Do Random Errors Explain Newsvendor Behavior?

Abstract

Previous experimental work showed that newsvendors tend to order closer to mean demand than prescribed by the normative critical fractile solution. A recently proposed explanation for this mean ordering behavior assumes that the decision maker commits random choice errors, and predicts the mean ordering pattern because there is more room to err toward mean demand than away from it. Do newsvendors exhibit mean ordering simply because they make random errors? We subject this hypothesis to an empirical test that rests on the fact that the random error explanation is insensitive to context. Our results strongly support the existence of context-sensitive decision strategies that rely directly on (biased) order-to-demand mappings, such as mean demand anchoring, demand chasing, and inventory error minimization.newsvendor model, task context, heuristics, random choice

A Review of Behavioral Decision Making in the Newsvendor Problem

“…This phenomenon is called "pull-to-center" (PTC hereafter) effect. Later on, several other studies also confirmed the similar too low/too high pattern (else, underordering and over-ordering) of choices in newsvendor experiment (Benzion et al 2008, Bolton and Katok 2008, Bostian et al 2008, Lurie and Swaminathan 2009, Kremer et al 2010, Moritz et al 2013. These studies provide insights into identifying biases and decision heuristics of individual decision making in newsvendor setting.…”

“…For the symmetric cumulative demand function about mean, a newsvendor will exhibit "pull to center" effect Schweitzer and Cachon (2000), Bostian et al (2008), Kremer et al (2010Kremer et al ( , 2014, Bell (1982Bell ( , 1985, Sugden (1982,1987) anchoring and insufficient adjustment bias tendency of decision maker to use a reference point, or 'anchor' and then insufficiently adjusting away from this anchor two alternative anchors:  mean anchor heuristic, and,  previous demand chasing heuristic Schweitzer and Cachon (2000), Benzion et al (2008), Bolton and Katok (2008) limited rationality of decision makers due to their cognitive limitations, insufficient information, time and cost constraints a boundedly rational newsvendor chooses an order quantity which is more satisficing rather than optimal, i.e. he chooses high yielding order quantities with larger probability Simon (1955), Su (2008), Kremer et al (2010), Katok (2011), (Ovchinnikov et al 2015) overconfidenc e bias (or, overprecision) tendency of decision maker to overestimate their own opinions due to precise information that they believe they have An overconfident newsvendor will underestimate the variance of demand and make suboptimal ordering decisions.…”

“…he chooses high yielding order quantities with larger probability Simon (1955), Su (2008), Kremer et al (2010), Katok (2011), (Ovchinnikov et al 2015) overconfidenc e bias (or, overprecision) tendency of decision maker to overestimate their own opinions due to precise information that they believe they have An overconfident newsvendor will underestimate the variance of demand and make suboptimal ordering decisions. Over-order in low profit-margin Croson et al (2008), Ren and Croson (2013), Ren et al (2017) Cui et al (2013) 4.1 Individual Decision-Making Biases Schweitzer and Cachon (2000) found that the PTC effect observed in their experimental data results were not consistent with alternatives like risk aversion preferences, loss aversion, waste aversion bias, stock out aversion, or underestimating opportunity costs (Table 3), instead they found two explanations for this behavior: (1) individual's preferences to reduce ex-post inventory error; and (2) subjects suffer from anchoring and insufficient adjustment bias.…”

“…Without touching upon the risk-neutral behavior of the decision-maker, this model predicted that the similar ordering behavior in newsvendor problem, observed in Bolton and Katok (2008), that is, the average order quantity converges toward mean demand. However, Kremer et al (2010) tested the descriptive validity of the work of Su (2008) in a controlled lab experiment and suggested that random error explanation is insensitive to the context of decision settings and could not exclusively explain the mean ordering behavior. In their lab experiment, they found that the group of participants facing standard newsvendor problem (operation frame) were observed to be significantly more biased toward mean ordering and demand chasing than the group of participants facing standard risky choice framework (neutral frame).…”

Blinded by Experience: Prior Experience, Negative News and Belief Updating

“…Work examining one or two period settings or multi-period, static environments finds that individuals behave irrationally for many reasons, depending on the context, including anchoring (Schweitzer and Cachon 2000), anticipated regret (Davis, Katok and Kwasnica 2011), demand chasing (Kremer, Minner and Van Wassenhove 2010), loss aversion (Davis, Katok and Santamaría 2014), and incorrectly incorporating relevant information (Croson and Donohue 2006;Kremer, Moritz and Siemsen 2011). Once one moves to a multi-period, dynamic setting a new challenge arises.…”