Levy-type stochastic integrals with regularly varying tails

Levy-type stochastic integrals M = (M(t), t >= 0) are obtained by integrating suitable predictable mappings against Brownian motion B and an independent Poisson random measure N. We establish conditions under which teh right tails of M are of regular variation. In particular, we require that the intensity measure associated to N is the product of a regularly varying Levy measure with Lebesgue measure. Both univariate and multivariate versions of the problem are considered.