From risk sharing to pure premium for a large number of heterogeneous losses (Q2656992)

Let \(\{X_1,X_2,\ldots,X_n\}\) be a collection mutually independent non-negative random variables representing monetary losses of \(n\) participants to an insurance pool. Denote by \(h_{i,n}(s)\) the amount participant \(i\in\{1,2,\ldots,n\}\) contributes to the pool, where \(s=\sum_{i=1}^{n}x_i\) is the sum of the realizations \(x_1,x_2,\ldots,x_n\) of random variables \(X_1,X_2,\ldots,X_n\). A fair \textit{risk sharing rule} is an allocation scheme such that, for all \(n\in\mathbb{N}\) there exist measurable functions \(h_{1,n},h_{2,n},\ldots,h_{n,n}\) satisfying two requirements: \[ \sum_{i=1}^{n}h_{i,n}(s)=s \ \text{ for }\ s\geqslant 0; \] \[ \mathbb{E}\left(h_{i,n}(S_n))\right)=\mathbb{E}X_i\ \text{ for }\ i\in\{1,2,\ldots,n\}. \] The first condition means that the total risk \(S_n=X_1+X_2+\ldots+X_n\) is entirely allocated among the \(n\) participants, and the second condition ensures actuarial fairness. Authors of the paper study the behaviour of three \textit{risk sharing rules} when the size of the pool becomes large.