Land Transaction Tax statistics: January to March 2020

These statistics are about LTT transactions that we (the WRA) have received by 20 April 2020.

Figure 1.1 below shows:

• the quarterly headline numbers for January to March 2020
   
• the first estimates of annual data for April 2019 to March 2020

Figure 1.1 also highlights the percentage change for these headline numbers against our first estimates for the same time periods a year earlier. We explain why these comparisons are made in the Section 1 of this release (‘Comparisons with the same period a year earlier’).

LTT statistics by time period and transaction type on StatsWales

By the close of 20 April 2020, the total tax due for 2019-20 was £264 million.

For some transactions, we are unable to provide any information other than the total tax due figure in the year, as there is a risk of revealing details of the individual transactions. These are rounded to the nearest million pounds for additional protection, and should only be included if seeking a value for total LTT revenue in the year.

However, excluding these transactions, the total tax due for 2019-20 was £233.9 million. In order to make comparisons with earlier periods and by transaction type on a like-for-like basis, it is this value we use (see Figure 2.3).

Despite a small fall in the number of transactions in 2019-20 (compared with the previous year), the total tax due increased by 3% on a like-for-like basis. This is mainly due to:

• upwards movement in residential property values and the associated tax due on each transaction
   
• an increased number of transactions paying the higher rates of residential tax

For April 2019 to March 2020, £64.6 million of additional revenue was raised from higher rates transactions. This is 8% higher than our first estimate for April 2018 to March 2019 (published in April 2019).

The coronavirus (COVID-19) crisis has had an impact on these figures, but it is a minor one. More detail about this is shown in Figure 2.1 in section 2 of this release.

These initial estimates for January to March 2020, and April 2019 to March 2020, are provisional. We will revise these estimates in the next statistics publication. This is because organisations are still able to report further transactions for March 2020. Organisations have 30 days after the date of the transaction to report them to the WRA.

If in future, we are given permission to publish more detail about the transactions with restricted detail, we will include further information (such as a less rounded figure, or the breakdown between residential and non-residential).

Adam Al-Nuaimi, Head of Data Analysis in the WRA, commented on these statistics:

Figure 1.1 above shows the quarterly headline numbers for LTT transactions relating to January to March 2020, and first estimates of annual data for April 2019 to March 2020. These show general increases in tax due from a year earlier, despite broadly static numbers of transactions.

Most transactions were received before the coronavirus (COVID-19) outbreak took hold, so any impact on these statistics is minor.

In this release, we’ve also presented information on the number of transactions received in April. This offers an early insight into a substantial decrease from the long-term average of around 1,200 per week, down to an average of around 500 per week so far in April. This will likely lead to a large decrease in the LTT revenues that will be shown in the statistics for April and May 2020, to be covered in our data-only releases over the next two months.

We’ll continue to monitor this decrease and update the data in these data-only releases. And we’ll comment in more detail in our next substantive release - our main annual statistics publication, currently planned for July.

Introduction of LTT

From 1 April 2018, LTT replaced Stamp Duty Land Tax (SDLT) on residential and non-residential property and land interests purchased in Wales. The tax rates and tax bands for LTT vary depending on the type of transaction.

LTT statistics are not fully comparable to previous SDLT statistics. This is because different rates and bands are used in LTT. The reliefs may also be different for the two taxes. For example, first time buyers’ relief applies to SDLT but not to LTT.

Value of LTT statistics

Timely information on activity in the property market is important for policy makers. When filing an LTT return about a property transaction, the organisation paying the return has 30 days after the effective date to submit and pay any tax due. LTT statistics therefore are relatively timely. 

Forecasting LTT revenues for Wales in future is an important use of LTT statistics. The Welsh Government and the Office for Budget Responsibility mainly do this.

Data available for LTT

All of the data used in this statistical release is available in a spreadsheet on the headline statistics page.

Annually, we also publish geographic datasets for LTT on the StatsWales website. This includes annual data by:

• Local authority
   
• Assembly Constituency (for residential transactions only)
   
• Level of deprivation, using the Welsh Index of Multiple Deprivation (for residential transactions only)
   
• Built up areas (for residential transactions only)

For data at the Wales level, we provide links to the relevant StatsWales datasets throughout this release.

