Direkt zum InhaltDirekt zur SucheDirekt zur Navigation
▼ Zielgruppen ▼

Humboldt-Universität zu Berlin -

Game Theory and Experimental Economics

GOR B-Conference

Humboldt-Universität Berlin, June 14-15, 2018

The aim of this conference is to strengthen rigorous and state-of-the-art research at the interface of (algorithmic) game theory and experimental economics with a focus on economic and managerial problems. The long term goal is to establish a specialized working group dedicated to these question within the German Society for Operations Research (GOR).

We invite contributions from researchers with a research focus on game-theoretic approaches in management, economics, computer science, mathematics, or related fields. The relevant methodology incorporates cooperative and non-cooperative game theory as well as experiments and field studies.

The main goal of this workshop is to connect researchers and research. We therefore encourage participants to present state-of-the-art review on active research for a diverse audience.

 

Registration

For participation, registration via mail to Max Klimm is required. Please mention if you want to give a talk. There is no participation fee. Lunch and Coffee at the workshop is provided but the dinner on June 14 is not covered.

 

Keynotes

Dorothea Kübler

Dorothea Kübler
Foto: David Ausserhofer
  • Department of Market Behavior (Director)
    Berlin Social Science Center (WZB)

Felix Fischer

Felix Fischer
Foto: Felix Fischer
  • School of Mathematical Sciences
    Queen Mary University London (QMUL)


 

Tentative Schedule

June 14, 2018

14:00

Arrival

14:30-15:00

Opening
Max Klimm (HU Berlin) und Guido Voigt (U Hamburg)

15:00-16:00

Coffee Break and Discussion

16:00-17:00

Plenary Talk 1
Felix Fischer (Queen Mary U)

Non-Truthful Position Auctions Are More Robust to Misspecification

17:15-17:40 Short Talk
Michael Becker-Peth (Rotterdam School of Management, Erasmus University)
Behavioral Inventory Management - How to develop Behavioral Newsvendor Models

19:30

Dinner
Restaurant Hackescher Hof

 

June 15, 2018

09:00-10:00

Plenary Talk 2
Dorothea Kübler (WZB Berlin)

How Lotteries in School Choice Help Leveling the Playing Field

10:00-11:00

Coffee Break and Diskussion

11:00-11:25 Short Talk
Louise Molitor (HPI Potsdam)

Schelling Segregation with Strategic Agents
11:30-11:55 Short Talk
Sebastian Schiffels (TU München)
Project management under information asymmetry: the impact of familiarity on Parkinson’s Law

12:00-14:00

Lunch Break and Discussion

14:00-14:25

Short Talk
Francis de Véricourt (ESMT Berlin)
Warning Against Recurring Risks: An Information Design Approach

14:30-14:55

Short Talk
Alexander Skopalik (U Paderborn)

On approximate pure Nash equilibria - Existence and Complexity

15:00-15:25

Short Talk
Laura Vargas Koch (RWTH Aachen)

Nash Flows with Spillback

15:30-15:55

Short Talk
Anja Schöttner (HU Berlin)

Delegating Pricing Authority to Sales Agents: The Impact of Kickbacks

16:00-17:00

Coffee Break and Discussion

17:00-17:30

Closing

 

 

Abstracts

Michael Becker-Peth: Behavioral Inventory Management - How to develop Behavioral Newsvendor Models

We summarize the literature on human decision making in the newsvendor model. In the newsvendor model, a decision maker faces stochastic demand and must determine the order quantity. Consistent findings in the literature are that people choose order quantities that are between expected-profit-maximizing quantities and mean demand and that they vary over time. We discuss how deviations of people’s orders from the expected-profit-maximizing quantities can be explained by decision biases and alternative utility functions. We also discuss how heterogeneity among people can explain between-subject variety, present alternative behavioral models, and identify potential areas for future research.


