Lies, damned lies and guns in Australia
Comparing mass shootings in Australia and the US
Happy New Year.
The recent mass shooting in Sydney has shocked and disquieted a nation. Three weeks ago, a father-and-son team of Islamic extremists approached a Hannukah celebration on Bondi Beach and began hurling pipe bombs and firing rifles and shotguns into the crowd. Despite the heroic efforts of police officers and bystanders to intervene, fifteen people were killed (along with one of the perpetrators) and dozens more injured. Sadly, this is just the latest item on the laundry list of anti-Semitic violence we’ve witnessed across the world over the last two years, from Washington D.C. to Manchester. To paraphrase the newly inaugurated mayor of New York City, the intifada has been successfully globalised.
I have many thoughts on this attack. I agree with many that the attack was likely precipitated (at least in part) by two consecutive years of nonstop baseless accusations of genocide levelled at Israel and the wider Jewish diaspora, fostering a discursive environment in which violence towards Jews seems like a reasonable course of action: it has been, to use the progressive expression, a game of stochastic terrorism. The attack provides yet another sad example of the fallacious reasoning of Jews who think they can keep themselves safe by loudly proclaiming themselves to be “good” Jews who support the cause of Palestinian independence: if these two murderers made sure to diligently ask each of their victims if they were pro- or anti-Zionist before gunning them down, this is the first I’ve heard of it. While Ahmed Al Ahmed’s bravery in disarming one of the shooters is to be applauded, I wholeheartedly agree with Yassine Meskhout that the media’s selective championing of him (and not Boris & Sofia Gurman, Gefen Bitton and Reuven Morrison, four equally courageous Jews who also attempted to intervene) is a cynical and transparent attempt to run interference on behalf of Australia’s Muslim community and deflect from the religious beliefs of the perpetrators.1
Lastly, I want to make one specific point, leaving zero room for ambiguity or misinterpretation. If you heard about this attack and found yourself thinking that it was only to be expected given Israel’s conduct in Gaza over the past two years — I would invite you to consider how disgusting and grotesque such a chain of reasoning would sound if made about any other religious or ethnic group (e.g. “The Christchurch mosque shootings were despicable, but what did those Muslims expect after 9/11?”). And if you still think it’s a valid argument, then there’s no other term for it: you are an anti-Semite, you should not pretend otherwise, and you should be treated accordingly. I am completely done trying to meet such bigots halfway or trying to understand where they’re coming from.
Now that I’ve gotten that off my chest, let’s turn to something a bit less controversial. What does this attack imply about the accessibility of firearms in Oz?
For the better part of the last thirty years, there’s been a stock narrative in Australian politics. For decades, gun violence was a major problem in the country, culminating in the infamous Port Arthur massacre of 1996, in which a profoundly disturbed young man murdered thirty-five people more or less indiscriminately in a resort town in Tasmania. In response, the Australian government decided that enough was enough and took steps to limit the number and type of weapons available to the public. In the short term, this took the form of a national gun buyback program, a twelve-month amnesty in which gun owners could return their weapons to a public body in exchange for a fee, upon which these weapons would be decommissioned. The buyback program was such a roaring success (so the narrative goes) that gun violence in Australia fell precipitously immediately afterwards, and mass shootings like Port Arthur are a thing of the past.
Like many of you, I first heard of this program when John Oliver of The Daily Show interviewed some of the program’s architects in 2013. Ever since, it’s been a common talking point among pro-gun control Americans that preventing mass shootings in the US would be as simple as following Australia’s example.
This narrative is powerful and seductive: the idea that mass shootings can be decisively prevented by a few pieces of legislation (and hence that American politicians opposed to similar legislation have blood on their hands); that mass shootings really are a case of a one-sided policy debate where we actually don’t have to weigh up competing values and balance trade-offs — we can just flick a switch and the problem disappears. The US is, after all, the only nation where mass shootings regularly happen, or so The Onion would have us believe: it’s intoxicating to imagine how easy they would be to prevent, if those darned Republicans would only get their priorities in order and stop taking back-handers from the NRA.
