National Accounts articles: Methods article on the implementation of the reclassification of English housing associations and other classification decisions

This article covers the changes to national accounts resulting from the implementation of the reclassification of English housing associations to the public corporations sector.

In this article the shorthand of “housing associations” is used to refer to private registered providers of social housing registered with the Homes and Communities Agency (HCA).

In October 2015, we reviewed the statistical classification of “private registered providers” of social housing in England (where registered refers to being registered with the HCA). We concluded that these bodies are public, market producers and should be reclassified to the public non-financial corporations (S.11001) sector for the purpose of national accounts and other Office for National Statistics economic statistics.

This reclassification was implemented in the Public Sector Finances publication in February 2016. This article covers the implementation of the reclassification in national accounts back to 1 July 2008, the start of Quarter 3 (July to Sept) 2008.

The article shows how data from Global accounts of housing providers, published by the HCA, were used to estimate national accounts assets, liabilities and transactions. These include methods used to estimate the impact of the reclassification on the financial balance sheet, gross operating surplus and capital expenditure (gross fixed capital formation and expenditure on non-produced assets (land under housing properties)). Figures for housing associations were added to sector accounts for public corporations and deducted from private non-financial corporations. There is no impact on national accounts aggregates for the UK.

The Global accounts of housing providers covers all housing associations in England managing or owning at least 1,000 social units. These providers own or manage 95% of social housing stock. The article describes how grossing factors were calculated to produce estimates for all housing associations registered with the HCA.

Impacts on the main statistics for financial years ending (FYE) 2009 to 2012 are provided.

These changes will be introduced when revised figures for the UK National Accounts, consistent with Blue Book 2017, are published in September 2017.

This article describes the only major sector reclassification to be implemented in Blue Book 2017; the reclassification of English housing associations to the public corporations sector.

2.1 The classification

In October 2015, we reviewed the statistical classification of “private registered providers” of social housing in England (where registered refers to being registered with the Homes and Communities Agency (HCA)). The review was undertaken to establish whether these providers of social housing should continue to be recorded as private non-financial corporations (S.11002) in Office for National Statistics (ONS) economic statistics or if the sector classification should change following legislative and regulatory changes brought about through the Housing and Regeneration Act 2008 (HRA 2008).

We concluded that these bodies are public, market producers and should be reclassified to the public non-financial corporations (S.11001) sector for the purpose of national accounts and other ONS economic statistics. The classification applies with effect from 22 July 2008; the date of enforcement of the HRA 2008.

A further review of the relevant legislation in September 2016 concluded that private registered providers should be classified as public corporations from 24 July 1996. This article covers the initial reclassification from July 2008; the historic reclassification from July 1996 to the end of June 2008 is not being implemented in Blue Book 2017.

Private registered providers are frequently part of wider groups, which include entities not registered with the HCA. For example, financial subsidiaries responsible for sourcing bond finance, or construction subsidiaries responsible for developing properties. These non-registered subsidiaries are not included within the reclassification.

Although the HCA regulates within England some of the providers registered with the HCA will also operate across the border in Scotland, Wales and even perhaps Northern Ireland. However, these operations outside England are almost always part of separate subsidiaries that are unregistered with the HCA but are registered with Scottish or Welsh regulatory bodies. From a classification point of view this is not important and the entire institutional unit is to be reclassified to the public non-financial corporations (S.11001) sector regardless of which territories it is operating in. Similarly, the entire institutional unit is to be reclassified regardless of the range of activities of that provider.

In this article the shorthand of “housing associations” is used to refer to the providers of social housing to which the classification applies.

2.2 Prior implementation in public sector finances

The reclassification of private registered providers with the HCA was implemented in the Public Sector Finances publication in February 2016 and the Excessive Deficit Procedure publication in March 2016. Reclassifications are implemented in public sector finances as soon as possible to ensure impacts feed promptly into main statistics such as the public sector deficit (public sector net borrowing) and public sector net debt.

Implementation takes longer in UK National Accounts because it is necessary to include the additional impacts on the private sector (removing housing associations from figures for private non-financial corporations) and counterparties to financial transactions outside the public sector (for example, financial institutions who issued loans to housing associations).

