Beware the ideas of IDS


I am a proponent of Negative Income Tax (NIT). I am because it can deliver social security via a basic income, because it is simple and cheap to administer, and because it is libertarian.

In this piece, I try to set out why implementing NIT in Britain does not require any radical surgery to the existing income tax system. I also point out that the disincentive effects of NIT are comparatively low. I suggest that the sanctions regime currently in vogue would be inappropriate under a fully developed NIT. I also briefly explain why the cost of implementing NIT, at least at first, is likely to be close to zero.

What is NIT?

It’s a very simple idea. For those individuals whose incomes fall below the personal allowance, the tax authority, instead of charging income tax, makes a reverse payment to top up the individual’s income. One of its original architects, Milton Friedman, thought the payment should be universal,  unconditional and be made without moral judgement. His idea was to replace complex and judgemental social security systems. The logic of income tax automatically and progressively claws back negative income tax payments when an individual’s income exceeds their personal allowance.

How would IDS and his followers view NIT?

Despite the libertarian origins of NIT, the IDS school of thought would seek to bring NIT, were it to be implemented in Britain, within the currently fashionable sanctions regime.  IDS  has already expressed his intention to bring in-work benefits within the sanctions regime. It is hence reasonable to suppose that IDS would wish a NIT system to be conditional and to include it in his beloved sanctions regime too. It is pretty clear from the high level of  coercion embodied in Universal Credit that IDS does not share the libertarian values of NIT.

How would a negative income tax look in Britain?

Surprisingly perhaps, it would look similar to the current income tax system in Britain – most people would notice no change to their income tax charge and to their net pay. Income tax, whether through deductions at source or through end of year self-assessment would operate as now,  without change to method or amount. Only individuals with incomes below the personal allowance would notice any difference.

Only a comparatively simple tweak to HMRC’s software would be needed to implement NIT. NIT would be simple, cheap and low risk to implement. Because radical surgery to the existing system would not be needed, a government should not find practical objections to NIT implementation.

So how would NIT differ from the current income tax system?

At the moment, the first £10,600 of an individual’s annual income is tax free – individuals pay income tax only on their annual income in excess of £10,600. The starting rate of tax for incomes above the £10,600 threshold is 20%. For the purpose of this explanation it is not necessary to consider the higher income tax rates (40% and 45%) as these are unaffected by NIT. In fact, NIT will make no difference to the income tax charge for anyone whose income exceeds £10,600, whether their marginal income tax rate is 20%, 40%, or 45%. Only those whose income is below £10,600 will be affected by NIT.


Freda’s annual income is £8,000.

Her tax charge under the current system is 20% x (£8,000 – £10,600) = – £520, which, because it is negative, means HMRC will not seek to collect income tax from Freda. So Freda pays no income tax and her post-tax income is £8,000.

In contrast, under NIT the negative result of -£520 produced by her tax computation  would trigger a payment to Freda of £520 from HMRC. Her post tax income would hence be £8,520 (£8,000 + £520). In effect, Freda’s wage is being topped up by the state through the tax system.

NB. 20% x (£8,000 – £10,600) can be written as 20% x £8,000 – 20% x £10,600

So why not stick to tax credits currently operated by HMRC?

The NIT calculation is very simple and should logically be preferred to the complex strictures and administrative expense of the tax credit schemes currently operated by HMRC. A further point is that IDS’s intention is to bring in-work tax credits into the benefits sanction regime via Universal Credit.  This will add further complication to an already over-complicated system.. An NIT could ease the burden on Universal Credit, which is already creaking under the weight of its own complexity. The implementation proposed here is modest and cautious. Over time, increasing confidence would hopefully impel a government to implement NIT comprehensively, perhaps with a view of complete replacement of other social security benefits.

Would a negative income tax have disincentive effects?

IDS et al hold that low income earners are discouraged from increasing their earned incomes because they are in receipt of tax credits and other means tested benefits. Increased use of sanctions (the withdrawal of benefits) has been IDS’s chosen method of countering these putative disincentive effects.  NIT would be perceived by IDS to have disincentive effects given the evidence. IDS et al would thus argue that receipt of NIT should be conditional and subject to sanctions should a recipient be perceived as “free-riding” on the revered tax-payer. IDS is not a libertarian.

