A bivariate nonlinear model perturbed by external white noises is investigated stochastically. Attention is concentrated on the transient properties before the nonequilibrium phase is achieved. Effects of both additive and multiplicative noises are found to weaken stability and to slow down transient processes. The critical exponent describing this slowing-down phenomenon near a noise-induced instability is estimated for various types of noises. Results derived with two versions of stochastic calculus are compared systematically.