Current IFREE-Funded Research Projects

IFREE Research Grants

Combinatorial Auction
Endogenous Group Formation via Unproductive Costs
FDA Dual Track program
Methods Grant
Mysteries of Morality
Software to Support Studies in Financial Economics
The Alliance Hypothesis for Human Friendship

Combinatorial Auction
A combinatorial auction is simply a process that can be implemented when multiple resources must be simultaneously allocated amongst competing users, and information concerning the values of the various possible uses, and the constraints impinging upon those uses is unknown to the central decision maker(s). For example, the demand for takeoff and landing slots at all major airports are highly related to the uses that competing airlines have in the daily routing of their stock of airplanes. Now imagine that congested airspace becomes a simultaneous reality at many major airports: various reports indicate this day will soon be upon us. How will the airport authorities decide which airlines can use their slots most efficiently? There is no effective alternative but a combinatorial auction, in which competing airlines can express their values for sending different planes on different routes, and then a computer figures out how to maximize the total value of all schedules awarded given the millions of possible alternatives.

In the past year David Porter and Stephen Rassenti were asked to write a review article for the Wiley Encyclopedia of Management Science and Operations Research. They researched the extant literature and summarized all that they could find concerning combinatorial auctions. There aren’t very many real world implementations, and as sophisticated an approach as it may be, running a combinatorial auction remains for the most part more art than science. Hundreds of alternative implementation rules are possible, and the cost of ferreting out the best of them is more than the real world cares to bear. With the help of a grant from IFREE and graduate students working at Chapman University they have delved into the world of agent based modeling to get some answers concerning combo auctions.

Because of their complexity, combo auctions are normally run in a sequence of multiple rounds that allow bidders to update their multi-item bids as the individual prices for different items in the auction rise non-uniformly. But on which items should the auctioneer raise the price and by how much? Not an easy question to answer and not the same answer for all situations, but the first IFREE sponsored study indicates that it is a bad idea to raise the prices on many related items simultaneously: typically, 30% is good. Furthermore, it seems the price increments should be fixed and not be adjusted in proportion to announced prices.

But why should you run an ascending price auction at all? The Dutch sell millions of dozens of flowers annually using descending price clock auctions, with many auctions lasting only a few seconds. It is well known in the field and in the laboratory that it is much harder for buyers to collude in their purchases when the second that someone says “I accept that price” the auction is over. However, to generalize the concept of a descending auction to multiple items with related usage values requires some nontrivial innovation. Supported by IFREE we have recently had a breakthrough in the second phase of our combo auction research. Using descending prices we can now generate more efficient allocations requiring shorter auction times in our agent based simulations. In the next phase of our IFREE research, we will program an interactive auction interface and bring this combo auction innovation into the laboratory to be tested by cash motivated human subjects.

Frequently, when high stakes are involved, stakeholders are reluctant to exchange multilateral back room negotiations (a comfortable and manageable environment where they understand the rules of the game) for an impersonal computer driven mechanism that is capable of exploring many more possibilities in the usage frontier than human conversation can possibly address. Research sponsored by IFREE to tune the implementation rules of combo auctions and produce laboratory evidence of reliable auction performance can lead to more comfort for stakeholders in adopting combo auctions.

Endogenous Group Formation via Unproductive Costs
Aimone, Jason A., Laurence R. Iannaccone, Michael D. Makowsky, and Jared Rubin

We demonstrate that unproductive costs facilitate group formation and mitigate free riding. We conduct an experiment that allows groups to form endogenously by having subjects reveal their willingness to “sacrifice” a fraction of their private productivity within a voluntary contribution mechanism public goods game. The sacrifice mechanism, previously identified in the theory of religious clubs, functions in the lab absent any group identity or shared doctrine. We find that groups which emerge from subject preferences for higher rates of sacrifice screen out free riders, attract conditional cooperators, increase contributions to the public good, and offer potential welfare gains for members.
Read the working paper

FDA Dual Track program
Stephen Rassenti and Dan Houser teamed with Bart Madden to propose experiments on the novel "Dual Tracking" approach Madden advocates for reforming the FDA drug approval process. While the FDA currently has absolute control of access to drugs, Madden's idea is to provide an access-to-drugs path around the FDA where patients, advised by their doctors, would take responsibility for deciding if late-stage experimental drugs offer a better opportunity for them compared to FDA-approved drugs.