The withdrawal rate, (the amount by which the state top-up is reduced for each additional £ of income) is  equal to the 20% starting rate of tax. This is far lower than the 65% withdrawal rate aimed for by Universal Credit. The NIT implementation proposed here thus compares favourably to IDS’s Universal Credit.. This suggests that sanctions should not be necessary to “motivate” NIT recipients to increase their incomes

How much would NIT cost?

Confining NIT (at least to start with) to recipients of market incomes and to pensioners then the answer is not much. This is because their benefits and tax credits would be reduced by the amount equal to the NIT payments received by them. NIT would be payable only where market incomes and pensions fall below £10,600, so many people would not be NIT recipients, The total social security bill and tax take would be identical as between the current arrangements and NIT.

So why implement NIT then?

A modest and cautious implementation has been recommended here so that the costs and risks of failure are low. Over time, more benefits of NIT can be captured. These benefits include

  1. reduced administration costs;
  2. better incentives;
  3. increased flexibility;
  4. a basic income;
  5. simplicity;
  6. reduced benefit costs;
  7. a more constructive role for DWP;  and
  8. less state coercion

Using the tax system to deliver a living wage

In a previous post, I proposed a formula for calculating income tax that would provide for a  negative income tax. I did so because I believe a negative income tax system would be simpler, fairer, more humane, technically more feasible , and cheaper to develop than the current Universal Credit project being rolled out by Her Majesty’s government. A negative income tax would be administered by HMRC and have at its core the following formula for calculating income tax:

T(n) = m∑Y(n) – c∑H(n) – ∑T(n-1)


m is the single marginal tax rate;

c is the hourly value of the tax shield

∑Y(n) is the cumulative market income received by the end of period n

∑H(n) is the cumulative hours worked by the end of period n

∑T(n-1) is the cumulative income tax paid by the start of period n.


The proposed tax system has a single marginal tax rate. The choice of this rate is a political as well as a technical decision (i.e., achieving the desired behavioural effects) An individual’s hourly tax-free pay is given by the grossed-up tax shield (c / m).

In this post, I want to show how a negative income tax system can deliver what many describe as a “living wage”. The figure of £10 / hour has been mooted as a living wage. I shall hence assume this figure for the living wage  in my illustration.


In this post, I will incorporate the UK’s national minimum wage (NMW), currently £6.50 per hour, into the model.

I will also use a marginal tax rate of 40%, a tax rate which is both comparable to and competitive with existing UK income tax rates (20%, 40%, and 45%). Hence m has been set to 40% or 0.4 in my illustrations.

I will assume a working week of 35 hours and an individual’s weekly hours will be capped at this level.

I will also show how the model can be used to generate a basic income. I will briefly set out my views on whether a basic income should be conditional or unconditional.

Calculating the tax shield

In this section, I will show how negative income tax can transform the minimum wage of £6.50 / hour into a living wage of £10.00 / hour.  To achieve this, the value of the tax shield (c) must be calculated with reference to the minimum wage and to the living wage of £10.00. The value of the tax shield (c) must be set so that the hourly negative tax payment makes up the difference of £3.50 

Ignoring the summations in the above f ormula (they are not necessary here), we have

T = m x (£6.50) – c = – £3.50

T= 0.4 x £6.50 – c = – £3.50

c = £6.10

The tax formula that transforms the hourly national minimum wage of £6.50 to an hourly living wage of £10 will thus be

T(n) = 0.4 x ∑Y(n) – £6.10 x ∑H(n) – ∑T(n-1)

or on a week 1 basis, the hourly tax charge will be:

T = 0.4 x £6.50 – £6.10

= – £3.50 (negative,  as required)


Lenny earns the adult minimum wage. He works 35 hours per week. His pay slip will show:

Market income (35 hours x £6.50/hour)                                       £227.50

Negative income tax  (35 hours x £3.50)                                     £122.50

Take home pay (35 hours x £10 hour)                                         £350.00


Lenny now receives £10 per hour.


The hourly post-tax income of someone whose market income is £10 would be transformed to £12.10.

