Reoffending and crime

Posted: May 24th, 2012  Author:   1 Comment »

Disturbing article in the Independent today explaining that the number of hardened criminals reoffending is at a record high while the number of first-time offenders has dropped. At the end of the month – we learnt today – will be releasing the data on justice outcomes which we will take a look at and decide whether and how to integrate. For all that, the general crime trend – recorded and estimated by the BCS –  has been falling gently for 10 years now.

In the States, that has been put down to better Policing, in the UK the answer is much less clear. I never really accepted that Policing skill and financial resource were unicausal to crime or the lack thereof.

But the point that these latest figures make clear, is that we still have much to learn about bringing down reoffending rates – short of keeping criminals in prison longer.


30 years of Recorded Crime versus the British Crime Survey

Posted: July 29th, 2011  Author:   No Comments »

Which of these is correct?













The answer has got to be both and neither of them.  I don’t think you’ll ever get 100% capture of all the crimes and I have my doubts about the British Crim Survey.  Still, there is a big gap between the two because the BCS allows for unreported crime which seems to have peaked in the mid to late 90s at around 12 million plus.


A long-term policy goal should be to keep narrowing the gap between the two.  That would I suggest require some radical innovation to the lower the cost and inconvenience in reporting crime i) by the victim and ii) the recording of the crimes by the Police and Insurance companies.


See the latest annual British Crime Survey here from where I extracted this chart, Figure 1.1 on page 15.

What’s missing from crime statistics data? Competition and innovation, not the ONS

Posted: June 7th, 2011  Author:   No Comments »

According to Jill Matheson, National Statistician of the Office of National Statistics, both the British Crime Survey (who’s shortcomings I have explained here) and the Police figures (which we use) were not properly covering crimes against children, businesses, fraud and cyber crime – and some anti-social behaviour were not included either in the figures.

So far so good, these are the sorts of things we need to know. But beyond that, apart from a few minor improvements, the report strikes me as not very imaginative or game-changing.

So for starters, a big hat-tip to The Thin Blue Line Blog who I’ve just corresponded with for a detailed list of how crime statistics can be gamed and how this does not seem to have been addressed.

And then there’s Recommendation no. 4 which wants a new quango – An Independent Advisory Committee. Oh dear. I thought we were all over quangos – the ultimate Blairite policy solution.

Finally, I have to say my greatest disappointment with the report came from  one of the conclusions, arguing that the solution was to move it out of the Home Office and put themselves – i.e. the ONS -  in charge of it. Isn’t that a bit of a naked bureaucratic power grab?

Granted, the Home Office has long come in for some well-justified stick. But not everyone believes the ONS is some alternative paragon of efficiency, accuracy and brilliance. Remember Len Cook?

To date, the ONS website is still impenetrable and I’m currently tired of waiting for a reply from a couple of weeks ago on a bog-standard socio-economic population breakdown of A,B,C1, C1, D and E.

So here’s what you need to know;

With all government data, crime included, the solution will never be to hand control of it from one monopoly provider to another. The answer is to let many developers use the raw data and to compete and innovate in its provision to the world.

The full report is available here.

How accurate is the British Crime Survey?

Posted: May 31st, 2011  Author:   3 Comments »

A number of times now, a few people have remarked to me that the British Crime Survey (BCS) is a much more accurate guage of crime than police recorded crime which is what we use.  They’re free to think that of course, but I’m always a bit surprised and disappointed to find out that they have no idea how the BCS is calculated. What is it about people who more fervently believe something they don’t actually understand?

Anyway, here is a brief explanation of the BCS compared to what we do;

All crime statistics have 3 components; the i) victim i) the crime and iii) the criminal.

What the BCS does is focus a select group of victims (i) and asks around 50,000 victims living in private households about the crimes they have experienced in the last year.
What we do is take the recorded crimes (ii) by the 43 Police Forces in England and Wales, their category, their location and feed all 500,000 into a database once a month.
So the BCS is about the perception of crime from a victims point of view – useful of course but a much smaller and less factual sample. What our data does is strictly evidential, unless you think the Police are making up the crimes, their location and category !