UK National Accounts are estimated on a quarterly basis. To simplify the implementation in national accounts the reclassification is assumed to have taken place on 1 July 2008 rather than the exact date of the change in legislation (22 July 2008). Moving the date by 3 weeks has a very small impact on the accounts and makes implementation much simpler. Thus housing associations are included in the private non-financial corporations sector in Quarter 2 (Apr to June) 2008 and move to the public corporations sector at the beginning of Quarter 3 (July to Sept) 2008.

2.3 Reclassification of private registered providers of social housing in Scotland, Wales and Northern Ireland

This article only covers the reclassification of “private registered providers” registered with the HCA (in England). In October 2016, following a review, we concluded that registered providers of social housing in Scotland, Wales and Northern Ireland are also public market producers and will be reclassified to the public non-financial corporations sector.

This reclassification was implemented in the Public Sector Finances publication in February 2017 and the Excessive Deficit Procedure publication in March 2017. The reclassification of private registered providers in Scotland, Wales and Northern Ireland will be implemented in the UK National Accounts at a later date.

2.4 Characteristics of housing associations

Private registered providers vary greatly in the services they offer. The majority of their activity is providing homes for rent, but many also provide homes for ownership. Some provide additional services such as supported care for the elderly and housing for older people, students or nurses.

The Homes and Communities Agency (HCA) publishes a list of all registered private providers on its website. There are about 1,500 private providers.

3.1 Homes and Communities Agency (HCA) global accounts

All housing associations (that is, registered private providers in England) are required to complete regulatory returns to the Homes and Communities Agency (HCA). Full details of the financial accounts of housing providers that own or manage at least 1,000 social homes are published in the annual Global accounts of housing providers.

The HCA publishes both an annual report and a detailed spreadsheet showing accounts for each provider. In the financial year ending (FYE) 2015 there were 329 providers who owned 95% of the total stock.

The HCA data provide a coherent and robust data source. They are derived from returns to HCA required by legislation and are linked to the individual audited accounts of each body. For this reason the data are believed to provide the most accurate picture of the assets, liabilities and financial transactions of housing associations.

The HCA data have already been used in implementing the reclassification of housing associations in the public sector finances. However, there are two issues with using HCA data to implement the reclassification in national accounts and public sector finances:

• the dataset only covers providers with at least 1,000 homes and so excludes smaller providers (although as noted previously these larger providers cover 95% of the total housing stock for housing associations)
• the data in the dataset are presented following business accounting principles and not those of national accounts

To deal with the first issue the UK National Accounts and Public Sector Finances publications use figures from the HCA global accounts data, which have been grossed up (revenue, expenditure and assets by 4.7% and liabilities by 2.6%) to provide some estimate of the impact of the remaining small providers. The methods used to derive these grossing factors are described in Section 4.

To deal with the second issue the National Accounts and Public Sector Finances publications have mapped data from the HCA global accounts to national accounts’ transaction and stock concepts. Details of this mapping are explained in Section 5.

3.2 Other data sources

Other potential data sources were reviewed but none have been identified which provide as complete and robust a set of data as that provided by the HCA global accounts.

The FAME dataset of accounts held by Companies House was considered, but unfortunately many of these smaller providers are classified as Registered Societies and accounts are not reported to Companies House and so the data are not held on FAME.

Registered Societies are registered with the Financial Conduct Authority but ONS does not have access to these data.

3.3 Data on capital grants from central government to housing associations

In addition to the global accounts data we also make use of administrative data on capital grants to housing associations to part finance development of social housing. The Department of Communities and Local Government (DCLG) issues capital grants to the Greater London Assembly (for housing associations in London) and the HCA (for housing associations based elsewhere in England).

The Greater London Assembly and HCA then allocate grants to individual housing associations. DCLG supplied us with data on the size of this capital grant for each financial year.

Since there is no readily available data source for the accounts of smaller providers, figures from global accounts have been grossed up using the same methods used for public sector finances. Grossing factors were derived using an alternative data source from the Homes and Communities Agency (HCA).

Every year the HCA publishes statistics on Private Registered Provider Social Housing Stock in England. All private registered providers, regardless of size, are required to fill in this return.

