 F 42.110 Dependent var.: Natural Logarithm of vinegar price per liter Variables Coeff. Ponti 0.172 0.059 2.918 0.004 Cirio -0.069 0.124-0.554 0.579 Fini 0.326 0.127 2.566 0.010 Monari Federzoni -0.044 0.090-0.484 0.629 Ortalli 0.035 0.089 0.395 0.693 Sasso 0.464 0.124 3.739 0.000 Cork type cap 0.170 0.084 2.026 0.043 Spray cap 0.305 0.129 2.360 0.018 Plastic bottle -0.073 0.119-0.617 0.537 Dark bottle -0.011 0.065-0.162 0.871 Back label 0.111 0.067 1.649 0.100 Logn-Size -0.941 0.075-12.511 0.000 High shelf (uppermost) 0.053 0.057 0.942 0.347 High shelf (eye level) 0.209 0.055 3.774 0.000 Low shelf (leg level) -0.069 0.082-0.833 0.405 Low shelf (bottom) -0.215 0.081-2.669 0.008 Logn-Facings -0.627 0.039-15.918 0.000 No variable indicating the type of vinegar shows significant effect on price The bottle size shows not only a strongest effect on price, but also the highest VIF (3.346). Regressing logn-size on the other independent variables yields an adjusted R 2 (0.693) The bottle size shows a significant negative correlation with both the bottle caps, the back label and the Sasso and Cirio brands, and a significant positive correlation with the plastic material, the Ortalli brand, the white wine, the red wine and the discolored types, the Famila Group and the discount outlet. 25 (13**) indip. var. hedonic price model Variables Coeff. Constant 3.264 0.475 6.878 0.000 Urban area 0.029 0.046 0.633 0.527 Famila Group -0.104 0.050-2.065 0.039 Hypermarket -0.083 0.053-1.568 0.117 Discount -0.116 0.095-1.230 0.219 Minimarket -0.078 0.070-1.114 0.265 Aged vin. 0.105 0.139 0.755 0.450 Organic vin. 0.034 0.096 0.354 0.724 Apple vin. -0.009 0.094-0.098 0.922 White wine vin. -0.085 0.096-0.888 0.375 Red wine vin. -0.114 0.098-1.171 0.242 Discolored vin. -0.143 0.173-0.827 0.409 Logn-Acidity -0.857 0.261-3.284 0.001 R 2 0.470 F 32.794 Dependent var.: Natural Logarithm of vinegar price per liter Variables Coeff. Ponti 0.216 0.063 3.412 0.001 Cirio 0.345 0.129 2.677 0.008 Fini 0.378 0.137 2.766 0.006 Monari Federzoni -0.004 0.097-0.044 0.965 Ortalli -0.128 0.095-1.351 0.177 Sasso 0.906 0.128 7.071 0.000 Cork type cap 0.764 0.075 10.215 0.000 Spray cap 0.990 0.126 7.865 0.000 Plastic bottle -0.393 0.125-3.138 0.002 Dark bottle 0.007 0.070 0.093 0.926 Back label 0.351 0.070 5.041 0.000 Logn-Size High shelf (uppermost) 0.022 0.061 0.352 0.725 High shelf (eye level) 0.219 0.060 3.670 0.000 Low shelf (leg level) -0.066 0.089-0.748 0.455 Low shelf (bottom) -0.235 0.087-2.712 0.007 Logn-Facings -0.623 0.042-14.673 0.000
Still some remarkable collinearity. Notable VIF for back label and the dark color. The bottle dark color exhibits important negative correlation with vinegars from apple, white wine, red wine, the discolored one, the two caps and the Ponti brand and positive with back label and the Fini, Monari Federzoni and Ortalli brands. The back label is significantly correlated with almost all the previous variables and also (negatively) with the sale in the discount outlets, the Sasso brand and the plastic bottle. Vinegar hedonic price estimates (1) 16** Variables B Percent Price Impact Constant 3.661 0.469 7.803 0.000 Urban area 0.033 0.047 0.693 0.488 Famila Group -0.114 0.051-2.237 0.025-10.89 Hypermarket -0.082 0.054-1.528 0.127 Discount -0.165 0.095-1.730 0.084 Minimarket -0.085 0.071-1.207 0.228 Aged vin. 0.107 0.141 0.761 0.447 Organic vin. 0.032 0.097 0.325 0.745 Apple vin. -0.224 0.076-2.930 0.003-20.30 White wine vin. -0.298 0.077-3.851 0.000-25.96 Red wine vin. -0.316 0.083-3.825 0.000-27.32 Discolored vin. -0.420 0.162-2.594 0.010-35.13 Logn-Acidity -0.911 0.264-3.456 0.001-91.09 R 2 0.457 F 33.449 Dependent variable: Natural Logarithm of vinegar price per liter : Vinegar hedonic price estimates (2) Variables B Percent Price Impact Ponti 0.272 0.063 4.336 0.000 30.94 Cirio 0.273 0.130 2.104 0.036 30.27 Fini 0.390 0.135 2.885 0.004 46.33 Monari Federzoni 0.069 0.096 0.720 0.472 Ortalli -0.073 0.095-0.765 0.444 Sasso 0.827 0.129 6.420 0.000 126.66 Cork type cap 0.771 0.072 10.778 0.000 115.56 Spray cap 1.000 0.125 7.988 0.000 169.73 Plastic bottle -0.524 0.124-4.221 0.000-41.21 High shelf (uppermost) 0.024 0.062 0.385 0.700 High shelf (eye level) 0.215 0.060 3.569 0.000 23.77 Low shelf (leg level) -0.061 0.089-0.677 0.499 Low shelf (bottom) -0.261 0.088-2.982 0.003-23.29 Logn-Facings -0.626 0.043-14.593 0.000-62.62 Low impact of service attributes linked to the point of sale o Only Famila chain 11% reduction on the average price o No rural/urban differences o No differences among point of sale types. But discount Intrinsic features o No effects of the aged and the organic attributes o Apple: the lowest reduction ( 20%) o Discolored: the highest reduction ( 35%) in comparison with the Modena PGI o Noticeable role of acetic acid: (9% price decrease for a 10% increase in its degree) shadow effect.
Brand 1 st Sasso 127% premium. An example of unquestionably strong (olive oil related) brand 2 nd Fini (vinegar specialized) 46% premium Ponti (vinegar specialized) and Cirio (multi-food) about 30% premium Bottle features still important Cork type cap prestige vinegar Spray cap innovative, simplifying the pouring quite high premium (170%) Plastic ( 41%) Merchandising attributes o Bottom and eye tiers: significant effects ( 23%, +24%) o Facings: 100% increase in the number of vinegar bottles 62.6% decrease in the price per liter o Both the facings and the shelf position choices obey to merchandising rules and the promotional strategy of the point of sale o Very low degree of correlation with the other independent variables o Noticeable contribution to the hedonic price model: they account for 186% of the total vinegar price variance. Concluding Remarks (2) Concluding Remarks (2) Similarities and differences relatively to other hedonic price models on food products Goodness of fit does not seem too high. Possible reasons: o Brand price strategies (107 brands) o Point of sale price strategies o Intra-brand price variability not fully explained by the intrinsic and extrinsic features Disproportion between intrinsic and extrinsic features in determining the vinegar price ( wine & olive oil) High contribution of the service-merchandising attributes to vinegar price (consumers scarcely aware of intrinsic quality attributes?)