Well, there is not an option for this, however, the computation of the p-value is based upon a normal approximation and is easy to compute by hand.
The computation could be done as follows:
Z = T / sqrt(Var),
where T = Kendall Tau correlation and
Var = 2*(2*N+5)/(9*N*(N-1)), and
N = Number of observations,
the two tailed approximate p-value is then 2 * NormalCDF(-|Z|).
The formula in STATISTICA for NormalCDF(x) is INormal(x,mean,sd). Here use INormal(-Abs(z), 0, 1).