The Standard and DPF are ‘debating’ the recently released crime statistics for last year. I’m inclined to agree with The Standard that the increase in violent crime is largely caused by an increase in the reporting of domestic crime and that this is basically a good thing.
But I can’t help noting this gem in Steve Pierson’s original article: its a graph proving that Labour is beating crime:
This is accompanied by the following quote:
Crime fell 1% in the last year from 1025 recorded crimes per 10,000 people to 1014. This is part of a pattern that has continued since National was booted from power and living conditions for the poor started to improve.
Kinda looks to me like crime started falling in 1997 three years before the National government was booted out and that the trend under Labour has been a rather erratic continuation of that decline. ‘There’s none so blind’ and all that.
it’s better than that, the entire line trends down from about 1992.
you’d need to see the 80s numbers to decide where it started.
Comment by Che Tibby — October 2, 2008 @ 6:15 am |
I think overall crime isn’t a very useful measure. Can you plot the violent crime category? I.e. murders, assaults, rapes, those kinds of things. Because overall crime is prone to be under reported (would you report a break-in these days?) but violent crime is less likely not to be reported.
Comment by Berend de Boer — October 2, 2008 @ 6:24 am |
would you report a break-in these days?
Yes – your insurance company doesn’t pay out if you don’t report it.
Comment by Danyl Mclauchlan — October 2, 2008 @ 6:32 am |
it’s better than that, the entire line trends down from about 1992.
you’d need to see the 80s numbers to decide where it started.
That’s interesting – almost every other western country in the world also saw a decline in crime during the 90’s.
Comment by Danyl Mclauchlan — October 2, 2008 @ 6:34 am |
That graph uses the oldest trick in the book — don’t show the zero for the vertical axis. That way you exaggerate the changes in favour of your argument.
There is another reason OVERALL crime is down. People are losing confidence in the Police.
Here’s an example. My neighbour came over to warn us about the people on the other side of her. One of them was discovered climbing in her window, when they thought there was no one home. She is moving out. She won’t go to the Police for fear of retribution.
Sadly, because I’m sure there are many good cops, the bad ones (including in national HQ) are causing the reputation of the Police to slide down faster than that graph.
Comment by David White — October 2, 2008 @ 6:37 am |
You have to report a burglary if you want to claim insurance. It’s paranoid to simply disbelieve the conclusions of the Police statisticians because they don’t match with your political objectives.
Che Tibby. It looks like that only because of the consistant decrease since 1999. If you look at the 1990s stats alone they start essentially flat, peaking when unemployment peaks.
David White. You don’t use zero on the axis when the value your are measuring is never going to approach zero. For instance, say we ere graphing changes in the surface temperature of the sun, it would be silly to have 0 axis. 0 is not some special number, wouldn’t demand that a grpah always include 1000. That said, when we are interested in levels relative to zero, you need zero on the axis. But what we’re interested in is the patterns of change, so you ‘zoom-in’ the axis to frame the range of values that have occured
Comment by Steve Pierson — October 2, 2008 @ 7:25 am |
@steve, that’s not the case. restricted figures between 1991 and 1999 would still trend down, just more gradually than between 2000 and 2008.
as i say, without the 80s figures you can’t tell if the 90s figures are the high-point of reported crimes.
but, to address your point, the entire graph is likely to reflect employment figures, and overall the trend is *down* since a high-water mark after the 4th labour government.
Comment by Che Tibby — October 2, 2008 @ 7:34 am |
“the entire graph is likely to reflect employment figures”
I’d be interested to see how it correlates with the proportion of males aged 18-25.
Comment by Neil — October 2, 2008 @ 8:07 am |
Freakonomics looked at crime stats and the downwards trend that started in the 90’s/ The book drew a correlation with the use of abortion. That is a generation of potential criminals never existed.
You may or may not agree with their view, but it highlighted that emphasising being tough on law and order is only part story, as well as rising perceptions of well being. Also rise of cheap goods from Asia helped remove secondary market for stolen goods at the pub. Not sure whether this still holds with the rise of P.
Comment by What would Hayek say — October 2, 2008 @ 8:17 am |
That’s interesting – almost every other western country in the world also saw a decline in crime during the 90’s.
Yeah, crime steadily rose in most western countries until the early 90’s, then suddenly dropped away. There’s been a LOT of research on why but I don’t think anyone has come up with anything conclusive yet.
Theories tested include:
- stronger/weaker economy
- higher/lower incarceration rates
- more/less police
- legalized abortion
- unemployment levels
- drug use
- gun controls
There’s statistical evidence that *some* of these are factors in *some* places, but none of them stack up as full explanations for the general trend.
Crime rates might just be the result of culture/society as a whole and too complex to explain with simple answers.
Funnily enough Gordon Brown has been claiming the same thing in the UK – that crime under Labour has dropped for the first time since WWII. But their trend is the same as ours – dropping from 1992, five years before Labour got in.
Comment by gazzaj — October 2, 2008 @ 8:18 am |
Police resolved 9685 more offences in the year, increasing the resolution rate to 47 per cent from 44.7 per cent in the previous period.
Wa-hey.
Comment by StephenR — October 2, 2008 @ 8:39 am |
The most reliable crime statistic is the murder rate, for a number of reasons.
And it shows the same trend as is expressed above; a peak around 1991-1992, and a slow decline from the mid-1990s which continues steadily under Labour.
Everything else is much more affected by reporting rates and understandings of what constitutes a rape or assault etc.*, and as a result can’t be reliably compared over time.
*I’m inclined to say that people are less tolerant of these things and that this is A Good Thing.
Comment by George Darroch — October 2, 2008 @ 5:01 pm |
Oh, and again, this is why The Standard’s constant ‘attack attack attack, Helen is wonderful’ modus operandi does them no good.
Real statistical work like this becomes propaganda (1999 as the turning point), and a very valid point is lost.
Comment by George Darroch — October 2, 2008 @ 5:10 pm |
“The most reliable crime statistic is the murder rate, for a number of reasons.”
Not in New Zealand. The numbers are too low and bounce around too much for any decent conclusions to be drawn. And come to think of it, things like improvements in trauma medicine and the firearms technology actually distort the general picture.
Random violent crime in first world countries is really strongly co-related with alcohol consumtion, both among perpetrators and victims. Increases in violence in New Zealand are almost certainly a factor of more liberal alcohol laws since 1999. So it’s as a result of policy decisions, but not really a symptom that the world is falling apart.
Comment by Guy Smiley — October 3, 2008 @ 5:47 am |
I should have clarified – I meant that the murder rate is most reliable in terms of reporting, and secondly as police investigative methods improve over time so to does the resolution rate (although annoyingly murder is lumped in with violent crime and I can’t find a separate resolution rate).
That the murder rate bucks like a pony with such a small sample should be obvious to everybody. Unfortunately it isn’t.
I agree with Guy about alcohol, and would say that young men (15-35) particularly, (and young women less so) having more significantly more disposable income also appears to be a factor.
People in my home community simply couldn’t afford to get as boozed as they do now. Obviously there is a point where more income does not mean more alcohol, but up to a certain point it holds true, and the results are obvious.
Comment by George Darroch — October 3, 2008 @ 11:09 am |
I also wonder if things like cellphones affect the murder rate. Now that everybody has a phone all the time, emergency crews arrive on scene a lot faster.
I’ve also read that the various offenses around threats/intimidation are a big part of the increase in violent crime and that the police feel that these are reported more because people now have access to phones when they’re walking around the city late at night, and they didn’t used to previously.
Comment by danylmc — October 3, 2008 @ 1:38 pm |