Summary statistics for the patent distributions of the 22 countries.

The best fit of the empirical data to a power law model is achieved with a power law exponent and cut-off . The estimated standard deviation in these parameters is also given. In the case of the standard deviation of , two estimates are given. The first, , is calculated using a bootstrap method via the Matlab code which accompanies [21]. The estimate is obtained via the analytic expression eqn. (3.6) of [21] which gives an estimate for the standard error , assuming that the underlying distribution is well fitted by a power law (i.e. the -value is large). The estimate is calculated via the bootstrap method. The values for indicate the “goodness of fit” of the empirical data to a power law model. Also given, is the number of applicants and the total number of patents held.