Housing summary measures analysis

This report presents analysis of the 15 Housing Summary Measures which form part of the new ONS Housing Statistics Portal, which is the platform for a range of official statistics and analysis on housing. New analysis of housing statistics will be added to the Portal over time, using the latest available data. The Housing Summary Measures provide a broad overview of the availability and affordability of privately owned and social housing for local authorities in England and Wales, and also looks at the affordability of private rented housing. These summary measures of housing can be used to identify the relationships between various measures of housing availability and affordability.

The Housing Summary Measures are either taken directly from previously published official statistics or calculated using previously published official statistics as a base. Where possible, consistent time series data for English and Welsh local authorities are presented and these are expressed as percentages, ratios or prices to enable comparisons across local authorities. The 15 Housing Summary Measures are as follows:

1. Privately owned housing stock as a percentage of total housing stock

2. Number of new dwellings built as a percentage of privately owned housing stock

3. Number of residential house sales as a percentage of privately owned dwelling stock

4. Social housing stock as a percentage of total housing stock

5. Vacant social housing as a percentage of social housing stock

6. Social housing shortfall as a percentage of social housing stock

7. Median house price

8. Median gross annual salary

9. Ratio of median house price to gross annual salary

10. Median monthly private rent

11. Median gross monthly salary

12. Median monthly rent as a percentage of median gross monthly salary

13. Average weekly social housing rent

14. 10th percentile gross weekly salary

15. Average weekly social housing rent as a percentage of 10th percentile gross weekly salary

For the 348 local authorities in England and Wales in 2014, median house prices¹ ranged from just under £1.2m in Kensington and Chelsea, to £75,000 in Blaenau Gwent. The median house price for England and Wales overall in 2014 was £194,955. Since 1995, the local authorities with the highest median house prices have tended to be in London and the South East, whilst local authorities with the lowest prices tended to be distributed across the other English regions and Wales.

In London in 2014, Barking and Dagenham had the lowest median house price for the all dwelling type category, at £215,000. Despite having the lowest median house price in London, this was still higher than 60% of all local authorities in England and Wales. In general, median house prices for West London boroughs were higher than for East London boroughs. Outside London, the areas with the highest median house prices were typically within close commuting distance of London, although other cities also had higher house prices than their surrounding areas.

Figure 1 shows that the difference in house price growth between median house prices in the most and least expensive 10% of small areas varied considerably. Since 1995, the price paid for houses in the least expensive 10% of areas in England and Wales has increased by three times the average 1995 price. Meanwhile, the most expensive 10% of areas have seen price rises of more than 4 times the average 1995 price.

Between 2004 and 2009, house price growth between the most and least expensive areas was similar, shown by the convergence of the lines on the chart. However, after the recession house price growth in the most expensive 10% of areas has far outstripped growth for the least expensive 10% of areas, which in 2014 had not recovered to pre-recession levels. These areas experienced the so called ‘double-dip’ in 2011 where average house prices decreased again whilst prices in the most expensive 10% of areas continued to recover.

The Annual Survey of Hours and Earnings provides estimates of average gross annual salaries for residents of local authorities in England and Wales. In 2014, local authorities in which residents had the highest average earnings were predominantly spread across London and its surrounding areas. These were also the areas where average house prices were highest. The chart below shows median house prices against median earnings to give an indication of housing affordability. Points on the chart which appear above the trend line can be considered relatively more affordable than average, while points that appear below the line are relatively less affordable.

Out of the local authorities where median earnings were highest, house prices also tended to be high. Figure 2 shows that there was a strong positive correlation between earnings and house prices. People who earn more are more likely to be able to afford to buy their own home, and this can increase the demand for owner occupied housing, which in turn pushes up average house prices.

Figure 2: Median salary and median house price by local authority

England and Wales, 2014

Source: Annual Survey of Hours and Earnings (ASHE) - Office for National Statistics

The local authority of Blaby in the East Midlands is an exception to this trend. In 2014, average gross annual salary in Blaby was £25,637 which is in the highest 20% of all local authorities. In England and Wales overall, median annual salary for all workers was £22,215. However, the median house price in Blaby, at £170,000 falls below the England and Wales average house price, making Blaby a relatively affordable area in which to own property. One explanation for this is that some of the residents of Blaby work in other local authorities in which average salaries for workers is higher, leading to an increased average income of Blaby residents relative to house prices. In Purbeck however, housing affordability was relatively low, where the median gross annual salary was £16,481 in 2014 while the median house price here was £247,500.

