Research in Progress

“Virtual Worlds and Real Economics”
Abstract: Virtual worlds or massively multiplayer online games provide an opportunity for a new methodology of economic research. These games attract tens of millions of regular users and generate billions of dollars in revenues.  In addition, they have features that mimic the real world in important ways, such as government control of the money supply and taxation, floating exchange rates, and rudimentary legal systems.  This paper sets the groundwork for a series of papers seeking to answer real-world economic questions through research in virtual worlds.  In order to establish the external validity of virtual world research, I perform a series of standard economic experiments in the virtual world, Second Life.  I run the Ultimatum Game, the Trust Game, and the Monty Hall problem on avatars recruited in the game.  I find that individual behavior in these games is indistinguishable from that found in real-world experiments.  Additionally, demographic data collected from subjects indicates that this is one of the most diverse subject pools in an economic experiment according to several measures. 

“Experimental Macroeconomics in a Virtual World”
Abstract: Macroeconomic theories are difficult to test explicitly due to the difficulty in observing or constructing counterfactuals.  Detailed data on a macroeconomy and natural or controlled policy experiments are required.  A unique aspect of virtual worlds is that they possess both qualities: complete observability of all transactions as well as the potential to manipulate economic rules such as monetary policy and tax policy.  Using a unique transaction-level data set from the virtual world Second Life, I am able to perform and evaluate a series of macroeconomic experiments that would be impossible in the real world.  Experimental manipulations include altering both monetary policy and tax policy at the economy-wide level.  This allows for a variety of experimental tests of macroeconomic predictions, including those pertaining to  money demand, Taylor rules for interest rate policy, and taxation.

“A Virtual Labor Market Experiment” (with Alain Cohn and Ernst Fehr)
Abstract: We construct a factory and employ workers in a virtual world in order to test several theories of labor and behavioral economics, including gift-exchange and efficiency wage theories.  The virtual world environment allows for an ongoing labor relationship with full-observability at a cost substantially below that in the real world.

“A Market for Justice” (with Daniel Chen)
Abstract: We empirically examine the impact of a more liberalized litigation funding scheme than currently exists in the United States.  Two recent innovations in Australia are considered: the ability to trade legal claims and the ability for law firms to go public.  We analyze data from Australian courts to determine the impact of litigation trading on settlement rates and amounts, time to settlement, volume of litigation, fees, establishment of precedent, and development of law.  Using this data we may test several potential theories of the impact of litigation trading.  We can then use these empirical findings to inform a theory of the impact of allowing publicly traded law firms.

“When Docs Snooze, Do You Lose? Medical Resident Work Hours and Patient Outcomes”
Abstract: In 2003, a national 80-hour weekly work limit was imposed on medical residents for the first time.  The frequently stated motives were to improve patient outcomes and resident effectiveness, but little is known about how to optimally balance the tradeoff between resident fatigue and patient handoffs.  This study makes use of a natural experiment in which New York State imposed a similar work hour limit in 1989.  Using a difference-in-difference and triple differences methodology with CMS data, I estimate that the New York rule change decreased 180-day mortality rates by 6-11%.  I further analyze preliminary data from the 2003 rule change to investigate whether the results generalize to a national sample.