After earning his Ph.D. in economics in 2012, Matt Backus took a somewhat unusual postdoctoral position. Instead of heading off to the ivy-covered walls of a university, he joined other academics at eBay Research Labs.

Backus, a new assistant professor of economics in the College of Arts and Sciences, has a research interest in the economics of and bargaining. He said the eBay experience allowed him access to massive amounts of data and an environment to test his theories in real-world applications. While there, he worked with top economists in the area of online commerce like Steve Tadelis of the University of California, Berkeley.

From that postdoc year, Backus gathered many research ideas, some of which are now working papers.

In one project, Backus studied whether first-time eBay bidders were less likely to return to the site when someone "sniped" their item during an auction. An auction sniper places a higher bid than the current highest public bid at the last possible moment, giving other bidders no time to outbid. Although the practice is controversial, it's not against eBay policy. Backus' work found that first-time buyers were indeed bothered by snipers, enough so that they may not return to the site.

For another paper, Backus collaborated with Henry Schneider from the Samuel Curtis Johnson Graduate School to study why the same item auctioned on the same platform will sometimes sell for a very different price. Economists call this phenomenon "price dispersion."

They hypothesized that some buyers are willing to spend more time searching for all available options – in this case multiple auctions of the item they want to buy. These buyers took the time to consider shipping costs and other terms of the sellers and entered more than one auction, thereby securing their item for a lower price.

In other cases, buyers who aren't willing to spend this extra time in research joined the first auction they found and paid a higher price.

Their research studied the impact of word counts on eBay listings, which affects search visibility, and found that sellers of DVDs who used more words in their title were more likely to sell their items, while the word "new" in a listing resulted in an ending price increase of 80 cents.

Backus also has papers in the works studying the behavior of people who bid frequently in dynamic online auctions, best offer bargaining on eBay, and the causal relationship between competition and productivity, which focuses on the ready-mix concrete industry.

"Bargaining is a fascinating area for empirical work because to date there has been so little data available," Backus said. "Often it is highly confidential, but even when it is not, the things we bargain over are hard to compare (think union contracts, company mergers, etc.). Understanding how we bargain in the real world is a foundational and open question in economics."

More information: To learn more about Backus and his work, see sites.google.com/a/umich.edu/mrb/

Provided by Cornell University