Average house price statistics themselves do not give a full picture of the affordability of privately owned housing, because these figures do not take into account average earnings which in part dictate how much people can afford to pay to buy their own home. Looking at the ratio of median house prices to median annual salary sheds light on the relative affordability of owner occupied housing across the local authorities of England and Wales. The ratios presented here were calculated using median house price data from ONS’ House Price Statistics for Small Areas and dividing them by the median total gross annual salary figures from the Annual Survey of Hours and Earnings.

Figure 3 shows that in 2014, Westminster had the highest ratio of house prices to annual salary in England and Wales, where house prices were 24 times greater than gross annual salary on average. This could therefore be described as the least affordable area in which to buy a house. The most affordable area in 2014 was Blaenau Gwent, where the average house price was four times greater than the average salary.

The chart shows the ratio of house price to salary for Westminster, the least affordable area in England and Wales, and for Blaenau Gwent, the most affordable area for most years in the series. The ratio for Westminster has increased over the last 13 years and can therefore be described as less affordable, while affordability for Blaenau Gwent has improved over this period. House prices here fell from a high of 5.34 times the average annual salary in 2007 to four times average salary in 2014. However, the affordability gap between the most and least affordable areas has steadily increased over this time.

This section uses data from the Valuation Office Agency to look at the geographical distribution of median monthly private rent by local authority for England. The method for producing these statistics provides a ‘snapshot’ of private monthly rent which does not enable comparisons over time, and so only data for 2014 have been presented in this analysis.

Data for Wales have not been included because they are not strictly comparable over time, between geographies or between property types due to variations in the composition of the sample used to calculate the statistics. However, to put Welsh private monthly rent in context, during 2014, the majority of rents recorded were for 2 and 3 bedroom properties where the median rents were £495 and £550 respectively. At a local authority level, the lowest median rent recorded during 2014 was £238.33 per month for a non self-contained single room in both Powys and Rhondda Cynon Taf. The highest median rent was £950 per month for a large four bedroom property in Cardiff.

Figure 4 shows that in 2014, the majority of the most expensive areas for renting privately were in London and the surrounding areas, as shown in the chart below by the London boroughs appearing on the upper end of the private rent distribution. Median monthly private rent was highest in Westminster (£2,383) and Kensington and Chelsea (£2,275). These were the only two local authorities in England to have a median monthly private rent of more than £2,000 in 2014. The areas with the lowest median monthly private rent in 2014 were Kingston upon Hull (£365) and Burnley (£395).

In English local authorities, the cost of private renting was very strongly correlated with average house prices overall, as shown in Figure 5. Areas that had the highest median house prices also had the highest median monthly rent prices. Again, this was particularly the case for areas within and just outside London which tended to have both the highest average house prices and the highest average private rents. The scatter plot below shows median monthly private rents against median house prices and it shows how strongly correlated the two indicators are.

Figure 5: Median monthly private rent and median house price by local authority

England, 2014

Source: Office for National Statistics and Valuation Office Agency

Data on median gross monthly salary are derived from the Annual Survey of Hours and Earnings for England and Wales. This section identifies the distribution of monthly salary by local authority.

The geographical pattern of median gross monthly salary follows that of the annual figures, with residents of London boroughs tending to have the highest earnings, as shown in Figure 6. Generally, monthly salary was highly correlated with the monthly cost of private renting, although, as with other housing summary measures, some local authorities did not show this trend. For example, in 2014, residents of Copeland, in the North West, earned an average salary of £2,051 a month which was in the top 25% of local authorities in England and Wales, whereas the monthly cost of private rent was £400 on average, which was in the bottom 10% of all local authorities.

By examining the ratio of median monthly private rent to median gross monthly salary, we can assess the relative affordability of renting privately in different areas in England.

The highest ratio of rent price to salary for which data are available was in the London borough of Westminster at 78.3%. In addition to this, the highest 18 areas were London boroughs. 25 areas had a median private rent equal to 50% or more of median monthly salary.

In its 2015 Summer Budget, the new Government announced plans to charge full market rent to social housing tenants who earn a household income (pre-tax) of more than £30,000, or £40,000 in London. According to the Institute for Fiscal Studies, this will affect around 10% of households in social rented accommodation. Taken together, the figures on average income and private rent prices suggest that households just outside London, where private rent prices are relatively high, will be most affected by these changes. Most of the previous government’s cuts to housing benefit were in the private rented sector, which accounts for 40% of housing benefit spending. The coalition reduced the maximum amount of rent that housing benefit can cover from the 50th percentile to the 30th percentile of rents in the local area. Early evidence suggests that this had little effect in pushing down rents, and hence resulted in most claimants paying more net rent.

