HMRC: improving the operation of Pay As You Earn

Comments from ACCA to HM Revenue and Customs (HMRC), September 2010.

Executive summary

Even though much of the operation of PAYE now takes place on computerised systems, the underlying process is still based upon protocols developed for handling paper forms which are not necessarily appropriate for newer technology. Where a more efficient modern alternative exists and can be implemented it should be.

The introduction of the new system has presented opportunities to radically overhaul the operation of PAYE, freed from the constraints of a primarily paper-based system. The potential benefits to all taxpayers and administrators of real time working are considerable, and while there remain issues to resolve, the objective is worthwhile and should be pursued.

ACCA is concerned that the proposals for Centralised Deductions are a step too far in the current climate. Creating a viable technical model will be difficult. Making it work in the real world will be harder. Persuading taxpayers that either step can be achieved is likely to prove impossible. The risk not simply to HMRC's reputation but to the economy as a whole if the centralised system were to fail on payday is incalculable, as households' automated banking transactions, based on availability of salary funds, would fail on a hitherto unimagined scale. Until and unless HMRC can point to a comprehensive history of reliable, and reliably operated, IT systems that operate faultlessly and consistently for a number of years Centralised Deductions, should not be adopted.

General comments

The smooth and effective operation of the PAYE system, responsible for collecting around 40% of the Government's total income, is fundamental to the health of the UK economy. Currently, confidence in the system is low as a period of change has highlighted historical issues. The overwhelming majority of taxpayers do not appear to understand the distinction between assessment and collection of tax, or that they may need to have some input into the process to ensure that HMRC and employers have the correct information to set the system up.

While there may be a case for system change, there is clearly a need for education of both taxpayers and employers. The current high profile of tax matters presents an opportunity for HMRC and the accountancy profession to remind all those involved in the tax system of their responsibilities to help promote the smooth running of the system, to the benefit of all. We make a number of suggestions later in this paper for changes to the presentation and operation of PAYE, including an increased emphasis on engaging small businesses in the system. ACCA would be happy to work with HMRC and other bodies to ensure that these messages are effectively targeted and appropriately delivered as a lower cost, lower risk way of improving PAYE compliance than some of the current proposals.

We set out below our responses to some of the questions posed in the consultation document.


HMRC would welcome views on the concept of Real Time Information, whether it would support the collection of tax through the PAYE system and the issues that would need to be addressed in putting it into effect.

Collection of tax through the PAYE system is already extremely efficient. Collection of exactly the right amount of tax through PAYE is more problematic, reliant as that is upon interaction with the assessment process. Correct collection of tax is not directly linked to frequency of information, although there is likely to be a strong correlation between the two. Correct assessment of taxes due is more closely linked to frequency of information, although the effort required to provide the quantity and quality of information needed must be taken into account. Given the huge upheaval that would accompany any revision to the basics of the PAYE system, alternative approaches to improving collection of tax should be explored to assess whether short and medium-term aims and benefits can be achieved more easily.

Any move to Real Time Information will, as HMRC recognise, require wholesale reform of the software products used by employers and payroll providers. However, the requirements of PAYE should not be considered in isolation. PAYE is not the sole reason for running payroll processes, and for example the impending implementation of the auto-enrolment strategy for pensions will mean changes to payroll processing for many employers who do not currently operate widescale pension provision through payroll.

Operating in real time for personal allowances would be straightforward but the position is far more complex for current gift aid donations and savings interest for those near the higher rate band, property rentals and expenses, self-employment profits, dividends attracting higher rate tax, personal pension contributions, restriction of married couples allowance depending on total taxable income, and any other taxable income of variable amounts. Fundamental to addressing any of these issues is the education of taxpayers. It is not reasonable to expect any computerised system to operate effectively unless it is given the correct information on which to work. Garbage in, garbage out is as true now as when it was first applied to IBM mainframes. If the information provided under real time reporting is out of date or incomplete, the assessment of taxes which HMRC use to update codes will remain incorrect, no matter how regularly it is updated. Until UK taxpayers recognise and act on that basis, the PAYE system will continue to blamed for problems it was never designed to solve.

Currently, the main problem area requiring employer intervention is the year end P11d process. As there is no proposal to remove or revise this aspect of the PAYE tax system, employers and HMRC would still face the prospect of an annual spike in queries arising from this element of their tax assessment and collection activities.

