Even with that qualification that is still crap logic. There are lots of guns in America, but the number of Americans who own guns are at all-time lows. The "number of guns in America" figure is deceptive because of all of the people who own dozens to hundreds of guns.

http://www.pewresearch.org/fact-tank/2013/06/04/a-minority-o...

Arresting somebody who owns one or two guns is really not inherently safer than arresting somebody who owns a few crates of guns, so the number of guns in the country is irrelevant. The number of gun owners is relevant.

The apparent upward trend in police militarization and the number of Americans that own guns simply do not correlate. The Economist's unfamiliarity with American culture (particularly American gun culture) explains how they could make this mistake.

It isn't because gun ownership is up, it is stating that police shootings occur because of the mere fact that there are already more guns in the USA than any other first world country, thus police are more likely to encounter criminals with guns. We also have the highest gun homicide rate of any first world country. The two are directly related.

> "it is stating that police shootings occur because of the mere fact that there are already more guns in the USA than any other first world country, thus police are more likely to encounter criminals with guns"

Police in America are less likely today to encounter guns than they were 40 years ago, but police shootings are apparently at all time highs. If their theory held water, that would not be the case.

The USA still has 300 million guns. This is far and beyond any other first world country, so yes this is in fact related to police in the USA more likely shooting first than they would in other countries which do not have such an insane saturation of guns like the USA.

When you have a hypothesis, you have to try making predictions with it and see if it fits real-world data. If it doesn't, then it is wrong. It doesn't matter how elegant or beautiful your hypothesis is, if it does not fit the data then it is wrong.

Your hypothesis does not fit the data.