T= 0.4 x (£10.00) – £6.10 = – £2.10 (negative)

Hence that individual would take home £10 – (- £2.10) = £12.10 for every hour worked at or below the 35 hour weekly cap on hours. So although the “differentials” between market incomes have narrowed the higher paid worker still receives more per hour than the minimum wage worker.

Tax free allowance

The break-even hourly income with an hourly tax shield of £6.10 and a marginal tax rate of 40% is £15.25. This means a worker with an hourly income of £15.25 will neither pay income tax nor receive a tax payment. This is equivalent to an annual tax free allowance of £27,755 (compared to the current allowance of £10,000 p.a.). Individuals with a a market income of above £15.25 per hour or £27,755 per annum would pay income tax of 40% on the excess.

Basic income

The parameters used in the formula produce an amount 0f £6.10 per hour for those who have no market income. This could be taken to be a basic income payable to anyone whose market income is zero.

T = 0.4 x £0.00 / hour – £6.10 / hour

= – £6.10 per hour (negative)

The issue is whether this fortunate by-product of negative income tax should be conditional or unconditional.

Unconditional basic income

An unconditional basic income would credit every citizen with 35 hours. The basic income would thus be received by every citizen of working age.  The basic income would amount to £213.50 (35 hours x £6.10 per hour).

Hourly income with unconditional basic income and living wage of £10/hour

Hourly income with unconditional basic income and living wage of £10/hour

Conditional basic income

A conditional basic income would credit some groups with 35 hours a week but not others. For example, the disabled, the infirm, and those with caring responsibilities could be credited.  This would automatically trigger a payment at HMRC of £213.50 to individuals falling in these groups. The unemployed would be credited with hours only to the extent of their hourly contributions to society via voluntary work and / or participation in counter hysteresis activities (such as internships, work placements, or work programmes).

Hourly income with a conditional basic income and a living wage of £10/hour

Hourly income with a conditional basic income and a living wage of £10/hour

Conditional or unconditional basic income?

My main concern is that every UK citizen should have access to honest and adequate income. For disadvantaged groups, income accessability, particularly with increased benefit conditionality, can often be an issue.

Provided plentiful opportunities for voluntary work and / or opportunities to engage in counter hysteresis activities are made available to physically and mentally capable individuals without a market income then basic income should be conditional along the lines described above. A reasonably generous income of £213.50 per week would accrue to such individuals if their engagement in these activities was for 35 hours a week. Lesser contributions would be rewarded proportionately.

Summary of the model

Basic income = £6.10 /hour. May be conditional or unconditional. Could replace unemployment benefit.

Minimum wage (statutory)  = £6.50 / hour.

Living wage (assumed) = £10.00 /hour.

Break-even wage = £15.25 /hour.

Marginal tax rate = 40%

Maximum hours of credit = 35

Some final points

The withdrawal rate for someone transiting from unemployment to minimum wage work is about 43% (0.4 x £6.50 / £6.10). This is substantially better than the proposed withdrawal rate of 65% planned for Universal Credit.

However, the withdrawal rate rises as the starting hourly wage rate rises. At a starting rate of £15.25 per hour the withdrawal rate is 100%. It is unlikely that such a high withdrawal rate would be a disincentive to accept work at this wage rate.

It seems likely that a negative income tax designed with the parameters outlined in this post would lead to a lower overall income tax take for the government. Some of this might  be due to a lower marginal tax rate of 40% for top earners compared to their current 45% and because the tax free allowance is more generous than the current £10,000 annual allowance. For this reason, it may be better to use income tax as a means of correcting the inequalities that arise from distortions in the labour market (eg directors effectively setting their own pay) rather than as a means of raising revenue.

Capping the corporation tax deductibility of market incomes to £27,779 per employee (the annual tax free allowance) may also help to mitigate reductions in the income tax take.

The redistribution of income arising from this variant of a negative income tax system is likely to have a stimulative effect on the economy. Poorer workers with a high propensity to consume relative to richer workers should increase demand and hence economic activity.

Self-employment and negative income tax


Previously, in “Is Universal Credit the wrong approach?”, I proposed that a thoughtfully designed negative income tax system would be more effective in delivering UC’s stated aims. I believe UC is fundamentally flawed, primarily because it does not address the UK’s low and falling wages. The feasibility of designing incentives to make work pay into UC, in an environment with low and falling wages, may be logically flawed. This doubt exists in addition to the current uncertainty about UC’s technical feasibility. In this post, I set out how negative income tax (NIT), the proposed alternative to UC, would apply to the self-employed.