Copeland had the lowest ratio with median private rent price at 21.95% of monthly salary. There were a further 70 areas with a ratio of rent price to income of less than 30%, none of which were in London. The lowest London borough was Bexley with median rent price equal to 40.4% of median monthly salary. Map 1 shows the geographical spread of the most and least affordable areas for private renting.

Map 1: Highest and lowest areas of median monthly private rent price as a % of median gross monthly salary

England, 2014

Source: Office for National Statistics and Valuation Office Agency

Notes:

1. Contains Ordnance Survey data © Crown copyright and database right 2015

This section looks at the average cost of social housing rent for local authorities in England and Wales using data from the Department for Communities and Local Government ¹ (for England) and Stats Wales ² (for Wales).

Newham in London had the highest average social rent price at £128.90 per week, just under double that of the cheapest, Pembrokeshire at £65.50. This is indicative of a general pattern of higher social rent costs in areas of London and the South East compared with other parts of England, as shown in Map 2. There were 104 areas with average social rent over £100 per week in 2014. These include all 32 London boroughs.

Map 2: Average weekly social housing rent: Highest and lowest 30 areas by local authority

England and Wales, 2014

Source: Communities and Local Government, Welsh Government

Notes:

1. Contains Ordnance Survey data © Crown copyright and database right 2015

The geographical difference in social rent price does not necessarily indicate geographical inequality in the affordability of social renting. We must first take into account residents’ average earnings. This is explored further in the analysis for the affordability of social housing measure (weekly average social housing rent as a percentage of 10th percentile gross weekly salary). There were no strong correlations between average social rent and social housing shortfall, social housing stock or vacant social housing.

Notes for 7. Cost of Social Rent

1. For England, data on weekly rent for Private Registered Provider (PRP) social housing have been used where available. Note that figures on weekly rent for local authority provided housing differ from PRP weekly rent figures.
2. Data used are average weekly rents in self-contained stock at social rent by provider type and year.

In order to analyse and draw conclusions about an area’s lowest paid residents, we can look at its tenth percentile gross weekly salary. In the same way that the median salary is the middle value in a series, the tenth percentile is the value ten percent of the way from the bottom number in a series. This can give an idea of the earnings of the lowest earning workers, who are among the most likely to live in social rented accommodation, in a standardised way across multiple areas.

Figure 7 shows that between 2002 and 2014, tenth percentile gross weekly salary increased overall in England and Wales with two short periods of decline in 2004 and 2010.

Much like annual and monthly salary, areas in which residents had higher tenth percentile gross weekly salaries were prevalent in London and parts of the South East and the East of England. There were 24 local authorities with a tenth percentile gross weekly salary below £100 a week. Areas where residents generally had lower incomes had a fairly even geographical spread across England and Wales.

Workers in Islington had the highest tenth percentile weekly salary at £247.20, over three times that of the lowest, Broadland (£80.20). Newham was the London borough with the lowest tenth percentile gross weekly salary of £111.10.

Areas of relatively low income did not correlate with areas of higher social housing stock or social housing shortfall. This further indicates that the home ownership and rental markets along with other economic factors are more prominent drivers in social housing prevalence and shortfall. Despite this, areas with the lowest tenth percentile incomes tended to have the lowest social housing costs. This is discussed further in the analysis section for weekly average social housing rent as a percentage of tenth percentile gross weekly salary.

By examining the ratio of average weekly social rent to the tenth percentile gross weekly salary we can assess the relative affordability of different areas in England and Wales. Using earnings data from the Annual Survey of Hours and Earnings and social rent data from the Department of Communities and Local Government (for England) and Stats Wales demonstrates the relative affordability of social renting.

The tenth percentile gross weekly salary is used in the construction of this measure in order to best represent individuals who are likely to live in social rented accommodation. Within the context of social housing, this has more relevance than median salary because social tenants are likely to earn less than the median income in a given area. These ratio data do not take into account rent reductions for those receiving Local Housing Allowance nor do they take into account other benefits received. This means that tenants in some local authorities appear to spend more on rent than is received in earnings on average, which is not necessarily the case. However, on a local authority basis, these figures do provide a useful indication of the general affordability of social housing relative to other local authorities.

Figure 8 shows that in general, the affordability of renting socially in England and Wales has decreased between 2002 and 2014. This is reflected in the fact that on average, a larger percentage of residents’ earnings was spent on rent in 2014 than in 2002.

