Printer Friendly

Association between market price and seller/market characteristics.

Over the past 40 years an extensive literature has emerged addressing association between price and quality. In an article surveying the first 35 years (1950-1985) of this literature, Geistfeld (1988) concluded that the evidence supporting association between price and quality to be weak. In a study examining the association between used car reliability and price, Ginter, Young, and Dickson (1987) found that reliability was not reflected in price. Bodell, Kerton, and Schuster (1986), using Canadian data, found a weak association between price and quality. Hjorth-Andersen (1987) concluded that it is not generally wise to purchase the most expensive model to reduce the risk associated with the purchase of a product.

Price/quality research has typically ignored the unique aspects of specific markets, differences types of retailers, and ancillary seller services. The importance of market characteristics was recognized by Sproles (1977) who noted that there may be variability in prices and models avaialable in local markets because of market-related factors such as size and number of competitive retailers, size of the trading area, and local economic conditions. In an analysis of three markets Geisfeld, Maynes, and Duncan (1980) found that markets differed with respect to number of seller for a given model and number of models for a given product; however, it was not clear if one market was systematically more "expensive" than another.

The possible effect of seller characteristic on the association between price and quality has been looked at in several studies. The evidence, however, is conflicting. Duncan (1981) noted that if seller characteristics were associated with price, one would expect hihg service stores to have consistently higher prices and low services stores to have consistenly lower prices for all models. Analysis of the data revealed this was not the case. Examining five store types Geistfield (1982) found average Spearman correlation coefficients between price and quality to rage from - .14 to .65. This suggest that seller characteristics may affect the association between price, and quality. Ginter, Young, and Dickson (1987) recognized the possible effect of seller characteristics when they noted the National Automobile Dealers Association (N.A.D.A.) used car prices could reflect elements other than reliability--perhaps seller services.

If seller/market characteristic are a significant part of the set of characteristics consumers purchase and are reflected in the price paid for a product, studies relating price of a good to measure of quality not reflecting seller/market characteristics may not reflect what happens in a given market. While the need to be sensitive to seller/market characteristics has been recognized in the price/quality literature, there has been no systematic effort to ascertain whether seller/market characteristics are reflected in product prices. The research reported here is a first attempt to directly focus on the association between seller/market characteristics and price.


When purchasing an item, consumer buy the product per se and a set of seller services. A "high service" retailer would provide, in addition to the product, an environment in which product repairs are easily and correctly made, returns are handled with mininal hassle, sales personnel are knowledgeable, etc. A "low service" seller may provide no more than the actual product.

Numerous seller characteristics exist that could affect market price of a given product model. These range from the types of services retailers offer to the physical environment of a store. In particular, they include

Return Policy. Offering flexible requirements with respect to proof of purchase and a relatively long-time period over which replacement/exchange is honored;

Warranties. Offering warranties directly or handling manufacturer warranties;

Repair Service. Providing access to service facilities;

Physical Access. Proximity to other stores, access to parking, and adeuacy of parking; and

Store Environment. Number of sales personnel, checkout facilities, forms of payment accepted, hours of operation, training of sales personnel, and depth of offering.

To the extent the seller services, affect information search costs and expected flow of services arising from a product, consumers should be willing to pay more a given model having more seller characteristics associated with it. If seller characteristics are unrealted to search cost or service flow, there should be no relationship between seller characteristics and the price consumers are willing to pay for a model from a particular seller.

Seller characteristics could also affect price from the cost side since high service sellers would be expected to have higher costs associated with the provision of services than would low service seller. To the extent that sellers are able to pass costs onto consumers, one would expect prices to be affected by seller characteristics.

Market characteristics describe the environment in which a store is situated. Number of sellers, size and growth of a market, and income of consumers in a market are several market characteristics. While these characteristics are independent of any specific seller, they can affect the price for which a product is sold in a given market. If there are many seller, prices are likely to be lower due to competition. Areas experiencing population growht could experience higher prices for a given product because market demand may be growing more quickly than number of sellers. High income areas will likely exhibit a greater demand for products because consumers have more discretionary income suggesting the possibility of higher prices.