But is this narrative true?
To begin with, proponents of gun buyback programs are generally careful to qualify the program’s success in Australia as one of preventing mass shootings, not of preventing gun deaths in general. As exhaustively catalogued by Handwaving Freakoutery, even if the Australian gun buyback program was exactly as successful at preventing mass shootings as its champions claim, that’s a somewhat peculiar metric to focus on, given what a small proportion of gun deaths these incidents represent. Despite the pro-gun control’s faction’s obsessive fixation on banning AR-15-style rifles, in any given year, the overwhelming majority of gun deaths in the West are caused by handguns.2 (Indeed, the Virginia Tech shooting, the deadliest school shooting and third-deadliest mass shooting in American history, was committed by a perpetrator who used handguns exclusively.) The fact that indiscriminate Columbine-style mass shootings attract so much media coverage is precisely because they are “man bites dog” stories: the ordinary churn of gang violence and murders within the family does not attract anything like this level of public interest.
Following the massacre at Bondi Beach, certain American conservatives have been quick to make an even more glaring criticism of the stock narrative. After all, if the gun buyback program was so successful in preventing mass shootings, how could this one even have occurred?
I wanted to get to the bottom of this, so I planned to do a quick and dirty analysis comparing gun deaths in Australia and the US. I expected to find that Australia has vastly fewer deaths from mass shootings, but that when comparing total gun deaths per capita, the difference between the two nations wouldn’t be quite as striking.
Instead, I came to an even more surprising conclusion: even just limiting my analysis to deaths by mass shooting since the gun buyback program, the difference between the US and Australia is nowhere near as stark as is widely believed.
First, a question of definitions. “Mass shooting” is not a term of art, and there are many competing definitions in use concurrently. At least one NGO defines it as an incident in which four or more people are shot at one time, even if none of them are fatally wounded. I don’t think this reflects how the term is used in common parlance, but to be as generous to the pro-gun control position as strikes me as reasonable, I’m using this database of mass shootings in the US since 1982, produced by Mother Jones, an outlet which has been accused of lots of things but never of being a shower of Second Amendment diehards. For the Australian data, I’ll be using the Wikipedia article listing mass shootings in the country over the last two centuries. Mother Jones’s definition of a mass shooting excludes any incidents with fewer than three fatalities (therefore excluding the Annunciation Catholic Church shooting in August 2025), so to compare like with like, I will be excluding any Australian incidents meeting that description.
As mentioned above, the Port Arthur massacre occurred in April 1996, and the subsequent gun buyback program ran for twelve months, from October 1996 to September 1997, retrieving 650,000 guns. If the program was as successful as is claimed, we would expect to see a sharp divergence in mass shooting deaths between Australia and the US, starting no later than September 1997. According to Wikipedia, since September 1997 there have been 12 mass shootings in Australia in which three or more people were killed (including Bondi Beach), claiming some 61 lives. Per Mother Jones, in the same period there were 136 mass shootings in the US which collectively claimed 1,006 lives.
To perform a per capita analysis, I will be simply taking the populations of the US and Australia in 1997, their equivalents at the end of 2025, and averaging them.
I’m sure it’s obvious that, per capita, the difference in the number of people killed in mass shootings in the two countries is not terribly large. By my count, between September 1997 and December 2025, a given American was only 21% more likely to be killed in a mass shooting than a given Australian was. Now, that’s certainly a metric that Australians have every right to be proud of — but it is not quite the vindication for the gun buyback program that many have claimed. The claim has typically been that the program in question prevented deaths from mass shootings almost entirely, not merely reduced them to 79% of the equivalent figure in the US.
Comparing the two countries over time in five-year buckets is similarly illuminating. Per capita, more Australians were killed in mass shootings between 1997—2006 than Americans were. However, unlike in the US, that figure stayed fairly flat throughout the 2000s and 2010s, right up until 2022, at which point it shot up, coincidentally at the same time that the American figure nearly halved. (From reading progressive news outlets, would you have gotten the impression that mass shootings in the US peaked in the period 2017-21 and then fell precipitously? I certainly wouldn’t have.)
