A reconsideration of the Kar solution for minimum cost spanning tree problems

Minimum cost spanning tree (mcst) problems try to connect agents effi ciently to a source when agents are located at different points in space and the cost of using an edge is fixed. The application of the Shapley value to the stand-alone cost game, known in this context as the Kar solution, has been criticized for sometimes proposing allocations that are outside of the core. I show that the situations where the Kar solution might be outside of the core are conceptually similar to glove-market games, which led to a criticism of the core by Shapley and Shubick (1969), as the notions of stability and fairness clash.