Francis de Véricourt: Warning Against Recurring Risks: An Information Design Approach

The World Health Organization seeks effective ways to alert its member states about global pandemics. Motivated by this challenge, we study a public agency’s problem of designing warning policies to mitigate potential disasters that occur with advance notice. The agency privately receives early information about recurring harmful events and issues warnings to induce an uninformed party to take costly preemptive actions. The agency’s decision about whether to issue a warning critically depends on its credibility, which we define as the uniformed party’s belief regarding the accuracy of the agency’s information. This belief is updated over time by comparing the agency’s warnings with the actual incidence of harmful events. The sender, therefore, faces a trade-off between eliciting a proper response today and maintaining her credibility in order to elicit responses to future adverse events. We formulate this problem as a dynamic Bayesian persuasion game, which we solve in closed form. We find that the agency must be sufficiently credible to elicit a mitigating action from the uninformed party for a given period. More importantly, the agency sometimes strategically misrepresents its advance information about a current threat in order to cultivate its future credibility. When its credibility is low (i.e., below a threshold), the agency downplays the risk and actually downplays more as its credibility improves. By contrast, when its credibility is high (i.e., above a second higher threshold), the agency sometimes exaggerates the threat. In this case, a less credible agency exaggerates more. Only when the agency’s credibility is moderate does it consistently send warning messages that fully disclose its private information about a potential disaster. These findings provide prescriptive guidelines for designing warning policies and suggest a plausible rationale for some of the false alarms or omissions observed in practice.


Felix Fischer: Non-Truthful Position Auctions Are More Robust to Misspecification
with Paul Dütting (LSE) and David C. Parkes (Harvard)

The area of mechanism design is concerned with the development of protocols that produce good outcomes given inputs from self-interested individuals. One of the stars of mechanism design theory is the Vickrey-Clarke-Groves (VCG) mechanism, which makes it optimal for participants to truthfully reveal their inputs. In the real world, however, the VCG mechanism is used with surprising rarity and the mechanisms we actually find are non-truthful. An example of this phenomenon are position auctions used by internet search engines like Google and Bing to allocate space for advertisements displayed next to genuine search results. Here only non-truthful mechanisms have ever been used, and what in theory looks like a mistake was a huge success in practice. An obvious advantage of the non-truthful mechanisms is that they use simpler payments, and it has been known for some time why this simplicity does not lead to chaos when participants decide strategically how to bid. We will see that it also comes with increased robustness to misspecification of the auction's bidding language.


Dorothea Kübler: How Lotteries in School Choice Help Leveling the Playing Field
with Christian Basteck (ULB) und Bettina Klaus (UNIL)

In a number of school choice systems, such as in the UK, New York City, and Berlin, lotteries for school seats have been advocated to desegregate schools and create diverse student bodies. We study lottery quotas in the two most common school choice mechanisms, namely the deferred and the immediate acceptance mechanism. Some seats are allocated based on merit (e.g., grades) and some based on lottery draws. We focus on the effect of the lottery quota on truth-telling, the utility of students, and the student composition at schools, using theory and experiments. We find that the lottery quota strengthens strategyproofness in the deferred-acceptance mechanism in the sense of reducing the set of equilibrium strategies that are not truth-telling. However, the lottery quota has no clear-cut effect on truth-telling in the immediate acceptance mechanism. These results are supported by the experiment. Moreover, the lottery quota leads to a more diverse school population in the experimental markets, as predicted.


Louise Molitor: Schelling Segregation with Strategic Agents
with Ankit Chauhan (HPI Potsdam) and Pascal Lenzner (HPI Potsdam)

Schelling's segregation model is a landmark model in sociology. It shows the counter-intuitive phenomenon that residential segregation between individuals of dierent groups can emerge even when all involved individuals are tolerant. Although the model is widely studied, no pure game-theoretic version where rational agents strategically choose their location exists. We close this gap by introducing and analyzing generalized game-theoretic models of Schelling segregation, where the agents can also have individual location preferences.
For our models we investigate the convergence behavior and the eciency of their equilibria. In particular, we prove guaranteed convergence to an equilibrium in the version which is closest to Schelling's original model. Moreover, we provide tight bounds on the Price of Anarchy.