Recap of the negative income tax proposal

1. Income tax would be assessed on Comprehensive Realised Income (CRI).

2. An individual’s tax-free allowance would be calculated according to the number of hours worked. The value of each hour worked would be £20 and be capped at 40 hours per week (2,080 per year).

3. A single rate of tax would be applied to comprehensive realised income, By way of example, and for philosophical reasons, the single tax rate has been set at 50%,

4. Carers, the disabled, and the infirm would be credited with 40 hours per week by HMRC. This would release  a payment of £800 per week (40 hours x £20/hour) to individuals falling in these categories. Any earnings would not disqualify recipients from this minimum income guarantee. However, earnings would be taxed at the marginal rate of 50%.

5.  Job seekers would be paid according to the hours they work part-time, voluntarily in the charitable sector, and in work placements. These hours would be notified to HMRC by the placement providers using HMRC’s Real Time Information reporting system. Payments would be accordingly triggered at HMRC. The placements would not be compulsory – job seekers would have choice as to the  number of hours and type of work they took on. The role of the DWP would be to find and arrange work placements in cases where the job seeker has been unable to do it for themselves.

6. The self-employed (sole-traders and partners), would be allowed a tax-free allowance equivalent to deemed hours of work. The imputed hours would be capped at 40 hours per week (or 2,080 per year), as for all other cases. The hours worked would be imputed by dividing the tax-adjusted earned profits by the standard value of an hour (£20) and multiplying the result by the marginal tax rate. This arrangement would be necessary since no third party could attest to the number of hours worked.

H(n) = m x Y(n)  ⁄ c


H(n) are the imputed hours for the period (capped at 52 x 40 = 2,080)

Y(n)  is the tax adjusted profit in period n

m is the marginal tax rate (set to 50% in the examples)

c is the standard value of an hour worked (set to £20 in the examples)

Example 1

Stephen enters self-employment at the start of the tax year. His (tax-adjusted) profits for the year come to £10,000. He has no other sources of income.

His imputed hours for the year would be

H(n) = £10,000 x 50% / £20 = 250

Hence Stephen’s tax charge for the year would be 

T(n) = 0.5 x £10,000 – 250 x £20 = £0

Stephen’s disposable income is hence £10,000 because his tax charge is zero. He has used up 250 hours of his annual allowance of 2,080 hours. He could have used the unused allowance to do part-time work and/or voluntary work. His income would have been supplemented had he done this.

Had Stephen also sold his house at a capital gain of £40,000, his income tax computation would be:

 T(n) = 0.5 x £50,000 – 250 x £20 = £20,000

Hence his post-tax comprehensive realised income would be £30,000. He is charged no tax on earned income but is charged 50% on his capital gain (an unearned component of comprehensive realised income)


Example 2

Anita enters self-employment at the start of the tax year. Her (tax-adjusted) profit for the year comes to £100,000. She has no other sources of income. Her imputed hours of work in year n are calculated as thus.

 H(n) = £100,000 x 50% / £20 = 2,500

The hours must be capped to 52 weeks x 40 hours per week = 2,080.

Her tax for year n would be:

T(n) = 0.5 x £100,000 – 2,080 x £20 = £8,400

Anita’s disposable income (take home pay) is hence £91,600. Her tax rate is 8.4% on her earned income. 

Had Anita also received rental income (i.e., unearned income) of £10,000 in the same year then her income tax computation would be:

T(n) = 0.5 x £110,000 – 2,080 x £20 = £13,400

Note that the unearned components of comprehensive realised income have again been taxed at a straight 50%. This feature, whereby an individual’s tax-free allowance is calculated on hours worked, assists in aligning the tax system with the “making work pay” agenda that underpins the government’s welfare reforms and UC.

NB. The parameters used in these examples, i.e., a working week capped at 40 hours, a marginal tax rate of 50%. and a standard hour set to £20, produces a tax-free allowance of £83,600 on earned income. This produces low average tax rates on middle earnings and is very generous by current standards. It is intended to incentivise. Of course, the average rate will approach the marginal rate of 50% as incomes rise.

