The economics of Universal Credit

Irregular UC and regular pay - A solution that also ends the 5 week wait?

I looked in some detail in a post last month - landlords-be-cautious-about-being-helpful/ - at the problems faced by people with regular earnings and regular outgoings, in particular rent, and how Universal Credit turns them into irregular income.

I expanded on that in my written evidence to the House of Lords Economic Affairs Committee enquiry into the economics of Universal Credit. That’s been published at: 2F837&slug=euc0012-written-evidence-gareth-morgan-ferret-information-systemsdocx .

On the basis of that, I was invited to give oral evidence to the inquiry on 17th March, with a particular brief to consider “How can this be resolved, does it mean changing the assessment period, and if so to what?”

A good question. I’ve talked previously about ways in which the irregularities might be smoothed, by averaging or adjusting the periods that were used; but always in very broad terms. This request spurred me to see if I could put some concrete ideas and numbers in place to demonstrate whether there might be a workable solution and how difficult it would be to implement.

So last weekend, I fired up Excel and set to work. Somewhat surprisingly, I think I came up with the basis for a workable solution. I set myself a couple of aims

I also took account of the ‘Johnson case‘[1] where the High Court found

On a proper interpretation of regulation 54 of the 2013 Regulations, read in context, the amount of the earned income of a claimant in respect of an assessment period is to be based on, but will not necessarily be the same as, the amount of earned income actually received in that assessment period. There will need to be an adjustment where, as in the present case,

the claimants actually received two months’ salary in one assessment period but the combined salaries do not, in fact, constitute earned income in respect of the period of time included in that assessment period. The defendant, therefore, erred in treating the combined salary for those two months’ as earned income in respect of that assessment period for the purposes of calculating the amount of universal credit payable.

Although the government have been given leave to appeal this decision, I also wanted to respect the

logic of the court’s findings, which meant that I excluded averaging over longer periods.

Unfortunately, the session was cancelled while I was en-route, but it has given me the chance to expand more on my thoughts.


A proposal

The existing data consists of that passed across to HMRC by the employer and HMRC’s calculation of tax and National Insurance (NI), which together provide the net income figure used by DWP. Other income and expenditure, such as pension contributions, fall outside this exercise.

The employer submits, using a payroll software package, a Full Payment Submission (FPS) to HMRC on or before the date an employee is paid. This provides details of the employee, their pay and deductions. Employers and pension providers have to send an FPS 'on or before' they make a payment to an employee.

The relevant fields, and field numbers, in the FPS are shown in table 1 below


A screenshot of a cell phone  Description automatically generated

Table 1


The employer submits gross taxable pay which HMRC passes across to DWP after deducting tax and NI. For my purposes, I have assumed that the pay is in respect of a period ending on the payment date. While there have been many concerns about employers incorrectly using the date in field 43, to record the date that they make their payment to the employee, the correct usage is to enter the date that pay is due. This removes problems caused by such issues as weekends and bank holidays. HMRC guidance is clear.

You must enter the usual date that you pay your employees, even if you pay them earlier or later. For example, if you pay your employees early because your usual payday falls on a Bank Holiday, you should still enter your regular payday.

Early reporting

You can send an FPS before your regular payday, for example if your payroll staff are going on holiday.[2]

These data items mean that I am able in my model to assess the dates for which pay is made, assuming that it ends on the date submitted. From that, I can determine the number of days falling within each assessment period (AP) for each pay period. That gives me the number of paid-days in total within each AP. I use a rather crude calculation for the day rate of:

A close up of a sign  Description automatically generated


Figure 1

This generates what I call the Day-Pay Rate (DPR), which is the daily net pay for that pay period. It’s important to stress that there may be different pay periods, varying in length or rate of pay, running within, or across, APs and thus each with potentially different DPR’s.

If HMRC collected the start and end dates of the period for which he has been made, it would enable a much more accurate assignment of earnings, for example, a common practice is to make a payment on the 25th of the month for a period covering from the 1st to the 30th. Many software products collect and use the dates, for assessing maternity pay dates for example.

The current system of assessing earnings within an AP generates irregularity from a regular pattern of earnings because of its inflexibility. It takes account only of the date paid. Because of that, within the calendar month, which may start on any day of the month, there can only be whole pay periods used within the assessment. Table 2 shows the rigid results that creates.




Pay period

Number of periods used


Four weeks or five weeks

Two weekly

Four weeks or six weeks

Four weekly

Four weeks or eight weeks


Zero, one or two months depending upon the

date of pay and date of UC claim

Table 2



Contrast that with table 3 which shows that it is possible, given existing data, to place the number of days within weekly pay periods into a UC AP.


A screenshot of a cell phone  Description automatically generated

Table 3


The matrix shows, in the columns, the days within each assessment period which are composed of days within pay periods and, in the rows for each pay period, how many DPR’s fall into each assessment period.

Table 4 shows the same for a monthly cycle.A screenshot of a computer  Description automatically generated


Table 4


This approach brings a number of advantages over the current system. Where people are regularly paid, and will therefore have consistent DPR, the assessment period will simply have that rate times the number of days in the period. Under the current system there can be anything from no days pay to 62 days pay used in the assessment. People are paid three monthly, six monthly or annually, as payroll systems permit, the situation can be even worse, as all the pay is compressed into one assessment period. Although employers have to calculate their own day rate sometimes, for example when people start or leave employment mid-month, this is not passed to HMRC.