When assessing the reduction in need for year end filings and report production, it must be remembered that even if the deduction of tax at source from employment income is streamlined, this will not settle all the affairs of any individual with non-employment income. For those completing a tax return they will still need to know how much taxable income they have received in the year, and how much tax they paid upon it. Employers will still need to provide annual returns of income, tax and benefits to their employees, and HMRC will still need to factor the results of those returns into the subsequent adjustments to codings under the Real Time Information system.

Annex A sets out the information that HMRC would need to be transmitted with each payment to an individual. HMRC would be interested in views as to how easy it would be to provide this information.

HMRC have highlighted 'hours worked' as an area of interest. The document does not specify the reason for this requirement, which is presumably related to monitoring NMW data (although it could also be of interest for Working Time Regulation compliance).

'Hours worked' is a veritable minefield of varying practices. Hours worked may not be the same as hours paid - overtime at enhanced rates, standby duties, split duty allowances, travelling time, piece work, bonuses etc can all affect the calculation. Where overtime hours are variable, it is common practice to pay them in the following week or month. Accordingly, HMRC will need to be clear on the purpose they will be putting the information to. If it is for NMW compliance then it will not be particularly relevant to those on higher salaries, such as highly skilled shift workers, whose circumstances might generate much of the complexity around variable overtime rates and the like. Imposing a requirement for strict compliance in these cases would create an unnecessary administrative burden, and if the NMW considerations can be addressed separately then this would be preferable.

In any event, if hours worked are to be reported then HMRC would need to provide clear and unambiguous guidance to employers on the interaction of different practices with the regulations.


HMRC would welcome views on the Centralised Deduction concept, the assumptions made about its benefits to all parties involved in the operation of PAYE and the issues it would raise and how they might be addressed.

ACCA is gravely concerned at the potential dangers of implementing the Centralised Deduction concept if the operation was anything less than flawless. Without wishing to prejudge the issue, it cannot be ignored that there have been no recent examples of the successful rollout of any nationwide government backed IT programmes. In fact, the opposite is the case with systems regularly failing to operate as expected, or even at all.

The teething problems with HMRC's own new National Insurance and PAYE Service (the "New Service") should serve as a more than adequate warning of the difficulties which can arise. Please note in this context that ACCA does not view the tax reconciliation exercise which has resulted in some six million taxpayers being notified of historic under or overpayments as being a fault with the computer system. To the extent that the year end reconciliation exercise has resulted in bad press for HMRC, this is simply a result of the management and publicity (or lack of it) surrounding the introduction of the New System over a period of years rather than days or weeks.

However, in early 2010, there were problems with the system itself. The primary impact on taxpayers of these problems with the New Service was the issuance of some 25m coding notices of which "a significant number" (according to the National Audit Office report of 20 July 2010) were incorrect. This caused considerable inconvenience and distress to many taxpayers, particularly among pensioners whose circumstances made them more susceptible to the errors. In direct financial terms the effect of applying incorrect tax codes has been to marginally reduce monthly income. While this is clearly not acceptable in the long term and must be reversed, the effect on the wider economy has been minimal and an argument could be made that these "teething problems" are part of the wider costs and must be borne for the greater good.

The potential impact under Centralised Deductions is far greater. The reaction if a significant proportion of the UK workforce were to have their monthly salary incorrectly calculated and paid is hard to imagine. The knock on effects of millions of workers being unable to service direct debit payments for mortgages and utilities are incalculable. Based on the implementation of previous government IT projects, the risk of such errors occurring is simply too great to justify the potential savings (considerable though they may be) that such a scheme could offer.

HMRC is interested in hearing proposals that could deliver a similar or better outcome through a different means to those outlined in this chapter.

The main concern about Centralised Deductions is not the concept of HMRC attempting to calculate tax deductions in real time. This would actually be a significant benefit for businesses, saving them from a major administrative burden. The concern is that by intercepting the salary payment instruction in real time and revising the level of funds remitted to the employee HMRC will place itself in a direct relationship with every employee in the country, and will be directly responsible for the amount of payment received by the employees.

This would inevitably lead to far higher levels of taxpayer contact with HMRC and the attendant resource and process requirements. In the current environment of restricted funding ACCA is concerned that HMRC would struggle to properly staff and equip the required teams. HMRC would also need to expand its 'back room' staff to operate the national payroll. Although the likely contraction of the private sector payroll industry may provide the skills base this will be scant compensation to those individuals who would face upheaval and relocation, or potentially even long term unemployment, as a result of the proposals.

An alternative method of operation would be for HMRC to receive monthly information, as already projected, and then (by the same secure means of communication) inform the employer of the precise amounts to be remitted first to the employee and second to HMRC. Existing payment structures would not need to be changed, and businesses would not be subject to the one off cash flow penalty of moving from the current payment in arrears model to a real time payment model.