Is Universal Credit the wrong approach?

“Work doesn’t pay” is the oft cited justification for Universal Credit (UC). It’s not clear that rolling up six separate benefits into one single payment will address this  problem, aka the unemployment trap. Wages in the UK are at best static and may have been falling in real terms over the last decade. Coupled with declining real wages has been the almost complete elimination of affordable council housing. Wages nowadays are insufficient to meet private housing rental costs which have soared due to shortages. So yes, it may pay claimants to remain on benefits, particularly Housing Benefit.

An alternative to UC is to integrate the tax system with benefits. Low earners, instead of paying income tax, would instead receive a generous negative tax payment. The incomes of low earners would be enhanced to the extent that they would no longer need benefits just to survive; in short, integrating tax and benefits could give low earners independence. But how can this be achieved? One framework to achieve this is a system of cumulative hourly averaging combined with comprehensive realised income (CRI) subject to a single rate of tax

Cumulative Hourly Averaging – the advantages

  • It rewards work – those who work will pay less tax on their comprehensive realised incomes

  • It can integrate benefits with tax – a generous negative income tax is facilitated

  • It supports flexible labour markets – the incomes of zero hours contract workers will be enhanced and less variable. There should be no need of a minimum wage.

  • It restores or strengthens the contributory principle.

Comprehensive Realised Income – the advantages

  • It addresses inequalities in wealth distribution – wealth transfers and other windfall gains are taxed at the same rate as earned income

  • It overcomes equitable objections to Inheritance Tax – legacies would be taxed according to the recipient’s circumstances and at the same rate as earned income

  • It reduces incentives to avoid tax – a single rate of tax can be applied to all sources of income

Cumulative Hourly Averaging – the Generalised Model

 The tax charge in period n would be calculated with the following formula:

T(n) = m∑Y(n) – c∑H(n) – ∑T(n-1)


T(n) = tax charge in period n

∑Y(n) = Cumulative comprehensive realised income received by the end of period n

∑H(n) =Cumulative hours worked by the end of period n

∑T(n-1) = Cumulative tax paid at the start of period n, and so ∑T(n) = T(n) + ∑T(n-1)

m = the marginal tax rate. I have used a single marginal rate of 50% applied to all components of CRI  without preference

c = the value of an hour of work. I have set this to £20 so as to produce a generous negative tax component which encourages work

 The Mechanics

Before illustrating how the tax calculation works, I have set out some parameters which I have used in the examples below. These are:

  • The standard working week has been set to 40 hours per week

  • The standard wage rate (or the value of an hour of work) is £20. Note this is an administrative value and has nothing to do with a minimum wage.

  • There is no minimum wage

  • A flat rate income tax of 50% applies to all components of comprehensive realised income without preferment. This is a limiting rate, meaning no one in work will pay tax at this rate however high their income.

  • Tax free personal allowances depend on cumulative hours worked and are valued at £20 per hour worked.

  • Tax free personal allowances are carried over to succeeding years, unlike the current “use or lose” system

  • No individual’s personal allowance can exceed 48 hours in a week

  • HMRC operates RTI (Real Time Information) so that incomes and changes in circumstances are reported as and when they occur.


Example 1

Freda starts work for 30 hours per week at £7 per hour (£210 per week). She has no other sources of income.

T(1) = = 0.5 x £210 – £20 x 30 = – £495

Freda will pay no income tax and instead will receive £495 under the negative tax mechanism. Her disposable income will thus be £210 + £495 = £705.  Not bad for a week’s work !

In the second week, Freda’s tax calculation is as follows:

T(2) = 0.5 x £420 – £20 x 60 – (-£495) = -£495

So again, Freda’s disposable income (take home pay) will be £210 +£495 = £705. Enjoy yourself, Freda. you deserve it! And so it will continue until Freda’s circumstances change.

Example 2

George starts work as a director of a large company. His monthly remuneration package comes to £60,000. His monthly hours of work are restricted to 208. He has no other sources of income

T(1) = 0.5 x £60,000 – £20 x 208 = £25,840

So George’s disposable income will be £60,000 – £25,840 = £34,160.