Where pay or periods are not regular, the DPR approach deals much more satisfactorily with changes in dates of pay or in amounts of earnings (see responsiveness later). This approach works particularly well for couples with different pay cycles where the current system may produce exaggerated differences.

Figure 2, below, shows the current Universal Credit entitlement for the couple in the previous post. They both work on the 2019/20 National Minimum Wage level of £8.21 an hour. One works for 35 hours a week, paid monthly, and the other works for 10 hours a week paid weekly. They have two children and pay rent to a social landlord of £150 a week.

The full-time worker is paid £1116.31 a month net and the part-time worker earns £355.77 a month net. Their rent works out to £650 a month. On the basis of those figures, they would be entitled to

£907.70 Universal Credit every month.

That would leave them with £1,779.78 after paying their rent. Having consistent income and outgoings would make their budgeting and planning much easier.

Under the current rules, using all income received during the assessment period, their earnings are treated as being irregular, because the weekly paid worker will have either four or five pay days in each period and the monthly worker will have no paydays one payday or two paydays, because they are paid on the last working day of the month and the Universal Credit claim was also made at the end of the month. The very large difference in earnings used, month-to-month, creates a consequential large difference in Universal Credit.

A screenshot of a cell phone  Description automatically generated

















Figure 2


Using a day-rate calculation instead, smooths the entitlement almost entirely as shown in figure 3. The small variations are due to the varying number of days in the AP.

A screenshot of a cell phone  Description automatically generated
   Figure 3


For reasons best known to themselves, but probably the assumption that housing costs are more predictable, Universal Credit takes rent into account by assessing an annual total and dividing by 12. For people with weekly rent cycles this can mean that they will have either four or five rent days in each AP and the Universal Credit assessment will use a figure slightly more than four weekly rent amounts.

Figure 4 shows the current and day rate income, from earnings and Universal Credit, that is left after meeting the rent payments in each AP.

A screenshot of a cell phone  Description automatically generated

Figure 4



It is clear that, even after the variations caused by the number of rent days, the resultant disposable income is much more even, easily understood and amenable to budgeting and planning.

Irregular earnings and periods

The day rate approach is not limited to regular periods or regular earnings. It works very well as a method of assessing periods of different lengths and of different, or no, earnings. The example in table 5 shows some output from my model with these variations.

Table 5


Different frequencies of payment are shown in the table, ranging from weekly to 3 monthly and including two weekly, calendar monthly and last working day payments. Each of these has a separately calculated day rate and the number of days in each assessment period, at the relevant day rate, is used to determine total net pay within the AP.

It is likely that there could be a smoothing effect from this approach, but varying earnings within a means tested benefit must be reflected in the amount payable.



The glaring problem of this approach is, of course, that the income, taken into account in an assessment period, may not actually have been paid, or even earned, yet.

That means that an assessment may need to use an estimated day rate for some of the assessment period.

That is unlikely to be difficult in practice where, as in most of these examples, pay rates and periods are regular but a move to a day-pay assessment would need to be applicable across the wide range of earnings and pay periods that can exist.

That will inevitably mean some form of reconciliation, where the estimated day rates differ from the actual amounts paid later. The estimated rate could be derived from a period of past earnings that was long enough to reflect patterns of employment, such as seasonality. In new employment, or changed pay or hours, the expected pay could be used. Even the worker’s expectations, particularly, for self employment could be used, as it is, effectively, in tax credits. Reconciliation could also take place over a more extended period.

This is the approach that has been used with Tax Credits for many years. Implementing it within the monthly AP can be expected to make the assessment more responsive to change. To an extent, this already happens, for example, employers use the previous 52 weeks’ pay to determine the value of a week’s holiday pay.

The issue is more difficult with self-employed earnings where it can be very difficult to attribute earnings to a particular period. In Universal Credit it’s also necessary to decide on a tax and NI figure that leaves net earnings in the assessment, unlike tax credits where gross earnings are used.

Universal Credit’s ‘all-income’ vs ‘all expenses’ approach also means that expenditure has to be assigned to an AP as well. A longer period of estimation and assessment may be the only practical approach.



Universal Credit, although intended originally to quickly reflect changes in earnings and circumstance, is poor at responding to such changes. It depends upon a single day’s snapshot of circumstances during the month, as well as its crude capture of all payments within that period.

Its harsh rigidity takes little account of possible patterns of life and work outside its monthly ideal. It is clear that attempting to impose this ideal on people, especially the self-employed, with irregular employment or earnings can be even more baffling than the effect it can have on those in regular circumstances.

Recognising and responding to such variances quickly will be particularly important over the next months as the coronavirus impacts continue.

One very important advantage that would come if the principle of future estimation, and subsequent reconciliation were accepted would be that the five-week wait would no longer be required. Estimations of earnings and income, including Statutory Sick Pay, New Style ESA and other support could be used as the basis for immediate assessment and rapid payment of Universal Credit.



Universal Credit is struggling. It struggles to meet the needs of those it supports who, rightly, are often baffled by the amounts they receive, even after they have succeeded in claiming and waited for their payment. The DWP is struggling with administration and systems. Those who provide support and advice to claimants are struggling with insufficient resources and poor access to officials.

The current crisis can only make things much worse, and support system will inevitably have to change to meet the needs of the time. My suggestion undoubtedly has flaws and difficulties but I do not think that they will be unsurmountable or difficult to solve. I believe that it is possible that this change to assessment methods could be made quickly, and relatively easily.

Many thanks to those who commented on the earlier draft of this, especially Ian Holloway of Cintra who generously provides payroll expertise.


21 March 2020