تصمیم گیری درباره مساله روزنامه فروش: یک مطالعه آزمایشگاهی در سطح ملی

In this paper, we conduct a laboratory experiment using the classic newsvendor problem to examine cross-national differences in inventory ordering patterns between Chinese and American decision makers based on a theoretical examination of the role of the Doctrine of the Mean in Chinese decision making. Drawing on the theory of context-dependent preferences (specifically extremeness aversion), we also revisit the flat-maximum hypothesis of Bolton and Katok [12], i.e., “thinning the set of order options leads to newsvendor decisions that achieve a higher proportion of maximum expected profit.” The results show that the “pull-to-center” effect is more prominent for Chinese than Americans, i.e., average order quantities of Chinese subjects are closer to the anchor of mean demand than those of American subjects. Furthermore, we find that thinning the set of order options such that the optimal order quantity is a middle option, not an extreme option in the choice set, leads to better performance in newsvendor decisions, which complements the flat-maximum hypothesis.

Cross-national differences have been well documented in judgment and decision making, including differences in probability judgments [1], [2] and [3] and risk preferences [4] and [5]. For example, Yates et al. [2] found that Chinese respondents showed greater overconfidence in the accuracy of their answers to general knowledge questions than Americans and Europeans, and they proposed that cross-cultural differences in education might account for such cross-national differences in overconfidence. Despite the prevalence of cross-national studies in judgment and decision making, there is little research in operations management that investigates cross-national differences in operations decisions. Nowadays business operations are set in a global environment and firms need to deal with suppliers and customers in global supply chains. For example, according to the World Trade Organization's annual report (1998, p. 36, http://www.wto.org/english/res_e/booksp_e/anrep_e/anre98_e.pdf), the production of a particular “American” car takes place across a number of different countries. Thirty percent of the car's value comes from Korea for assembly, 17.5% from Japan for components and advanced technology, 7.5% from Germany for design, 4% from Taiwan and Singapore for minor parts, 2.5% from the United Kingdom for advertising and marketing services and 1.5% from Ireland and Barbados for data processing. Also, about 70% of the products sold on Wal-Mart's shelves are made in China [6]. At the national level, trade between the United States and China has been dramatically increasing since China joined the WTO in November, 2001. According to the US Census Bureau, China was the second largest trading partner for the US in 2007 (specifically, the US imported goods worth nearly 321.5 billion dollars from China in 2007). These above examples indicate that operations decisions within a global supply chain may come from parties with different backgrounds. Since cross-national issues can affect procurement, production and distribution in global operations, purportedly optimal solutions predicted by theories in operations management are unlikely to apply across decision makers from different countries. In addition, ignorance or misperception of cross-national differences may lead to substantial operational inefficiencies in international businesses. Thus, it is important to study behavioral differences in decision makers from different countries and understand the effect of cross-national differences in global operations. In this research, we aim to examine cross-national differences in the newsvendor setting. Recently there has been a growing interest in behavioral operations management (see reviews in [7], [8] and [9]). In this stream of research, a number of papers have employed the classic newsvendor problem to study decision makers’ behavior by using laboratory experiments and present inconsistencies between theoretical predictions and empirical observations in the newsvendor setting [10], [11], [12], [13], [14], [15], [16], [17] and [18]. It has been widely observed that average order quantities systematically deviate from the optimal expected profit-maximizing quantity and actually fall between the mean demand (e.g., the center of the uniform demand interval) and the optimal order quantity, i.e., subjects order too few when they should order more and vice versa. This is called the “pull-to-center” effect in [15]. Several decision biases have been proposed to explain the “pull-to-center” effect in the newsvendor experiments. Specifically, Schweitzer and Cachon [10] proposed two decision biases in their study, i.e., a preference for minimizing ex-post inventory error, and the anchoring and insufficient adjustment bias. Su [17] proposed the boundedly rationality bias, in which a boundedly rational decision maker is prone to errors in the newsvendor experiments. Croson et al. [18] proposed the overconfidence bias, i.e., subjects tend to underestimate demand variation in the newsvendor decisions. As discussed above, the “pull-to-center” effect has been replicated in various studies [10], [11], [12], [13], [14], [15], [16], [17] and [18]. However, almost all of these studies used subjects from western countries (e.g., the United States) for their experiments. Prior literature suggests cross-cultural differences in decision making between Western and Eastern countries [5], [19], [20], [21] and [22], for example, the United States and China are two representative countries with different cultures and decision makers from these two countries may behave differently in many aspects. Thus, in this study, we chose to examine possible differences in ordering patterns between Chinese and Americans in the newsvendor setting. The main contribution of this paper is two fold. Firstly, to the best of our knowledge, our work is the first cross-national study in behavioral operations management using laboratory experiments. More specifically, from the perspective of cross-cultural differences between American and Chinese people, we conjecture that the “pull-to-center” effect is more prominent for Chinese than Americans, i.e., average order quantities of Chinese subjects are closer to the anchor of the mean demand than those of American subjects. The “pull-to-center” effect has been widely observed in the newsvendor experiments using American subjects. However, the ordering pattern is not known for Chinese decision makers in newsvendor decisions. Building on the theory of the Doctrine of the Mean, we hypothesize that Chinese tend to exhibit a more prominent “pull-to-center” effect than Americans in newsvendor decisions. This hypothesis is examined in Study 1. Secondly, we identify conditions under which decision makers may perform better in newsvendor decisions. Previous experimental studies in the newsvendor problem suggest that decision makers nearly always incur inefficiencies by ordering a quantity different from the expected profit maximizing quantity. Bolton and Katok [12] propose that thinning the set of order options improves decision makers’ performance in the newsvendor problem, which they called a flat-maximum hypothesis. 1 Through a laboratory experiment, they find that reducing the number of ordering options does not necessarily result in better performance for newsvendor decisions. On the basis of the theory of context-dependent preferences, we propose that extremeness aversion explains why the flat-maximum hypothesis was not supported. Specifically, we hypothesize that thinning the set of order options such that the optimal order quantity is a middle option, not an extreme option in the choice set, leads to better performance in newsvendor decisions. Study 2 is conducted to test this hypothesis. The rest of the paper is organized as follows. Section 2 presents Study 1 to examine cross-national differences in ordering patterns between Americans and Chinese from a cross-cultural perspective. Section 3 describes Study 2, which revisits and complements the flat-maximum hypothesis in newsvendor decisions. We discuss some managerial implications and limitations of the study, and conclude the paper in Section 4.

