Monthly Archives: November 2017

UK Prisoners by Sex: The MoJ Report


  1. Introduction and Background
  2. mra-uk analyses – summary
  3. The MoJ Report – Headline Findings
  4. The Raw MoJ Data
  5. The MoJ Regression Analysis
  6. Race versus Sex
  7. Sex Offending
  8. Other Sources of Gender Disparity
  9. Conclusion

1. Introduction and Background

Well, well, what do we have here? A report from the MoJ itself on gender disparity in prison sentencing. I’ll get to that. First, a brief reprise.

It was in February 2014 I first wrote an article, on my older site, about men being treated more harshly than women by the judiciary in regard to imprisonment. In the process I discovered I was not the only one to have noticed. Philip Davies, MP for Shipley, had spoken about the matter in Parliament on 16th October 2012 (2:30pm). I later posted a precis of the argument on this site on 21st November 2014. It was to become the most visited article on this site (excluding the Home Page and “Introduction”).

I addressed various challenges to my (admittedly crude) analysis in this post which followed in March 2015. Later I presented an updated analysis, here, based on indictable offences only, which I now regard as more robust. Both the original analysis and the updated analysis were presented in a public talk held at UCL on 25th February 2016 (see the Powerpoint slides or watch the video).

Then, on the 8th July 2016 Philip Davies spoke again, at ICMI16, on the matter of the different treatment of men and women as regards prison sentencing. Having used the expression “feminists want to have their cake and eat it too”, Davies’s talk was subsequently subjected to the full force of rigorous, rational, feminist rebuttal in the form of selfies on social media of young women eating cake. Slightly less bonkers, though no more accurate, were a rash of newspaper articles by feminists. I noted only a couple – those by Sophie Walker, leader of the Women’s Equality Party, and by Baroness Jean Corston, author of the infamous 2007 report on women in prison. These articles, ostensibly refuting what Philip Davies had said, turned out to be a litany of falsehood (see also this).

I had no intention of revisiting this issue again…..until I stumbled, quite by accident, on this…

The Ministry of Justice Report  “Associations between being male or female and being sentenced to prison in England and Wales in 2015

The report uses data for calendar year 2015. Judging by this web page it was published on 24th November 2016 – later than the sequence of speeches and analyses I summarised above. In particular, the MoJ published this analysis a few months after the ridiculous ‘cake selfies’ and the championing of untruth by Sophie Walker and Baroness Corston. One wonders if there is any link. The rest of this blog, indeed the point of this blog, is to summarise the MoJ’s own findings.

2. mra-uk analyses – summary

But first, the headline findings of my own analyses are shown in Table 1. Recall my analyses quantified three factors,

  • Imprisonment disparity: the probability of a convicted man being sentenced to prison divided by the probability of a convicted woman being sentenced to prison for the same category of offence.
  • Sentence length disparity: for men and women both sentenced to prison for the same category of crime, the ratio of their average sentence lengths (men/women).
  • Parole disparity: the average fraction of prison sentence actually served by male prisoners divided by the average fraction for female prisoners.

My claim was that the overall sentencing disparity consists of the product of those three factors. Based on indictable offences only, the overall disparity was a factor of 3.6 in 2013 (but appeared to be increasing).

Thus, if women were treated as harshly as men, there should be 3.6 times as many women in prison. Or, if you prefer, if men were treated as leniently as women, 26 male prisoners out of every 36 would not be there.

My analyses included a crude consistency check: by factoring the overall disparity by the ratio of the number of convictions for men compared with women, the resulting factor (around 20 to 22) reproduces the actual gender ratio of prisoners.

Table 1: Disparity analyses on this and associated sites

Item Based on all offences

(original analysis)

Indictable offences only

(revised analysis)

Male:Female Convictions 3.4 6.2
Imprisonment disparity 3.4 1.9
Sentence length disparity 1.6 1.7
Parole disparity 1.1 1.1
Product(1) 20 22
Overall disparity(2) 6 3.6

(1)Product of all four factors results in an estimate of the ratio of men:women in prison

(2)Product of the three disparity factors gives the overall disparity

3. The MoJ Report – Headline Findings

So, now to the MoJ’s 2016 report. The MoJ looked only at imprisonment disparity, i.e., how much more frequently convicted men are sentenced to immediate imprisonment than women convicted of the same offence. The MoJ did not look at disparity in sentence length or parole.

