Asymptotics for the finite-time ruin probability in a discrete-time risk model with dependent insurance and financial risks*.

We consider a discrete-time risk model with insurance and financial risks. Within period <italic>i</italic> ≥ 1, the real-valued net insurance loss caused by claims is the insurance risk, denoted by <italic>X</italic>_{<italic>i</italic>}, and the positive stochastic discount factor over the same time period is the financial risk, denoted by <italic>Y</italic>_{<italic>i</italic>}. Assume that {(<italic>X, Y</italic>), (<italic>X</italic>_{<italic>i</italic>}<italic>, Y</italic>_{<italic>i</italic>}), <italic>i</italic> ≥ 1} form a sequence of independent identically distributed random vectors. In this paper, we investigate a discrete-time risk model allowing a dependence structure between the two risks. When (<italic>X, Y</italic> ) follows a bivariate Sarmanov distribution and the distribution of the insurance risk belongs to the class ℒ(γ) for some γ > 0, we derive the asymptotics for the finite-time ruin probability of this discrete-time risk model. [ABSTRACT FROM AUTHOR]