In 2014 there was only a weak positive correlation between tenth percentile gross weekly salary and average social housing costs. This suggests that social housing was not relatively affordable in all areas in England and Wales, as shown in Figure 9. A broad pattern of relatively more affordable areas can be seen in the Midlands and Wales, with relatively less affordable local authorities in London and southern areas of England. In 158 local authorities across England and Wales, residents who earned the tenth percentile income or less could expect to spend at least 75% of their income on social housing rent.

Home ownership is often viewed as a long term aspiration of families and households, and areas where home ownership is common are often thought of as more affluent than those areas where private renting and social housing is more common.

Dwelling stock data for England and Wales are available from the Department for Communities and Local Government and Stats Wales respectively. For English local authorities, dwelling stock figures include both owner occupied housing and privately rented housing as one category, and so the data presented here for England and for Wales combine these two categories in order to enable like for like comparisons. However, care should be taken when drawing conclusions from these figures because the percentage of privately owned dwellings may not necessarily indicate the percentage of owner occupied housing, as the percentage of privately rented accommodation accounts for part of this figure.

In recent years, the percentage of privately owned housing in England and Wales rose from 81.1% in 2011 to 82.4% in 2014. By contrast therefore, the prevalence of social housing has decreased over the period. The extent to which the private rented sector has driven these changes is not shown here, although previous research, using Census data, found that the prevalence of owner occupied housing fell for the first time in a century in the decade between 2001 and 2011, while the proportion of people living in private rented accommodation rose from 12% to 18%. Given that this trend has continued in recent years in England and Wales, this suggests that increases in the privately owned housing stock shown in the chart below, were mainly driven by increases in private renting.

Figure 10 shows that in Southwark, residents were the least likely to privately own a home out of all local authorities in England and Wales, with 56.5% of all the housing stock being privately owned in 2014. Figure 10 shows that the local authorities where the proportion of privately owned dwellings was lowest were in London, whereas those with the highest prevalence were outside of London. Unsurprisingly, the local authorities with the highest proportions of residents owning in 2014 generally had a low proportion of social housing stock. These recent trends in privately owned housing may not necessarily be a reflection of simply the availability of privately owned housing, but more specifically the availability of affordable housing. The affordability of housing is described in the analysis of the other housing summary measures presented here.

The supply of housing is largely driven by the completion of newly constructed dwellings, and this in turn contributes to changing house prices. Housing construction statistics are therefore important when considering the likely impact on prices of new housing developments (which tend to have higher average prices than existing houses). This is particularly important given the Government’s aim to increase the number of houses built. Official statistics on housing can show that the relationship between house building and house prices is somewhat more complex than the logical assumption that areas in which more homes are built have lower house price inflation than areas with fewer new builds.

Figure 11 shows that the UK overall has seen a decline in the number of new dwellings built since the post-war boom in house building. The largest year on year falls in house building were during the most recent recession, where in 2009 the number of completed dwellings fell by 39,740 compared with the previous year when the number of completed dwellings was relatively high. However, before the recession, gradual declines in house building were seen year on year, albeit with a few exceptional years in the late 1980s and mid-1990s. Since the most recent recession, house building has not recovered to exceed 150,000 completed dwelling per year. Moderate increases in 2012 and 2014 were seen where the number of newly built dwellings did increase on the previous year, but these may not have been large enough increases to bring about slower house price growth, and did not match pre-recession levels. A summary of UK level housing statistics is available in the UK Perspectives article.

The number of residential house sales in an area can provide an indication of activity in the housing market. However, some areas have a larger housing stock, and more house building, than others. Therefore, looking at the number of house sales as a percentage of the privately owned housing stock is a more helpful measure of housing turnover because it accounts for variations in the size of the housing stock.

The overall turnover of residential housing varies considerably by area and Figure 12 shows that it has increased in recent years for Englnd and Wales overall. Excluding the extremely high and low turnover areas of the City of London and the Isle of Scilly respectively, the range of housing turnover in 2014 was between 3% of the privately owned housing stock in Carmarthenshire, and 8.5% in Wandsworth. Many of the local authorities where housing turnover was highest in 2014 are within commuting distance of London, and this may reflect not just a buoyant housing market but also an active jobs market as people move in and out of London and its surrounding areas for work. Understandably, areas with a higher proportion of privately owned housing had a relatively higher number of house sales.

Also linked to housing turnover is the affordability of privately owned housing. In 2014, the areas with the highest housing turnover also tended to have the highest median house prices and residents with the highest annual salary. This suggests that in areas where income is relatively high, the demand for privately owned housing is greater than areas where residents’ income is lower.