Sebastian Schiffels: Project management under information asymmetry: the impact of familiarity on Parkinson’s Law

Projects involving different decision makers for planning and execution are common, however, their interaction bears the risk of poor performance. Information asymmetry and the well-established phenomenon that work expands so as to fill the time available for its completion (Parkinson’s Law) makes the project planning even more challenging. We study the impact of familiarity, which is known to improve project performance in numerous situations, on Parkinson’s Law. To address our research question, we consider a model that reflects the interaction of a planner symbolizing the management of a company and of a worker symbolizing the project executors of the company. Based on this framework, we set up an experimental study providing support that, first not all time planned too long is exploited by workers, second this effect is more pronounced in a setting with familiarity than without, third managers decision in the familiarity treatment indicate higher trust levels towards the workers. Interestingly and in contrast to the situation without familiarity managers seem to greatly underestimate exploitation, thus familiarity decreases performance.


Anja Schöttner: Delegating Pricing Authority to Sales Agents: The Impact of Kickbacks

We investigate a situation where a firm can delegate pricing authority to a sales agent who is better informed about customers' valuations for the product than the firm. When the agent has pricing authority, customers may offer kickbacks to the agent to obtain a discount. The firm can prevent such collusion by designing the agent's performance pay and his degree of pricing authority appropriately, but may prefer not to do so if impeding collusion is too costly. We derive optimal collusion-proof contracts, describe conditions under which collusion arises in equilibrium, and study the optimal interaction between delegation, incentive pay, and the firm's installed auditing technology.


Alexander Skopalik: On approximate pure Nash equilibria - Existence and Compexity

Pure Nash equilbria are a natural and convincing solution concept that does  not require agents to randomize. Unfortunate, it does a Nash equilibrium in pure strategies does not always exist in finite games unlike the well known Nash equilibrium in mixed strategies.
In addition, both solution concepts have been criticized from a computer science perspective due to their computational intractability.
We introduce and discuss the notion of approximation in the context of pure Nash equilibria to alleviate the issues of non existence of pure equilibria and the computational intractability of pure equilibria.


Laura Vargas Koch: Nash Flows with Spillback
with Leon Sering (TU Berlin)

Modeling traffic in road networks is a widely studied but challenging problem especially under the assumption of selfish driver. There are various approaches with different advantages and disadvantages. One of them is the deterministic queuing model. The basic idea is to model traffic by a continuous flow that travels over the time through a network, in which the arcs are endowed with transit times and capacities. If more flow uses an arc than the capacity allows, a queue builds up and the infinitesimal small flow particles need to wait in line in front of the bottleneck. It has been shown that a dynamic equilibrium, also called Nash flow over time, can then be described by a sequence of static s-t-flows with specific properties, so-called thin flows with resetting. Until now, it was not possible to handle upper limits on the length of the queue. We extended the model by allowing a storage capacity for each arc that bound the total amount of flow on it. Thus, an arc can become full which causes queues on previous arcs and blockage of intersections, in other words we have spillback.
This is a phenomenon we regularly observe in real traffic situation and
has to be considered in dynamic traffic assignment. For this extended model, we prove the existence of Nash flows over time, which again can be described by a sequence of static s-t-flows, which we call spillback thin flows.

 

 

 

Venue

School of Business and Economics
Foto: Constanze Haase
  • School of Business and Economics
Heilig-Geist-Kapelle
Foto: Stefan Müller
  • Heilig-Geist-Kapelle

 

The workshop takes place at the Heilige-Geist-Kapelle located within the main building of the School of Business and Economics of HU Berlin. The address is

Spandauer Straße 1
10178 Berlin
Germany

Closest public transport stations are

  • S-Bahnhof Hackescher Markt
  • S+U Alexanderplatz

From Berlin Hauptbahnhof go to the upper level and take an S-Bahn (green S sign) eastbound to Hackescher Markt.

From Berlin Tegel Airport take the TXL bus to Spandauer Str./Marienkirche. (If you are short on time, change to the S-Bahn at Hauptbahnhof and proceed as above.)

From Berlin Schönefeld Airport take the S9 to Hackescher Markt.

For public transport directions in Berlin visit BVG.

 

 

Sponsors

German Societiety for Operations Research (GOR)

Einstein-Center for Mathematics (ECMath)

 

Organization

Max Klimm (HU Berlin)

Guido Voigt (U Hamburg)