Timing of and revisions to LTT statistics

The diagram on the key quality information page explains the timing of LTT statistics. We present provisional estimates for March 2020, the quarter January to March 2020, and the year April 2019 to March 2020. We present revised estimates for periods before this. We will revise the provisional data in future. Not all tax returns for these periods may yet have been received.

In future, we may continue to revise statistics for earlier periods, to account for any amendments to transactions. In particular, this will be due to higher rate refunds being made for several years after the date of the original transaction.

Comparisons with the same period a year earlier

There can be seasonal patterns in the property market, with higher levels of activity generally seen in the summer and autumn, and lower levels in winter and spring. These effects are also seen in the levels of transactions and tax due. Therefore, it can helpful to compare the current period with data for the same period a year earlier

However, in our monthly and quarterly statistics, we are gradually revising downwards the tax due for earlier periods. This is because of higher rate refunds being paid out in each month (for higher rates residential transactions which were effective in earlier periods, back to April 2018).

The value for January to March 2019 will have already been subject to some of this downward revision, whereas the equivalent figure for 2020 will not yet. Also, in future, there will be some upward revisions to the values for January to March 2020 due to late transactions.

Therefore in this release, we compare:

• January to March 2020 data against our first provisional estimates for January to March 2019 (which we published in April 2019).
   
• April 2019 to March 2020 data against our first provisional estimates for April 2018 to March 2019 (which we also published in April 2019).

This provides for the fairest comparisons over time.

Key quality information and glossary pages

Please see the separate glossary and key quality information while reading this statistical release:

• We define relevant terms in the glossary as they are used in this release.
   
• On the key quality information page, we describe how Land Transaction Tax statistics meet the Code of Practice for Statistics and the dimensions of value, trustworthiness and quality.

Properties or land sold more than once

These statistics relate to transactions which were effective in particular month, quarter or year. A property or piece of land may have been sold more than once in that time. If so, it would feature multiple times in the statistics.

For example, in 2019-20, our best estimate is that between 4% and 5% of transactions involved a piece of land or property which has been sold more than once in the year.

In March 2020, we released an update on publishing WRA statistics due to coronavirus (COVID-19). In this update, we stated that in our April 2020 release for LTT, we would look at any potential impacts of coronavirus (COVID-19) on our statistics.

Figure 2.1 above shows the total number of transactions submitted to the WRA in each 7-day period. For example, the bar ’18.4’ shows the number of residential and non-residential transactions submitted to the WRA from 18 April 2020 to 24 April 2020 (inclusive).

The grey dashed line shows the average weekly number of transactions submitted between January and December 2019 (residential plus non-residential transactions).

Please note that Figure 2.1 shows data by submitted date. This differs from effective date, which is the main date measure we use for analysis in this release.

Figure 2.1 shows the large peak in transactions submitted in the last working week before Christmas 2019. Then the number of transactions submitted drops to a much lower level in the two-week period for Christmas holidays, in line with the previous year.

Weekly transactions submitted in January to March 2020 were around the expected levels based on the early part of the previous year. However, the weekly number of transactions submitted has dropped sharply in April following the coronavirus outbreak, particularly in the residential sector.

This drop in the number of transactions received relates in small part to transactions that would have been effective in March, and hence there is a small impact on the revenues shown in this release.  However, the main impact of the decline is a reduction in transactions that would otherwise have been effective in April.

This means that in future editions of this release, and in the data only updates we release in the interim, that there will be a substantial decrease in revenues shown for April 2020, and for at least the first quarter of 2020-21. We will update this chart in each of the next two data-only updates and provide further comment alongside the chart.                                              

LTT statistics by time period and transaction type on StatsWales

By the close of 20 April 2020, we received details of 13,280 notifiable transactions with an effective date in January to March 2020. This is 1% higher than our first estimate for the same period in 2019.

Provisionally, there were 61,190 transactions effective during 2019-20. This is 1% lower than our first estimate for 2018-19 (published in April 2019).

In January to March 2020, 89% of transactions were residential and 11% were non-residential. These are similar percentages to previous three-month periods.

In future, we expect to receive details of a small number of further transactions for January to March 2020.

LTT statistics by time period and transaction type on StatsWales

By the close of 20 April 2020, the total tax due for 2019-20 was £264 million (shown in Figure 1.1).