Figure 13 shows that terraced houses were the most commonly sold type between 1995 and 2014, while flats and maisonettes were the least commonly sold type in general. This reflects the size of the housing stock by type. During the most recent recession, the number of sales for all house types reduced markedly. In 2008, the number of sales fell from the previous year for all house types. The largest fall was for terraced houses (50% below the 2007 figure of 391,729). In 2010, a recovery in housing market activity was underway, with all house types having more sales than in the previous year. The figures for the most recent complete year, 2014, shows that the recovery was gathering pace, with all house types having more sales than since before the recession, although housing market activity overall had not returned to pre-recession levels in 2014.

In 2014, local authorities in London had some of the highest percentages of social housing (the top seven were all London boroughs). Southwark was the area with the highest percentage of social housing at 43.5% of the overall dwelling stock, much higher than the average for all English and Welsh areas (17.4%). Areas with a high percentage of social housing tended to be urban and inner city areas, although Merthyr Tydfil, Blaenau Gwent and Torfaen in Wales had social housing prevalence in the top 20% of all local authorities.

In areas with the largest percentages of social housing, average weekly salary was also shown to be high. We may expect areas of high social housing prevalence to correlate negatively with average earnings, but this is not the case. Figure 14 below shows that the areas where social housing was most prevalent had some of the highest average earnings according to the Annual Survey of Hours and Earnings, yet this was not indicative of the overall pattern and can partly be explained by the relative affluence in many London boroughs.

When comparing the availability of social housing with other housing indicators at the local authority level, there are no strong correlations. This could be due to other factors such as local economies, job opportunities and demographic factors driving social housing prevalence. Data at a more granular geographic level may reveal stronger relationships between key indicators.

Data on social housing vacancies are available for England from the Department for Communities and Local Government, and for Wales from Stats Wales. Despite the number of social housing vacancies being generally low, a small number of areas had high vacant social housing relative to England and Wales overall (1.4%). In 2014, Blackpool had the highest percentage of vacant social housing (5.9%) out of all the local authorities in England and Wales. Less than a fifth of all areas had more than 2% vacant social housing.

Generally, the prevalence of vacant social housing was not strongly correlated with the prevalence of social housing overall. Areas with a relatively high demand for social housing did not have a lower prevalence of vacant social housing, and vice versa. For example, in 2014 social housing in Southwark comprised 43.5% of the entire housing stock (the highest percentage in England and Wales) but the prevalence of vacant social housing in Southwark was similar to the England and Wales level overall (1.4% of all social housing). This could partly be caused by a higher turnover of social housing in areas where it is more prevalent, making vacant social housing, even if temporarily being refurbished, more likely.

Figure 15 shows that the local authorities where vacant social housing was most prevalent were also areas where the cost of buying a house was lower. This suggests that the areas where pressure on social housing is highest are the more expensive areas. Private registered providers (Registered Social Landlords in Wales) of social housing may be more likely to expand their housing portfolio in areas where purchasing a house is relatively more affordable, which in turn can lead to a higher vacant housing stock in these areas at times.

The shortfall in social housing can be defined as the number of households on a local authority’s social housing waiting list minus vacant social housing, expressed as a percentage of the overall social housing stock. This section contains local authority data for England only because statistics on social housing waiting lists are not available for Wales.

The areas with the largest shortfall in social housing were Medway, Solihull and the Isle of Wight (239%, 143% and 123% respectively). The shortfall in social housing in Medway was more than twice the social housing stock in 2014.

All areas had a shortfall of social housing except three: Allerdale, Wyre and Milton Keynes (-1.7%, -1.5% and -0.7% respectively). These areas were unique in that they reported waiting lists of zero households in 2014, despite having waiting lists of more than 1,000 in years prior. This means that any vacant social housing represents a surplus rather than a shortfall, which was exhibited by all other areas.

The uptake in the Government’s Right to Buy scheme is also a factor in the overall social housing stock for England and Wales. From the turn of the millennium up to 2013/14, 398,170 social dwellings have been added to the stock (which includes both new builds and the acquisition of existing buildings), while there have been 457,757 sales of social housing under the Right to Buy Scheme over the same period, which shows two distinct phases visible in Figure 16: The first seven years seeing Right to Buy sales exceed the development of new social housing, and the last eight years in which social housing development has outstripped Right to Buy sales. Although in 2014, there were slightly more sales under the Right to Buy scheme than there were additional social houses provided (in England only). The recent trend implies that the size of the social housing stock has been increasing over the last few years, however, in the UK overall, it has fallen by more than 8% between 2000 (5.39 million dwellings) and 2013 (4.96 million dwellings). This suggests that the changing usage of social housing has contributed to pressure on availability, which may in turn have consequences for the number of households on social housing waiting lists.