The MoJ’s headline finding was “Under similar criminal circumstances the odds of imprisonment for males were 88% higher than for females”, i.e., an imprisonment disparity factor of 1.88. This compares with my revised analysis based on indictable offences of 1.9, see Table 1 – remarkably close agreement.

It is important to note that the MoJ have controlled for offence type, albeit across a rather coarse set of only 20 crime categories. However, they have also carried out a logistic regression analysis, which provides a rather more sophisticated means of controlling for other variables than my crude analysis.

One would hope that the MoJ will analyse the sentencing length data at some point, in order to estimate the overall disparity as I have done. So far, it looks like they will come to a very similar conclusion.

However, the MoJ note – as I did too – that there could be systematic differences in offending severity within offence categories. Essentially this is a recognition that 20 categories may provide insufficient granularity. But the MoJ would hardly have issued this report if they regarded its findings as misleading.

Many people would rationalise the disparity simply by assuming male offending to be more severe than female offending in every crime category. They will be especially prone to this assumption in the case of violent crimes (though I am unaware of any evidence to support it – such an assumption would be prejudice, not fact). And it is harder to cling to this belief in the case of, say, fraud, handling stolen goods, drink-driving, jumping bail, etc., where physical strength or violence is not an obvious feature. Moreover, Starr’s analysis of USA data uses far greater granularity in offence categorisation and produces a remarkably similar imprisonment disparity (about 2). She remarks,

One obvious question is whether the crimes differ in ways not captured by the arrest offense codes. The arrest offense is not a perfect proxy for underlying criminal conduct, and if it overstates the severity of female conduct relative to that of men, that might explain some of the observed disparity……Unobserved differences naturally cannot be ruled out, but there are good reasons to doubt that they explain much of the observed disparity.  First, the observable covariates are detailed, capturing considerable nuance.  They include not just the 430 arrest codes and the multi-defendant flag….but also additional flags based on the written offense description.  Second, the disparities are similar across all case types (and across arresting agencies), suggesting it is not a matter of a few crimes being “worse” when men commit them.  Such differences would have to be prevalent across a variety of crimes and agencies to explain the result.  Third, there is some reason to believe unobserved divergences between the arrest offense and actual criminal conduct may bias disparity estimates downward.”

Some of the detailed results of the MoJ’s analysis follow – looking firstly at their raw data and then at the results of their regression analysis.

4. The Raw MoJ Data

  • Females were less likely to be imprisoned than males (8% of those sentenced versus 18%).
  • The most common offence for both sexes was violence against the person (VAP)*, accounting for 22% of women convicted and the same percentage of men convicted. However, whilst 20% of men convicted for VAP were imprisoned, only 7% of women convicted for VAP were imprisoned.
  • *For women, theft also accounted for 22% of convictions.
  • The second most common offence for both men and women was drink-driving, accounting for 15% of women’s convictions and 12% of men’s. However, whilst about 700 men were imprisoned for drink-driving, seemingly no women were (though “0%” may be rounded).
  • 31% of men convicted of fraud or forgery were imprisoned, compared with 18% of women.
  • 23% of men convicted for handling stolen goods were imprisoned compared with 10% of women.
  • 63% of men convicted for domestic burglary were imprisoned compared with 35% of women
  • 13% of men convicted of public order or harassment offences were imprisoned compared with 4% of women.
  • 20% of men convicted for vehicle-related theft were imprisoned compared with 6% of women.
  • 4% of men convicted of welfare fraud were imprisoned compared with 2% of women. (This was a rare instance of the absolute number of women offenders exceeding that of men).
  • 11% of men convicted of absconding/jumping bail were imprisoned compared to 7% of women.

