Empty Homes Part 1
About this series
This article is part of an on going investigation into how empty and second homes affect the UK housing market and what it means for people in different parts of the country and the market overall. The data is not all in and the quality of it is variable, there are still some ongoing discussions with various local authorities about whether they will/want to/have the ability to give me the data so things can change from week to week. As a result I will be publishing my results and current thoughts on what I am seeing. The goal is to work towards a publishable piece of research maybe it gets there maybe it doesn't but there will definitely be a lot of insight into the British housing market and the number, value and distribution of low use homes throughout the country.
Housing prices in the UK can be very high, especially in London and areas of natural beauty. This is the first part of an ongoing analysis that looks at how many low use homes there are in various parts of the UK and tries to find out what effect they have on the market overall. This first article describes how the data is gathered and the challenges found in making large numbers of unusual Freedom of Information requests.
In late 2015 after moving back to the London after many years outside the UK, it felt like any journey in a taxi would end up with the driver saying house prices were so high because of rich foreigners buying up all the property as an investment and leaving it empty. Initially I didn't pay much attention to what they were saying however over time colleagues and friends started talking about how only a few of the apartments where they lived seemed to be lived in and the flats were owned by some foreign investment vehicle, what's more newspapers started publishing stories on how empty homes were driving up prices in the capital. Over the last few years foreign ownership of properties has been getting steadily more press see for example recent articles from The FT, The Telegraph, The Guardian, The Independent. In spring 2016 when I started thinking more about the effects of empty homes in London, I started to notice other articles about the effect rich British people were having on rural parts of the UK (See here for example) it seemed that they were really the same problem viewed from different perspectives.
Whilst the conversations with taxi drivers, colleagues and newspaper articles all seemed to be saying the same thing, a lot of the evidence seemed anecdotal or lacking hard numbers. The question was how to find some quantitative basis for evaluation, I decided to combine council tax data and the standardised ONS geography Local Super Output Area (LSOA). Using this information I would be able to map local authorities by how many low use homes were in that area and combine it with other data such as Land Registry price paid data to find a relationship between percentage low use and the price of housing.
Are there a lot of empty homes?
Across the county as a percentage of housing stock, not really. The charity Empty Homes says that there are about 200,000 long term empty homes in the UK (that is homes that have been empty for 6 months or more) out of 27 million homes overall. In some ways the number of low use homes isn't the point as much as where they are and in what concentrations, if low use homes were spread evenly across the country it wouldn't be a problem but if they concentrate in certain areas the market can become distorted affecting locals who actually want to live in the area. What's more concentrations of high levels of low use homes in expensive areas have a very different effect to if they are in cheap undesirable areas.
What is a low use home?
Low use homes are properties which are not a main residence, but are being used as either a second home or kept empty.
In this post
This post will discuss the method and broad take aways from trying to gather and analyse this data.
The method for this project is quite involved as the data gathering is much more complex. Although the total number of empty homes is available at local authority level the much more granular data needed for this investigation is not. What's more local authorities don't have data at LSOA level and most haven't heard of LSOA. This creates difficulties obtaining the data as all requests have to be done using a Freedom of Information request (FOI).
Whats's the Deal with the FOI?
A freedom of information request allows you to ask the government for information that isn't publicly available and which doesn't contain sensitive government information or that may breech the Data Protection Act. It is designed to increase government transparency and a lot of the most common FOIs are eventually turned into standard releases. FOIs can be used to ask some very obscure and difficult to answer question or just ask about radioactive monsters. In order to stop resources being used unnecessarily on FOI requests they're time limited, if a request takes more than a certain number of hours or requires more than an upper limit of money the request can be refused. As no council has LSOA data the FOI would be rejected as taking too much time
So how did you get the data then?
In order to not have the FOI rejected on the basis of time required, a template needed to be developed without having seen the structure of the data and which would work for all local authorities in the UK. The good news is that the data required was very simple, a postcode and a tax class. All councils hold this data and it can be extracted simply in two columns, knowing this we could create a template as below.