Table 48 of this return shows the distribution of private registered providers by the size of the registered provider. In the financial year ending (FYE) 2015 over 95% of all social housing units were managed by providers with more than 1,000 social housing units, which would be included in global accounts. Medium-sized providers (250 to 999 social housing units) accounted for 3% of stock, leaving 1,111 small providers.

4.1 Grossing up transactions or balance sheet items that would apply to all providers regardless of size

Current revenue or expenditure is assumed to be proportional to the number of units managed regardless of the size of the provider. This is also assumed to apply to balance sheet items for cash assets and other accounts payable or receivable. For these transactions the estimate in global accounts is grossed up by (100 divided by 95.5) equals 1.047.

4.2 Grossing up transactions and stocks related to capital expenditure, debt and interest paid on debt.

Not all smaller housing associations excluded from global accounts are the same.

Very small providers, managing or owning fewer than 250 units, will be significantly different from those recorded in global accounts. In particular they are less likely to have outstanding loans or be raising new debt to develop new properties. In both public sector finances and national accounts we have assumed that these providers have negligible debt at the time of the reclassification and are not involved in developing properties (so do not acquire more debt).

Medium sized providers (managing or owning 250 to 999 units) will be more similar to the smaller providers included in global accounts (managing 1,000 to 2,500 units). They may well have debt and be investing in new properties.

The estimated debt for medium sized providers was calculated in the following way:

• the detailed spreadsheet data for individual housing associations was used to identify providers managing 1,000 to 2,500 units (the smallest included in global accounts)
• these data were used to calculate the average debt per unit managed
• estimated debt for medium sized providers equals average debt multiplied by the number of units managed by medium providers

The grossing factor for debt equals the sum of (global accounts debt plus estimated debt for medium providers) divided by (global accounts debt).

This grossing factor (1.026) was also used to gross up interest paid on the debt.

As with most administrative data sources the Homes and Communities Agency (HCA) global accounts data are not presented using national accounts concepts. So each national accounts stock or transaction needs to be mapped to global accounts data.

Mahajan (2013) provides a useful summary of the main differences between national accounting and business accounting concepts.

The main changes as applied to global accounts are shown in Table 1.

Table 1: Mapping global accounts to national accounts concepts

Source: Office for National Statistics

Notes:

1. Details of the methods used to estimate the consumption of fixed capital for Housing Associations will be published in the Capital Stocks bulletin in November 2017.
2. The treatment of Social Housing Grants for Housing Associations in National Accounts is due to be reviewed by the ONS Economics Statistics Classification Committee later in 2017.
3. For more information about the valuation of liabilities in Global Accounts, including changes introduced as part of the adoption of FRS102 see HCA Global accounts 2016.

In each case the data were removed from the private non-financial corporations sector (S.11002) and added into the public non-financial corporations sector (S.11001). As a result there is no impact on the UK economy as a whole.

Implementing the reclassification in the national accounts includes identifying how much of each change to output and expenditure from private non-financial corporations to public corporations should be allocated to each industry.

In national accounts all businesses are assigned an industry based on the Standard Industrial Classification: SIC 2007.

The Structure and Explanatory notes for SIC 2007 explain how bodies should be allocated to SIC 2007 classes. In the case of enterprises covering a number of functions the classification for national accounts purposes should be based on the contribution of activities to gross value added.

Housing associations make most of their money from renting out property so this rule on gross value added would assign them to industry 68.2 Renting and operating of housing association real estate. This would imply that all changes to the components of gross domestic product (whether based on production, income or expenditure approaches) should be assigned to industry 68.2. However, this may not reflect the reality of data on housing associations currently included in estimates for private non-financial corporations.

The Office for National Statistics Inter-Departmental Business Register (IDBR) records the industry of all businesses. In some cases this may be based on the part of the business employing the most staff rather than the contribution to gross value added. A few housing associations operate large care homes employing substantial numbers of staff so will be assigned to industry 87 Residential care activities. Survey data for these housing associations will have fed into estimates for industry 87, not industry 68.

It was therefore important to assign some of the changes in the current account to industry 87.

