Detailed findings

Introduction

1.1 The Wales Act 2014 devolved the collection of taxes on landfill disposals and land transactions to Wales from 1 April 2018, replacing Stamp Duty Land Tax (SDLT) and UK landfill tax. The Welsh Revenue Authority (WRA) collects and manages the Welsh Land Transaction Tax (LTT) and Landfill Disposals Tax (LDT) for the Welsh Government and produces statistics on the revenue collected.

1.2 These statistics provide high quality, relevant and trusted information on LTT and LDT revenue. Welsh Government uses the statistics to monitor the amounts of LTT and LDT tax revenues collected, against those due. The statistics form part of HMRC Revenue and Customs (HMRC) UK statistics covering similar taxes. The statistics are also used by Welsh Treasury and OBR to produce fiscal forecasts and by local authority planning for waste management and monitoring their progress towards meeting net-zero targets.

1.3 In requesting this assessment, the statistics team at WRA is demonstrating its commitment to producing LTT and LDT statistics that meet the standards required of National Statistics and the Code of Practice for Statistics.

1.4 We judge that these statistics meet the highest standards of the Code of Practice for Statistics and can be designated as National Statistics.

User engagement and public value

1.5 The public value of LTT and LDT statistics is evident, supported by their wide use in the preparation of national fiscal forecasts and in environmental, waste management and housing policy. WRA’s active engagement with its key users and continuous efforts to improve the public’s understanding and accessibility to the statistics are examples of best practice.

1.6 Statistics relating to the devolved LTT and LDT meet the continued needs of users since the taxes were devolved to WRA in 2018. Their equivalents for UK and Scotland are produced by HMRC^[1][2] and Revenue Scotland^[3][4], respectively.

1.7 The statistics are used by OBR to forecast LTT and LDT revenues (with support from Welsh Government), and by OBR to forecast equivalent UK income streams. Policy analysts in Welsh Government use the statistics to assess the impact of LDT, and changes to the rate of LDT, on environmental targets, while local authorities use the statistics for monitoring and designing policy to assist with waste management. Other bodies, including HMRC and National Parks, use the statistics for triangulation (LDT statistics) and to assess market trends (LTT transaction types) respectively.

1.8 There is broad public interest in these statistics, for example, during the devolution of these taxes and subsequently on LTT developments^[5]. Welsh Government is currently consulting on planning legislation and policy for second homes and short-term holiday lets, due to conclude on 22 February 2022. Welsh Government is also consulting on local variation to LTT rates for second homes, due to conclude on 28 March 2022.

1.9 WRA is dedicated to engaging closely with the users of its statistics. In meetings we held with users as part of this assessment, across both the LTT and LDT statistics, they reported that WRA’s commitment to understanding their needs and preferences in pursuit of public value was impressive. Users reported that WRA statisticians had a comprehensive understanding of their outputs and that they had often already considered potential avenues for development.

1.10 Since the devolution of the taxes in April 2018, WRA consulted closely with key users and stakeholders to ensure best practice was used throughout the planning process for the statistics. When users presented new ideas for breakdowns and different publication formats of the statistics, WRA’s team was consistently receptive to these ideas. Users reported that WRA successfully integrated user feedback into its work at early stages of planning.

1.11 One key example of effective early dialogue with expert users is WRA’s selection of the reporting date to use for the Land Transactions Tax. The ‘effective’ date^[6] and the ‘submitted’ date^[7] were options for the LTT reporting, as well as deciding upon a suitable ‘cut-off’ date^[8] within the month to extract data. Close consultation with users helped inform WRA’s decision to use the effective date as the standard. The Welsh Treasury and OBR prefer that the definition used for monthly transactions relate to the effective date because this enables an estimate of revenue collected to be produced, consistent with the forecast of future revenue.

1.12 Some of HMRC’s statistics for the UK use the ‘submitted’ date. Therefore, to aid the comparability of LTT statistics across nations WRA has included some comparable statistics based on ‘submitted date’ in its recurring annual statistical release. WRA provided commentary (see Figure 2.8, April 2020 to March 2021) to explain why the differences between the transactions by ‘submitted’ and ‘effective’ date arose, which helps reassure users of the quality of the data.

1.13 WRA also demonstrated excellent commitment to meeting user needs in a timely manner when the statistics team received a request for a breakdown for the LTT statistics by national park. WRA investigated the feasibility of the request and consulted users to understand how the data could best be formatted to meet their needs for planning for new housing and to better understand the nature of transactions at a national park rather than local council area level.

