How artificial intelligence will affect the future of retailing

Artificial intelligence (AI) will substantially impact retailing. Building on past research and from interviews with senior managers, we examine how senior retailing managers should think about adopting AI, involving factors such as the extent to which an AI application is customer-facing, the amount of value creation, whether the AI application is online, and extent of ethics concerns. In addition, we highlight that the near-term impact of AI on retailing may not be as pronounced as the popular press might suggest, and also that AI is likely to be more effective if it focuses on augmenting (rather than replacing) managers’ judgments. Finally, while press coverage typically involves customer-facing AI applications, we highlight that a lot of value can be obtained by adopting non-customer-facing applications. Overall, we remain very optimistic as regards the impact of AI on retailing. Finally, we lay out a research agenda and also outline implications for practice.     

Contemporary Extensions of the Rasch Model

This paper will introduce, discuss and illustrate two contemporary extensions of theRasch model: the one parameter logistic model (Verhelst and Glas, 1995) and theMultidimensional Rasch model (Hoijtink et al., 1999). Using data with respect tothe measurement of schizotypy (Vollema and Hoijtink, 2000) the most importantfeatures of both models will be illustrated. For the one parameter logistic modelthese include: a (discrete) discrimination parameter for each item; a test for itembias; and, estimation of the location of a person on the (latent) trait that is beingmeasured. For the multidimensional Rasch model these include: specification ofthe model; and, model selection. All analyses presented in this paper can be executedusing either OPLM (Verhelst et al., 1995), TESTFACT (Wilson et al.,1984) or ConQuest (Wu et al., 1998). At the end of the paper some features ofmodels and software that have not been discussed will be summarized.     

Misperceiving the value of information in predicting the performance of others

Economic models typically allow for “free disposal” or “reversibility” of information, which implies non-negative value. Building on previous research on the “curse of knowledge” we explore situations where this might not be so. In three experiments, we document situations in which participants place positive value on information in attempting to predict the performance of uninformed others, even when acquiring that information diminishes their earnings. In the first experiment, a majority of participants choose to hire informed—rather than uninformed—agents, leading to lower earnings. In the second experiment, a significant number of participants pay for information—the solution to a puzzle—that hurts their ability to predict how many others will solve the puzzle. In the third experiment, we find that the effect is reduced with experience and feedback on the actual performance to be predicted. We discuss implications of our results for the role of information and informed decision making in economic situations. Electronic Supplementary Material Supplementary material is available in the online version of this article at http://dx.doi.org/10.1007/s10683-006-9128-y. JEL Classification C91, D83     

Does the spatial representation affect criteria weights in environmental decision-making? Insights from a behavioral experiment

The need and interest to consider cognitive and motivational biases has been recognized in different disciplines (e.g. economics, decision theory, risk analysis) and has recently reached environmental decision-making. Within this domain, the intrinsic presence of a spatial dimension of both alternatives and criteria calls for the use of maps throughout the decision-making process to properly represent the spatial distribution of the features under analysis. This makes spatial Multi Criteria Decision Analysis (MCDA) a particularly interesting domain to explore new dimensions of cognitive biases. This study proposes a behavioral experiment aimed at discovering to what extent the spatial visualization (i.e. maps) of criteria versus the non-spatial one (i.e. tables) can bias the weight elicitation phase of a spatial MCDA process. The experiment simulates a very common analysis in environmental and land use planning: land suitability analysis. Our findings show that there are significant consequences on how important we perceive a certain criterion to be, depending on whether it is represented as a map or as a table among a mix of maps and tables. Indeed, the map representation of the same criterion leads to higher weights attributed to that criterion compared to the table representation. Visualizing the same information as a map or as a table, although technically equivalent, is thus not psychologically equivalent for Decision Makers. The results of this experiment are expected to have implications for spatial decision-making processes, by generating better awareness on the impacts of map-mediated land suitability analysis.     

