اثر حق بیمه قیمت منفی در بازار آنلاین-تاثیر رقابت و ارزشمندی خریدار در استراتژی های قیمت گذاری فروشندگان با مختلف سطوح شهرت

Motivated by the contradictory findings in literature regarding whether high-reputation sellers enjoy a price premium over low-reputation sellers, this paper examines the pricing strategies of sellers with different reputation levels. We find that a negative price premium effect (i.e., a high-reputation seller charges a lower price than a low-reputation seller) exists due to: (1) the presence of both informed and uninformed buyers, which makes sellers follow mixed pricing strategies. It is then possible for a high-reputation seller setting a lower price than a low-reputation seller. Moreover, when the proportion of informed buyers exceeds a certain threshold, the expected price of a high-reputation seller is even lower than that of a low-reputation seller; (2) the competition among the sellers, which reduces the high-reputation sellers' prices but increases the low-reputation sellers' prices. Consequently, a high-reputation seller is more likely to charge a lower price than a low-reputation seller when the competition intensifies. Our empirical findings also support our theoretical results on the negative price premium effect.

Online markets attract a lot of sellers due to the low entry and operational costs [38]. For example, in the “Electronics” category of BizRate.com, a famous price comparison shopping website, there are 3830 retailers, and more than 50 distinct retailers offering “Canon PowerShot SX210 IS 14.1 Megapixel Digital Camera — Black”.3 However, it is not easy to inspect seller identity as well as product quality in online markets. The sellers are often hidden under the masks of meaningless electronic IDs [19]. At the same time, payment and delivery for the products are also separated [2]. These online market characteristics create chances for opportunistic behaviors, such as non-delivery, identity theft, and miscellaneous fraud [15] and [18]. In the year 2009 alone, the Internet Crime Complaint Center (IC3)4 website received 336,655 complaint submissions, corresponding to a $559.7 million dollar loss [18]. Fortunately, current information technologies help reduce these risks and facilitate buyers to infer seller quality through various reputation mechanisms, such as buyer ratings and reviews, feedback systems, online discussion forums, etc. [3], [22], [35] and [39]. It is commonly believed that buyers are likely to pay price premiums to high-reputation sellers, so the high-reputation sellers should charge relatively high prices [3], [22], [25], [38] and [39]. However, some studies find the reverse. For example, Ba et al. [4], [5] and [6] identify the “adverse price effect,” which shows a seller may decrease her price when her recognition level increases. Baylis and Perloff [10] show that “good” internet retailers of digital cameras and scanners charge relatively low prices and provide superior services, while “bad” internet retailers charge relatively high prices and provide poor services. Motivated by these contradictory findings in literature, this study aims to understand the pricing strategies of sellers with different reputation levels, and examine whether, and under what conditions, does a “negative price premium effect” occur (i.e., a high-reputation seller charges a lower price than a low-reputation seller). Note that this is different from the “adverse price effect” studied in Ba et al. [4], [5] and [6], which refers to the phenomenon that when the low-recognition seller's recognition increases, both the low- and high-recognition sellers cut their prices [6]. In this paper, we first build a theoretical model to study the effect of competition. We extend Varian's sales model [36] in two ways: to allow sellers to have different reputation levels (the benchmark model); and to allow more than one seller with the same reputation level (the competition model). We find that the negative price premium effect exists due to: (1) the co-existence of informed and uninformed buyers, which makes it impossible for sellers to set their prices following pure strategies. When the proportion of informed buyers exceeds a certain threshold, a high-reputation seller even sets a lower price than a low-reputation seller on average; (2) the competition among the sellers, which makes a high-reputation seller reduce the prices while a low-reputation seller increase the prices. So the negative price premium effect is more likely to occur when the competition intensifies. We also collect field data from BizRate.com. Our empirical testing supports our theoretical findings on the negative price premium effect. To the best of our knowledge, this is one of the first few papers which study the negative price premium effect. It theoretically explains the negative price premium effect from the perspective of buyer informativeness, which is an extremely important factor in the economics of information [21] and [27]. Specifically, we show that sellers may play mixed strategies, so there is no simple and fixed relationship between seller reputation and pricing. Our study offers an explanation to the contradictory findings in the literature. The rest of this paper is organized as follows: we review relevant literature in Section 2, and present the main theoretical model in Section 3. To better understand the impact of competition, we present a benchmark model in Section 3.1 in which there is only one seller for each reputation level, and relax this assumption in Section 3.2. We present the empirical study in Section 4 and conclude in Section 5.

In this paper, we study the pricing strategies of sellers with different reputation levels in online markets. We identify three factors, that is, seller reputation, competition, and buyer informativeness, which impact a seller's pricing strategy. We find that a negative price premium effect exists due to: (1) the co-existence of informed and uninformed buyers in the market, which makes sellers' prices follow mixed strategies, so that it is possible for a high-reputation seller to set lower prices than a low-reputation seller. Furthermore, when the proportion of informed buyers exceeds a threshold, a high-reputation seller may even set lower prices on average than a low-reputation seller. (2) The competition among sellers. Competition has different effects on sellers with different reputation levels. The competition makes the high-reputation sellers reduce their prices while the low-reputation sellers increase their prices. Therefore, the negative price premium effect is more likely to occur when the competition intensifies. Our empirical data also confirm our theoretical findings of the negative price premium effect. This paper offers an explanation on the contradictory findings in literature [3], [10], [23] and [38] regarding whether and when a high-reputation seller enjoys a price premium. The existing explanations of such mixed findings include whether the reputation score is properly calculated (e.g., whether the reputation score is calculated by the number of positive ratings, or the difference between the number of positive ratings and negative ratings) [25], or whether a proper regression model is used (e.g., OLS or Tobit) [24]. Different from these findings, our results show that there is no simple and fixed relationship between seller reputation and pricing. Sellers may play mixed pricing strategies when informed buyers coexist with uninformed buyers, so a negative price premium effect may occur. This also indicates that, simply attributing price dispersion to seller differentiation may not be sufficient. This research also extends our understandings on the effect of competition. The common knowledge is that the competition reduces seller price. However, our model shows that, a low-reputation seller may give up the competition for informed customers and increase their prices when competition intensifies, a “counter-intuitive” phenomenon in online market. This research also offers practical implications. First, a high-reputation seller may seek a mixed strategy in pricing, such as offering discount from time to time, when the competition is fierce or when they serve many informed buyers. Second, our findings encourage sellers to adjust their pricing strategies across different product categories based on the product value, or, the proportion of informed buyers. This research is not without limitations. First, it is difficult to empirically measure the buyer informativeness, i.e., the proportion of informed buyers. According to the findings in the literature [3] and [25], we assume that it is less likely for online buyers to invest in searching for low value products (such as DVDs). In other words, this study uses product category as a proxy of buyer informativeness. It would be helpful to develop a more formal approach. Second, our theoretical model only considers four players, i.e., two high-reputation sellers vs. two low-reputation sellers. A real online market is more complex. Furthermore, there may be other explanations on the negative price premium effect: the sellers may have different cost structures caused by economies of scale or advertising cost [6]; the low-reputation sellers may charge relatively high prices in order to signal their quality, etc. It will be interesting to study the impact of these factors and empirically test these effects. Future study may focus on the following extensions: (1) collecting data in more product categories, or from multiple websites, to verify our theoretical findings; (2) developing a more formal empirical model to better understand seller pricing strategies; and (3) developing a theoretical model of a large number of sellers with their reputation randomly distributed, to simulate the online markets more accurately.