1.14 WRA statisticians consistently demonstrate openness to user input, even in cases where users had limited historical engagement with WRA or where they considered themselves occasional users. This has undoubtedly strengthened the LTT and LDT statistics publications.

[1] HMRC Landfill Tax Statistics

[2] HMRC SDLT Statistics

[3] Revenue Scotland Landfill Tax Statistics

[4] Revenue Scotland Land and Building Transaction Tax Statistics

[5] For example, https://businessnewswales.com/first-land-transaction-tax-data-released-for-all-welsh-local-authorities/ and https://nation.cymru/news/im-thinking-of-selling-up-second-home-owner-complains-about-tax-hike/

[6] This is when the tax becomes liable to be paid, usually when a transaction is completed on a property.

[7] This is the date when a tax return is submitted to the WRA

[8] LTT returns can be submitted at any time. When producing monthly, quarterly and annual LTT statistics, WRA uses the third Monday of the month as the ‘cut-off’ date. New returns or amendments received after this time are not included in LTT statistics published in that month.

Enhancing accessibility

1.15 WRA is continuously looking to improve its approach to publication and the accessibility of its statistics. Initially WRA published reports in a PDF format. However, to improve the accessibility, since June 2019, WRA has transferred its published statistical documentation to a HTML format. This is best practice and in line with guidance published by the Government Digital Service.

1.16 A more recent example of WRA’s commitment to accessibility includes its use of a message board on the web platform Basecamp. Using feedback from the message board, and in line with Government Statistical Service guidance, WRA has reformatted its publication spreadsheets. Those spreadsheets have been evaluated against a new draft accessibility checking tool developed by the Office for National Statistics and found to be compliant, with the comment that the WRA’s spreadsheets are among the best examples considered to date. The changes are set to be made across future publications to improve the legibility of the spreadsheets for all users.

1.17 When WRA began publishing annual LTT statistics by local area in July 2019, parts of the media coverage included misleading claims about the statistics. Common inappropriate interpretations included:

1. the assertion that the proportion of properties that attracted the higher rate band was representative of the total housing stock in Wales, rather than only those sold in-year (less than 5% of the total stock); and
2. that all transactions charged at the higher rate related to second homes, rather than including buy-to-let properties, seasonal residencies and homes bought to be developed or renovated for onward resale.

1.18 WRA was proactive in monitoring the interpretation of its reports and developed a communication strategy to help aid public understanding of its statistics. WRA took timely steps to address the above common misinterpretations by working directly with journalists to help avoid future misinterpretation. In addition, WRA consulted with expert users of the statistics for guidance on the future wording of the LTT reports.

1.19 WRA produced and published a very clear article which explains common misunderstandings relating to LTT statistics and how to avoid them. Subsequent media coverage of the LTT statistics has improved considerably as a result of WRA’s media and communications strategy.

Public value – Collaboration with HMRC

1.20 WRA and HMRC collaborate closely to integrate the devolved LTT statistics into aggregate UK statistics. HMRC administers Stamp Duty Land Tax (SDLT) in England and Northern Ireland. Transactions data for these nations are loaded onto HMRC systems and are extracted to be used in producing the SDLT statistics.

1.21 Although WRA publishes estimates of LTT transactions totals each month, the WRA date of publication follows that of the HMRC publication ‘Property transactions in the UK’, so that the useful statistics available to HMRC are lagged by around a month. Until July 2019, HMRC estimated the latest month by grossing up weekly lists of LTT transactions from WRA. However, the delay in the data reaching HMRC statisticians meant that estimates of the LTT data for the latest month available to the HMRC statisticians were not sufficiently up to date for the grossing to be reliable. From August 2019 HMRC took data directly from the most recent WRA publication with a one-month lag, and estimated the missing most-recent month by a simple extrapolation of the year-on-year growth in England and Northern Ireland and applied to the transaction numbers in Wales.

1.22 In general, before the COVID-19 pandemic, HMRC’s August 2019 method produced small and generally acceptable differences between published HMRC UK property transaction statistics and WRA LTT statistics. COVID-19 however has caused greater volatility in the property market. As a result, during the pandemic, differences between WRA and HMRC figures worsened to unacceptable levels.

1.23 Consequently, in June 2021, WRA proposed a revised method of LTT data supply that included WRA estimating for the most-recent month based on partial returns, and set up a data sharing agreement with HMRC. The revised method allowed HMRC to use the latest available estimates of residential and non-residential LTT transactions completed during the latest full month directly. Following review of a 3-month trial period between July 2021 and September 2021, HMRC and WRA agreed to transition to the proposed LTT estimation methodology permanently from October 2021. Additionally, WRA agreed to directly supply HMRC with historic revisions for residential and non-residential transactions dating back to April 2018 in order to better align each series.

