Problem Description
Two neighboring landowners each possess a parcel of land. The combined value of these parcels is significantly greater than the sum of their individual values. For instance, a hotel could be built on the combined land, whereas only single-family homes are feasible on the individual parcels.
The landowners must decide how to maximize their profits: sell their land separately or form a coalition and sell the combined land as a single parcel. If they choose the latter, they face the challenge of fairly and efficiently allocating the revenue, especially given that the bundled value exceeds the sum of the individual values. This allocation problem is further complicated by the unique characteristics of each parcel. For example, while one parcel might be larger, it could be constrained by existing development. The smaller parcel, on the other hand, might offer desirable features like a park view.
How can the landowners balance these competing factors, size and amenities, when dividing the proceeds from a joint sale?
The same problem arises in other contexts, such as spectrum auctions, shared sequencing, MEV taxes, refund rules and combinatorial orderflow auctions. For example, suppose sequencing Unichain and Arbitrum blocks together generate more revenue than they would separately, thanks to the efficient extraction of cross-domain MEV. How should that additional revenue be divided between the two domains in a way that is incentive-compatible for both roll-up operators? A natural solution would be to use the revenue that both operators would generate on their own; however, this approach is weak against shill bids. And so, can we design shill-proof and efficient mechanisms in this context?