To decide how much to allocate, the list of housing associations in the global accounts was matched against the Inter-Departmental Business Register to identify the proportion of business turnover recorded on the IDBR for housing associations in industry 87. This analysis found that about 90% of turnover was for housing associations assigned to industry 68 and 10% for those assigned to industry 87. So for current account transactions, 90% of the change was assigned to industry 68 and the remaining 10% to industry 87.

In the case of capital expenditure all gross fixed capital expenditure on dwellings in national accounts is automatically assigned to industry 68 so there was no need to split the impact in this case.

Housing associations finance the development of new properties using a combination of surpluses from operations, capital grants and borrowing from banks and the financial markets. Table 2 shows the mapping of global accounts data to national accounts transactions for the financial balance sheet.

Global accounts does not include information on holdings of assets or liabilities for equity and investment funds shares or units (AF.5). Information is available on net pension assets or liabilities but separate figures for pension assets and liabilities are not available.

The recording of pension liabilities in national accounts does not follow the treatment in company accounts. Global accounts does not include information on insurance assets or liabilities.

Some housing associations included liabilities for financial derivatives within their returns for other long-term creditors, but data from the Homes and Communities Agency (HCA) Quarterly Survey of Private Registered Providers suggest the contribution is small for these early years. For these reasons no changes to AF.5, AF.6 or AF.7 have been implemented in national accounts.

In the absence of information on valuations, changes in assets or liabilities (flows) were derived as the difference between values for successive financial years.

Since there are no data available on disposals and acquisitions by quarters, flows are assumed to be evenly spread throughout the financial year. Quarterly figures are obtained by dividing financial year figures by four.

7.1 Bond and loan liabilities

The estimated figures for long-term liabilities need to be split into the national accounts balance sheet items for bonds (AF.3) and long-term loans (AF.42).

Whilst global accounts does not collect this information the Homes and Communities Agency Quarterly Survey of Registered Private Providers does include information about types of debt.

This publication includes data from the same private registered providers as global accounts (with a few very small exceptions) so is a good source of information to allocate long-term liabilities to different classes of debt.

Data from this survey have been used to estimate the percentage of housing association debt (long-term loans plus short-term loans plus bonds) that is held as corporate bonds. This increased from 13% of debt in financial year ending (FYE) 2012 to 31% in FYE 2016. No information is available on the bond or loan split for FYE 2009 to FYE 2011 so the split for these years was assumed to be the same as in FYE 2012. This is likely to have resulted in an overestimate of the amount of bond debt in these earlier years, with a corresponding underestimate of liabilities from long-term loans. The estimated bond or loan split has no impact on the total amount of long-term liabilities or total interest paid by housing associations on this debt.

7.2 Adjustments to global accounts data to take accounts of differences in reporting by housing associations

Since the implementation of the reclassification in the Public Sector Finances publication in March 2016 there has been a further opportunity to review the contributions of data for individual housing associations to the estimate of net debt. As is common with administrative data collections not all data providers were consistent in their recording of liabilities. One large group of housing associations included substantial long-term liabilities within figures for other current liabilities up until FYE 2013. To improve estimates of loan and bond liabilities in national accounts an adjustment was calculated to reallocate some of these “current liabilities” to long-term liabilities.

In FYE 2014, this group started to report these long-term liabilities correctly. In this year they reported a stock of £150 million current liabilities.

The adjustment assumed that “other current liabilities”for this group of housing associations was £150 million for all years prior to FYE 2014. Any additional liabilities included in “other current liabilities” were assumed to be long-term liabilities and were split between long-term loans and corporate bonds.

Details of the impact of these adjustments on public sector net debt are shown in Table 3. Accuracy of reporting has improved in recent years so most adjustments are to figures for the earlier years.

Table 4 shows how the gross operating surplus for housing associations is derived from the income and expenditure account in global accounts.

Current expenditure for housing associations is split into two national accounts transactions: compensation of employees (D.1) and intermediate consumption (P.2). An estimate was produced for compensation of employees and intermediate consumption derived as a residual.

Payments to staff form a large part of current expenditure for housing associations and need to be recorded separately in national accounts.