Data quality

1.24 WRA ensures that the data used to produce its statistics are of the highest quality by employing a robust approach to Quality Assurance, which includes evolving plans for data assurance that meets with the widening needs of users. Confidence and trust in WRA statistics is supported by a comprehensive transparent approach to data sharing with for instance HMRC and Natural Resources Wales (NRW), the treatment of revisions, and the communication of uncertainty to users.

1.25 Throughout our interviews with users, they consistently reported confidence in WRA data quality and the methods used to produce the LTT and LDT statistics. Where potential for data errors has been identified, for example mismatching postcodes and local authority area declarations, WRA has introduced processes to address the source of the errors. Here, for example, WRA cross-references the postcode with the local authority selected on the tax return using ONS’s Postcode Directory.

1.26 WRA assesses LTT and LDT data for quality in line with OSR guidance on the quality assurance of administrative data (QAAD). This includes a clear evaluation of the risk categories that each of the constituent components of the statistics pose to the overall quality of the statistics which are published in the Key Quality Information for both the LTT and LDT.

1.27 When the statistics production process was being developed, WRA staff produced comprehensive systems to assure the quality of the tax data. Statisticians use Reproducible Analytical Pipelines to check, for example, the internal consistency of tax submissions. WRA operates a relatively mature system of automated data quality assurance.

1.28 WRA produces a detailed Annual Reports and Accounts of its operations which includes an analysis on the effectiveness of LTT and LDT collection. The quality of the LTT and LDT statistics depends on effective operational performance. The Key Quality Information and annual review allows users to further understand the quality of the data. For example, users are provided with detailed information on the quantity of LTT that is at risk from, for example, land that requires mixed tax treatment, or from differences in tax treatment compared with WRA’s tax calculator.

1.29 As part of ensuring that the right amount of tax is paid, WRA’s operational teams run checks across a variety of data sources such as HM Land Registry, Ordnance Survey and Companies House for purpose of checking land use for LTT. In the case of the LDT, on-site inspections are completed at landfill sites and close relationships are maintained with landfill site operators to ensure reliable reporting is achieved. These activities further contribute to assurance about the quality of the statistics.

1.30 WRA is also considering avenues to further strengthen its quality assurance. For example, WRA is actively scoping the opportunities for coordination with NRW to develop a rolling check of refuse volumes, to diagnose potential data mismatches earlier in the data collection process.

Revisions and uncertainty communication

1.31 Both the LTT and the LDT statistics are initially published as provisional and are subject to planned revisions. WRA has published high-quality revisions explanation for both outputs in its quarterly bulletins and in WRA’s policy for publishing official statistics. This includes a clear explanation of how provisional statistics are indicated as such. The policy document also explains how often commentary on revisions is published as well as where to find it for each of the LTT and LDT statistics.

1.32 In the case of LTT statistics one cause of planned revisions is the gap between the date of publication for the monthly statistics and the date at which filing is required for the tax – returns can be made up to 30 days after the effective date without penalty. WRA revisions analysis shows that revisions to the number of transactions is similar each month but if the transactions are high in value then revisions will be higher and revisions subsequently become smaller between the second and third estimates.

Independence, orderliness and data governance

1.33 WRA statistics are produced independently and on an orderly basis. Statistical production is supported by robust data governance and a professional approach to the management of its resources and the development of its staff.

1.34 WRA has published a policy for publishing official statistics which outlines the approach it takes to ensure that its statistics meet the standards of clarity and transparency required for users to trust their outputs. WRA’s documentation is thorough, high quality and underpins the trustworthiness of its statistics.

1.35 Taxpayer data are confidential, so WRA takes steps to prevent disclosure of sensitive data. WRA communicates its approach to managing disclosure risk through its quarterly bulletins (See Table 1b, footnote 5, July to September 2021). In our interviews with users, they stated that they were aware of the impact of disclosure control on requests for more-granular data, such as the need for transaction suppression in the case of small numbers of transactions dominating categories.

1.36 WRA’s statistics team encourages continuous professional development. WRA’s employee records include up-to-date training logs covering areas such as data analysis, leadership and managing a quality service. The statisticians report that they feel encouraged to participate in training and development as part of their role, demonstrating WRA’s commitment to staff development.

1.37 The roles within the team are clearly defined and the team’s workplan sets out the timeline and responsibilities for publishing the statistics.