Electric irons were the product selected for study. The primary reasons for focusing on small electric appliances were (1) purchase price and listed store price are generally the same and (2) many stores sell small electric appliances.

Data and Sample

Data were collected in Columbus, Ohio, during 1988. The sample consisted of 28 stores. All of the stores were located within the area defined by the outerbelt circling the city. Data collection included an inventory of all models of electric irons sold in each store including brand, model number, price, and sale price where applicable. Information with respect to seller characteristics included store return policy, warranties for appliances sold in the stores, information with respect to repair/replacement of defective products, proximity to other stores, ease of access to store from parking facilities, adequacy of parking, hours open for business, types of payment accepted employee training, and number of personnel. The market characteristic collected was 1980 mean census tract income for the tract in which a specific store was located.

Price information on iron models is reported in Table 1. The table contains maximum and minimum (nonsale) prices and relative range for each model. Relative range--the difference between maximum and minimum model prices divided by minimum price (Geistfield, Maynes, and Duncan 1980)--reflects the extent to which maximum price exceeds minimum price for a given model, with larger values indicating

TABLE 1 Market Price Variation for Electric Irons in Columbus, Ohio, 1988
 Maximum Minimum
Model $ $ Range N
 1 59.99 34.99 0.71 13
 2 49.99 43.99 0.14 2
 3 39.99 29.99 0.33 11
 4 37.99 19.99 0.90 11
 5 34.99 30.49 0.15 2
 6 31.99 24.94 0.28 5
 7 29.99 16.99 0.77 5
 8 29.99 26.99 0.11 5
 9 24.99 18.99 0.32 6
 10 24.99 16.93 0.48 8
 11 23.99 14.99 0.60 5
Total 64

a greater price range. From Table 1 it can be seen that, holding product quality constant, substantial price variation existed in the market. Of the 11 models of electric irons, four had relative ranges exceeding 0.5 indicating that the most expensive store's price was 50 percent greater than the least expensive store's price for the same model.

The distribution of seller characteristics over 28 stores is presented in Table 2. Twenty-two stores required proof of purchase for exchange of a good. The majority of stores (18) allowed more than 90 days for return of a defective product, and 11 of the 28 stores provided their own warrantly or handled manufacturer warranties. Fifteen of the 28 stores provided repair facilities either at point of purchase or at a store-supported central repair facility.

Nineteen stores were locted in either a strip shopping center with an anchor store or were freestanding. The majority of stores (21) had more than adequate parking facilities. Seven of the 28 stores provided specialized sales staff. Stores accepted an average of 4.8 methods of payment (cash, checks, credit cards, store credit cards, etc.). Operating hours per week ranged from 48 to 90 hours with a mean of 73.3 hours. The average number of electric iron models sold in stores was 6.3.


To determine whether seller characteristic are reflected in product price, price variation based on differences in product quality per se must be controlled. Because no measure of product quality was available, this was accomplished by using normalized price, Z-price scores, as the price measure. Z-price scores were calculated by subtracting store price of a given model from mean market price of the model and dividing the difference by the standard deviation of mean market price of a model.

Z-price = P /--P std. dev. (i = 1, 2, . . . , N models of irons)

(j = 1, 2, . . . . , K stores)



[Z-price.sub.ij] = normalized price of model i sold in store j,

[P.sub.ij] = price of model i sold in store j,

[P.sub.i] = mean price of model i in the market, and

[std. dev..sub.i] = standard diviation of the mean market price of model i.

Z-prices facilitate comparison of model prices such that product quality does not confuse the analysis. Average market price of a given model reflects what consumers, on average, are willing to pay for the product. Price of a given model from a given store reflects not only what consumers are willing to pay for the product itself, but also the extent to which a store's customers are willing to pay for a particular seller's services or the extent to which a seller's costs are reflected in price. The difference between average market price for a model and price for which a model is sold in a specific store is the component of price which reflects seller/market characteristics.