The national accounts concept, compensation of employees (D.1) is defined as the total remuneration, in cash or in kind, payable by an employer to an employee in return for work done by the latter during an accounting period. (Eurostat 2013, paragraph 4.02).

Estimates for businesses in the public corporations sector are normally derived from figures on expenditure on wages and salaries in published accounts. An additional 15% is added for employers’ social contributions to give the figure for compensation of employees.

The global accounts data do not include information on wages and salaries paid to staff. Whilst this information is recorded in business accounts submitted to Companies House, it would be impractical to extract for individual housing associations and is not available at all for those who do not publish business accounts. So an alternative method was used combining two data sources:

• the Homes and Communities Agency (HCA) collects information on the number of full-time equivalent staff employed by each large housing association; these data were adjusted to account for the small housing associations that do not submit data to the HCA and then combined with Office for National Statistics (ONS) data and other HCA data to produce the estimates required to implement this reclassification in the public sector employment statistics
• the ONS Public Sector Employment Team provided estimates of the number of full-time equivalent staff by industry (68 Housing association or 87 Residential care) and geographical location (London or outside London)
• ONS also collects data on earnings and hours paid in the Annual Survey of Hours and Earnings (ASHE):
  • statistics for mean gross hourly earnings were multiplied by 37 (full-time equivalent hours a week) and 52 (weeks in a year) to estimate the mean salary per full-time equivalent member of staff in each calendar year
• multiplying the number of full-time equivalents (FTE) from the public sector employment data by the average salary per FTE gives an estimate of the annual wage and salaries bill for the housing associations included in global accounts

Since average salaries for staff based in London are higher than elsewhere in England and those for employees in industry 68 (Housing associations) are higher than for industry 87 (Residential homes), figures were estimated separately for these four groups and then aggregated to obtain estimates for industry 68 and 87. Estimates for these four combinations were compared with figures extracted from published accounts from housing associations in these groups and were found to be comparable.

We advise users of earnings statistics from ASHE to use the median as an estimate of average earnings to avoid the disproportionate impact of high hourly rates from a few high-earning individuals in the sample. Initial estimates of total expenditure on wages and salaries for housing associations were based on median hourly earnings. However, these median-based estimates were found to underestimate total expenditure on wages and salaries when compared with published accounts for a sample of housing associations.

Importantly the estimate of compensation of employees is only used to split current expenditure into compensation of employees and intermediate consumption. It does not impact on gross operating surplus, nor other main fiscal aggregates such as net borrowing.

10.1 Estimating total capital expenditure

Housing association capital expenditure on a national accounts basis comprises:

• capitalised expenditure on existing fixed assets
• less asset sales
• plus new capital expenditure on housing properties, investment properties and other fixed assets

Up until 2016, global accounts did not collect data on capital expenditure. So an alternative method had to be developed for implementation in both public sector finances and national accounts (Table 5).

The method assumes that new housing association capital expenditure is financed by a combination of net cash flow from operating activities, property income, capital grants and increased net borrowing.

Net cash flow from operating activities is calculated as the sum of the global accounts operating surplus, plus tangible asset depreciation less expenditure on capitalised repair and maintenance costs.

The global accounts operating surplus includes profit from the sale of current assets (homes developed for market sale and first tranche shared ownership schemes). In national accounts these are treated as fixed assets not current assets. The sale of these assets is treated as a disposal of fixed assets in the national accounts estimate of capital expenditure.

Sales of fixed assets recorded in global accounts are treated as disposals of fixed assets (GFCF) in national accounts. Since these capital receipts are not included in our estimate of money available to finance new capital expenditure, we do not need to deduct them from our estimate of GFCF.

10.2 Adjusting figures for capitalised expenditure on existing fixed assets

Looking closely at figures for capitalised expenditure on major repairs in Table 6 reveals a large increase between financial year ending (FYE) 2011 and FYE 2012.

In FYE 2012, the reporting of capitalised expenditure in global accounts changed with the adoption of component accounting as outlined in the Statement of Recommended Practice (SORP) 2010. Some expenditure that had previously been classified as major repairs was assigned to capitalised expenditure.