HMRC would of course lose the benefit of paying money directly to itself, and would remain dependent upon tax paying employers to remit the deductions to HMRC. ACCA acknowledges that structural laws, which make it more difficult to achieve undesirable activity, can be a more effective way of dealing with systemic issues such as failure to pay over PAYE deductions than statutory commands, which simply impose a (separate) penalty for indulging in the undesirable activity.

In order to improve the behaviour of taxpayers without resorting to the expensive and risky option of centralised deductions, HMRC could consider the specific types of taxpayer who are most likely to default on their PAYE obligations and introduce measures tailored to improve their compliance. The introduction of security bonds for employers with a poor compliance history has already been proposed for the UK, and ACCA would in principle support such a mechanism, subject of course to appropriate safeguards.

Statistics suggest that the population which seems to experience the most difficulty complying with the system, and give rise to the highest levels of irrecoverable PAYE taxes, is small business. Research conducted in the US into similar issues with their comparable system of deductions (Fogg, Keith, In Whom We Trust (October 28, 2009). Creighton Law Review, Forthcoming; Villanova Law/Public Policy Research Paper No. 2010-06. Available at SSRN (please see the External links box at the bottom of this page) indicates that in many cases the reasons for failure to account properly for tax payments relate to the business owners' preferred hierarchy of payments. The tax authority, being relatively remote, ranks behind suppliers ('friends') and other money making opportunities which hold out the promise of a return on investment which would allow the business owner to clear their tax debts and retain some element of profit. In order to improve compliance, the author suggests a number of measures, of which those applicable to the UK situation would include education of tax payers, incentives to small business to pay tax on time, require the payment of bonds (already proposed) and transfer at least some part of the burden of unremitted PAYE directly onto the business owners.

These four measures do to some extent already exist in the UK, but all could be strengthened in order to improve compliance at a far lower cost, and without the attendant risk of mass failure of salary payments, than Centralised Deductions.

The process of setting up a PAYE scheme could be modified to include far more explicit education for employers as to their responsibilities, and the consequences of non-compliance, than is currently the case. Making a named individual responsible to HMRC as a contact point should increase their level of engagement, and also reduce the effort needed for HMRC to chase late payments.

The modelling of any incentive to pay tax on time would need to be carefully analysed to avoid too great an impact on the Treasury, and might in any event be considered undesirable. Taxpayers should not need any further encouragement to simply comply with their legal obligations; the existence of punishment for failure to comply should be adequate.

Transferring the burden of unremitted PAYE onto business owners is already possible in certain limited circumstances, but the mechanisms are complicated and the process opaque to the average business owner. In order to effectively motivate small business owners/officers, the legislation could impose upon the named responsible individual (see first idea above) a requirement to "comply or explain" with the PAYE remittance requirements. Failure to either a) pay over the tax or b) contact HMRC to explain to their satisfaction why payment was late would result in liability for that individual's income tax due under PAYE transferring directly to them. The individual would by definition have sufficient knowledge of HMRC and the relevant processes to understand the impact of such a transfer. The real time nature of the process would remove the current delays in HMRC noticing underpayments, so communication would be far swifter. Obviously the system would require safeguards around changing the responsible officer, and of course situations where the responsible individual had no tax liability due under PAYE (eg a "minimum fine of £X or the PAYE due").

HMRC would like to engage with all groups of people who could be affected by this concept to discuss the practical advantages and issues to be addressed.

Centralised Deductions would inevitably impose an administrative cost on all businesses as part of the change in systems. However, it is likely that a more significant cost for many businesses would be the cashflow impact of moving from the current model of PAYE deductions being remitted to HMRC in arrears to a "real time" payments model. Even the most basic analysis demonstrates that the cash movement from business to HMRC in the year of changeover would be measured in billions of pounds, probably in excess of £10bn and possibly over £20bn. However fragile the government's medium term finances may be, the private sector needs to be in a far stronger state than it is now to survive a cash raid of such magnitude.

As with Real Time Information, there will still be many cases where the system will not turn PAYE into an accurate means of assessing the individual's tax. The same list of variable payments will be able to cause discrepancies between the tax assessed/collected by Centralised Deductions.


HMRC would like to hear views on data, resilience and availability issues raised in chapter 6.

The success of a project of this type would depend upon the availability of clean data. ACCA remains to be convinced that HMRC could guarantee the quality of its databases in time to operate Centralised Deductions within a reasonable timeframe.