The guy makes some very good points, but I have to question the value he places on the U.S. having six times as many metropolitan populations of 250,000 or greater than England and Wales. That would absolutely affect the total number of murders, but it wouldn't necessarily affect the murder rate. If the murder rates (murders per 1,000 people) are lower in large English and Welsh cities than they are in large U.S. cities, then that part of his argument evaporates unless we are to believe that somehow the murder rate is increased in a given American city simply because five other cities like it also exist. Unless I'm missing something, the number of cities in a country should be unimportant to the murder rate in those cities.
If we're to draw anything from the comparison, wouldn't we also have to know the percentage of people in each country that live in large cities?