Date: 9 March, 2018 Friday
Place: Faculty of Business Administration,
Umit Berkman Seminar Room
“An Experimental Study of Newsvendor Decisions under Incomplete Information”
Ayşe Kocabıyıkoğlu, Ph.D. in Management, Sabancı University
We present an experimental study of newsvendor decisions under incomplete information. We observe that orders deviate from normative benchmarks when decision makers have incomplete information, and this tendency is stronger when information about the demand distribution is not available. This pattern is reflected in profits: subjects capture at most 81% of earnings they could have incurred if they ordered the normative quantity in high-margin settings; the corresponding percentage is 50% in low-margin settings. Our results also suggest decision variability is not affected by the type or amount of information available. Comparison with benchmarks with full and no information reveal, when the underlying demand distribution is not known, the availability of price and cost information significantly improves decisions, whereas when demand information is available, not knowing price or cost does not hurt decisions. We also consider the impact of information on profits; in our study the improvement in profits with any type of information (price, cost, demand) depends on the information set available to the decision maker, and more information does not necessarily lead to better performance.