For some transactions, we are unable to provide any information other than the total tax due figure in the year, as there is a risk of revealing details of the individual transactions. These are rounded to the nearest million pounds for additional protection, and should only be included if seeking a value for total LTT revenue in the year.

However, excluding these transactions, the total tax due for 2019-20 was £233.9 million. In order to make comparisons with earlier periods and by transaction type on a like-for-like basis, it is this value we use (see Figure 2.3).

This value is 3% higher than our first estimate for 2018-19 (published in April 2019). The corresponding changes for residential tax due and non-residential tax due were a 7% increase and a 7% fall, respectively.

£54.4 million tax was due for transactions with an effective date in January to March 2020. This represents a 3% increase when compared with our first estimate of tax due for January to March 2019 (published in April 2019). This was comprised of a 16% increase in residential tax due and a 18% fall in non-residential tax due (over the same time periods).

LTT statistics by time period and transaction type on StatsWales

The value of property taxed in January to March 2020 was £2.6 billion. For the year April 2019 to March 2020, the value of property taxed was £12.2 billion.

Separately, the rental value for newly granted non-residential leases was £236 million in January to March 2020. This value was £1.3 billion for the year April 2019 to March 2020.

LTT statistics by time period and transaction type on StatsWales

The numbers of residential transactions by effective month has varied greatly since April 2018. The monthly trend in residential transactions in 2019-20 has broadly tracked that of the previous year. There are some differing peaks between years, in some cases due to there being five Fridays in certain months, rather than four. Figure 10.8 in section 10 of our annual statistical release shows that nearly half of transactions have an effective date that is a Friday.

In March, we see an increase from the previous month (February) in non-residential transactions. This may be expected, as it is common for non-residential leases to be renewed at the end of the financial year.

LTT statistics by time period and transaction type on StatsWales

As may be expected, similar trends are seen in the monthly tax due as are seen in the monthly counts of transactions. However, the generally higher revenues for residential transactions in 2019-20 can be partly explained by higher rates refunds that have yet to be claimed for 2019-20 transactions. Figure 1.1 presents numbers which better takes account of this effect.

There is greater volatility in the monthly series for non-residential transactions. They also make up a larger share of total tax due than the share of the number of transactions.

LTT statistics by transaction type and transaction description on StatsWales

The value of the properties associated with conveyances and transfer of ownership during January to March 2020 was £2.5 billion (not shown in Figure 2.7).

Most transactions in January to March 2020 were associated with a conveyance or a transfer of ownership. This figure was 93% for residential transactions and 61% for non-residential transactions.

A new lease was granted in 37% of non-residential transactions (compared with 2% of residential transactions).

Similar percentages are seen for previous three-month periods.

Please note that all analyses in section 3 exclude the additional transactions with restricted detail (shown in Figure 1.1 in the ‘Main points’ section of this release).

LTT statistics by time period and residential transaction value on StatsWales

For each tax band, Figures 3.1 and 3.2 show the quarterly trends in the number of residential transactions and tax due. There are six residential tax bands. We have combined the largest three bands here to show results for properties purchased for more than £400,000.

Similar seasonal trends for the tax bands can be seen in both numbers of transactions and tax due. Most of the tax bands show a fall in transactions and tax due in January to March (compared with the preceding October to December).

However, there is greater volatility in the trends when considering the tax due on properties in the higher value bands. For properties purchased for more than £400,000, the tax due in October to December 2019 was the highest quarterly value seen to date. This was also the case for properties purchased for between £250,000 and £400,000.

LTT statistics by time period and residential transaction value on StatsWales

Just over three fifths of residential transactions were within the first tax band (purchase price £180,000 or lower). Although the main tax rate on residential transactions of up to £180,000 is 0%, these transactions still accounted for around a sixth of total residential tax due, which relates to the higher rates residential component of the tax.

Combining the fourth, fifth and sixth bands (purchase price of greater than £400,000), these accounted for only 5% of transactions. However, the tax due for these transactions accounted for 37% of the total residential tax due.

Please note that all analyses in section 4 exclude the additional transactions with restricted detail (shown in Figure 1.1 in the ‘Main points’ section of this release).