Variables used in the study are listed in Table 2. A set of dummy variables, listed in the first part of the table, was created to represent most of the seller characteristics included in the study. For these variables, yes or existence of the "nonbase' situation is reflected by the value of the variable being set equal to one; the variable being set equal to zero indicates the stated condition is not met. For example, RETURN is 1 when a store accepts returned items without question and 0 when a store has specified conditions to be met before it accepts returned items.

Type of store was introduced into the analysis as a control variable because it may affect the price of a given model of electric iron independent of seller/market characteristics. Large retailers may be able to capitalize on economies of scale from purchasing and selling strategies not available to smaller retail establishments. Mass merchandisers included K-mart and Gold Circle. J.C. Penney was included with department stores, and no Sears stores were included because at the time of the study they did not sell national-brand irons.

The following model was estimated using the SAS general linear model algorithm.



In interpreting results of this study, several limitations should be recognized. First, data were obtained in the Columbus, Ohio, market and results may not be generalizable to other markets. Second, the analysis assumes a linear relationship between characteristics and price variation, and only the main effects of these variables were included in the model. To the degree to which a linear specification is inappropriate and/or the effect of the predictor variables are not independent, coefficients underestimate the underlying relationship.

The unit of observation was models sold in specific stores; therefore, if a given store sells ten models of irons, ten normalized prices from the store entered into the regression--one for each model. Estimated coefficients presented in Table 3 are beta coefficients. The set


of independent variables explained 65 percent of the variation in normalized price. The equation F-statistic (6.57) was significant at the .05 level. Examination of the correlation matrix suggested colinearity was not a problem.

Allowing consumers more than 90 days to return a defective product or providing/handling warranties resulted in significantly different prices relative to the omitted categories (customer handling of warranty and 30-90 day return). Unquestioned returns and provision of store-based or centrally based repair facilities did not significantly affect price.

Stores that were freestanding, in a strip center with an anchor, in a small shopping mall, or had more than adequate parking did not have electric iron prices significantly different from those charged by stores not having these attributes. Number of payment methods accepted by a store did not significantly affect price; however, operating hours were significantly related to price. Product depth was not found to be related to product price, while provision of personnel with specialized training was found to be associated with price.

The market characteristic 1980 mean census tract income was not significantly related to price. Department stores (a control variable) had significantly different prices relative to the omitted category (mass merchandisers and others).


The question posed previously is whether seller/market characteristics affect market price. Of the 12 seller/market characteristics, four had a statistically significant effect on price. In addition, one of the control variables significantly affected price.

While it goes beyond the literal purview of this paper, it would be informative to briefly consider the signs of the statistically significant coefficients. It is not surprising that stores handling or providing warranties, and number of operating hours had a positive effect on price. What is surprising is the negative effect on price of store willingness to accept returns beyond 90 days and whether store personnel receive specialized training. Intuition suggests that if these factors affect price at all, they should be associated with higher prices.

The negative effect on price when stores provide or handle warranties could reflect that few warranty claims are made for electric irons. Because there are few claims, consumers would see seller provided warranty services as having little or no value and would be unwilling to pay a higher price for an unneeded service. Pushing this reasoning a bit further, it is possible that consumers perceive stores offering such services as "high price" sellers resulting in these sellers having to sell at a reduced price to attract customers.

Specialized training for store personnel would be of value to consumers purchasing complex, expensive products. Electric irons clearly do not fall into this category. As argued, stores providing such services may have to offer especially attractice prices for products such as electric irons to offset the belief that knowledgeable salespersons are associated with high prices.


The most significant conclusion of this research is that it is not appropriate to dismiss the effect of seller/market characteristics on local market prices. What are the implications of this finding for future research on association between price and quality?

When assessing association between price and quality at a highly aggregated level (e.g., national), market/seller characteristics are not of interest because they are typically associated with a specific market. This suggests that average selling price is an appropriate price measure because it reflects product quality per se--the averaging process removes the effect of any particular set of seller characteristics on price. However, when focusing on specific markets, this study suggests that seller/market characteristics should be incorporated into the analysis.

