We develop a model of unforeseen contingencies. These are contingencies that are understood by economic agents — their consequences and probabilities are known — but are such that every description of such events necessarily leaves out relevant features that have a non-negligible impact on...

We consider a contracting problem between a principal who wants to be informed about relevant stochastic processes and an expert who claims to know which process will generate the data. The data generating process is known to belong to a given class.

Hume (1748) challenged the idea that a general claim (e.g. "all swans are white") can be validated by empirical evidence, no matter how compelling. We examine this issue from the perspective of a tester who must accept or reject the forecasts of a potential expert. If experts can be skeptical...