LTT statistics by time period and non-residential transaction value on StatsWales

Figures 4.1 and 4.2 show that there were 1,470 non-residential transactions effective in January to March 2020 with £16.0 million tax due.

In each three-month period since April 2018, around 50% to 70% of the tax due has been contributed by transactions with a non-rental value greater than £1 million. In each three-month period, around 10% to 30% of the tax due has been contributed by the rental value of non-residential properties.

LTT statistics by time period and non-residential transaction value on StatsWales

Figure 4.3 shows that in January to March 2020, 4% of non-residential transactions had a non-rental value of more than £1 million. These transactions accounted for 69% of the non-residential tax due (Figure 4.4).

Figure 4.3 also shows that for 31% of non-residential transactions in this period, a rental value was associated with the property (which contributed to the tax paid on the transaction).

The rental value of non-residential properties accounted for 12% of the total non-residential tax due (Figure 4.4).

Please note that all analyses in section 5 exclude the additional transactions with restricted detail (shown in Figure 1.1 in the 'Main points' section of this release).

Taxpayers can claim reliefs on both residential and non-residential transactions. Reliefs reduce the amount of tax due when certain conditions are met. More than one relief can be applied to a single transaction.

Reliefs may reduce the tax due:

• to zero, known as a full relief
   
• or by a certain percentage or amount, known as a partial relief

LTT statistics on reliefs by measure and transaction type on StatsWales

Figure 5.1 shows that there were 350 transactions in January to March 2020 with reliefs applied to them that reduced the associated tax due. This is lower than the previous quarter.

On average, there are around 120 reliefs claimed in each three-month period which had no impact on the tax due. These reliefs are excluded from Figure 5.1. Many of them have been reported unnecessarily by the organisations completing the tax return. As an example, some of these apply to low value residential transactions. Indications are that they are due to a perceived but mistaken need to claim first time buyer relief (which applies for the predecessor tax, but not to LTT). This is known following queries raised with several agents asking why tax reliefs have been claimed where there is no impact on value of the tax. Further information about this category of reliefs is provided in Example 4 in our key quality information.

This example also describes some adjustments that have been made to more correctly identify the value of tax relieved associated with these transactions. We expect further adjustments in future and we therefore expect to revise Figure 5.1 in future.

In previous statistical releases, we excluded linked and relieved transactions from Figures 5.1 and 5.2. This was so that we could carry out further analysis on these transactions. We have now carried out this analysis and have a reasonable level of confidence in the quality of these data. We have now added linked and relieved transactions into Figures 5.1 and 5.2, revising all data back to April 2018. This has added:

• around 60 to 70 transactions each quarter into Figure 5.1
   
• an average of £4 million to £5 million each quarter into Figure 5.2

LTT statistics on reliefs by measure and transaction type on StatsWales

Figure 5.2 shows that in January to March 2020, the total value of reliefs claimed was £11.4 million. This is less than half of the value in the same three-month period in 2019.

For each three-month period, the numbers of reliefs claimed on residential transactions was higher than for non-residential transactions. However, in most of the three-month periods, non-residential transactions contributed around 70% to 90% of the total value of reliefs claimed.

Further data on reliefs, including quarterly data by type of relief, is available on StatsWales at the link above.

LTT statistics on higher rate refunds by original transaction date on StatsWales

When a taxpayer claims a refund for higher rates residential LTT, the original transaction is amended to a main rate residential LTT transaction. The data in this release is adjusted for any refunds approved by WRA up to and including 20 April 2020.

Figure 6.1 shows that 690 higher rate refunds were claimed for transactions effective in 2019-20, with £5.6 million refunded to taxpayers.

Taxpayers have up to three years to sell their previous main residence and claim a refund. Therefore, all the values in Figure 6.1 will continue to be revised upwards in future editions of our statistics. This will also to lead to the total tax due in other tables and charts reducing.

The number and value of refunds presented for January to March 2020 are lower than for earlier periods. This is because compared with earlier periods, not enough time has passed since the transaction was effective for many of the relevant taxpayers to sell their previous main residence and claim their refund.

For all higher rates transactions, the WRA asks the question whether the taxpayer intends to reclaim the higher rates element in future. It will take several years before we know how likely someone is to claim based on their stated intentions (it can take up to three years to make the claim). But we do currently know that around 70% of those who do claim answer this question in the positive.

