The Empirical Research On The Impact Of Sellers'Reputation On Sales In The Online Exchange

The inherent temporal and physical separation between buyers and sellers in impersonal online marketplaces poses product quality uncertainty and seller quality uncertainty as a result of asymmetric information, which leads two information problems:the pre-contractual problem of adverse selection and the post-contractual problem of moral hazard. The sellers possess more information than buyers about the quality of the products. Buyers and sellers are self-interested parties with incongruent goals, and buyers can not fully monitor sellers' behavior. As a result of quality uncertainty, online marketplace may become a lemon market.Online feedback mechanism is an effective way to handle the problems as a signaling and incentive mechanism. The sellers'reputation profiles function as a signal of trustworthiness. Good feedback profile provides a signal of trustworthiness to potential buyers, while poor feedback profile may scare buyers away. In addition, the feedback mechanism provides sellers incentives to guard their good feedback profile. Repeat sales and price premium create incentives for sellers to act cooperatively. Game theory analysis suggests that self-interested agents (sellers) tend to cooperate given that payoffs from cooperation are higher than that from cheating in repeated transactions. In short, feedback mechanism discourages dishonest and opportunistic behaviors, and it is an effective way to build trust in online shopping market.There has been a great amount of papers published on the topic of the impact of sellers' reputation on auction price in the English auctions. This paper studies the feedback mechanism in a different research context, the fixed-price online exchange marketplace. Data from electronic market and the Internet has many advantages over other data sources, and it is much more reliable and richer than other data sources.Thus, the data of this research is from a real website, taobao, which is the largest online marketplace in China. Data collection is always a time-consuming and labor intensive job, so a spider program is developed to crawl the data from the web pages. The spider collected about 100,842 listings involved eighty kinds of products.The products whose descriptions contain the key words, but are not the targeted goods or the products which are not brand-new and sealed are eliminated. Finally, we have 97,387 valid observations.The transaction data was collected automatically by a "spider" program, obtaining huge dataset. Research data include product name, sales, posted price, reputation score, reputation rank, membership of Business Alliance Program and Consumer Protection Program.A log-log model may reduce heteroskedasticity compared to a linear model.Both sales and reputation score are logarithmic transformed.The correlation between sales and price, sales and reputation indicators is statistically significant, however, some independent variables are correlated, as a result, a formal multicollinearity test is conducted.Most of sellers sell nothing as a result of high prices or low reputation. The dependent variable, sales, is censored at zero. The relationship between sales and reputation indicators or price resembles a broken line instead of a straight line. Tobit regression is used to fit a broken line instead of a straight line, thus it can handle censored data. As a result, OLS regression and Tobit regression are both used to analyze the data, however, the results show that Tobit regression is more suitable.Through OLS and Tobit regression, we found that the effect of reputation rating score on sales is statistically significant. Both product value and product type affects buyers' purchase involvement. Given the greater risks inherent in the exchange of expensive products, buyers would seek more trustworthy sellers with whom to conduct business.Online goods are divided into four categories (search, experience-1,experience-2,and credence) based on quality uncertainty. Buyers are willing to transact with sellers with stellar reputation profiles when they buy products involving greater quality uncertainty.Feedback mechanism can prevent electronic market from becoming a lemon market effectively, leaving a healthier market with a variety of prices and service qualities.