- Create a mapping of post codes to LSOA (current mappings are out of date)
- Make an Excel file with two sheets: the council pastes its data into the first sheet, which then looks up the LSOA on the second sheet.
- Make a template for each of the councils in England and Wales so their computers don't crash when they look up the LSOA.
- Clearly mark that the post codes be deleted to avoid potentially breaching the Data Protection Act.
What data was being requested?
Initially I asked for the council tax exemptions which contain houses that are exempt from council tax due to being empty. However, it became apparent that the data of interest was in fact the types of discounts. A law change in 2013 meant that local authorities were no longer obliged to give a discount on second homes. Luckily most of them still keep the category, what's more homes that are empty for two years or longer are often charged a levy in order to discourage this kind of behaviour. Currently the data being requested is both the discounts and the exemptions since they are getting the information anyway getting both provides a rich data set for further analysis.
What is the data being combined with?
The data is being combined with the LSOA and geographical data that was used to visualise the brexit blog. However we are also bringing in data from the Land Registry, which has records of all prices paid for properties going back to 1995. The Land Registry has the full address and price paid, this allows us to get a rough estimate of how much properties in each area are worth.
What are you going to do with it?
That's the million pound question. At the moment as the data is coming in and being organised, just visualising it looking at the problems and seeing if there are any immediate correlations. However hopefully this will give a clue as to where to go next.
So far over 50 FOIs have been made and data has been received from about 30 local authorities, the total process up to now has taken 9 months. The majority of this data has come from London where the main focus has been however I have also received data back from Cornwall in the west of England, 2 authorities in Suffolk in the east of England as well of 4 of the councils that make up the 6 councils of Cumbria which covers the Lake district, an area in the north of England famous for its scenic countryside and very popular with tourists and holidaymakers.
Below is the current map of London, not all the boroughs are available, and requests are ongoing to get hold of this data to complete the map. Some councils are working on it, some councils are ignoring the requests and a few councils are (sensibly) concerned about data protection. What became clear was that small deviations between councils have large effects in the readability of the maps. Southwark and Tower Hamlets are in the process of demolishing some council estates and rebuilding. The resulting emptiness makes the rest of the map looks like it has very few low use homes.
Adjusting for some boroughs having extremely high levels of empty homes due to reconstruction doesn't really help either, as then the whole map ends up a mess of different shades and finding the patterns becomes challenging as can be seen by the figure below. One thing that does come out of this is that it looks like the boroughs may be reporting (or we may be interpreting) different kinds of information. Notice just south of the river in the centre of the map there is a block of darker land surrounded by lighter land, this is Lambeth, it seems unlikely that there is actually a major difference in home usage between Lambeth and its neighbouring boroughs so further investigation as to why this is showing up is necessary and probably being careful comparing the differences between boroughs for now.
Currently the results show a large range of differences with low use homes varying from a few hundred in some boroughs to over 10,000 (Kensington and Chelsea).
There are two main things to take away from this analysis. The first is that there are some pretty stunning differences in the amount of empty homes between regions in London. The second is that getting data out of councils when they are not sure what it is you are asking for is very challenging. The data collection process has taken about 9 months and is not yet complete, there have been months of back and forth with individual councils debating whether the request breaches the Data Protection Act and whether it is possible to get the data at all.
Currently it doesn't seem like a great idea to try and analyse all the data from the local authorities together, so the next blog post will focus on individual analysis or some very big picture thoughts.
Although it has been a struggle to get the data and often very frustrating, it has also been clear how hard the people at the council work and the willingness of some to discuss such a strange request has made the process a lot easier (A big thanks to Lambeth's FOI team). It also was a reminder of how easy it is to make mistakes when you are trying to do something new, I messed up not understanding how different the data councils store is and had to go back to councils who had been very cooperative and ask for the same data but in a different way, this didn't always go down well, and when both sides make mistakes the situation can get even messier.
In the end the FOI process requires careful planning and thought, lots of patience, peer pressure and generally any sort of arm twisting you can think of to keep the council in contact with you and making progress on the request.
The next post will take a closer look at what is going on in individual Boroughs of London and trying to see whether there are any broader trends that are worth noting.
If you want to see the code see the repo on github