The last few decades have seen the development of mechanism design theory as a powerful toolbox for designing desirable economic or social institutions for the allocation of resources. Its applications have successfully led to major breakthroughs in many areas including auctions and selling mechanisms, pricing and revenue management, regulations and taxes, public good provision, voting procedures and political system, and so on. Due to theoretical obstacles, however, the current mechanism design theory largely assumes that the participating agents are fixed.
With his Venture Fund project, titled ‘Mechanism design with varying population’, Joosung Lee tackled this problem by looking at the conflict of interest that occurs when people interact with others, and studied how the conflict can be aligned in different situations.
The project produced two research papers:
- 'Sequential Search Auctions with a Deadline' co-authored with Daniel Li (Durham University) and submitted to American Economic Journal: Microeconomics.
- This paper focuses on a dynamic auction in which a seller invites numbers of bidders over time. It studies the strategic interactions not just between the seller and the bidders, but also between the already-invited buyers and the potential buyers. The paper shows that the seller invites more and more buyers and also decreases the reservation prices as time passes. But by balancing the two different effects - decreasing the price and increasing competition - the seller can maximise their revenue.
- 'Referral Incentive Design' co-authored with Eugene Jeong (University of Bristol), in preparation for submission
- This paper concerns an abstract situation, where people’s social network is their own private information and discusses how to incentivise people to bring others into it although it would increase competition or congestion. It proposes a multilevel mechanism, which finds a desirable outcome for the whole society and allocates the cost and benefits in a fair way. The applications include public goods provision and referral marketing.
The auction model was further developed for a general dynamic decision-making problem in which the decision-maker chooses their effort level over time. The main techniques and results for such an abstract model were drafted in a separate paper, 'Dynamics of Optimal Search Intensity'. This new approach provides a rigour statistical foundation for the choice of search intensity, while the existing models focus on binary search rule - that is, whether to keep searching or stop. It was presented at multiple conferences and seminars and the search-intensity model has been further developed for a new ongoing project, 'An Endogenous Time Preference Model'.