In this paper, we conduct a laboratory experiment using the classic newsvendor problem to examine cross-national differences in ordering patterns between Chinese and American decision makers based on a theoretical examination of the role of the Doctrine of the Mean in Chinese decision making. Drawing on the theory of context-dependent preferences (specifically extremeness aversion), we revisit the flat-maximum hypothesis of [12]. The results show that the “pull-to-center” effect is more prominent for Chinese than Americans, i.e., average order quantities of Chinese subjects are closer to the anchor of mean demand than those of American subjects. Furthermore, we find that thinning the set of order options such that the optimal order quantity is a middle option, not an extreme option, in the choice set, leads to better performance in newsvendor decisions, which complements the flat-maximum hypothesis. Our study sheds some light on global operations in the sense that firms may need to understand differences in procurement patterns to successfully manage various functions in global supply chains. Our experimental findings are limited to the newsvendor setting and more research is needed to help advance our understanding of cross-national differences in global operations. Nonetheless, our general approach suggests that cultural differences are likely to arise. For example, suppose a US retailer opens a chain store in China and implements vendor managed inventory (VMI) with a Chinese supplier to manage one of the high-profit products. The retailer's business decisions, e.g., shelf arrangement and promotions, depend on her understanding of the inventory policies being used. The retailer may assume that the Chinese supplier manages the inventory in a way similarly to how she does. However, our findings suggest that the Chinese supplier is likely to fulfill the inventory differently than what the US retailer has expected. For example, the US retailer may arrange too much shelf space for this product, or assign too large of a sales force/advertising budget to manage this product based on her own projections of inventory replenishment and associated costs and profits. In this research, we base our findings on the theory of the Doctrine of the Mean. One potential limitation of this study is that there might be some other factors which could play a role in cross-national ordering differences. For example, differences in the demographic information between Chinese and American respondents might also have an impact on their ordering difference. As mentioned before, professional working experience has been shown not to be a critical factor in behavioral newsvendor decisions in [13]. However, this might not be true for other demographic indicators, such as gender, race and age. Finally, as mentioned before, we are, to our knowledge, the first to conduct a cross-national study in behavioral operations management using laboratory experiments. According to the major findings on newsvendor decisions in our paper, we believe that there may also be significant cross-national differences in decision making in other emerging areas of behavioral operations management, such as behavioral issues in supply chain contracting, sourcing and auctions. Related studies on these topics, perhaps using a verbal protocol approach to examine operational decisions as in Gavirneni and Xia [40], would be interesting and merit further research.