This new path is Dual Tracking. The aim of the team's research proposal is to demonstrate why and how dual tracking makes things better. The research team has developed sophisticated software that places participants within a complex dynamic system that allows experimenters to discover relationships between information flows and health-related decisions.

Methods Grant
Daniel Houser's Report

One of the issues that interests me is knowing how to know – scientifically – that a particular type of market institution is more likely than others to achieve our goals, say for pollution abatement. We know how to obtain scientific data from the laboratory, but how can we learn from that data what works best? With luck it is easy when one of the markets far outperforms the others. Sometimes, though, the differences can be subtle. When they are, experimental economists have a set of standard tools they use to determine which market performs the best. As part of the IFREE methods grant, my student Li Hao and I (both at George Mason University) have recently developed new formal approaches to determining which market works best, not only between two but also among many, and have further demonstrated that this new approach is better – by quite some distance – than the standard approaches in use today. In light of the importance of knowing which market-oriented institution performs best in relation to the goal, and in view of the substantial improvements in accuracy and reliability offered by our new procedure, we are highly optimistic that our approach will be rapidly embraced by not only experimental economists but also the broader scientific community. We are pleased that the leading journal Experimental Economics has requested a revision of our first paper on this topic, “Adaptive Procedures for Nonparametric Tests: Seven Decades of Advances”. In addition, we are finishing a second methods paper, “Getting it Right the First Time: Belief Elicitation with Novice Participants”, which promises to provide better information about the beliefs that underlie people’s actions in key market environments. Finally, Li Hao and I are currently working to develop powerful adaptive procedures for matched samples, an important extension of our first paper into additional environments of substantial interest for empirical scientists.

Another topic of substantial methodological interest is to understand how humans’ use of language has impacted economic discovery, particularly as it relates to specialization and exchange. In a recent paper published in the American Economic Review, Erik Kimbrough, Bart Wilson and Vernon Smith (all scholars at Chapman University) provide evidence that the use of language – even the appearance of particular words – can have dramatic implications for the economic discovery process. This fascinated me, and in my own work with former-student Erte Xiao (now a professor at Carnegie Mellon University) we developed a new approach to classifying natural language. Our idea is related to so-called “human-computing”, used for example by Google to help label the internet with their “image labeling” game. The idea behind the game is simple: two people are anonymously matched and cannot communicate, but are shown the same picture. They are then told to start entering possible labels for the picture. If the two people come up with the same label, it is used as a “keyword” for that picture. Our approach for the classification of natural language messages is similar in spirit, and we find that this approach to learning about natural language messages generates systematically more reliable information than other leading alternatives. It is an exciting approach and we continue to develop its extensions. We are optimistic that our approach is valuable and its use will become widespread. Indeed, we are excited that our first paper on this topic “Classification of Natural Language Messages Using a Coordination Game”, was selected as the lead article for Volume 14 (2011) of the prestigious journal Experimental Economics. Further, I am happy to report that my student Jason Aimone recently received a grant from the National Science Foundation to pursue methodological studies that extend Erte and my initial IFREE-funded research on this topic.

