February 24th, 2016, 1:04 pm
When interpolating, you have a set of (say) times t_i, a set of quantities x_i (for instance, discount factors) and you simply interpolate values between them.When bootstrapping, you have the t_i and a set of f(x_0, x_1, ..., x_i) (for instance, swap rates, which depend on a series of discount factors before maturity). You don't want to interpolate the f_i directly; instead, you want to find the underlying x_i and interpolate those. So you start from f_0 and find the corresponding x_0, then you tackle f_1 and find the corresponding x_1 given the x_0 you calculated previously, and so on iteratively until the last x_n.