Asymptotically unbiased estimation of autocovariances and autocorrelations for panel data with incidental trends

We consider the estimation of autocovariances using panel data with incidental trends under double asymptotics. The conventional autocovariance estimator suffers from a bias whose value is approximated by twice the long-run variance. We propose a bias-corrected estimator.     

分析师的盈利预测偏差与本地优势

分析师在金融市场中起着重要的信息中介作用,其信息来源及获取方式会直接影响信息处理的质量。本文从分析师对企业年度盈利预测的偏差出发,基于对分析师与公司地域关系的实证研究,发现分析师在预测那些所在地与分析师隶属的机构同属一个省份的公司时,预测结果更加准确,这种本地优势与公司高成长性及公司国有性质显著相关。研究还发现,我国分析师的本地优势与个人本身的乐观性并无关系,可能更多与信息优势相关。     

收入不平等、政策偏向与最优财政再分配政策

本文构建的理论模型探讨了收入不平等、政策偏向与最优财政再分配之间的关系。模型显示:政策偏向是收入不平等与财政再分配恶性循环的决定因素。当政策偏向于穷人时,最优财政再分配系统的选择是提高税率,扩大对穷人的转移支付;当政策偏向于富人时,最优财政再分配系统的选择是降低税率,减少对穷人的转移支付。依据中国数据的实证研究发现:(1)中国收入不平等不断恶化,针对恶化的不平等而采取的财政再分配政策效果差,没能扭转不平等恶化趋势,政策偏向严重。(2)税收再分配政策偏向富人,使富人税负相对轻于穷人,居民收入不平等加剧;偏向于穷人的转移支付多为消费券(物),导致越扶越贫。为此,需要从调整所得税和转移支付政策等方面采取相关对策。     

微观企业全要素生产率测度方法的比较与应用

微观企业TFP通常是不可观测的,它的估计面临内生性、联立性等诸多问题。在系统分析微观生产函数估计所面临主要问题的基础上,经过蒙特卡洛模拟法,比较了微观生产函数最新估计方法的优劣,发现广泛应用的OP、LP和ACF参数估计方法,倾向于高估微观企业的TFP,而De Loecker、GNR和联合估计的TFP测度方法能够得到TFP的稳健估计,增加值类生产函数估计倾向于高估微观企业的TFP,并且高估TFP的离散程度。将前沿TFP测度方法用于中国生产要素错误配置问题的研究,发现要素再配置具有显著的TFP促进效应,但是低于Hsieh和Klenow(2009)政策模拟的估计结果。据此,提出以微观企业TFP开展实证研究的对策建议。     

计量模型中的加总偏误与内生性：一种数值模拟方法

理论文献已经证明了细分数据模型的内生性和加总过程产生的内生性是加总偏误的根本原因。但是由于内生性涉及误差项与回归量之间的相关性问题,试图通过实证方法审视这两类内生性对加总偏误的影响变得比较困难,而数值模拟却是一个较为理想的方法。在数值模拟中,我们通过控制随机变量的分布形式以及随机变量之间的相关程度,进而对两类内生性因素产生的加总偏误进行全面细致的考察。本文的研究将为加总偏误的内生性解释提供有力证据。     

THE BIAS OF THE MULTIPLIER MATRIX WHEN SUPPLY AND USE TABLES ARE STOCHASTIC

The literature on stochastic input–output (I–O) analysis has paid considerable attention to the bias in the Leontief inverse. This paper extends previous studies by assuming supply and use tables (SUTs rather than I–O tables or input coefficients matrices) to be stochastic. This is a natural starting point because SUTs have become the basic data sources for I–O applications. In a Monte Carlo simulation experiment, a given SUT is randomized in two different ways and the effects are determined for eight different multiplier matrices. The analysis is carried out for Spain, Italy, the Netherlands, Germany and Finland, using their SUTs for 2006. The findings indicate that, in general, biases are statistically significant but negligibly small. This corroborates earlier findings obtained for stochastic I–O tables.