T(2) = 0.5 x £120,000 – £20 x 416 – £25,840 = £25,840

So long as George’s circumstances don’t change his monthly disposable income will remain at £34,160.


Example 3

Mark Anthony is a notorious playboy who has never done a day’s work in his life. His very rich father bequeaths Mark Anthony £500,000 in his  will. Mark’s tax liability will be:

T(1) = 0.5 x £500,000 – £20 x 0 = £250,000

In the second  period, Mark Anthony decides to do some voluntary work. He registers 30 hours with HMRC.

T(2) = 0.5 x £500,000 – £20 x 30 – £250,000 = – £600

Mark Anthony is rewarded for his voluntary work to the tune of £600 via the negative income tax mechanism. Work pays! Even unpaid work.


Example 4

Lois has a severe disability that limits the amount of work she can do in a week to 10 hours. She is paid £3 per hour (there is no minimum wage). HMRC credits Lois with 30 additional hours per week to compensate her for the hours she is unable to work through no fault of her own. She has no other sources of income. In week 1 her tax will be calculated thus:

T(1) = 0.5 x £30 – £20 x 40 = – 785

So Lois’s disposable income in week 1 will be £30 + £785 = £815. This is an example of how institutions can be used to compensate disadvantage.

In week 2, Lois receives a pay rise of £2 per hour to £5 per hour. Her tax calculation in week 2 will be

T(2) = 0.5 x £80 – £20 x 80 – (-785) = -£775

Lois’s disposable income in week 2 has risen to £50 + £775 = £825.  A pay rise of £20 per week has resulted in Lois’s disposable income increasing by £10 and the state subsidy falling by an equivalent amount.

Example 5

Tony is a homeless alcoholic, without work or income. He is offered 40 hourly work credits per week which will trigger weekly payments of £800 via the negative income tax mechanism provided he undergoes treatment for his alcoholism at a residential clinic. In Tony’s case, the  weekly payments are paid directly to the clinic instead of to Tony, Tony will need to price himself into employment when his treatment is completed, a task made easier absent a minimum wage.

Information requirements

HMRC will need to keep an up-to-date record of every taxpayer/claimant and changes in their circumstances. This is not as onerous or as intrusive as might at first appear; RTI, which requires employers to submit details of employee  hours and pay in “real time”,  has already been introduced. The road has already been dug.

 Here is a logical record of the information and processing that HMRC would need to collect and process for an employee. The particular employee shown in the record has had a particularly turbulent time, starting as a highly paid director, becoming unemployed, suffering disability, inheriting a sizeable estate, etc, all in six weeks!. The record is not intended to be of a typical employee but is illustrative of how income tax would work under cumulative hourly averaging with CRI .

Table of events


One thing writing this post has taught me is how difficult it is to design a safety net which both protects and incentivises.  A safety net  which is too generous removes incentives to return to, or to get into, work. A safety net built around stick and no carrot is cruel and damaging to individuals. I fear that Universal Credit, with its vicious sanctions regime and its heartless treatment of the sick and disabled, falls into this latter category. I venture to suggest that the negative income system outlined above, albeit with its fault lines, would be more effective than the proposed UC project should the latter ever go live. The system outlined above is certainly kinder than UC. Having said this, it may be that the proponents of Unconditional Basic Income win the day – UBI sidesteps the tension between incentive and protection. Perhaps this is the way to go.


It would be remiss of me not to acknowledge the brilliant work of Douglas Bamford in the field of taxation and philosophy. Douglas very kindly gave me sight of his then forthcoming book in advance of its publication. His ideas on cumulative hourly averaging have very obviously informed this piece, as has his idea of using comprehensive income as a tax base. His book is entitled Rethinking Taxation – An Introduction to Hourly Averaging. ISBN 9781907720918.

Any errors or sub-standard work contained in this piece are mine, and mine alone.

Addendum 2 September 2014

1. Quite rightly, it has been pointed out to me that the negative tax proposal outlined above does not say much about unemployment support. To answer this, I suggest work placements should be available for all jobless people, which they can choose to take up. There should be no compulsion as to participation or as to the type of placement. If a jobless person can arrange a placement of their choice, say in a museum, then so be it. The only requirement would be the readiness and agreement  for the placement provider to submit the hours worked to the HMRC as registered hours. The registration of hours worked each week under RTI reporting would then trigger a payment via the negative income tax mechanism to the worker in the same way as for other employees.