An obvious objection to the above graph is that the enormous Australian spike in the period 2022-25 is almost entirely attributable to the Bondi Beach massacre (although not solely: in 2022, two shootings claimed nine lives between them), and that outlier events like this can unfairly skew our results. This is a reasonable objection: after all, of the 61 deaths we’re counting, 16 were caused by a single incident, or one-twelfth of the total number of incidents. But interestingly, the same is true of the US: of the 1,006 deaths in that dataset, a full ten per cent are attributable to exactly two incidents (Las Vegas and the Pulse nightclub attack). If we exclude the most lethal one-twelfth of all incidents,3 the picture between the two nations still looks much the same:
Once again, excluding these extreme outlier events, a given American was 15% more likely to be killed in a mass shooting than a given Australian. A significant difference, but not the all-or-nothing disparity that proponents of gun buyback programs would have us believe.
My analysis has vindicated the gun buyback program in one regard, however. It’s been nearly thirty years since the program was completed, and if we compare Australian mass shooting deaths per capita before and after the program came into effect, the difference is indeed quite dramatic:
When viewed strictly in local terms, the gun buyback program must be judged a huge success. In the period immediately after the program’s introduction, mass shooting deaths per capita fell by a whopping ninety per cent, and even after this recent tragedy in Bondi Beach, the figure in the current period is only a quarter of its peak in the 1980s. The architects of any government policy that can cause this kind of seismic impact are to be commended.
But this commendation is to be qualified. The program did not eliminate mass shootings entirely, and nearly thirty years on, Australia’s mass shooting deaths per capita are still only about one-fifth lower than in the United States. “The Australian government recognised that the country had a serious problem with mass shootings, took steps to correct it, and ended up with 20% fewer deaths per capita than the US” could hardly be further removed from how the program is typically conceived of in the popular imagination.
In response to the massacre at Bondi Beach, prime minister Anthony Albanese has already announced a third gun buyback program (there was a second program in 2003, targeting handguns). Of course he has: after a highly visible public atrocity like this, a country’s leader must be seen to do something, and it makes perfect sense to go back to the well and re-implement a policy which was so successful in its first iteration. I certainly hope that it will have the desired effect and that further tragedies like that three weeks ago will be prevented.
But I am sceptical that this program will have anything like the impact of the 1996-7 program. I predict that, via regression to the mean, mass shooting deaths per capita will quickly return to the baseline they stood at throughout the 2010s, and the program will be judged a success on that basis. But expecting the same policy to cause mass shooting deaths to fall by ninety per cent twice in the space of thirty years is the product of unchecked optimism.
Inevitably, the Wikipedia page on the attack is sure to include an entire section about a spike in the harassment of Australian Muslims since the shooting. While I’m of course sympathetic, something about this framing strikes me as rather suspect. There are times when the Norm MacDonald references practically write themselves:
The day after Charlie Kirk was assassinated, I overheard a Canadian woman expressing disagreement with his opinions about the right to bear arms, saying something to the effect that “if you want a handgun for home defense, fair enough, but you don’t need a big AR-15 rifle”. All I could think was: wow, you have that exactly backwards.
Meaning that I’m excluding the Las Vegas Strip massacre, the Pulse nightclub massacre, the Virginia Tech massacre, the Sandy Hook Elementary massacre, the Texas First Baptist Church massacre, the El Paso Walmart mass shooting, the Robb Elementary School massacre, the Maine bowling alley and bar shootings, the Marjory Stoneman Douglas High School shooting, the San Bernardino mass shooting and the Binghamton shootings.
As you point out, mass shootings is a silly indicator of gun violence. AI tells me overall the US has a 30x higher per capita gun homocide rate than Australia...
Your prediction about the effect of regression to the mean strikes me as spot on. It's such a frequent problem and tough to combat.