In this paper we introduce different measures for the strength of the co-movement between dependent random variables. In a first part, we show how the distribution of the sum can be used to characterize a random vector. We give an interpretation in a utility framework and using distorted expectations.
The distribution of the sum and the theory of comonotonicity are the main ingredients to construct a set of dependence measures, called herd behavior measures. These measures can be used to derive indices for the co-movement between stock prices. Each index is model-free and based on option price data. It represents the expectation of the market about future co-movement between stock prices.