Well, I'm not quite as pessimistic as 19-3Ben, but his general skepticism is admirable.
Getting good data in this case is difficult. The least unsatisfactory data I could find was from
http://www.gunpolicy.org/firearms/region/australia, and this data was cobbled together from several sources.
Here is their homicide rate per 100,000, from 1990 to 2012, presented in a trended Individuals and Moving Range control chart. The purpose and design of this chart is to detect real change in a process.
[/URL][/IMG]
And the interpretation of the chart is:
1. Australia has had a stable and predictable decline in homicide rate during the years 1990-2012, with no statistically interesting deviations but one, in the 2004 and 2005 time frame (points 15 and 16).
2. There was no statistically interesting decline in homicides in the period immediately following Australia's seizure of semi-auto firearms.
3. Something interesting happened in 2004 and probably 2005. The reported homicide rate declined. This might be a real decline, or it might represent a change in the measurement system. Remember, the data are cobbled together from various sources. At any rate, by 2006 the data were back to the established trend line.
4. There is no evidence in the data to support the notion that the Australian gun ban actually did anything at all to reduce the total homicide rate. If the ban had been effective, a drop in homicide rates should have immediately followed, and it did not. The party proposing change has the burden of proof, and there just isn't any evidence in favor of the Australian ban.
5. Without banning semi-autos, the US experienced a very similar drop in homicide rate over the same period. Both countries reduced homicides by 50%. It is reasonable to believe that factors other than gun control have influenced both countries.