The tax return pre-population data feeds, allowing agents to access PAYE data directly from HMRC’s system, will soon celebrate their first birthday. Matt Bailey examines how they performed during their first tax season, and looks at what improvements are possible in the future.
The general consensus seems to be that the pre-population API feeds provide useful information. Accountants can access employment data directly from HMRC’s system, and can use these figures to check pay and tax records received from clients when completing a tax return. Data is also available for private pension income, certain state benefits, class 1 and class 2 national insurance, and basic details concerning marriage allowance.
How did the system perform during its first tax season? And was the data both readily available and accurate?
Use of the system
HMRC say that between April 2017 and mid-February 2018, they responded to 16.4 million pre-population (or pre-pop) requests. They provided data for nearly a million taxpayers, and approximately 10,000 accountants used the service.
There were some issues with capacity constraints in the peak period. This meant some users received error messages when the servers could not cope with the number of requests, an issue HMRC say they are tackling for future years.
The capacity problems may, at least in part, have been caused by poor software design, with HMRC saying some systems were making multiple requests for the same data. Their figures suggest there were an average of 16 API requests per taxpayer, a high rate bearing in mind the data very rarely changes.
Has sufficient data been available?
Most PAYE data seems to have been available via the pre-pop system, but it took some time for it to become visible.
These delays are understandable bearing in mind the way the system works. HMRC receives a huge amount of PAYE information from employers, and this needs to be processed and checked and any over- or under-payments corrected. And while basic pay and tax data is reported to HMRC in real time, P11D data only needs to be reported by 6th July. So many checks are postponed until this data is received.
Once the figures are checked and verified, they are copied over to the separate MTD system and made available to the pre-pop service. It may be September (or even later) before this happens for some taxpayers.
As well as time delays, HMRC also deletes taxpayer data once a tax return has been received. This may cause issues to accountants looking to amend tax returns that have already been submitted.
Has data been reliable?
While the API figures are a useful resource for accountants, there have been some issues with accuracy.
The most obvious problem is HMRC’s records may sometimes be wrong. If the primary figures in the main PAYE database are incorrect, the same flawed data will be copied across to the pre-pop system.
Another problem occurs when a client has more than one employment with the same employer during the tax year. HMRC provides all the data, but there is no reliable way of picking it apart into the individual periods. HMRC recommends combining this data into one instance of employment, but this won’t be a direct match for the P45 and/or P60 figures supplied by the client.
Finally, some pension providers report pay and tax data to HMRC in the same way as employment income. This means certain pension income appears as a stand-alone employment record, while others are reported (in aggregate) in the private pension income figure.
What enhancements would we like to see?
The pre-pop system could be improved by making figures available earlier, and by adding other types of data to the pre-pop menu.
Currently the main earnings figures (the most useful element) are only made available once the data has been “reconciled” (i.e. checked). This process seems to be beholden to the P11D deadline in July, and means records may not be available six months into the tax year even though HMRC knew the main pay and tax figures in April.
Accountants would find it useful to see the main earnings data in May or June, on the understanding that any P11D data (and possible amendments) may follow later.
Accountants would also like to see a breakdown of the private pensions income figures so they can match them against the more granular data provided by the client.
It would also be useful to add more types of data added to system. Dividends and bank interest would be helpful but may be too difficult for practical reasons (e.g. accounts in slightly different names, joint ownerships etc). But other types of data (e.g. state pension and student loan figures) would be more straightforward to share and would be useful additions to the system.
So has the pre-pop service been a success?
Overall, the new pre-pop system has proven to be popular with accountants. It provides a new way of checking client data, helping to make tax returns more accurate and future enquiries and/or penalties less likely. And the system has come though its first busy period with a few bruises, but generally intact.
But it’s fair to say the data is not perfect, and the accountant should manually check pre-pop figures before importing them into tax returns.
About Matt Bailey
Founder of Azura Cloud Systems (Gbooks), the leading cloud-based tax and accounts system.