Britain’s crime wave is real – and our data can’t keep up

We are in the midst of a crime wave: shoplifting and snatch theft are the highest on record. There were over 500,000 shoplifting offences last year and over 116,000 phones stolen last year in London alone. Headline crime has also increased by 7% since last year. This is driven, chiefly, by large increases in fraud.

This is unsurprising to most of us – how much shoplifting have you witnessed recently; how many people do you know that have had their phone or bike stolen in the past two years? Despite this, multiple political commentators chose to deny this reality. There was a concerted pushback that this ‘increase in crime’ narrative was simply false. The arguments for this have gone through a few distinct phases and they tell us how badly the British public may be served by official data. 

Last year, political commentators were focusing on the Crime Survey for England and Wales (CSEW) from 2023 to show Britain’s crime rates were reducing. The CSEW is the best picture of crime we have: it relies on asking individuals if they were a victim of crime in the last year. This is, in many ways, preferable to police reported crime as that number reflects a lot of potential variables. To take one example: low crime areas with effective police forces may get more reports simply because victims expect them to do something – while worse crime areas with ineffective forces might see fewer cases reported as the locals consider there to be no point.

This year, with the most recent CSEW figures showing an increase in crime (driven by fraud and theft), commentators have been quick to dismiss concerns again. This time, they have chosen to carve out fraud or ‘white collar crime’ and are focusing entirely on murder rates and other serious violence (which is down according to our most recent CSEW data).

There are serious problems with doing this. If a population is complaining that they see lawlessness around them – and the data we do have starts to bear this out – it is not sensible to point to a different category of lawlessness and to say ‘well, that’s doing better’. This approach also severely discounts differences in places across the country. I am sure crime is decreasing rapidly in many towns across this country; in some areas, it is almost certainly the case that you are more at risk than you were a few years ago.

Regardless, this entire debate is based on the best figures we have currently: the CSEW. But these figures are – in a word – poor. Shoplifting is still only an appendix, and is only police reported instances – this is because shoplifting is, apparently, a ‘victimless crime’. Considering less than 1% of reported shoplifting ends in a suspect being identified, I do not think the majority of shoplifting is being reported. The true figure must be astronomical.

Similarly, tourists are victims of crime but are not captured in the data, and nor are witness accounts of those crimes. More concerningly, response rates for the CSEW have collapsed. The year ending March 2025 edition of the CSEW claimed a response rate of 46%. In the 10 years prior to the COVID-19 pandemic, the response rate fluctuated between 70 and 76%.

The problems go deeper still. The CSEW sexual offences overview, which is based on six months of interviews, has seen an unweighted base reduction from ~25,000 in 2005/6, to ~3,000 in 2021/22. This is over an eight times reduction, and the reduced response rates now mean the data is not designated as National Statistics.

Other than the ‘prevalence’ of sexual assault provided in the main CSEW dataset, 2021/22 is the latest available detailed sexual offence data. Though the CSEW is often regarded by political commentators as the best available measure of crime in the country, this statistical degradation means it reveals nothing about rates of sexual offences now. Therefore, police recorded crimes are the only option remaining for assessing sexual offence rates in the UK – which are currently, according to these figures, increasing. These estimates are, despite their flaws, published as accredited official statistics, but police recorded crimes lost their accredited designation due to concerns regarding reliability, thereby creating difficulties for anyone attempting to rely on our national crime statistics.

This is doubly true when the CSEW does not record a number of crime types which appear, at least anecdotally, to be increasing.

This is not to say that we must await perfect data, but it is to say that dismissing the concerns of a population on the basis of this data is extremely unwise. It is also clear that our data collection on crime requires a thorough improvement. The quality of our data becomes more important the more vital the decisions made on it are – and this is true not just for crime. Our Labour Force Survey figures are no longer accredited statistics, and our official statisticians are under fire. Similarly, for all the influence of the Office for Budget Responsibility’s reports, do we know enough about the success rate of the OBR’s predictions – where is this recorded? Is it better than the baseline?

There are many pictures of life across Britain that official data does not bear out. We need to look at that data carefully before dismissing the decline that the public sees all around them.

 – the best pieces from CapX and across the web.