Additional data on higher rate refunds

Another useful way of presenting data on higher rates refunds is to use the date when the refund was approved by the WRA. A dataset using these dates, and the effective date of the original transaction, can be found on the StatsWales website at the link below:

LTT statistics on higher rate refunds by original transaction date and refund approved date on StatsWales

Further information on the refund payments made to taxpayers, by the month in which they were made, can be found in Table 6a of our monthly and quarterly statistics.

In the main, these additional data are provided to support forecasting requirements.

In 2019-20, the WRA received £233.5 million in LTT payments. These figures are less than those reported in Figure 2.3 as they relate to the payments received in each month (often referred to as ‘on a cash basis’). This differs from earlier tables in this release which are based on transactions that were effective in the month.

There is a difference in April 2018 as the WRA only started collecting the tax in that month. Therefore, no payments relating to transactions effective in earlier months were relevant.

The highest monthly receipts seen to date were in December 2019 (£30.6 million).

We look here at the effect of the regular revisions made to Land Transaction Tax statistics. We analyse the differences between the first, second and third estimates published for a month. This is for both the number of transactions and the tax due.

For example, we have published three estimates for January 2020. We published the first estimate on 21 February, published the second estimate on 20 March and the third estimate on 30 April.

Figure A1 shows that higher levels of revisions can generally be seen in the earlier months that the WRA began collecting LTT. This is particularly the case for the tax due for transactions with an effective date in April 2018, where there was a 30% increase in the estimate of tax due (from the first to the second estimate for the month). A larger revision in April 2018 was expected because the familiarity of the system to users would have been lower, and also because an earlier cut-off date in the following month was used to extract the data. Nevertheless, the 30% figure in terms of tax due is considerably higher than the equivalent figure for the number of transactions (10%). It is explained by a few larger transactions with an effective date late in April 2018 that were not reported to WRA until later in May 2018 (before the 30 day filing limit, but after the cut-off date for the April publication).

Figure A1 also shows the levels of revisions have generally decreased over time. Since October 2018, the revisions between the first and second monthly estimates have generally been between 0 and 3%. Recent exceptions were:

• June 2019 when the tax due was revised up by 9% between the first and second estimate
   
• September 2019 (tax due was revised upwards by 16%)
   
• January 2020 (tax due was revised upwards by 27%)

These exceptions are generally due to a small number of returns arriving towards the end of the 30-day notification period.

The lower level of revisions generally seen now is likely to be due in part to an increasing familiarity with the system amongst solicitors and conveyancers completing the returns. It is consistent with a general decrease in the time taken for returns to be filed with the WRA over the same period (not shown in tables or charts).

There may also be seasonal effects in revisions to the data. For example, we saw higher revisions for the July 2018 estimates than the months around it. However, we do not see any obvious seasonal patterns in data for 2019 or so far in 2020. We will require at least another year’s worth of data to properly assess any data seasonality.

We made a downward revision to the non-residential tax due for April 2019 (between the first and second estimates). A non-residential transaction effective in April 2019 was entered incorrectly as being overly large and was later revised.

Revisions between second and third published estimates

In a spreadsheet published alongside this statistical release, Tables A1 and A2 show the difference between first, second and third published estimates for a month.

We see relatively small increases between the second and third estimates for a month. In general, this is also the case for the later estimates for a month (not shown in the tables). However, falls may be seen in the second, third and later estimates of tax due for a month. This is because the data are shown net of any refunds for higher rate residential transactions. These refunds may be claimed several years after the effective date of the original transaction. We analyse refunds in section 6 of this statistical release.

In future, we may consider applying a grossing factor to the first estimates for a month. This may help reduce the revisions required to the first estimate for a month. With the volatility shown in the data to date, it is likely we will need several years of LTT data in order to calculate appropriate grossing factors.

In general, we see larger revisions in the data on non-residential transactions than for residential transactions. This reflects the more volatile nature and often larger size of non-residential transactions.

Our key quality information page describes how our Land Transaction Tax statistics meet the Code of Practice for Statistics and the dimensions of value, trustworthiness and quality.

We define relevant terms in the glossary as they are used in this release.

We would be grateful for your feedback on these statistics, to help us improve them. Please contact us using the details below.