Macroseismic intensity has recently attracted attention as a tool for validating probabilistic seismic hazard assessment (PSHA) studies or as an alternative method for PSHA in countries where the historical catalog is much longer than the instrumental one. In Italy, the new seismic hazard map was recently produced using the Cornell–McGuire approach in terms of the peak ground acceleration characterized by a 10% exceedance probability for an exposure time of 50 yr (Amax). We compare this map with an alternative one, produced using a different approach based on a nonparametric and zonation-free statistical analysis of local seismic histories. In this case, results are expressed in terms of the maximum intensity corresponding to an exceedance probability of not less than 10% for an exposure time of 50 yr (Iref). In order to compare the two maps, we selected 1401 control sites, where local seismic history includes at least 10 intensity values relative to felt effects documented during past earthquakes. The values of Amax and Iref at these sites have been ranked in the respective domains. The spatial distribution of rank differences of Amax and Iref values shows a strong correlation with the seismogenic zoning used in the calculation of PSHA following the Cornell–McGuire approach. This suggests that the adopted zoning could be incomplete (some further “hidden” sources may exist) and too rough to capture actual seismogenic sources. Because more detailed zoning is prevented by the amount of data available, the results obtained suggest the preference of zonation-free approaches for seismic hazard assessment in Italy. Furthermore, among the possible zonation-free approaches, those that offer better exploitation of local information about the effects of past earthquakes would be preferred.