And so on.

One of the most frequently repeated canards beloved of feminists is that women are more likely to be imprisoned for a first offence. I have discredited that claim previously but the MoJ does the job again. Only 5% of convicted women with no previous convictions or cautions were sent to prison, compared with 11% of men. Similarly, only 9% of convicted women with one or more previous convictions or cautions were sent to prison, compared with 19% of men.

However you cook it, cut it, or break it down, men are sent to prison substantially more often than women.

5. The MoJ Regression Analysis

Multivariate logistic regression models were built to account for the associations between sex, officer-identified ethnicity, age, offence group, previous criminal history, and being sentenced to prison, compared with being sentenced to another type of disposal such as a community order, suspended sentence order, or discharge. This allowed the associations between sex and imprisonment to be examined under similar criminal circumstances.

Under similar criminal circumstances the odds of imprisonment for males were 88% higher than for  females.

The odds ratios (men:women) for imprisonment in various crime categories resulting from the regression were,

  • Violence against the person: 2.68
  • Fraud and forgery: 2.23
  • Theft (non-motor): 1.35
  • Drug import/production: 3.62
  • Drug possession/supply: 2.19
  • Handling stolen goods: 2.27
  • Public order and harassment: 2.67
  • Shoplifting: 1.35

The last of these kills another feminist canard, namely that women are treated more harshly for shoplifting (though I have refuted it previously).

6. Race versus Sex

Finally I note that the MoJ analysis identifies that BAME individuals are more likely to be sent to prison than whites. Thus, BAME men are 50% more likely to be imprisoned than white men, whilst BAME women are 28% more likely to be imprisoned than white women. However, both these race effects are smaller than the gender effect. Overall you are 88% more likely to be imprisoned as a man than a woman. This confirms Mustard’s findings from USA data that gender sentencing disparity is greater than race disparity. Yet, whilst David Lammy was commissioned to author a report on the impact of race on imprisonment, Parliament would not dare do the same on the basis of sex – because the sex disparity is being driven by government policy. But if the race issue is indeed a social timebomb, then the sex issue must be a bigger timebomb. It is, but refusing to recognise its existence is the only stance which is politically acceptable.

7. Sex Offending

It has been noted before that sex offending, in particular, is a special case which is hardly to be represented by other categories of crime as regards gender differences. However, the difference in the treatment of sex offenders does not invalid the claim that men are treated more harshly than women. On the contrary, it is the most extreme example. Just how extreme is illustrated by this recent post. In round terms there are 12,000 men in prison for sex crimes but only 100 women. This ratio of sex prisoners in no way reflects the true incidence of sex offences.

8. Other Sources of Gender Disparity

The MoJ report considered only the relative likelihood of imprisonment. My analyses also considered prison sentence length and parole. However, neither analysis has considered the other three potential sources of gender disparity, namely,

  • Arrest
  • Prosecution
  • Conviction

It is likely that gender bias exists also at these stages in the criminal justice process – perhaps especially at the arrest stage (without which a potential offender does not go forward to any later stages). There is a valid concern here given that it has become increasingly common for police to be ‘trained’ by organisations which openly expound ideologies of extreme gender bias. Concerns that there could be institutional bias in decisions to prosecute are also hardly allayed by the ideological commitment of the DPP, Alison Saunders.

Consequently, whilst the gender bias evident in sentencing data alone is already large enough to provoke incredulity in some quarters, the truth is likely to be that existing attempts to quantify the bias are gross underestimates. With 21 times more men than women in prison there is plenty of scope for the bias to be considerably larger than my current best estimate of a disparity of x3.6, and perhaps my earlier estimate of x6 may ultimately turn out to be closer.

In any case, the 2016 MoJ report is to be welcomed as adding further support to the existence of considerable gender bias in imprisonment.

9. Conclusion

There is decreasing scope to doubt the fundamental correctness of claims of huge gender-bias in imprisonment.

People can throw accusations of misogyny at the likes of Philip Davies or myself, but future complaints about these claims need be directed to the Ministry of Justice.