An important issue, then, is how could one incorporate seller/ market characteristics into an assessment of association between price and quality? One way to accomplish this would be to use Consumer Reports as the source for product quality and to conduct a retail store survey to collect price and seller characteristic data. Market data could be taken from appropriate secondary sources. Using price as the dependent variable with product quality and seller/ market characteristics as independent variables, one could use multiple regression to investigate association among market price, product quality, and seller/market characteristics. The product quality measure could be developed by using a set of dummy variables reflecting the quality quartile (quintile, decile, etc.) into which a particular model falls. This type of quality measure is necessary because most quality data provided by Consumer Reports are ordinal. A study of this type would not require the normalization of price as in equation 1 because product quality would be explicitly accounted for in the analysis. An important element to consider when identifying products is to select ones recently reported in Consumer Reports or in another consumer product test publication. If one did not do this it would not be possible to incorporate a measure of product quality into the study.

In addition to examining association between price and quality, additional work is needed on the effect of seller characteristics on price. This work should be based on an expanded product set that takes into consideration a wide range of product functions and price levels. Future research should also account for sales volume because using store characteristic data not weighted by sales volume treats a store selling one of a given model the same as a store selling 100 of the same model.

In conclusion, while the research reported is preliminary, it provides evidence that seller/market characteristics affect market price of products. This suggests that when assessing the degree of imperfection in local consumer markets by examining association between price and quality, researchers need to be sensitive to seller/market characteristics and account for them in a way that insures the price/ quality relationship reflects the actual situation.


Bodell, Richard W., Robert R. Kerton, and Richard W. Schuster (1986), "Price as a Signal of Quality: Canada in the International Context," Journal of Consumer Policy, 9: 431-444.

Duncan, Greg J. (1981), "The Dynamics of Local Markets: A Case Study of Cameras," The Journal of Consumer Affairs, 15(1, Summer): 64-74.

Geistfeld, L. V. (1988), "The Price Quality Relationship: The Evidence We Have, the Evidence We Need," in The Frontier of Research in the Consumer Interest, E. Scott Maynes et al. (eds.), Columbia, MO: American Council on Consumer Interests: 143-172.

Geistfeld, L. V. (1982), "The Price-Quality Relationship--Revisited," The Journal of Consumer Affairs, 16(2, Winter): 344-346.

Geistfeld, L. V., E. Scott Maynes, and Greg J. Duncan (1980), "Informational Imperfections in Local Consumer Markets: A Preliminary Analysis," in Advances in Consumer Research, Volume VIII, Jerry C. Olson (ed.), Ann Arbor, MI: Association for Consumer Research: 180-185.

Ginter, James L., Murray A. Young, and Peter R. Dickson (1987), "A Market Efficiency Study of Used Car Reliability and Prices," The Journal of Consumer Affairs, 21(2), Winter): 258-276.

Hjorth-Andersen, Chr. (1987), "Price as a Risk Indicator," Journal of Consumer Policy, 10: 267-281.

Sproles, George B. (1977), "New Evidence on Price and Product Quality," The Journal of Consumer Affairs, 11(1, Summer): 63-77.

Loren V. Geistfeld is Professor, Department of Family Resource Management, The Ohio State University, Columbus, OH; and Rosemary J. Key is Assistant Professor, Department of Consumer Economics and Housing, Cornell University, Ithaca, NY. Senior authorship is not assigned.

Comments by Peter Dickson and reviewers of this Journal were most helpful to the development of this paper.
COPYRIGHT 1991 American Council on Consumer Interests
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Geistfeld, Loren V.; Key, Rosemary J.
Publication:Journal of Consumer Affairs
Date:Jun 22, 1991
Previous Article:Industry characteristics and consumer dissatisfaction.
Next Article:Extended warranties: consumer and manufacturer perceptions.

Related Articles
Structuring the perfect package.
TQM and the market structure.
Gains in Stock Market and Homeowners' Equity Propelled Housing Market During First Quarter, C.A.R. Housing Survey Reports.
What Determines a House's Selling Price?
Rising Mortgage Interest Rates, Multiple Offers Impacted Housing Market In First Quarter, C.A.R. Survey Reports.
Survey Shows Buyers and Sellers Use Technology and Want Personal Services.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters