National Statistical

This month, we are making improvements to the UK House Price Index (HPI), our official measure of average house prices, to improve the accuracy of new build estimates and bring our initial provisional estimates closer to our revised final figures. 

Measuring house price inflation can be a challenging process, as we have explained in a previous blog. This is especially evident when we are looking to produce accurate initial estimates of house price inflation for new build properties. 

Why are new builds more challenging? 

We use key property attributes information, such as floor area and the number of rooms, alongside the transaction price to calculate UK HPI estimates.  

For new builds, some of this information is sometimes unavailable for recent sales because processing new build sales and creating a new record for a new property takes longer than updating an existing property record.  

This makes it harder to produce an accurate initial estimate for new builds while we wait for additional attribute information about those properties to become available.  

It is a bit like estimating the price of a car without knowing the engine size. You can make a reasonable initial estimate based on the information you do have, but you can improve your estimate when you find out the engine size. 

In the same way, as more property attributes information becomes available in the following months, we use this to improve our initial estimates, revising our UK HPI estimates for 12 months following the first provisional estimate. 

Under our previous method, early estimates for new build prices have tended to be too high, with estimates revised later as the missing information arrives. Headline UK HPI estimates are also subject to the same revision period, but since new builds account for just ~9% of sales, revisions to new build estimates affect headline UK HPI estimates to a lesser extent.  

This new methods improvement reduces the over-estimation previously observed in early estimates for new builds, improving our initial estimates in UK HPI.  

What is changing? 

Our improved approach, implemented on 20 August 2025, means that if some property details are missing, we will estimate the values of the missing data using a technique called imputation. This involves using known information from similar nearby properties to estimate (impute) a value for the missing property information.  

As an example, imagine a flat sells this month, but we don’t yet know how many rooms it has. However, similar flats recently sold on the same street and we know how many rooms they each have. Instead of waiting for the missing information to become available, we now use data from these similar neighbouring flats to make an evidence-based estimate of the number of rooms. We do this using a well-established technique called nearest-neighbour imputation. 

When the missing information becomes available in subsequent months, imputed values are replaced with the true value for that property, to use in revised UK HPI calculations.  

Our analysis has shown that imputing missing values in this way improves the accuracy of our initial estimates, reduces early over-estimation of new build estimates and hence reduces overall revision size from initial estimate to final revised estimate. 

What you will notice 

• More accurate and realistic early provisional estimates for new build prices. 

• A larger-than-normal downwards revision to new build estimates for recent months in today’s release (20 August 2025), as the improved methods are implemented. 

• Smaller overall revisions (1st provisional estimate to 13th revised estimate) to new build estimates in future releases, since early estimates will be closer to the final picture. 

• Greater confidence in early estimates and transparency; our Quality and methodology and About the UK HPI guidance has been updated to explain the improvement.  

Next steps 

These improvements are part of our commitment to develop our key economic and population statistics by making improvements in data collection and statistical production. These changes allow us to provide a more timely picture of the property market as cost of living challenges remain relevant and topical. 

We will continue to monitor revisions to the UK HPI, and intend to conduct a broader review of methods to ensure UK HPI statistics remain timely, useful and as accurate as possible. 

For more details, see our housing bulletin, Quality and methodology and About the UK HPI guidance.