2. People who have caring responsibilities, either for children or for aged parents, should receive hourly credits equal to the standard working week  (40 hours according to the parameters used in my examples).

3. Profits on the sale of houses, even if a house in question is a Principal Private Residence (PPR), should be brought into the Taxable Comprehensive Realised Income calculation. Currently, the gain on sale of a PPR is exempt from taxation.

Are benefit sanctions being applied consistently, as claimed by the DWP? Statistical evidence suggests not


In the UK, unemployment benefit is known as Job Seekers Allowance, or JSA.  Claimants for this benefit must demonstrate to the government’s Department of Work & Pensions (DWP) that they are actively seeking work. Failure to demonstrate an active job search may result in withdrawal of JSA by way of penalty. The penalty, which may last for three years, is known as a sanction. The ostensible purpose of a sanction is to penalise a claimant who is not actively seeking work.

Are sanctions being applied unlawfully?

In the last two years, the frequency with which sanctions have been used has increased noticeably. Some commentators suggest the increasing use of sanctions has not been justified. Anecdotal evidence of sanctions being applied for trivial reasons, or without sufficient or just cause, support the case that they are being misused. Further, there is now speculation that DWP is working to targets and that sanctions are being applied to meet these targets. In short, the sanctions, it is suggested, are being used to meet purposes separate from those set out in legislation and approved by Parliament. If true, this would mean, of course, that sanctions are being applied unlawfully by DWP. An independent review of DWP’s use of sanctions, we have been told, is now to take place to investigate these matters.

Some preliminary thoughts

In advance of the review, I have undertaken a quick statistical analysis using sanctions data compiled by DWP for the 18 months April 2011 to October 2012. My analysis is limited to discovering whether sanctions are being applied consistently from region to region in England. The regions I selected for the study are the six Metropolitan Counties together with Inner and Outer London. Hence eight regions form the sample in my study. My conclusion, although based on imperfect data, is that a rough but statistically significant north-south gradient exists, with a lower frequency of sanctions in the northern regions.


The variation in sanctions frequency between region is not consistent with the hypothesis that claimant behaviour is independent of region. So what is causing this variation? If sanctions targets were being set at regional level without the centre’s knowledge, then we might expect to find statistically significant variation in the regional sanctions frequencies. This is indeed what this brief study finds – the regional variation uncovered by this study is statistically significant at the 0.05% level. So the results of this study are consistent with the hypothesis of sanctions policies and targets being determined regionally, at least in some of the regions.


It is hoped that the promised government review will use a quantitative approach and will have access to better data than did this study. A quantitative and well resourced  study, perhaps using a similar approach as used by this short study, may produce more definitive conclusions.


1.  Sanctions data  for period 1/4/11 to 21/10/12 sourced from DWP Information, Governance & Security Directorate, via @AnitaBellows12 to whom a hat-tip is hereby given

2. Claimant Count  for January 2013 sourced from ONS website


A Chi Squared test was used to test for independence between region and sanction count. The 6 Metropolitan Counties were selected for the sample on the ground of speed and homogeneity. Inner and Outer London were added in to make the number of regions up to 8.

The” actual” sanctions count for the month of January 2013 was derived from from the sanctions count  for the 18 months April 2011 to October 2012 compiled by DWP. No sanctions figures could be found for January 2013. To arrive at a sanctions count for January 2013 the 18 month sanctions count was divided by 18 for each region.  There is no pretence that the resulting January sanction figure is precise, but the conclusion is unlikely to be different because of this. The purpose of this study is merely to discover, even if tentatively, whether there is a regional effect on sanction frequencies. The data, even though imperfect, are adequate for this limited purpose.

The actual claimant count came from the January 2013 ONS release. This  month was chosen because the claimant count was supplied in a way geographically consistent with the sanction count (ie the figures for both series were analysed into the same regional definitions). The claimant count was used to calculate the expected sanctions count for a region.

The calculated test statistic came out at 78. The probability of obtaining a test statistic of this magnitude with 7 degrees of  freedom under the hypothesis of independence between region and sanctions frequency is virtually zero.

Sanctions 3