Stern Center for Research Computing

New York University • Leonard Stern School of Business

Role of endogenous consumption in the counter-cyclical pricing

Minjung Kwon is a fifth year Doctoral Candidate in Marketing focusing on quantitative marketing and empirical industrial organization. Her research interests include forward-looking consumer choices, dynamic structural models, and marketing for products with seasonal demands. Her current research investigates the role of price promotion on consumers’ purchasing and consumption decisions for seasonal products, aiming to explain the firm’s marketing mix strategies. A prevalent pricing strategy in seasonal marketing is to increase price promotions despite high volume of demand during the peak season. This observation goes against the standard price movement when facing high volume of demand. Her research links the firm’s motive for counter-cyclical pricing with the primary demand effects of promotions. By employing the dynamic structural model, her work presents consumer’s rational and strategic behavior posited to optimize when, what, and how much to buy and consume. In order to empirically quantify the effects, she uses extensive consumer panel data sets, which record individual households’ purchasing details at the UPC-level, and which allows her to infer the trajectory of households’ inventory.

Minjung initially approached the Stern Center for Research Computing (the Center, hereafter) mainly to overcome her local machine’s overload from computational burdens. Now, she has become a big user of the Center’s system, including Guistatistics and the Stern Grid, and she has greatly benefitted because of the following reasons: first, her data is big in volume, requiring remote storage as well as remote processing. To this end, the Center responded to all her requests with programing languages such as MATLAB, SAS, and R. She has also used the Guistatistics computer to run jobs interactively through its GUI during the program-writing stage. Second, due to the complex structure of her model, she often suffers from overwhelming computation time. The Stern Grid system not only releases her local machine, but also reduces the computation time by enabling concurrent computing.