Mysteries of Morality
Peter DeScioli, Robert Kurzban

Evolutionary theories of morality, beginning with Darwin, have focused on explanations for altruism. More generally, these accounts have concentrated on conscience (self-regulatory mechanisms) to the neglect of condemnation (mechanisms for punishing others). As a result, few theoretical tools are available for understanding the rapidly accumulating data surrounding third-party judgment and punishment. Here we consider the strategic interactions among actors, victims, and third-parties to help illuminate condemnation. We argue that basic differences between the adaptive problems faced by actors and third-parties indicate that actor conscience and third-party condemnation are likely performed by different cognitive mechanisms. Further, we argue that current theories of conscience do not easily explain its experimentally demonstrated insensitivity to consequences. However, these results might be explicable if conscience functions, in part, as a defense system for avoiding third-party punishment. If conscience serves defensive functions, then its computational structure should be closely tailored to the details of condemnation mechanisms. This possibility underscores the need for a better understanding of condemnation, which is important not only in itself but also for explaining the nature of conscience. We outline three evolutionary mysteries of condemnation that require further attention: third-party judgment, moralistic punishment, and moral impartiality.

DeScioli, P., & Kurzban, R. (2009). Mysteries of morality. Cognition, 112, 281-299.

In the second paper, we find that people use omissions strategically to avoid punishment.
DeScioli, P., Christner, J., & Kurzban, R. (2011). "The omission strategy". Psychological Science, 22, 442-446.

In the last paper, we find support for the theory that people judge others' omissions less harshly than others' commissions because commissions provide stronger evidence of wrongdoing. Evidence is important because it allows people to coordinate their condemnation decisions with others.
DeScioli, P., Bruening, R., & Kurzban. R. (2011). "The omission effect in moral cognition: Toward a functional explanation". Evolution and Human Behavior, 32, 204-215.

Software to Support Studies in Financial Economics
David Porter

Much of the software that we have been using for the past 20 years to examine price behavior in financial markets is now outdated. The standard software for call markets in bubble environments no longer operates with the XP operating system used in our labs. In addition, we do not have a standard double auction for the asset trading (“bubbles”) environment that allows for forecasting and information conditions (sunspot messages). Lastly, extending the market software to micromarket structure conditions with noise traders, liquidity traders and earnings surprises is crucial for future experiments.

We are now embarking on a series of experimental inquiries that requires resuscitating the old software design and also requires a serious update of its capabilities. In particular, we are planning to investigate:

• The effect of naked short sales on price volatility, and bubble phenomena.
• Using call markets to clear trades, with continuous double auction trading in futures on the call price.
• Passive vs active investing style management of funds

• Affecting Policy by Manipulating Prediction Markets: Experimental Evidence
• Are Under- and Over-reaction the Same Matter? A Price Inertia based Account
• Can Manipulators Mislead Market Observers?
• Durability, Re-trading and Market Performance
• Gradual Information Diffusion and Asset Price Momentum
• Reaction to Public Information in Markets: How Much Does Ambiguity Matter?
• The Effect of Reliability, Content and Timing of Public Announcements on Asset Trading Behavior

The Alliance Hypothesis for Human Friendship
Peter DeScioli, Robert Kurzban

Exploration of the cognitive systems underlying human friendship will be advanced by identifying the evolved functions these systems perform. Here we propose that human friendship is caused, in part, by cognitive mechanisms designed to assemble support groups for potential conflicts. We use game theory to identify computations about friends that can increase performance in multi-agent conflicts. This analysis suggests that people would benefit from: 1) ranking friends, 2) hiding friend-ranking, and 3) ranking friends according to their own position in partners’ rankings. These possible tactics motivate the hypotheses that people possess egocentric and allocentric representations of the social world, that people are motivated to conceal this information, and that egocentric friend-ranking is determined by allocentric representations of partners’ friend-rankings (more than others’ traits).

DeScioli, P., Kurzban, R., Koch, E. N., & Liben-Nowell, D. (2011). "Best friends: Alliances, friend ranking, and the MySpace Social Network" in PERSPECTIVES ON PSYCHOLOGICAL SCIENCE, 6, 6-8.

DeScioli P, Kurzban R (2009) "The Alliance Hypothesis for Human Friendship". PLoS ONE 4(6): e5802. doi:10.1371/journal.pone.0005802