In national accounts, expenditure is defined as capital expenditure if it improves the existing fixed asset beyond ordinary repairs or maintenance, otherwise it is classified as current expenditure.

Figures for capitalised expenditure on existing fixed assets before FYE 2012 needed to be increased and expenditure on major repairs reduced to provide better estimates of capital and current expenditure for national accounts purposes.

In FYE 2012 and subsequent years, 77% of major repair costs in global accounts were capitalised. This figure was used to produce revised estimates of capitalised expenditure in earlier years (Table 6).

This adjustment has the same impact on both capital expenditure and gross operating surplus so has no impact on public corporations’ net borrowing.

10.3 Splitting capital expenditure into gross fixed capital formation and acquisitions less disposals of non-produced assets.

In national accounts, capital expenditure is split into two components:

• gross fixed capital formation (GFCF)
• expenditure on non-produced assets (NP)

GFCF consists of resident producers’ acquisitions, less disposals, of fixed assets during a given period plus certain additions to the value of non-produced assets realised by the productive activity of the producer or institutional units. Fixed assets are produced assets used in production for more than 1 year, such as building, machinery, transport equipment, or intellectual property such as software or the results of research and development (Eurostat, 2013, paragraph 3.124).

GFCF is further split by type of asset. For housing associations the vast majority of GFCF is acquisitions or disposals of housing properties (properties built for social rental, shared ownership sale and market sale) and investment properties (properties built for market rent). There is also a small amount of gross fixed capital formation on other fixed assets. Analysis of existing national accounts data for industry 68 showed that the majority of GFCF for other fixed assets is assigned to buildings.

Non-produced assets (NP) consist of assets that have not been produced within the production boundary, and that may be used in the production of goods and services (Eurostat, 2013, paragraph 3.184).

New capital expenditure on housing properties needs to be split into expenditure on the dwelling (which is GFCF) and expenditure on the land on which it sits (which is a non-produced asset).

Balance sheet valuations of housing properties, other fixed assets and land were used to estimate the percentage of capital stock in each category. It was assumed that new capital expenditure was proportional to the amount of existing capital stock.

10.3.1 Expenditure on other fixed assets

Other fixed assets accounted for about 1% of the existing capital stock. Estimated GFCF for other fixed assets was less than £100 million a year, which was assigned to the buildings asset.

10.3.2 Expenditure on dwellings

Recent work on the national balance sheet estimated that about 62% of the cost of dwellings between FYE 2009 and FYE 2012 was the value of land underneath them.

Each year about 62% of the remaining new capital expenditure was allocated to non- produced assets.

The remaining new capital expenditure was allocated to gross fixed capital formation for dwellings. All capitalised expenditure on existing assets was allocated to GFCF for dwellings.

A summary of the impact of the reclassification for the financial years ending (FTE) 2009 to FTE 2012 is shown in Table 7.

The reclassification was implemented in the Public Sector Finances publication in February 2016 and the Excessive Deficit Procedure publication in March 2016.

National accounts estimates for public corporations’ net borrowing are very similar to initial estimates published in public sector finances.

Estimates of public sector net debt are higher due to adjustments to long-term liabilities (Section 7.2).

We would like to thank staff at the Homes and Communities Agency for providing detailed advice on the global accounts data used in the implementation of the reclassification in the national accounts.

Classification of ‘private registered providers’ of social housing registered in England

Implementation of the reclassification in Public Sector Finances

Classification of “private registered providers” of social housing registered in Scotland, Wales and Northern Ireland

Global Accounts of Housing Providers published by the Homes and Communities Agency

FAME dataset produced by Bureau van Dijk

‘Private Register Provider Social Housing Stock in England’ published by the Homes and Communities Agency

Mahajan, S (2013). Challenges of using Company Accounts based data in National Accounts, ONS, UK

Future review of treatment of Social Housing Grants for Housing Associations

The Standard Industrial Classification (SIC 2007)

Homes and Communities Agency Quarterly Survey of Registered Private Providers

Public Sector Employment, UK

Annual Survey of Hours and Earnings, UK

European System of Accounts, ESA 2010

National Accounts articles: Changes to the National Balance Sheet for the Blue Book, 2017